Modflowsurfact V3 PDF

HYDROGEOLOGIC, INC.
MODFLOW-SURFACT
SOFTWARE LICENSE AND LIMITED WARRANTY
This legal document is an agreement between you, the end user, and HydroGeoLogic, Inc. (hereafter “HGL”). BY OPENING
THIS SEALED DISK OR TAPED PACKAGE, YOU ARE AGREEING TO BECOME BOUND BY THE TERMS OF THIS
AGREEMENT, WHICH INCLUDES THE SOFTWARE LICENSE, LIMITED WARRANTY AND ACKNOWLEDGMENT.
IF YOU DO NOT AGREE TO THE TERMS OF THIS AGREEMENT, DO NOT OPEN THE DISK PACKAGE. PROMPTLY
RETURN THE UNOPENED DISK PACKAGE AND THE OTHER ITEMS (INCLUDING WRITTEN MATERIALS,
BINDERS, OR OTHER CONTAINERS, AND HARDWARE, IF ANY) WHICH ARE PART OF THIS PRODUCT.
In order to preserve and protect its rights under applicable laws, HGL does not sell any rights in HGL’s SOFTWARE. Rather,
HGL grants the rights to use HGL SOFTWARE by means of a SOFTWARE LICENSE. HGL specifically retains title and
copyrights to all HGL software.
HGL SOFTWARE LICENSE
1. GRANT OF LICENSE. In consideration of payment of the license fee which is a part of the price you paid for this
product, and your agreement to abide by the terms and conditions of this License and Limited Warranty, HGL, as Licensor, grants
to you, the Licensee, a nonexclusive right to use and display this copy of an HGL software program (hereinafter the
“SOFTWARE”) on a single COMPUTER (i.e., with a single CPU) at a single location so long as you comply with the terms of
this License. If the single computer on which you use the SOFTWARE is a multiuser system, the Licensee covers only one user
at a time on that single system. HGL reserves all rights not expressly granted to Licensee.
2. OWNERSHIP OF SOFTWARE. As the LICENSEE, you own the magnetic or other physical media on which the
SOFTWARE is originally or subsequently recorded or fixed, but an express condition of the License is that HGL retains title and
ownership of the SOFTWARE recorded on the original disk copy(ies) and all subsequent copies of the SOFTWARE, regardless
of the form or media in or on which the original and other copies may exist. This License is not a sale of the original
SOFTWARE or any copy.
3. COPY RESTRICTIONS. This SOFTWARE and the accompanying written materials are copyrighted. Unauthorized
copying of the SOFTWARE, including SOFTWARE which has been modified, merged, or included with other software, or of
the written materials is expressly forbidden. You may be held legally responsible for any copyright infringement which is caused
or encouraged by your failure to abide by the terms of this License.
Subject to the restrictions above, and if the SOFTWARE is not copy protected, you may make one (1) copy of the SOFTWARE
solely for backup purposes. You must reproduce and include the copyright notice on the backup copy.
4. USE RESTRICTIONS. As the LICENSEE, you may physically transfer the SOFTWARE from one computer to another
provided that the SOFTWARE is used on only one computer at a time and a written undertaking to that effect is provided by the
LICENSEE prior to such transfer. You may not electronically transfer the SOFTWARE from one computer to another over a
network. You may not distribute copies of the SOFTWARE or accompanying written materials to others. You may not modify,
adapt, translate, reverse, engineer, decompile, disassemble, or create derivative works based on the SOFTWARE, and/or modify,
adapt, translate, or create derivative works based on the written materials without the prior written consent of HGL. You may
not remove any proprietary notices, labels, or marks on the SOFTWARE or accompanying documentation.
5. TRANSFER RESTRICTIONS. This SOFTWARE is licensed only to you, the LICENSEE, and may not be transferred
to anyone without the prior written consent of HGL. Any transferee of the SOFTWARE shall be bound by the terms and
conditions of this License and Limited Warranty. In no event may you transfer, assign, rent, lease, sell or otherwise dispose of
the SOFTWARE on a temporary or permanent basis except as expressly provided herein.
6. TERMINATION. This License is effective until terminated. This License will terminate automatically without notice
from HGL if you fail to comply with any provision of this License. Upon termination you shall destroy the written materials and
all copies of the SOFTWARE, including modified copies, if any.
7. UPDATE POLICY. HGL may create, from time to time, updated versions of the SOFTWARE. At its option, HGL
will make such updates available to LICENSEE and transferees who have paid the update fee.
LIMITED WARRANTY
HGL warrants to their original LICENSEE that the disk(s) on which the SOFTWARE is recorded is free from defects in materials
and workmanship under normal use and service for a period of ninety (90) days from the date of purchase as evidenced by a copy
of the receipt. FURTHER, HGL HEREBY LIMITS THE DURATION OF ANY IMPLIED WARRANTY(IES) TO THE
RESPECTIVE PERIODS STATED ABOVE. SOME STATES DO NOT ALLOW LIMITATIONS ON DURATION OF AN
IMPLIED WARRANTY, SO THE ABOVE LIMITATION MAY NOT APPLY TO YOU.
HGL’s entire liability and your exclusive remedy as to the disks(s) and/or the recording of the SOFTWARE thereon and/or
accompanying written materials, shall be, at HGL’s option, either (a) return of the purchase price or (b) replacement of the disk
which does not meet HGL’s Limited Warranty and which is returned to HGL postage prepaid with a copy of the receipt. If failure
of the disk and/or SOFTWARE has resulted from accident, abuse or replacement, disk will be warranted for the remainder of
the original warranty period or thirty (30) days, whichever is the longer.
DISCLAIMER OF WARRANTY
THE SOFTWARE AND ACCOMPANYING WRITTEN MATERIALS (INCLUDING INSTRUCTIONS FOR USE) ARE
PROVIDED “AS IS” WITHOUT WARRANTY OF ANY KIND. FURTHER, HGL DOES NOT WARRANT, GUARANTEE
OR MAKE ANY REPRESENTATIONS REGARDING THE USE, OR THE RESULTS OF USE, OF THE SOFTWARE OR
WRITTEN MATERIALS IN TERMS OF CORRECTNESS, ACCURACY, RELIABILITY, CURRENTNESS, OR
OTHERWISE. THE ENTIRE RISK AS TO THE RESULTS AND PERFORMANCE OF THE SOFTWARE IS ASSUMED BY
YOU. IF THE SOFTWARE OR WRITTEN MATERIALS ARE DEFECTIVE YOU, AND NOT HGL OR ITS DEALERS,
DISTRIBUTORS, AGENTS, OR EMPLOYEES, ASSUME THE ENTIRE COST OF ALL NECESSARY SERVICING,
REPAIR, OR CORRECTION.
EXCEPT AS EXPLICITLY STATE HEREIN, HGL GIVES NO WARRANTIES OR CONDITIONS, WHETHER EXPRESSED,
IMPLIED, STATUTORY OR OTHERWISE WITH REGARD TO THE SOFTWARE, ACCOMPANYING
DOCUMENTATION, DISKS(S) ON WHICH THE SOFTWARE IS COPIED AND HGL SPECIFICALLY DISCLAIMS ANY
IMPLIED WARRANTIES OR CONDITIONS OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE.
NO ORAL OR WRITTEN INFORMATION OR ADVICE GIVEN BY HGL, ITS DEALERS, DISTRIBUTORS, AGENTS OR
EMPLOYEES SHALL CREATE A WARRANTY OR IN ANY WAY INCREASE THE SCOPE OF THIS WARRANTY AND
YOU MAY NOT RELY ON ANY SUCH INFORMATION OR ADVICE. THE WARRANTY GIVES YOU SPECIFIC LEGAL
RIGHTS. YOU MAY HAVE OTHER RIGHTS WHICH VARY FROM STATE TO STATE.
NEITHER HGL NO ANYONE ELSE WHO HAS BEEN INVOLVED IN THE CREATION, PRODUCTION OR DELIVERY
OF THIS PRODUCT SHALL BE LIABLE FOR ANY DIRECT, INDIRECT, CONSEQUENTIAL OR INCIDENTAL
DAMAGES (INCLUDING DAMAGES FOR LOSS OF BUSINESS PROFITS, BUSINESS INTERRUPTION, LOSS OF
BUSINESS INFORMATION, AND THE LIKE) ARISING OUT OF THE USE OR INABILITY TO USE SUCH PRODUCT
EVEN IF HGL HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES AND/OR COULD HAVE FORESEEN
THE POSSIBILITY OF SUCH DAMAGES.
The License and warranty shall be governed by the laws of the State of Virginia and shall be for the benefit of HGL, its successors
and assigns. The LICENSEE consents to exclusive jurisdiction in the state and federal courts of the State of Virginia.
ACKNOWLEDGMENT
BY OPENING THIS SEALED DISK OR TAPED PACKAGE YOU ACKNOWLEDGE THAT YOU HAVE READ THIS
LICENSE AND LIMITED WARRANTY, UNDERSTAND THEM AND AGREE TO BE BOUND BY THEIR TERMS AND
CONDITIONS. YOU ALSO AGREE THAT THE LICENSE AND LIMITED WARRANTY ARE THE COMPLETE AND
EXCLUSIVE STATEMENT OF AGREEMENT BETWEEN THE PARTIES AND SUPERSEDE ALL PROPOSALS OR PRIOR
AGREEMENTS ORAL OR WRITTEN, AND ANY OTHER COMMUNICATION BETWEEN THE PARTIES RELATING TO
THE SUBJECT MATTER OF THE LICENSE OR LIMITED WARRANTY.
Should you have any questions concerning this License or this Limited Warranty, or if you desire to contact HGL for any reason,
please contact in writing: MS-VMS Sales, HydroGeoLogic, Inc., 1155 Herndon Parkway, Suite 900, Herndon, VA 20170, USA.
MODFLOW-SURFACT SOFTWARE (VERSION 3.0)
OVERVIEW: INSTALLATION, REGISTRATION,
AND RUNNING PROCEDURES
BY:
HYDROGEOLOGIC, INC.
HERNDON, VA 20170, USA
TEL: (703) 478-5186
FAX: (703) 471-4180
E-MAIL: [email protected]
WEB PAGE: www.hgl.com
COPYRIGHT © 1996, BY HYDROGEOLOGIC, INC.

ALL RIGHTS RESERVED.
NO PART OF THIS PUBLICATION MAY BE REPRODUCED
OR TRANSMITTED IN ANY FORM OR BY ANY MEANS,
ELECTRONIC OR MECHANICAL, INCLUDING
PHOTOCOPY, RECORDING, OR ANY INFORMATION
STORAGE AND RETRIEVAL SYSTEM, WITHOUT
EXPRESS PERMISSION FROM HYDROGEOLOGIC, INC.
CONTENTS
Page
Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii
Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-1
1.1 Introduction And Version History . . . . . . . . . . . . . . . . . . . . . . . . . . 1-1
1.2 System Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-6
1.3 Software Acquisition and Support . . . . . . . . . . . . . . . . . . . . . . . . . . 1-6
1.4 The SURF Packages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-6
1.5 The ACT Packages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-8
1.6 Installation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-8
1.7 List of Files . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-9
1.8 Memory Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-10
1.9 Virtual Memory Use . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-10
1.9.1 Increasing Swap File Size Under Windows 3.x . . . . . . . . . . . . 1-10
1.9.2 Specifying a Swap File under DOS . . . . . . . . . . . . . . . . . . . . 1-10
1.10 Getting Started . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-11
1.10.1 Protection of MODFLOW-SURFACT . . . . . . . . . . . . . . . . . . 1-11
1.10.2 Running MODFLOW-SURFACT — BACKGROUND . . . . . . . 1-11
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-15
Figure 1 Sub-directory structure for example flow problems

provided by the installation procedure . . . . . . . . . . . . . . . . . . 1-9
Table 1 MODFLOW-SURFACT Input/Output Default File Extensions . . 1-13
Contents i
M:\Projects\MODFLOW Manuals\SURFACT and HMS Ver.3\Vol I for Surfact only\SURFACT overview_new.wpd
OVERVIEW: INSTALLATION, REGISTRATION,
AND RUNNING PROCEDURES
ABSTRACT
This document provides a brief overview of procedures for installation,

registration, and operation of MODFLOW-SURFACT Software (Version 3.0). It is
intended for users who acquire the MODFLOW-SURFACT code on a stand-alone basis.
Included in this report is a brief outline of the code’s capabilities, hardware requirements,
and running procedures under Microsoft DOS and Windows.
OVERVIEW
1.1 INTRODUCTION AND VERSION HISTORY
HydroGeoLogic, Inc., is pleased to present our MODFLOW-SURFACT Version
3.0, a fully integrated flow and transport code, based on the U.S. Geological Survey
(USGS) groundwater modeling software, MODFLOW (McDonald and Harbaugh, 1988).
MODFLOW-SURFACT constitutes additional modules to MODFLOW, to improve on its
robustness and increase its physical simulation capabilities to include complex saturated-
unsaturated subsurface flow analysis, and contaminant fate and transport calculations
within a comprehensive, well established simulation framework. The groundwater flow
and solute transport modules of MODFLOW-SURFACT are detailed in Volumes 1 and
2 of the code documentation. This document provides a brief overview of procedures for
code installation, registration, and operation. It is intended for users who acquire
MODFLOW-SURFACT on a stand-alone basis.
The first public release of MODFLOW-SURFACT was Version 1.2, in 1996.
The flow packages incorporated into Version 1.2 of MODFLOW-SURFACT extend the
applicability of the original MODFLOW code to handle complex field problems. These
packages (generally referred to as SURF packages) provide:
• Ability for MODFLOW to readily accommodate conditions of
desaturation/resaturation of aquifer systems.
• Ability for MODFLOW to perform axi-symmetric analysis.
• Improved ability for MODFLOW to simulate multi-layer wells.
• Additional robust numerical and matrix solution techniques for complex problems.
• Ability for MODFLOW to simulate prescribed-ponding-recharge and seepage-face
boundary conditions.
• Ability for MODFLOW to perform transient simulations efficiently using an
adaptive time-stepping scheme.
Version 2.1 was then released in 1998 which provides additional physical flow modeling
capabilities and robust numerical schemes which include:
• Rigorous saturated-unsaturated modeling of water flow.
• Rigorous saturated-unsaturated modeling of air flow.
• Newton-Raphson linearization for robustness.
Recognizing that few contaminant transport codes compatible with MODFLOW
are available, HydroGeoLogic has developed and integrated our new Analysis of
Contaminant Transport (ACT) modules into MODFLOW-SURFACT. The transport
modeling features of Version 1.2 include:
• Advective-dispersive transport of up to five species in the water.
• Linear/non-linear retardation.
• First-order radioactive decay and biochemical degradation in soil and/or water,
including generation of transformation products.
Overview 1-1
Version 2.1 provides additional physical transport modeling capabilities which
include:
• Mass partitioning into all phases of a saturated-unsaturated system with advective
+ dispersive transport in the active phase and diffusive transport in the inactive
phase of a two phase system.
• Mass partitioning from an immobile (single or multi-species) NAPL phase into the
other two phases (air and water) with transport of contaminant (by advection +
dispersion in the active phase, and diffusion in the inactive phase) and subsequent
depletion of the NAPL phase.
This document presents Version 3.0 of MODFLOW-SURFACT. Enhancements
to Version 2.2 include:
• Variable anisotropy option - This option allows for horizontal anisotropy to be
included on a cell-by-cell basis. Note that MODFLOW requires the anisotropy
to be uniform within each model layer.
• Vertical conductivity input option - With this option, the vertical conductivity of
grid blocks is input, and the saturated leakance between layers is computed
internally. Note that MODFLOW requires the saturated leakance between layers
as input.
• Other packages developed by the USGS for later versions of MODFLOW
including MODFLOW-96 and MODFLOW-2000 have also been incorporated into
this version of MODFLOW-SURFACT. Packages incorporated from
MODFLOW-96 include the Transient Leakage Package (TLK), the Direct Solver
Package (DE4), the Stream Routing Package (STR1), the Hydraulic Flow Barrier
Package (HFB), the Interbed Storage Package (IBS), the Constant (time-varying)
Head Package (CHD), the General Finite Difference Package (GFD), and the
Flow and Head Boundary (time-varying) Package (FHB). Packages incorporated
from MODFLOW 2000 include the Reservoir Package (RES1), the Steam Flow
Routing Package (SFR1), the Lake Package (LAK 3), the Gauge Package
(GAGE), and the Evapotranspiration Segments Package (EST1).
• The additional packages, LAK3, SFR1, GAGE, TLK, STR1, HFB, ETS1, CHD,
FHB, and RES1 are documented in their respective USGS documents. When
input for these packages has been altered in the MODFLOW-SURFACT
framework (ETS1, FHB), the input documentation for those is provided in this
document.
• When using the LAK3 package as implemented here, the value of LAYCON
should be set to the unconfined options (LAYCON = 3, 13, 23, 33, 43) for the
LAK 3 package to work correctly with the BCF Packages herein. A warning to
this effective is reflected in the output listing file if this condition is not met.
• Rewetting of the LAK3 package has been modified herein such that the water level
in the lake is set just slightly above the lake bottom upon rewetting. The original
algorithm of setting the lake level to the average value of adjacent groundwater
heads can create lake levels much higher than the lake bottom upon rewetting, thus
artificially creating significant amounts of water in the lake during the rewetting
step. The original LAK3 package is documented by Merritt and Konikow, 2000.
1-2 Overview
• The Gage package default output file names and extensions are fname.001,
frame.002, etc., up to the number of distinct files opened by the user for gage
information output, where frame is the default filename. Alternately, the NAME-
file convention may be used to open these files.
• The FHB package had been modified to include an option providing for step
functions of head or flow versus instead of interpolation between the tabular input
time values for the various FHB nodes.
• The Recharge Package (RSF4) has been modified to include zonal input of
recharge with zonal recharge values provided via a separate recharge time-series
(RTS) file, that is independent of the stress-period setup for varying boundary
conditions.
• The ET packages EVT and ETS1 (referred to as ET2 in MODFLOW-SURFACT)
have been modified to include zonal input of ETmax values provided via a separate
ET time-series (ETS) file, that is independent of the stress-period setup for
varying boundary conditions.
• Links have been provided to the SALTCHEM module of the UNSATCHEM
(Simunek et al, 1996) software package for including the ion equilibrium and
kinetic chemistry routines developed at the U.S. Salinity Laboratory of the
Agricultural Research Service. The SALTCHEM modules may be obtained as a
DLL from the ARS to be used with MODFLOW-SURFACT.
• An optional PCG5 package has been included, that has more options for
decomposition of the matrix and includes reduced system ordering schemes. This
solver is the MATIS matrix solver developed by Dr. Peter Forsyth at the
University of Waterloo, Ontario, Canada. It is a proprietary package distributed
optionally with MODFLOW-SURFACT with only the ORTHOMIN acceleration
scheme. This solver can be up to 20 times faster than the PCG4 solver due to its
advanced computational techniques.
• Curvilinear grid option - This option allows for non-rectangular grids in the areal
plane. Hence, DELR and DELC for each node within a layer is required as input,
with multiple layers stacked vertically below the first. Note that the original
MODFLOW grid was rectangular in the areal plane.
• Additional moisture retention input options for unsaturated zone flow simulations
including bi-modal relative permeability curves and options for tabulated input.
• For air phase flow simulations an additional input option for initial water heads
within the system is provided. Using this option, the water heads may be supplied
on a cell-by-cell basis, representing certain water flow conditions. The original
option only allowed for input of a water-table elevation across the domain which
was used to calculate head values for water assuming hydrostatic equilibrium in
the vertical direction.
• The LAK2 package, developed by Greg Council at GeoTrans in 1996 has been
implemented with permission from the author. Input instructions and
implementation details of this package are available from GeoTrans.
• An observation node package has been included to list the breakthrough of head
or contaminant concentrations through time, at any prescribed observation
location(s).
Overview 1-3
• MODFLOW96 style input is allowed by creating a NAME file which opens the
input files for all relevant packages of a simulation.
• Modifications for using MODPATH with the FWL4, FWL5 and ATO4 Packages.
This enhancement extracts time step information to a binary *.mph file (when
using the ATO4 package) for use with a modified version of MODPATH. The
original version of MODPATH utilizes only time step information from the OC
package and does not recognize FWL4 and FWL5 cell by cell output from
MODFLOW-SURFACT.
• Documentation of MODPATH is available in the original USGS report (Pollock,
1994). The current document discusses the enhancements and provides the
necessary input instructions for the modified MODPATH.
• Branched Chain Decay option - This option allows for branched chain decay
computations with appropriate stotiometry. The original option allowed for only
a straight chain decay with one unit of parent decaying to one unit of daughter
product.
• Transport simulations with the evapotranspiration (EVT) package of MODFLOW
is enhanced. The previous version allows for water to leave the system due to ET,
with contaminant mass being left behind. The enhancement includes providing
a root uptake factor that determines how much contaminant mass is removed with
the ET flux, thus enabling removal of contaminant due to ET, to varying degrees
prescribed by the user. This functionality has also been extended to the
evapotranspiration segments package included here from MODFLOW 2000.
• An option IUNCAD provided in the BTN file for transport simulations accounts
for adsorption occurring only in the saturated portion of an unconfined grid-block
for groundwater flow simulation cases.
• Mass balance printout can now be performed at user-defined intervals independent
of printing of information for heads and concentrations when the adaptive time-
stepping package is used for transient simulations. (See flags IBUDFL and
ICONBD for flow and transport mass balances respectively, in the AT04
package).
• Cell-by-cell mass balance printout option is provided (see flag IUNITMB in the
AT04 package) for absolute and normalized mass balances at each cell.
• An optional dual domain transport module has been included to provide dual
porosity simulation capabilities for aggregated or fractured settings.
• An option, NOMATRIX has been added to solution of the transport equation and
input via the BTN file. When turned off (NOMATRIX = 0), the matrix of
coefficients is written to disk for subsequent use when necessary. When
NOMATRIX =1, this matrix is not written to disk and is computed whenever
needed. Depending on CPU and disk speeds of the computing environment, one
may be faster than the other.
• A density-dependent transport module is available as an exclusive option, for
saltwater intrusion analyses in coastal systems or other settings where the flow-
field is influenced by the density of contaminants.
• The adaptive time-stepping package (ATO) has been modified to include options
for restart of a simulation from some in-between point of the dataset.
1-4 Overview
• The compressible storage term of the air flow equation has been corrected. The
term was expressed as:
which is corrected to be:
The effect of this change is only felt on transient air flow simulations. Further this
effect is significant only for simulations without significant forced air movement
where the incorrect version of the term allows for much more rapid equilibration
to steady-state, than the correct form.
• The optional flag SS/TR of MODFLOW 2000 is included to allow for
combinations of stead-state and transient simulations during any stress-period.
This is detailed in Section 3.4.2 of the Transport chapter which discusses input to
the BAS package.
• Residual saturations during air-phase flow simulations were not implemented
correctly. For non-uniform Sy arrays, the execution would terminate with a
‘square-root domain error’. This bug has been fixed.
• A non-linear reactions (RCT) package has been implemented within the transport
module, which is a multiphase extension of the reaction package of the RT3D
code. If dual porosity or dual domain transport is considered, the reactions may
occur within either or both domains, as selected by a flag in the RCT input file.
• The case of transport in a transient unsaturated flow-field wherein the specific
yield (Sy) values used for flow differ from the porosity (M) values used for
transport requires the storage term to be expanded. This is detailed in Section
2.2.1 of the Transport chapter, and options are provided for phase allocation of
the remaining pore space. In previous versions, this remaining pore space was
allocated to water in the saturated zone and to air in unsaturated portions of the
domain.
• Input for the concentration at prescribed head nodes in the HCN package was
being over-ridden by zero for certain cases. This bug has been fixed.
• A bug in the cross-dispersivity option (ICROSS=1) for transport simulations has
been fixed.
• Additional input data consistency checks have been added.
All modules of MODFLOW-SURFACT are fully compatible, and a mix-and-match of
modules from the original MODFLOW, and from MODFLOW-SURFACT may be used
in a single simulation. The transport capabilities are also fully flexible with complexity
being added to the system as needed. Thus, a transport simulation can include the highest
level of complexity of multi-phase, multicomponent transport of reactive substances in a
dual porosity setting.
Overview 1-5
The modular structure of MODFLOW is maintained, and modules for the
enhanced features are called from the main program. New formulations and computational
schemes implemented into the additional packages have been verified using analytical
solutions and other numerical models. A complete compatibility is maintained between
the additional packages of MODFLOW-SURFACT and the original MODFLOW code.
Input preparation for MODFLOW-SURFACT is straightforward and follows the
MODFLOW format structure. Input data for flow modeling is essentially that required
by the original MODFLOW, and minimal additional data is required for the
enhancements. Transport simulations make use of all MODFLOW data sets, and
duplication of the supplied input information is eliminated. Transport parameters are
supplied to the code via an additional data set, and boundary conditions are implemented
through the MODFLOW boundary data sets.
1.2 SYSTEM REQUIREMENTS

MODFLOW-SURFACT is compiled specifically for an 80386/80486-processor
based microcomputer. It will not run on an IBM-AT or IBM-XT or compatible. The full
system requirements are summarized as follows:
• Pentium CPU
• at least 64 megabytes (Mb) of Extended Memory
• Windows 98 or higher.
1.3 SOFTWARE ACQUISITION AND SUPPORT
MODFLOW-SURFACT can also be compiled for Weitek 1167, 3167 (Abacus),
or 4167 math coprocessor chips. To obtain a Weitek version of MODFLOW-SURFACT,
or versions compiled for other operating platforms, please contact:
HydroGeoLogic Software Sales
1165 Herndon Parkway, Suite 900
Herndon, VA 22070, USA
Phone: (703) 478-5186
Fax: (703) 471-4180
E-Mail: [email protected]
On-line technical support and customized training on usage of MODFLOW-SURFACT
are also available. Please call for further details.
1.4 THE SURF PACKAGES

MODFLOW-SURFACT contains several new flow packages which
complement/supplement the original version of MODFLOW. These packages, generally
referred to as SURF packages, include:
• BCF4 This is the new block-centered flow (BCF) package which replaces
the earlier BCF packages in MODFLOW. This package incorporates
an axi-symmetric simulation option and a variably saturated
simulation option with pseudo-soil water retention functions to enable
robust simulations of desaturation/resaturation cases. This package
includes the capabilities of earlier BCF packages developed by the
USGS to maintain backward compatibility. Thus, the schemes of the
original BCF package of McDonald and Harbaugh (1988), the BCF2
package developed by McDonald, Harbaugh, Orr and Ackerman
(1991), and the BCF3 package developed by Goode and Appel (1992)
are included in the BCF4 package. Version 2.1 incorporates
1-6 Overview
additional simulation capabilities into the BCF4 package to create the
BCF5 package. These include unsaturated zone moisture movement
simulation capabilities using Richards Equation, and simulation of air
movement in a saturated-unsaturated system, for air sparging or
vacuum extraction analyses.
• FWL4 This is the fracture-well package which enhances MODFLOW’s
capability of modeling wells, and is added to existing packages.
• RSF4 This is the recharge-seepage face boundary condition package that
enables simulations of seepage faces and recharge to low-conductivity
unconfined systems.
• ATO4 This is an adaptive time stepping and output control package which
can supersede MODFLOW’s time stepping routines and output
control.
• PCG4 This is a preconditioned conjugate gradient (PCG) solver package
which may be used as an alternative to other solver packages
developed for MODFLOW. The PCG4 package is a robust
alternative to the previous PCG1 package developed by Hill (1994).
• NRB1 This is a Newtown-Raphson linearization package with a backtracking
scheme for step control. This package can greatly improve robustness
of simulations for unconfined or unsaturated fluid flow cases.
• OBS This is the observation node package which provides time varying
information at specified observation nodes. Nodal heads are output
with this package for flow simulation and contaminant concentrations
for transport simulations.
The above packages have been developed by HydroGeoLogic, Inc., to provide
MODFLOW with extended capabilities. Other packages, originally developed by the
USGS to extend the capabilities of MODFLOW, are also available. These packages are
also incorporated in MODFLOW-SURFACT and include the following:
• STR Streamflow-Routing Package developed by Prudic (1989).
• IBS The Interbed-Storage Package developed by Leake and Prudic (1991).
The IBS Package allows MODFLOW to simulate the effects of
aquifer compaction due to dewatering. The IBS Package also contains
the CHD (time-variant specified-head) Package which allows constant
head cells to vary with time. This overcomes one limitation of the
original version of MODFLOW.
• HFB Horizontal Flow Barrier developed by Hsieh and Freckleton (1993).
HFBs are used to simulate the effects of low-permeability fault zones
or slurry walls.
• FHB Flow and Head Boundary Package documented in Leake and Lilly
(1997). The FHB Package allows for implementation of time-varying
flux and head boundary conditions, regardless of the stress period
setup of a simulation.
• SFR Stream-Flow Routing Package documented in Prudic et al. (2004).
The SFR package allows streamflow routing and includes links to the
Lake (LAK3) and Gage (GAGE) packages.
Overview 1-7
• ETS Evapotranspiration Segments Package documented in Banta (2000).
The ETS package allows for segmented input of the ET function thus
allowing for a generalized function of ET flux with depth.
• LAK3 The Lake Package documented in Merrett and Konikow (2000). The
LAK3 package represents lakes interacting with the groundwater and
includes the GAGE package and links to the SFR package.
The user may incorporate his or her own custom packages into MODFLOW-
SURFACT. For customization of MODFLOW-SURFACT, please contact the software
sales department as directed in Section 1.3 of this overview.
1.5 THE ACT PACKAGES

MODFLOW-SURFACT contains the ACT modules that provide the capability to
perform single-species and multi-component transport analyses. The ACT modules consist
of transport computational subroutines compatible with the flow modules as well as
packages for transport simulation input and output.
• BTN1 This is the Basic Transport package which reads the basic model
parameters (e.g., dispersivities, retardation coefficient and
degradation rates) required for the transport simulation, and
performs contaminant transport analyses for single component or
multicomponent contaminants, with chain decay. In addition,
contaminants may transfer mass into all fluid phases present in the
system including an immobile NAPL phase which can deplete due
to contaminant leaching, or increase in saturation as a result of
complex chain decay reactions.
• PCN1 This is the Prescribed Concentration boundary condition input
package which reads first-type (prescribed concentration)
boundary condition data.
• HCN1 This is the Prescribed Head-Concentration boundary condition
input package which reads the boundary condition data for those
cells where the hydraulic head is prescribed in the respective flow
simulation.
1.6 INSTALLATION
MODFLOW-SURFACT should be installed under the Windows operating system.
The installation process is very similar to many other Windows applications. To install,
follow the steps described below.
1. Insert Disk 1 of the MODFLOW-SURFACT distribution disks into one of your
CD drives.
2. From the Program Manager window, click on the File item from the horizontal
menu underneath the window title.
3. Choose the Run item from this menu. You will be presented with a dialog box
entitled Run.
4. There will be a text entry box with the heading Command Line:. If the CD drive
you are using is D, for example, then enter the following:
D:SETUP
and press ENTER.
1-8 Overview
5. The SETUP program will ask you to select the Hard Disk Drive and the directory
name where the MODFLOW-SURFACT files will be stored. The default is
C:\MODFLOW_SURFACT. Either enter a new path or click on Start install
to continue with the installation in the default directory. The SETUP program will
check your system to make sure that you have enough disk space.
6. After creating the necessary directories and expanding the files, SETUP will
provide an option to update your "AUTOEXEC.BAT" file. If this option is
selected, SETUP will add the directory name to your PATH statement so that you
may run MODFLOW_SURFACT from any directory on your computer. If you
do not want SETUP to modify your AUTOEXEC.BAT file, you should modify
the PATH yourself.
7. The last step involved in the installation of MODFLOW-SURFACT is rebooting
your computer so that the changes made to the "AUTOEXEC.BAT" file become
effective. After rebooting your computer, start Windows and you will find that
SETUP has created a MODFLOW-SURFACT program folder.
1.7 LIST OF FILES

The MODFLOW-SURFACT distribution CD contains the
MODFLOW-SURFACT program executables and sample input data files for some of the
problems documented in the code documentation. Other example data files can also be
requested from HydroGeoLogic, Inc. The SETUP program will create the necessary
directories on your hard disk and will install the MODFLOW-SURFACT executables and
associated files. In addition to the input data files, each sub-directory will contain a
“readme” file explaining the contents of the directory. The structure of sub-directories
containing example flow and transport problems is shown in Figure 1.1.
EXAMPLES
*
*——VOL_1
* *
* *——SEC2_4_1
* * *
* * *—CASE_A
* * *
* * *—CASE_B
* *
* *——SEC3_4_1
* *
* *——SEC4_3_2
* *
* *——SEC5_3
Fig 1. Sub-directory structure for example problems provided by

the installation procedure.
Overview 1-9
1.8 MEMORY MANAGEMENT
MODFLOW-SURFACT allocates the memory required by data arrays and lists
at execution time in a one-dimensional array called the “X” array. The total memory
space needed by the “X” array depends on the number of model grid cells and on the type
and number of packages required in a simulation. In the original MODFLOW code, the
dimension of the “X” array is declared in a static manner, i.e., at compile time.
Therefore, if the simulation problem required additional memory the user would have to
modify the Fortran source code and recompile the code with increased “X” array
dimension. This limitation has been overcome in MODFLOW-SURFACT by using
dynamically allocatable arrays. MODFLOW-SURFACT allocates and de-allocates
memory dynamically during the execution time based on the memory requirements of the
problem. This ensures the most efficient use of computer memory and allows simulation
problems of virtually unlimited size to be handled without code modification, subject only
to available computer memory.
1.9 VIRTUAL MEMORY USE

If enough random access memory (RAM) is unavailable, the user can specify a
swap file to provide virtual memory. Brief instructions to create and define a swap file
and its size are provided below. For additional information, the user should refer to
corresponding Operating System Manuals and/or Compaq Visual Fortran V.6.1 Manuals.
1.9.1 Increasing Swap File Size Under Windows 3.x
The Microsoft Windows operating system uses a default swap file. If additional
memory is required, the swap file size can be increased by following the instructions given
below:
1. Double Click on the “Main” icon in the “Program Manager”
2. Double Click on the “Control Panel” icon in the “Main” program group
3. Double Click “386 Enhanced” icon in the “Control Panel”
4. In the “386 Enhanced” window click on the “Virtual Memory” button
5. Click on the “Change>>” button in the “Virtual Memory” window
6. Increase the size of the swap file in the appropriate field in the “Change>>”
Window
1.9.2 Specifying a Swap File under DOS
Under the DOS operating system a swap file can be setup by setting the
DOS4GVM environment variable. To set the environment variable, enter the following
at the command prompt:
set dos4gvm=deleteswap virtualsize #value
and press ENTER.
Note that virtualsize #value is the swap file size in Kilo-bytes. As an example, to
create an 8-megabyte swap file, the second argument would be “virtualsize #8195”. To
set the environment variable permanently, incorporate the above command in your
AUTOEXEC.BAT file.
1-10 Overview
1.10 GETTING STARTED
The following instructions give you a step-by-step introduction to running
MODFLOW-SURFACT. By following these instructions you will be able to:
1. Use the MODFLOW-SURFACT hard-lock protection system.
2. Understand how to use MODFLOW-SURFACT.
1.10.1 Protection of MODFLOW-SURFACT
A hard-lock is included in your MODFLOW-SURFACT package. Before you
can use MODFLOW-SURFACT, you must put the hard-lock in the USB port on your
computer.
If you choose to uninstall MODFLOW-SURFACT from one machine and install
it on another, the hard-lock is easily moved from one computer to another.
1.10.2 Running MODFLOW-SURFACT — BACKGROUND
MODFLOW-SURFACT can be executed in two different modes depending on the
supplied command line argument. The following modes are available:
Argument Description
1 BAS file format
3 NAME file convention for opening input and output files of a simulation.
In the two modes, MODFLOW-SURFACT will read the Basic (BAS) Package file
name (for argument “1”) as a second command line argument and assumes the default
filenames for the remaining files, or the entire filename of the NAME file (for argument
= “3”) which opens all other input and output files required by the simulation. The
default naming convention uses the file extension to distinguish each of the various input
files used in the simulation. All files, however, share the same name. The default file
naming convention used by MODFLOW-SURFACT is shown in Table 1. It is required
that all the input data files are located in the same directory and have been given the
appropriate default names.
If no command line arguments are provided, MODFLOW-SURFACT will require
a file called “file.id”. This file provides the above mentioned options (Modes 1 or 3) and
the file names (BAS or NAM) needed for the simulation. Thus, a user may use a “file.id”
file that provides one of the two modes of operation or provides the mode information
using command line arguments using DOS as described below.
(a) Using “file.id”
Under Windows, MODFLOW-SURFACT uses the “file.id” file to determine the
mode of operation. For option 1 the “file.id” file contains the following text:
1
drive:\directory\basic_package_file_name
For option 3, the “file.id” file contains the following text:
3
drive:\directory\NAME_file_name
Overview 1-11
(b) Using command line options using DOS
Under DOS, the user can run MODFLOW-SURFACT in any mode described
previously, for example:
To execute MODFLOW-SURFACT with Option 1, type the following at the
command prompt:
MSFT 1 drive:\directory\basic_package_file_name
To execute MODFLOW-SURFACT with Option 3 type the following at the
command prompt.
MSFT 3 drive:\directory\NAME_file_name
where drive:\directory\basic_package_file_name represents the complete path and name
of the basic package input file, using the default naming convention shown in Table 1.
For example:
MSFT 1 c:\MSVMS\EXAMPLES\VOL_1\SEC3_4_1\PRB_1.BAS
will run the problem with default filename PRB_1.
An example of using the NAME file convention (similar to the method of opening
files in MODFLOW 96) is:
MSFT 3 c:\MSVMS\EXAMPLES\GV42\GV42.NAM
where the name-file, GV42.NAM appropriately opens all input and output files needed for
the simulation. As an example, the NAME file may contain the following lines:
LIST 6 gv42.out
BAS 1 gv42.bas
BCF 11 gv42.bcf
GHB 17 gv42.ghb
RCH 18 gv42.rch
SIP 19 gv42.sip
OC 22 gv42.oc
PG4 23 gv42.pcg
ATO 32 gv42.ato
DATA 33 Lay1.Kx
DATA (BINARY) 34 gv42.hds
DATA (BINARY) 35 gv42.rar
where the first entry in each row denotes the type of file being opened, the second entry
is the FORTRAN unit number on which this file is attached, and the last entry denotes the
appropriate filename. The type of file being opened is denoted by LIST for the output
listing file; by the appropriate two or three letter acronym for each of the packages (as
noted in Table 1); or by the word DATA or DATA (BINARY) for other files that contain
ASCII or binary arrays. For instance, in the above example, the hydraulic conductivity
array for layer 1 is read from a separate file called Lay1.Kx which is opened on Fortran
Unit 33. Note that Table 1 entries are ordered in the same manner as entries in the IUNIT
array of the BAS package.
1-12 Overview
Table 1. MODFLOW-SURFACT Input/Output Default File Extensions
CELL-BY-CELL
OUTPUT NAME-
FILE file
PACKAGE NAME FLOW TRANSPORT acronym
Basic Flow (BAS) *.BAS -- -- BAS
Block-Centered Flow (BCF4) *.BCF *.CBB -- BCF
Well (WEL) *.WEL *.CBW *.TBW WEL
Drain (DRN) *.DRN *.CBD *.TBD DRN
River (RIV) *.RIV *.CRV *.TRV RIV
Evapotranspiration (EVT) *.EVT *.CBE -- EVT
Transient Leakage (TLK) *.TLK CBT -- TLK
General-Head Boundary (GHB) *.GHB *.CBG *.TBG GHB
Recharge-Seepage Face Boundary *.RCH *.CRC *.TRC RCH
(RSF4)
Strongly Implicit Procedure (SIP) *.SIP -- -- SIP
Direct Solver (DE4) *.DE4 -- -- DE4
Slice-Successive Overrelaxation *.SOR -- -- SOR
(SSOR)
Output Control (OC) *.OC -- -- OC
Preconditioned Conjugate *.PCG -- -- PG4
Gradient (PCG4)
Streamflow-Routing (STR) *.STR *.CS1 -- STR
*.CS2
Preconditioned Conjugate *.PG2 -- -- PG2
Gradient 2 (PCG2)
Horizontal-Flow Barrier (HFB) *.HFB -- -- HFB
Basic Transport (BTN1) *.BTN -- *.TBT BTN
Lake 2 Package by GeoTrans *.LAK CLK -- LAK
(LAK)
Interbed-Storage (IBS) *.IBS *.CBI -- IBS
Time-Variant Specific-Head *.CHD -- -- CHD
(CHD)
Fracture Well (FWL4) *.FWL *.CFW *.TFW FWL
Adaptive Time-Stepping and *.ATO -- -- ATO
Output Control (ATO4)
Prescribed Concentration *.PCN -- *.TPC PCN
Boundary (PCN1)
Prescribed-Head-Concentration *.HCN -- *.THC HCN
Boundary (HCN1)
Overland Flow (OLF)1 *.OLF *.OL1 *.OL2 OLF
1
Channel Flow (CHF) *.CHF *.CH1 *.CH2 CHF
Overview 1-13
CELL-BY-CELL
OUTPUT NAME-
FILE file
PACKAGE NAME FLOW TRANSPORT acronym
1
Analytical Infiltration (PRL) *.PRL *.PR1 -- PRL
Observation Node (OBS) *.OBS *.OBW Channel flax on OBS
*.OBV
Generalized Finite Differences *.GFD *.CGF -- GFD
(GFD)
Flow and Head (time-varying) *.FHB *.CFB -- FHB
Boundary (FHB)
Interception of Precipitation and *.IPT *.ESI -- IPT
Evapotraspiration (IPT)
Reactive Transport (RCT) *.RCT -- TRC RCT
Saltchem Module (SLC) *.SLC -- TSL SLC
2
Link-AMG (LMG) *.LMG -- -- LMG
Preconditioned Conjugate *.PG5 -- -- PG5
Gradient 5 (PCG5)
Reservoir (RES) *.RES *.CRS -- RES
Multi-node Fracture Well *.WL5 *.CW5 *.TW5 WL5
(FWEL5)
Stream-Flow Routing (SFRI) *.SFR *.CSF -- SFR
Lake 3 Package of USGS (LAK3) *.LK3 *.CL3 -- LAK3
Gage Package (GAGE) *.GAG -- -- GAG
Evapotranspiration Segments *.ET2 *.CEZ *.TEZ ET2
Package (ETS1)
Main Output *.OUT -- --
Head Output *.HDS -- --
Drawdown Output (Saturation *.DDN -- --
output for unsaturated fluid flow
simulations)
Concentration Output *.CON -- --
Default 2-D Real array Input *.2DR -- --
Default 2-D integer array input *.2DI -- --
* respresents the root file name of the basic package. The same name is used with different file extensions for all the
remaining input/output files when MODFLOW-SURFACT is invoked in Non-Interactive mode.
– indicates that the corresponding file is not created.
1
Not available in MODFLOW-SURFACT.
2
The AMG solver is not available.
1-14 Overview
REFERENCES
Banta, E.R., 2000. MODFLOW-2000, the U.S. Geological Survey modular groundwater
model - Documentation of packages for simulating evapotranspiration with a
segmented function (ETS1) and drains with return flow (DRT1), U.S. Geological
Survey Open-File Report 00-466, 127 p.
Goode, D.J. and Appel, CA, 1992. Finite-difference interblock transmissivity for
unconfined aquifers and for aquifers having smoothly varying transmissivity: U.S.
Geological Survey Water-Resources Investigations Report 92-4124, 79 pp.
Hshieh, P.A. and Freckleton, J.R., 1993. Documentation of a Computer Program to
Simulate Horizontal-Flow Barriers Using the U.S. Geological Survey’s Modular
Three-Dimensional Finite-Difference Ground-Water Flow Model, U.S. Geological
Survey, Open-File Report 92-477, 32 pp.
Hill, M.C., 1994. Preconditioned Conjugate-Gradient 2 (PCG2), a Computer Program
for Solving Groundwater Flow Equations, U.S. Geological Survey Water
Resources Investigations Report 90-4048, 43 pp.
HydroGeoLogic, Inc., 1996. MODFLOW-SURF (Version 1.2). The HydroGeoLogic
Enhanced Version of the U.S. Geological Modular Groundwater Flow Model,
Code Documentation Report.
Leake, S.A., and M.R. Lilly, 1997. Documentation of a computer program (FHB1) for
assessment of transient specified flow and specified head boundaries in
applications of the modular finite-difference ground-water flow model
(MODFLOW), U.S. Geological Survey Open-File Report 97-571, 50 p.
Leake, S.A., and Prudic, D.E., 1991. Documentation of a Computer Program to
Simulate Aquifer-System Compaction Using the Modular Finite-Difference
Ground-Water Flow Model, U.S. Geological Survey, Book 6, Chapter A2, 68 pp.
McDonald, M.G., and Harbaugh, A.W., 1988. A Modular Three-Dimensional Finite-
Difference Ground-Water Flow Model: U.S. Geological Survey Techniques of
Water-Resources Investigations Book 6, Chapter A1.
McDonald, M.G., Harbaugh, A.W., Orr, B.R., and Ackerman, D.J., 1991. A method
of converting no-flow cells to variable-head cells for the U.S. Geological Survey
Modular Finite-Difference Ground-Water Flow Model: U.S. Geological Survey
Open-File Report 91-536, 99 pp.
Merrett, L.M., and L.F. Konikow, 2000. Documentation of a computer program to
simulate lake-aquifer interaction using the MODFLOW groundwater flow model
and the MOC3D solute transport model, U.S. Geological Survey Water-Resources
Investigations Report 00-4167, 146p.
Prudic, D.E., L.F. Konikow and E.R. Banta, 2004. A new stream-flow routing package
to simulate stream-aquifer interaction with MODFLOW 2000, U.S. Geological
Survey Open-File Report 2004-1042, 95 p.
Prudic, D.E., 1989. Documentation of a Computer Program to Simulate Stream-Aquifer
Relations Using a Modular Finite-Difference Ground-Water Flow Model, U.S.
Geological Survey, Open-File Report 88-729, 113 pp.
References 1-15
Simunek, J., D.L. Suarez, and M. Sejna, 1996. UNSATCHEM Software Package for
Simulating the One-Dimensional Variability Saturated Water Flow, Heat
Transport, Carbon Dioxide Production and Transport, and Multi-Component
Solute Transport with Major Ion Equilibrium and Kinetic Chemistry, Version 2.0,
U.S. Salinity Laboratory, ARS, U.S. Department of Agriculture, Riverside, CA,
Research Report No. 141.
1-16 References
DOCUMENTATION
VOLUME I: FLOW MODULES
BY:
HYDROGEOLOGIC, INC.
TEL: (703) 478-5186
FAX: (703) 471-4180

NO PART OF THIS PUBLICATION MAY BE
REPRODUCED OR TRANSMITTED IN ANY FORM OR
BY ANY MEANS, ELECTRONIC OR MECHANICAL,
INCLUDING PHOTOCOPY, RECORDING, OR ANY
INFORMATION STORAGE AND RETRIEVAL SYSTEM,
WITHOUT EXPRESS PERMISSION FROM
HYDROGEOLOGIC, INC.
CONTENTS - VOLUME I: FLOW MODULES
Page
ABSTRACT
CHAPTER 1: INTRODUCTION
1.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-1
1.2 Attributes of New Flow Packages in MODFLOW-SURFACT . . . . . . . . 1-2
1.3 MODFLOW-SURFACT Operation and Input Options . . . . . . . . . . . . . 1-3
1.4 Document Organization and Usage Guide . . . . . . . . . . . . . . . . . . . . . 1-4
CHAPTER 2: THE BLOCK-CENTERED FLOW (BCF4) PACKAGE

2.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-1
2.2A Formulation of the Unconfined Flow Equation Using a Variably Saturated
Flow Formulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-2
2.2B Formulation of Variably Saturated Water Flow Equations . . . . . . . . . . . 2-4
2.2C Formulation of Variably Saturated Air Flow Equations . . . . . . . . . . . . 2-5
2.3A Verification Example for Unconfined Flow . . . . . . . . . . . . . . . . . . . . 2-9
2.3B Verification Example for Unsaturated Flow of Water . . . . . . . . . . . . . . 2-10
2.3B.1 Problem 1 - Transient Infiltration in an Unsaturated Vertical Soil
Column . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-10
2.3B.2 Problem 2 - Two-dimensional Flow in an Unsaturated
Rectangular Soil Slab . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-12
2.3C Verification Example for Air-Phase Flow Simulation . . . . . . . . . . . . . . 2-12
2.4 Application Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-16
2.4.1 Problem 1—Steady-State Simulation of Pumping in an Unconfined
Aquifer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-16
2.4.2 Problem 2—Transient Simulation of Pumping and Recovery in an
Unconfined Aquifer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-19
2.5 Other Extensions to the Block-Centered Flow Package . . . . . . . . . . . . . 2-30
2.6 Input Instructions for the BCF4 Package . . . . . . . . . . . . . . . . . . . . . . 2-32
2.6.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-32
2.6.2 BCF4 Input . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-32
2.6.3 Explanation of Fields Used in Input Instructions . . . . . . . . . . . . 2-37
CHAPTER 3: THE FRACTURE-WELL (FWL4) PACKAGE

3.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-1
3.2 Formulation of Fracture-Well Equations . . . . . . . . . . . . . . . . . . . . . . 3-2
3.3 Verification Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-4
3.4 Application Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-6
3.4.1 Problem 1—Fully Three-Dimensional Simulation of Withdrawal
from an Unconfined Aquifer . . . . . . . . . . . . . . . . . . . . . . . . 3-6
3.4.2 Problem 2—Quasi Three-Dimensional Simulation of Withdrawal
from a Three-Aquifer System . . . . . . . . . . . . . . . . . . . . . . . . 3-9
3.5 Input Instructions for the FWL4 Package . . . . . . . . . . . . . . . . . . . . . 3-21
3.5.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-21
3.5.2 FWL4 Input . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-21
Contents i
M:\Projects\MODFLOW Manuals\SURFACT and HMS Ver.3\Vol I for Surfact only\Vol I - Chapter 1 and TOC - SURFACT.wpd
CHAPTER 4: THE RECHARGE-SEEPAGE FACE BOUNDARY (RSF4) CONDITION
PACKAGE
4.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-1
4.2 Formulation of the RSF4 Boundary Condition Package . . . . . . . . . . . . . 4-1
4.3 Rapid, Transient Recharge Conditions . . . . . . . . . . . . . . . . . . . . . . . . 4-2
Verification and Application Examples . . . . . . . . . . . . . . . . . . . . . . . . 4-2
4.4 Verification and Application Examples . . . . . . . . . . . . . . . . . . . . . . . . 4-3
4.4.1 Problem 1—Flow to Parallel Drains in an Unconfined Aquifer . . . 4-3
4.4.2 Problem 2—Seepage Through a Square Embankment . . . . . . . . . 4-3
4.5 Input Instructions for the Rsf4 Package . . . . . . . . . . . . . . . . . . . . . . . 4-9
4.5.1 RSF4 Input . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-9
4.5.2 Transient Recharge Time-Series File . . . . . . . . . . . . . . . . . . . . 4-12
CHAPTER 5: THE ADAPTIVE TIME-STEPPING AND OUTPUT CONTROL (ATO4)

PACKAGE
5.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-1
5.2 Formulation of Adaptive Time-Stepping Used in the ATO4 Package . . . . 5-2
5.3 Verification Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-2
5.4 Input Instructions for the ATO4 Package . . . . . . . . . . . . . . . . . . . . . 5-10
5.4.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-10
5.4.2 ATO4 Input . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-10
5.4.3 Explanation of Fields Used in Input Instructions . . . . . . . . . . . 5-11
CHAPTER 6: THE PRECONDITIONED CONJUGATE GRADIENT (PCG4) PACKAGE

6.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-1
6.2 Formulation of Standard PCG Scheme . . . . . . . . . . . . . . . . . . . . . . . . 6-1
6.3 Input Instructions for the PCG4 Package . . . . . . . . . . . . . . . . . . . . . . 6-2
6.3.1 PCG4 Input . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-2
CHAPTER 7: THE NEWTON-RAPHSON LINEARIZATION WITH BACKTRACKING

NRB1 PACKAGE
7.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-1
7.2 Formulation of Newton Raphson Scheme . . . . . . . . . . . . . . . . . . . . . . 7-1
7.3 Input Instructions for the NRB1 Package . . . . . . . . . . . . . . . . . . . . . . 7-3
CHAPTER 8: OBSERVATION NODE (OBS1) PACKAGE

8.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-1
8.2 Input Instructions for the OBS1 Package . . . . . . . . . . . . . . . . . . . . . . . 8-1
8.2.1 OBS1 Input . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-1
8.3 Output of OBS1 Package . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-2
CHAPTER 9: REFERENCES
ii Contents
FIGURES
Page
Figure 2.1 Dimensionless Drawdown Versus Dimensionless Time
Relationships Showing Comparison of Numerical and Analytical
Solutions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-10
Figure 2.B1 Hydraulic Head Profile at 0.4 and 0.8 days for Transient
Infiltration in an Unsaturated Vertical Soil Column. . . . . . . . . . 2-12
Figure 2.B2 Hydraulic Head Profiles at .508d for Flow in an Unsaturated
Rectangular Soil Slab. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-13
Figure 2.C1 Steady-State and Transient Distribution of Air Pressure Near
Sparging Well for Air-Phase Flow Simulation.. . . . . . . . . . . . . 2-16
Figure 2.2 A Schematic Diagram of the Unconfined Aquifer System. . . . . . 2-17
Figure 2.3 Steady-State Head Distributions in Layer 1 - Comparison between
Case (a) and Case (b). . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-19
Case (a) and Case (b). . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-20
Case (a) and Case (b). . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-21
Figure 2.6 Applied Pumping Stress at Each Well for Transient Analysis of
the Unconfined Aquifer System. . . . . . . . . . . . . . . . . . . . . . . 2-23
Figure 2.7 Head Distributions in Layer 1 at the End of Pumping (First)
Period (t=4 yr) Showing Comparison between Previous and New
Simulation Options. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-24
Figure 2.8a Head Distribution in Layer 3 at the End of Pumping (First)
Period (t=4 yr) Obtained Using the Previous MODFLOW
Option. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-25
Figure 2.8b Head Distribution in Layer 3 at the End of Pumping (First)
Period (t=4 yr) Obtained Using the New Variably Saturated Flow
Option in MODFLOW-SURFACT. . . . . . . . . . . . . . . . . . . . . 2-26
Figure 2.9 Head Distribution in Layer 1 at the End of Recovery (Second)
Period (t=16 yr) Showing Comparison Between the Previous and
New Simulation Options. . . . . . . . . . . . . . . . . . . . . . . . . . . 2-27
Figure 2.10 Head Distributions in Layer 3 at the End of Recovery (Second)
Period (t=16 yr) Showing Comparison between the Previous and
New Simulation Options. . . . . . . . . . . . . . . . . . . . . . . . . . . 2-28
Figure 2.11 Temporal Variation in Head at Three Observation Points Showing
Comparison between the Previous and the New Simulation
Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-30
Comparison between the New Simulation Option in MODFLOW-
SURFACT and STAFF3D. . . . . . . . . . . . . . . . . . . . . . . . . . 2-31
Figure 3.1 Physical Situation (a) and Conceptualization (b) of a Fracture
Well . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-3
Contents iii
Figure 3.2 Verification of the Numerical Solution (FWL4 Package) with the
Analytical Solutions of Theis (1935) and Papadopulos and Cooper
(1967). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-5
Figure 3.3 A Schematic Diagram of the Unconfined Aquifer System for
Fully 3-D Simulation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-7
Figure 3.4 Pumping Operation in the Well for a Fully 3-D Simulation. . . . . . 3-8
Figure 3.5 Temporal Variation of Water Level in the Well for a Fully 3-D
Simulation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-10
Figure 3.6 Temporal Variation of Total Well Flux for a Fully 3-D
Simulation: Comparison Between Prescribed Flux (FWL4
Package) and Prescribed Head (GHB Package) Simulation
Results. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-11
Figure 3.7 A Schematic Diagram of the Three-Aquifer System for a Quasi 3-
D Simulation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-12
Figure 3.8 Pumping Operation in the Well for a Quasi 3-D Simulation. . . . . 3-13
Figure 3.9 Temporal Variation of Head at Well Nodes in a Quasi 3-D
Simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-15
Figure 3.10 Temporal Variation of Well Fluxes of Aquifers for a Quasi 3-D
Simulation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-16
Figure 3.11 Vertical Profiles of Heads in the Aquifers in the First Stress
Period at t=2 yr. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-17
Figure 3.12 Vertical Profiles of Heads in the Aquifers at the End of First
Stress Period (t=3 yr). . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-18
Figure 3.13 Vertical Profiles of Heads in the Aquifers in the Second Stress
Period at t=4.5 yr. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-19
Figure 3.14 Vertical Profiles of Heads in the Aquifers in the Second Stress
Period at t=7.5 yr. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-20
Figure 4.1 Schematic Diagram Depicting Flow to Parallel Drains in an
Unconfined Aquifer. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-3
Figure 4.2 Discretization Scheme Used for the Simulation of Flow through
Parallel Drains. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-4
Figure 4.3 Comparison of Numerical and Analytical Solutions for the
Parallel Drain Problem (Recharge = 0.01 ft/d). . . . . . . . . . . . . 4-5
Figure 4.4 Transient Movement of the Water Table for the Parallel Drain
Problem (Recharge = 0.02 ft/d). . . . . . . . . . . . . . . . . . . . . . . 4-6
Figure 4.5 Steady Seepage through a Square Embankment. . . . . . . . . . . . . 4-7
Figure 4.6 Simulated Water Table and Base Pressure Head Profiles. . . . . . . 4-9
Figure 5.1 A Schematic Diagram of the Unconfined Aquifer System. . . . . . . 5-3
Figure 5.2 Applied Pumping Stress at Each Well for Transient Analysis of
the Unconfined Aquifer System. . . . . . . . . . . . . . . . . . . . . . . 5-4
Figure 5.3 Head Distributions in Layer 1 at the End of Pumping (First)
New Time-Stepping Schemes. . . . . . . . . . . . . . . . . . . . . . . . . 5-6
iv Contents
Figure 5.4 Head Distributions in Layer 3 at the End of (First) Stress Period
(t=4 yr) Showing Comparison between the Previous and New
Time Stepping Schemes. . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-7
New Time-Stepping Schemes. . . . . . . . . . . . . . . . . . . . . . . . 5-8
New Time-Stepping Schemes. . . . . . . . . . . . . . . . . . . . . . . . 5-9
Contents v
MODFLOW-SURFACT
VOLUME I: FLOW MODULES
ABSTRACT
MODFLOW-SURFACT is a fully integrated groundwater flow and solute
transport code developed by HydroGeoLogic, Inc., based on the widely popular U.S.
Geological Survey modular three-dimensional (3-D) groundwater flow modeling code,
MODFLOW. This volume (Volume I of the code documentation) describes new flow
modules added to MODFLOW to enhance its groundwater flow modeling capability and
computational robustness. The physical and mathematical concepts underlying the new
enhancements are detailed, and implementation of these concepts into the modular
structure of MODFLOW is discussed. Example problems are provided to verify the code
and acquaint the user with its applications. The reader is expected to be familiar with the
original MODFLOW documentation (McDonald and Harbaugh, 1988).
MODFLOW uses the block-centered finite-difference approach to simulate
groundwater flow. Fully or quasi 3-D simulations of confined and unconfined layers may
be performed. Through its new flow packages, MODFLOW-SURFACT enhances the
schemes for performing unconfined simulations to rigorously model
desaturation/resaturation of aquifers and overcome numerical difficulties encountered with
previous versions of MODFLOW. MODFLOW-SURFACT provides an option for
discretizing the domain using an axi-symmetric geometry for efficient simulation of
pumping tests, baildown/recovery tests, etc. External stresses normally allowed by
MODFLOW include constant head, constant flux, areal recharge, evapotranspiration,
drains, and streams. In addition, MODFLOW-SURFACT provides a rigorous well
withdrawal package, unconfined recharge boundary conditions, and seepage face boundary
conditions. Finally, MODFLOW-SURFACT includes options for adaptive time-stepping
and output control procedures and an additional Preconditioned Conjugate Gradient (PCG)
solution package. Version 2 of MODFLOW-SURFACT contains additional capabilities
which include rigorous saturated-unsaturated moisture movement simulation capability,
air flow simulation capability, and a Newton-Raphson linearization package for improved
robustness.
The new flow packages of MODFLOW-SURFACT are written in FORTRAN 77
and compiled using the Compaq Visual Fortran Professional Edition 6.6® compiler. It will
run without modification on any Pentium based PC with a minimum recommended 64Mb
of RAM.
1.1 GENERAL
The U.S. Geological Survey Modular Flow Model (MODFLOW) is probably the
most popular groundwater code being used worldwide as a simulation tool to support
groundwater investigations. MODFLOW has been greatly enhanced by HydroGeoLogic,
Inc., to allow the user to handle complex field problems in a more robust and efficient
manner. The enhanced code is named MODFLOW-SURFACT (i.e., MODFLOW with
new flow and contaminant transport packages). The SURF acronym given to the new flow
packages denotes the following:
• S - robust and efficient (matrix solution, nonlinear iterations with adaptive
time- stepping, output control, and axi-symmetric) Schemes
• U - rigorous treatment of Unconfined flow
• R - non-ponding Recharge and seepage face boundary conditions
• F Fracture well representation to provide rigorous treatment of well
-
conditions.
Seven new modular packages have been developed by HydroGeoLogic, Inc., and
added to the original MODFLOW code (McDonald and Harbaugh, 1988). These packages
are as follows:
(1) BCF4 - The Block-Centered Flow Package which includes an axi-
symmetric analysis option and provides a rigorous treatment of
unconfined flow using a variably saturated formulation with
pseudo-soil functions. This package also includes rigorous water
and air flow simulation capabilities for Version 2.
(2) FWL4/
FWL5 - The Fracture-Well Packages (FWL4 and FWL5) provide a
rigorous treatment of well withdrawal conditions using one-
dimensional (1-D) fracture tube elements to emulate a well.
(3) RSF4 - The Recharge-Seepage Face Package which accommodates non-
ponding recharge and seepage face boundary conditions.
(4) ATO4 - The Adaptive Time stepping and Output control Package which
conducts the transient solution using an aggressive time stepping
scheme with automatic generation and control of time steps as
well as better output control.
(5) PCG4 - The Preconditioned Conjugate Gradient Package which contains
a new PCG matrix solver as a robust alternative to the existing
Strongly Implicit Procedure (SIP) and Slice Successive Over-
Relaxation (SSOR) solvers in MODFLOW.
(6) NRB1 - The Newton-Raphson linearization package with backtracking
increases robustness for unconfined or unsaturated flow (of air or
water) simulations.
(7) OBS1 - The observation node package for recording hydrographs at select
nodes.
These additional flow packages in MODFLOW-SURFACT are documented
herein. Provided in the following chapters are their brief descriptions, formulations,
verification and application examples, input instructions, and sample data files. Users of
this document are assumed to be reasonably familiar with the original MODFLOW code
Chapter 1: Introduction 1-1

and its documentation. Those who are unfamiliar with MODFLOW should consult the
following reference:
McDonald, M.G., and A.W. Harbaugh (1988), Techniques of Water-
Resources Investigations of the United States Geological Survey, Chapter
A1: A Modular Three-Dimensional Finite-Difference Ground-Water Flow
Model, Book 6, Modeling Techniques.
1.2 ATTRIBUTES OF NEW FLOW PACKAGES IN MODFLOW-SURFACT
The new flow packages and enhancements provide the following key benefits described
below.
• Handling of complete desaturation and resaturation of grid blocks - The original
USGS MODFLOW’s BCF packages had difficulties in drying/rewetting of grid
blocks during transient simulations.
• Accurate delineation and tracking of water table position, taking into account flow
in the unsaturated zone, delayed yield, and vertical flow components - The
original USGS MODFLOW’s BCF package ignores these features.
• Axi-symmetric flow simulation option - The original USGS MODFLOW’s BCF
package ignores these features.
• Automatic and correct apportioning of the total flow rate of a multi-layer well to
the well nodes - The original USGS MODFLOW’s WEL package required apriori
approportioning of well pumping from multiple aquifers or multiple model grid
layers.
• Accommodation of wellbore storage and overpumped wells - The original USGS
MODFLOW’s WEL package fails when withdrawals demanded are larger than the
supply from a well.
• Handling of non-ponding or prescribed ponding recharge conditions - The original
USGS MODFLOW’s RCH package ignores these features..
• Handling of seepage face boundary conditions - The original USGS MODFLOW’s
RCH package ignores these features.
• Adaptive time-stepping schemes with automatic generation and control of time
steps - The original USGS MODFLOW’s OC package ignores these features.
• Better organization and control of simulation output - The original USGS
MODFLOW’s OC package ignores these features.
• Robust and efficient PCG matrix solution option.
• Capability of modeling unsaturated moisture movement.
• Capability for non-equilibrium preferential unsaturated flow using bi-modal
hydraulic conductivity functions.
• Capability of modeling unsaturated air movement.
• Robust and efficient Newton Raphson linearization option.
• Variable anisotropy option - This option allows for horizontal anisotropy to be
included on a cell-by-cell basis. Note that MODFLOW requires the anisotropy
to be uniform within each model layer.
• Vertical conductivity input option - With this option, the vertical conductivity of
grid blocks is input, and the saturated leakance between layers is computed
internally. Note that MODFLOW requires the saturated leakance between layers
as input.
• Other packages developed by the USGS for later versions of MODFLOW
including MODFLOW-96 and MODFLOW-2000 have also been incorporated into
1-2 Chapter 1: Introduction

this version of MODHMS. Packages incorporated from MODFLOW-96 include
the Transient Leakage Package (TLK), the Direct Solver Package (DE4), the
Stream Routing Package (STR1), the Hydraulic Flow Barrier Package (HFB), the
Interbed Storage Package (IBS), the Constant (time-varying) Head Package
(CHD), the General Finite Differences Package (GFD), and the Flow and Head
Boundary (time-varying) Package (FHB). Packages incorporated from
MODFLOW 2000 include the Reservoir Package (RES1), the Stream-Flow
Routing Package (SFR1), the Lake Package (LAK3), the Gage Package (GAGE),
and the Evapotranspiration Segments Package (ETS1). Note that LAYCON
should be set to the unconfined options (LAYCON=3, 13, 23, 33, 43) for the
LAK3 package to work correctly with the BCF packages herein. A warning to
this effect is reflected in the output listing file if this condition is not met.
• Modifications for using MODPATH with the FWL4, FWL5 and ATO4 Packages.
This enhancement extracts time step information to a binary *.mph file (when
using the ATO4 package) for use with a modified version of MODPATH. The
original version of MODPATH utilizes only time step information from the OC
package and does not recognize FWL4 and FWL5 cell by cell output from
MODHMS: Note that MODPATH can be used only with the subsurface portions
of MODHMS since surface water modules are not included in MODPATH.
• Curvilinear grid option - This option allows for non-rectangular grids in the areal
plane. Hence, DELR and DELC for each node within a layer is required as input,
with multiple layers stacked vertically below the first. Note that the original
MODFLOW grid was rectangular in the areal plane.
• The LAK2 package, developed by Greg Council at GeoTrans in 1996 has been
implemented with permission from the author. Input instructions and
implementation details of this package are available from GeoTrans.
• An observation node package has been included to list the breakthrough of head
or contaminant concentrations through time, at any prescribed observation
location(s).
All modules of MODFLOW-SURFACT are fully compatible, and a mix-and-match of
modules from the original MODFLOW, and from MODFLOW-SURFACT may be used
in a single simulation. The modular structure of MODFLOW is maintained, and modules
for the enhanced features are called from the main program. New formulations and
computational schemes implemented into the additional packages have been verified using
analytical solutions and other numerical models. A complete compatibility is maintained
between the additional packages of MODFLOW-SURFACT and the original MODFLOW
code. Input preparation for MODFLOW-SURFACT is straightforward and follows the
MODFLOW format structure. Input data for flow modeling is essentially that required
by the original MODFLOW, and minimal additional data is required for the
enhancements. Transport simulations make use of all MODFLOW data sets, and
duplication of the supplied input information is eliminated. Transport parameters are
supplied to the code via an additional data set, and boundary conditions are implemented
through the MODFLOW boundary data sets.
1.3 MODFLOW-SURFACT OPERATION AND INPUT OPTIONS
MODFLOW-SURFACT may be operated in a stand-alone mode or as part of a
complete package that includes a graphical user interface. The first option is relevant if
the user acquires MODFLOW-SURFACT by itself or wishes to perform simulations
outside of the user interface system.
This section focuses on the first option (i.e., running MODFLOW-SURFACT in
the stand-alone mode). Before running the code, the user will need to create input data

files for a given simulation problem. To construct input data files for groundwater flow
simulation, two key steps should be taken:
1. Refer to the original MODFLOW User’s Manual (McDonald and Harbaugh,
1988) and create input files for the relevant flow packages following the
instructions in the manual.
2. If usage of new SURF packages is desirable, then proceed to the relevant chapters
of this document and edit the corresponding data files following the additional
input instructions for the selected packages.
In the event that the user has existing MODFLOW input files, only step 2 is necessary.
1.4 DOCUMENT ORGANIZATION AND USAGE GUIDE
This document is organized into nine chapters. The remaining chapters and their
purposes are outlined below.
Chapter 2 presents the BCF4 Package which contains new variably saturated and
axi-symmetric flow formulations. Included in the chapter are formulation descriptions,
verification and application examples, and input instructions for the BCF4 Package. Since
BCF4 supersedes the previous BCF Packages, Chapter 2 provides the user with the most
general instructions for setting up the Block-Centered Flow input file in IUNIT(1).
Chapter 3 presents the FWL4 AND FWL5 Packages which incorporate new
schemes for rigorous treatment of multi-layered wells and special well conditions (e.g.,
wellbore storage and overpumping). This chapter should be consulted by the user who is
interested in handling multi-layer wells rigorously and accounting for special well
conditions.
Chapter 4 presents the RSF4 Package which incorporates new schemes for
handling recharge boundary conditions (with prescribed or zero ponding elevations) and
seepage face boundary conditions. This chapter should be consulted by the user who deals
with unconfined flow problems with such complex boundary conditions.
Chapter 5 presents the ATO4 Package which incorporates adaptive time-stepping
schemes with automatic controls of time step sizes, as well as simulation output controls.
These schemes allow the simulation to be performed more efficiently. Chapter 5 should
thus be read by the user who is interested in such options.
Chapter 6 presents the PCG4 Package which incorporates Preconditioned
Conjugate Gradient (PCG) and Orthomin matrix solution schemes. This chapter should
be consulted by the user who is interested in trying out the PCG solver.
Chapter 7 presents the NRB1 package for Newton-Raphson linearization. Note
that the PCG4 package needs to be used with the NRB1 package since it is the only solver
available in MODFLOW capable of handling an unsymmetric matrix.
Chapter 8 presents the OBS1 package to record the head variation in time of
specified observation nodes.
Chapter 9 provides a list of references cited in the previous chapters.
Also included in the document are five Appendices (A through E) containing
additional information and input data files for the example problems described in Chapters
2 through 5. The user may elect to use these files to perform test runs and develop insight
into the new capabilities provided by MODFLOW-SURFACT. Four additional
appendices (F through I) provide input instructions for expanded modules of MODFLOW-
SURFACT/MODHMS.

2.1 GENERAL
The MODFLOW code utilizes a block-centered finite-difference approach to solve
the groundwater flow equation. The flow domain is discretized into rows, columns, and
layers such that each node represents a rectangular block of porous material which is
referred to as a cell. A node in the resulting finite-difference grid represents either a no-
flow, variable-head, or constant-head cell, and any hydraulic property associated with a
cell is specified with respect to the corresponding node.
For unconfined flow simulations, the BCF1 package of the original MODFLOW
code (McDonald and Harbaugh, 1988) converts a desaturated variable-head cell into a no-
flow cell. This results in permanent exclusion of parts of the flow domain from the model
simulation. The inability of MODFLOW to revert these cells to variable-head cells in case
of any recovery of potentiometric levels may cause the code to produce misleading or
erroneous simulation results.
An attempt was made by McDonald et al. (1991) to reconvert no-flow cells into
variable-head cells to allow resaturation of the cells if needed. Their rewetting scheme
was implemented in the BCF2 Package (Block-Centered Flow Package, Version 2). The
scheme utilizes heads at the neighboring cells to determine whether to reconvert a no-flow
cell into a variable-head cell. Unfortunately, the BCF2 rewetting option is prone to
convergence and stability problems during rewetting or when withdrawals dry up the
respective cells (McDonald et al., 1991). This is because the ad hoc procedure used to
reactivate dry cells can seriously violate flow and mass conservation principles. In
addition, the harmonic transmissivity averaging scheme used in the code becomes
physically inconsistent when the head-dependent transmissivity at a neighboring cell
reduces to zero. A host of problems associated with the BCF2 rewetting package, along
with a host of alleviation schemes, have been documented by Goode and Appel (1992) who
added the BCF3 package to MODFLOW to provide alternative procedures for averaging
transmissivities. In view of the serious difficulties encountered with the above-mentioned
rewetting procedures, a rigorous approach that satisfies flow continuity requirements and
allows free movement of the water table in unconfined layers, without any kind of forced
conversion, is necessary. HydroGeoLogic, Inc., developed such an innovative approach
by using a 3-D variably saturated flow formulation with pseudo-soil water retention
functions automatically generated to reduce the unsaturated flow problem to one of seeking
the water-table level (i.e., the elevation where pressure head is zero or atmospheric) in a
cell. The formulation has been designed to provide accurate delineation of the water table
and capture the delayed yield response of an unconfined system to pumping and recharge.
Data requirements for the simulation are unchanged, since pseudo-soil relations are used
instead of constitutive relations of the real soils. Key advantages of our approach include:
• A rigorous treatment of 3-D flow (i.e., without the Dupuit assumptions) which
does not require turning cells on and off.
• A robust numerical scheme with good convergence and stability characteristics and
the desired capability to handle complete desaturation and resaturation of cells.
Using the finite-difference method, the variably saturated formulation with pseudo-
soil functions is programmed into the existing BCF3 Package of MODFLOW and the new
version is referred to as the BCF4 Package. The new option of the BCF4 Package
computes the interblock conductance as a product of the weighted harmonic mean of the
block hydraulic conductivities, relative permeability, and mean flow area. The use of the
Chapter 2: The Block-Centered Flow (BCF4) Package 2-1

M:\Projects\MODFLOW Manuals\SURFACT and HMS Ver.3\Vol I\R03-05.355_2-3.wpd
variably saturated formulation with pseudo-soil functions in the BCF4 Package and its
applications are discussed in the following sections.
The BCF4 package also provides the option to perform an axi-symmetric analysis
using an axially symmetric cylindrical coordinate system. This option avoids the use of
fully 3-D formulation via Cartesian (x,y,z) coordinates, thereby leading to considerable
savings in the computational effort. The axi-symmetric analysis option is appropriate for
certain situations (i.e., single well flow) where an axi-symmetric geometry is acceptable.
The modular nature of MODFLOW and the structure of input data preparation are
preserved in the development of the BCF4 Package. Data requirements for the new
capabilities are the same as for the original MODFLOW and no additional parameter
values are necessary. To maintain backward compatibility and allow the user to compare
various computational schemes, the BCF4 Package also incorporates all the options of the
previous BCF3 and BCF2 Packages. In this chapter, the user is systematically introduced
to the new capabilities of the BCF4 Package through a brief discussion of the formulation,
verification, and demonstrative example problems.
2.2A FORMULATION OF THE UNCONFINED FLOW EQUATION USING A

VARIABLY SATURATED FLOW FORMULATION
The 3-D movement of water in a variably saturated system can be expressed as
(Huyakorn et al., 1986)
(1)
where:
x, y, and z are Cartesian coordinates (L);
Kxx, Kyy, and Kzz are the principal components of hydraulic conductivity along the
x, y, and z axes, respectively ( );
krw is the relative permeability, which is a function of water saturation;

h is the hydraulic head (L);
W is a volumetric flux per unit volume and represents sources and/or sinks of
water ( );
N is the drainable porosity taken to be equal to the specific yield, ;
Sw is the degree of saturation of water, which is a function of the pressure head;
Ss is the specific storage of the porous material ( ); and
t is time (T).
For a fully saturated medium (i.e., Sw = 1.0), the relative permeability is unity and
equation (1) reduces to:
(2)
2-2 Chapter 2: The Block-Centered Flow (BCF4) Package

Equation (2) is the basic groundwater flow equation used in the development of
MODFLOW (McDonald and Harbaugh, 1988). Hence, the variably saturated flow
equation reduces to the conventional groundwater flow equation below the water table and
in confined systems. Pseudo-soil relations are used to define the functional relationships:
Sw = Sw (R) and krw = krw (Sw) where R is the pressure head defined as R = h+z, with z
being positive in the vertically downward direction. Note that the block-centered finite-
difference approximation of equation (2) contains interblock hydraulic conductivities,
which are obtained as a weighted harmonic mean of nodal values of the participating cells.
Therefore, for the BCF4 package modeling option, the user specifies horizontal saturated
hydraulic conductivities (Kxx and Kyy) and top and bottom elevations of each cell. The
interblock saturated hydraulic conductivities are multiplied by the relative permeability to
yield interblock effective hydraulic conductivities. The vertical interblock leakance
(vertical hydraulic conductivity Kzz divided by vertical distance between nodes )z) is input
directly by the user, as was done in all previous BCF Packages. These leakances may be
computed using various methods discussed in the original MODFLOW documentation
(McDonald and Harbaugh, 1988). The nonlinearity in the variably saturated flow equation
is treated using the modified Picard method (Celia et al., 1990). Similar to the previous
BCF Packages, the user inputs the dimensionless primary storage coefficient as SF1. The
specific storage (Ss) used by equation (1) is calculated internally by dividing the storage
coefficient SF1 by the block thickness.
In summary, the variably saturated formulation with pseudo-soil functions
embedded in the BCF4 Package includes two improved schemes over the previous
packages: (1) A new approach to the unconfined flow problem which treats desaturation
and resaturation as an integral part of the governing equation but does not require soil
water retention data; and (2) the harmonic averaging of saturated hydraulic conductivities
which is shown in Appendix A to be more appropriate than harmonic averaging of
effective (head-dependent) transmissivities. The new option does not inactivate dry cells,
but calculates hydraulic heads required to transmit the recharging water through the
unsaturated zone. Note that the new BCF4 Package can be used for fully 3-D, quasi 3-D,
and axi-symmetric simulations.
For an axi-symmetric flow simulation, the BCF4 Package utilizes the following
form of the saturated groundwater flow equation:
(3)
where r and z are the radial and vertical coordinates of the axi-symmetric cylindrical
coordinate system, and Krr and Kzz are the principal hydraulic conductivities along the
radial and vertical axes, respectively.
When the axi-symmetric option is invoked, the BCF4 Package regards the radial
coordinate (r) as the horizontal x-coordinate, while the vertical coordinate (z) is unchanged
and the horizontal y-coordinate is made inactive by setting the number of rows of grid
blocks (NROWS) to 1. Note that the axi-symmetric configuration of any grid block ()xi,
)zk) is generated by rotation of the block about the z-axis.

2.2B FORMULATION OF VARIABLY SATURATED WATER FLOW EQUATIONS
The 3-D movement of water in a variably saturated subsurface system is
expressed by equation (1). In order to solve the variably saturated flow problem, it is also
necessary to specify the relationships of relative permeability versus water phase
saturation, and pressure head versus water phase saturation. Two alternative functional
expressions are used to describe the relationship of relative permeability versus water
saturation. These functions are given by (Brooks and Corey, 1966):
(4a)
and (van Genuchten, 1977):
(4b)
when n and ( are empirical parameters and Se is the effective water saturation defined as
Se=(Sw-Swr)/(1-Swr) with Swr being referred to as the residual water saturation. The Brooks-
Corey expression produces a similar curve to the van Genuchten function when the
parameter n is chosen such that n = 1 + 2/( (van Genuchten, 1980).
A further enhancement of simulation capability is provided by inclusion of a
piecewise - continuous hydraulic conductivity function by following Mohanty et al (1997),
to include a bimodal description of preferential flow pathways in the system. Thus, if the
saturation in the system is less than a critical value S*, the van Genuchten function of
equation (4b) is used. However, when saturation is above S*, an exponential conductivity
curve is applied that increases conductivity rapidly near saturation, as follows:
(5a)
where S* is the critical or breakpoint saturation value where flow changes from capillary
dominated to non-capillary dominated and vice versa, kr* is the relative permeability
corresponding to S*, and * is a fitting parameter representing effective macroporosity or
other structural features contributing to non-capillary dominated flow. Other functional
forms may be accommodated by tabular input option also provided.
The relationship of pressure head (R = h-z, where z = vertically upward
coordinate) versus water saturation is described by the following function (van Genuchten,
1977, 1980):

(6)
where " and $ are empirical parameters, hc is the capillary head defined as (hap-R), where
hap, the pressure head in air is taken as atmospheric (=0), and Swr is the residual water
phase saturation. The parameters $ and ( are related by ( = 1-1/$. The Brooks-Corey
and van Genuchten functions for the moisture retention and relative permeability
characteristics can be measured in the laboratory for a given soil.
The BCF4 package invokes variably saturated flow simulations using real soil
functions [equations (4) or (5) with equation (6)] via an index IREALSL in the BCF4 input
file. For the van Genuchten soil functions, additional inputs are required for the
parameters ", $, and Swr. If the Brooks Corey relative permeability function is requested,
the Brooks Corey parameter n is also required as input. The remaining parameters are
identical to those discussed in section 2.2 for 3-D, cross-sectional, or axi-symmetric
situations.
2.2C FORMULATION OF VARIABLY SATURATED AIR FLOW EQUATIONS

The flow of air in a variably saturated porous medium may be expressed as
(Huyakorn et al., 1994)
(7)
where xi, xj (i, j = 1, 2, 3) are Cartesian coordinates, koij is the intrinsic permeability
tensor, kra is the relative permeability to air, µa is the dynamic viscosity of air, Ca is the
pressure in the air phase, Da is the density of air, Wa is the volumetric flux of air per unit
volume of porous medium representing sources and/or sinks of air, and Sa (=1-Sw) is the
air phase saturation.
For solution of equation (7) within the framework of MODFLOW-
SURFACT/MODHMS, it is convenient to transform the equation into MODFLOW type
variables. Equivalent definitions to pressure head and hydraulic head are therefore
required for the air phase as
(8)
where hap is the pressure head of the air phase, Ma is the potential in the air phase, Dw is
the density of water, g is gravitational acceleration, and haM is the potential head or total
head in the air phase. Further, by definition, the hydraulic conductivity Kij may be
expressed as:
(9)

The first term on the left hand side of equation (7) may therefore be manipulated
using equations (8) and (9) and re-written as
(10)
The last term in equation (7) may be manipulated for the air phase in an analogous
manner as was done by Bear (1979, p. 91) for the water phase to give
(11)
where Ssa is a volumetric specific storage to air defined analogously as:
(12)
where "s is the compressibility of the soil matrix and $a is the compressibility of air. Ssa
is not a tangible parameter for most aquifer systems, and further manipulation is required
to pose the equation in terms of Ss, the specific storage of the porous material, defined as
(Bear, 1997)
(13)
where $w is the compressibility of water. Equations (12) and (13) yield
(14)
Using equations (10), (11), and (14) in (7) gives the air phase flow equation in
terms of MODFLOW type variables and basic parameters as:
(15)
Note that the last term of equation (15) is the compressible storage term which is
significant in transient simulations that investigate subsurface air follow due to atmospheric
pressure fluctuations. The denominator of this term is close to unity, however the entire
form has been retained for completeness. Effects of density variations on air flow have
been tested against analytical solutions of Shan, 1995.
The van Genuchten capillary pressure function of equation (6) provides the
saturation distribution within the porous medium with the constraint that Sw+Sa=1.
Relative permeability to the air phase is expressed by the Brooks Corey relationship as:
(16)

where Sea is the effective air saturation defined analogously to the effective water saturation
Se, [see equation (6)] as:
(17)
Note that the Brook-Corey parameter n has to be greater than 2 for physically
realistic kr functions when air is the active phase.
Relative permeability to the air phase is expressed by the van Genuchten
relationship as:
(18)
When solving for the air phase flow equation only, an assumption has to be made
with regards to the water phase in the system, since water is treated as a passive phase.
Two options are provided in MODFLOW-SURFACT/MODHMS for treatment of the
water phase when solving for air phase flow. As a first option, water is assumed to be in
hydrostatic equilibrium with a constant hydraulic head, equal to the initial elevation of the
water table, ZWT (x,y), at any x, y location in the domain.
The pressure head in the water phase, hwp (=R used previously) at any elevation
z within the domain in therefore given as:
(19)
where water pressure, hwp equals the ambient air pressure (=0) at the water table.
The capillary head required in equation (6) [hc=hap-hwp] is therefore obtained by
using (19) and (8) as:
(20)
As a second option, the ambient water pressure may be obtained for equation(6) from a
steady-state solution to the water flow equation over the same domain. Thus
recharging/discharging systems may be accommodated during air phase flow calculations
by providing correct air and water saturations under ambient air conditions.
The density of air at system pressure conditions is obtained from the ideal gas law
as:
(21)
where Dstd is the standard air density at atmospheric pressure conditions, Catm; is the
absolute pressure of air in the system; and it is assumed that the apparent molecular weight
of air is the same for system conditions, as it is for standard conditions. Re-arranging
equation (8) and expressing the air pressure under absolute pressure conditions gives

(22)
where the reference pressure Patm occurs at the reference elevation z=0 (i.e, at the
datum). Density effects of air on the system potential are usually small and one may
prescribe Patm for mean sea level conditions regardless of the datum elevation, without
significant loss in accuracy.
Substituting (22) into (21) and re-arranging gives the air phase density in terms of
the MODFLOW-SURFACT/MODHMS variables as
(23)
Summarizing the formulation for variably saturated air phase flow, MODFLOW-
SURFACT/MODHMS uses equation (15) to govern air flow behavior subject to capillary-
saturation relations of equation (6) and relative permeability behavior of either equation
(16) or equation (18). Capillary head is determined from equation (20) with a passive
water phase assumption of instant hydrostatic equilibrium for water, or as hi = hap-hwp with
the ambient hwp as input (obtained from a steady-state solution to Richards equation).
Finally, air phase density is computed from equation (23).
The BCF4 package invokes the air phase flow formulation using the index
IREALSL in the BCF4 input file. In addition to the parameters required for soil moisture
relations and relative permeability functions, air phase flow simulations also require input
for Dw, µw, µa, $w, $a, Catm, g, and ZWT (x,y). The remaining simulation input parameters
are identical to those discussed in section 2.2 for 3-D, cross-sectional, or axi-symmetric
situations.
2.3A VERIFICATION EXAMPLE FOR UNCONFINED FLOW

The BCF4 package is verified by comparing numerical results with the analytical
solution of Neuman (1972) for well flow in an unconfined aquifer that exhibits a
significant delayed yield response to pumping.
The problem considered involves a fully penetrating well pumping at a rate of
3658 from a homogeneous and isotropic unconfined aquifer. The initial saturated
thickness of the aquifer is 40 ft, and the well is screened over the lower 30 ft portion of
the aquifer thickness. The aquifer properties are as follows:
Hydraulic conductivity, K = 21 ft/d
Specific yield, Sy = 0.132
Specific storage, Ss = 1.2 x 10 –4 ft –1
Modeling Approaches and Results

The above flow problem is solved using both the fully 3-D and the axi-symmetric
simulation options provided in the BCF4 Package. Numerical results obtained are
compared with the analytical solution.
For the fully 3-D simulation run, the model area is a square of dimensions
2,000 ft x 2,000 ft with the well at the center. To take advantage of symmetry, we

discretize only one quadrant of the flow domain and consider one quarter of the well
discharge, Q. The 3-D grid consists of 20 layers, each 2 ft thick with 21 rows and 21
columns. The horizontal grid spacings ()x and )y) vary from 0.5 ft to 100 ft. The value
of VCONT (vertical hydraulic conductivity divided by inter-layer distance between two
adjacent nodal layers) is calculated as 10.5 . A constant primary storage coefficient
(SF1, defined as specific storage times block thickness) is specified as 2.4 x 10 –4 for each
layer. The well discharge (Q/4 for a quadrant grid) is distributed uniformly to 15 well
nodes at row 1, column 1 of layers 6 through 20. No-flow boundary conditions are
assumed for the outer lateral and bottom and top boundaries. The Strongly Implicit
Procedure (SIP) Package is used to solve the matrix equation.
For the axi-symmetric simulation run, the model domain is a cylinder of radius
2,000 ft and 40 ft thickness. The grid consists of 20 layers of 2 ft thick, 21 rows and 1
column. The horizontal (radial) grid spacings ()x / )r) vary from 0.5 ft to 100 ft. The
full well discharge Q, is distributed uniformly to the 15 well nodes at row 1, column 1,
of layers 6 through 20. The remaining information is the same as that given for the fully
3-D simulation run.
A comparison of drawdown versus time relationships is made for three selected
observation points located at a distance of 31 ft from the well and at the centers of bottom,
middle and top nodal layers. The dimensionless drawdown sD and dimensionless time ty
plotted in Figure 2.1 are defined as:
sD = 4BKbs/Q
and
ty = Kbt/
where b is the initial saturated thickness, and the remaining symbols are as defined
previously.
The numerical results from the fully 3-D and the axi-symmetric simulations are
compared in Figure 2.1 with the Neuman analytical solution. For the bottom and middle
observation points, the numerical and analytical solutions are in excellent agreement. For
the top observation point, the two numerical solutions are in good agreement, but deviate
from the analytical solution at early time values. This deviation may be due to the impact
of the simplifying assumption of constant saturated thickness and the approximate water-
table boundary conditions used by the analytical solution. Note that our numerical solution
is based on a rigorous variably saturated flow formulation.

2.3B VERIFICATION EXAMPLE FOR UNSATURATED FLOW OF WATER
Two example problems are provided to demonstrate verification and application
of the unsaturated zone water flow option in the BCF4 package. The first problem
considers transient infiltration of water in a very dry soil column. The second problem
considers two dimensional flow in an unsaturated rectangular soil slab. These problems
are compared with numerical solutions available for the respective situations.
2.3B.1 Problem 1 - Transient Infiltration in an Unsaturated Vertical Soil
Column
This problem, detailed by Huang et al., (1996), concerns one-dimensional
infiltration in an unsaturated vertical soil column, 160 cm long, discretized into 80 layers
2 cm thick. The initial head throughout the domain was -104 cm representing extremely
dry conditions, and a constant infiltration flux of 50 cm/d was applied at the top at t=0.
A typical coarse-textured soil was analyzed for up to 0.8 days with the following hydraulic
properties:
Saturated hydraulic conductivity, K = 100 cm/d
Specific yield (i.e., porosity), Sy = 0.45
Specific storage, Ss = 1×10-5
van Genuchten " = 0.2 cm-1
van Genuchten $ = 5
Residual water saturation, Swr = 0.11
The van Genuchten functions were used for both moisture retention as well as
relative permeability.
Modeling Approach and Results

The PCG4 solver was selected to solve the system of matrix equations with a
convergence tolerance of 0.05 cm. Adaptive time stepping was performed using the AT04
package with an initial )t=0.2×10-7 d, a maximum )t=0.01 d, a time-step multiplier of
1.3 and a time step dividing factor of 2.7. Figure 2.B1 shows the predicted hydraulic head
profile at 0.4 d and 0.8 d. These results compare very well with results obtained by using
the finite-difference and finite-element models of Huang et al. (1996) and HydroGeoLogic
(1992). This problem also demonstrates the ability of the code to handle highly non-linear
situations and extremely dry conditions.

Figure 2.1 Dimensionless Drawdown Versus Dimensionless Time Relationships
Showing Comparison of Numerical and Analytical Solutions.

2.3B.2 Problem 2 - Two-dimensional Flow in an Unsaturated Rectangular
Soil Slab
This problem concerns two-dimensional unsaturated flow in a rectangular soil slab
with sectional dimensions of 15 cm horizontally and 10 cm vertically. A constant nodal
spacing of )x=1 cm and )z=1 cm was used to discretize the domain. The initial head
throughout the domain was -90 cm. The top 4 cm along the left edge of the domain were
prescribed a boundary head of 6 cm, and the right edge of the domain was maintained at
a prescribed head of -90 cm. The remaining boundaries (i.e., the top boundary, the
bottom boundary, and the lower 6 cm of the left edge) were no flow conditions. Hydraulic
properties of the soil were as follows:
Saturated hydraulic conductivity, K = 1 cm/d
Specific storage, Ss = 1×10-14
van Genuchten " = 0.005 cm-1
van Genuchten $ = 2
Brooks Corey parameter, n = 2
The soil was assumed to be homogeneous and isotropic. The van Genuchten
relationship was used for the moisture retention curve, while the Brooks Corey
relationship was used to define the relative permeability function.
The PCG4 solver was selected to solve the system of matrix equations with a
convergence tolerance of 0.01 cm. An initial time step size of 0.01 days was enlarged by
a factor of 1.2 up to a maximum time step size of 0.05 days. Computed values of
hydraulic head along the base and top of the flow region are plotted in Figure 2.B2 for a
time value of 0.508 days. These results compare very well with results obtained by using
the finite-element model of HydroGeoLogic (1992).
2.3C VERIFICATION EXAMPLE FOR AIR-PHASE FLOW SIMULATION

The air-phase flow simulation option of the BCF4 package is verified by
comparing numerical results with those obtained by using a comprehensive multi-phase
flow simulator (Huyakorn et al., 1994).
The problem considered here is detailed by Panday et al. (1994). This problem
was selected to study the flow dynamics of an air-sparging process. Relevant field site
information given by Marley and Magee (1992) is used for this study. The site of concern
has been contaminated by gasoline from several leaking underground storage tanks in both
the vadose and saturated zones. The average depth to the water table is -2.4 m below land
surface. The contamination covers an area of -450 m2. A preliminary assessment of the
sparging influence radius and air velocities in the vicinity of an injection well may be
obtained by considering variably saturated air flow only (i.e., assuming passive liquid
phases). A rectangular domain of 30 m × 20 m × 3 m was considered and discretized
into 900 nodes having uniform nodal spacings of )x = 2m, )y = 2m, and )z = 0.5m.
The initial conditions correspond to prevailing atmospheric pressure in the vadose zone
(haM=0). The sparging process was created by injecting air at a steady rate of 10-3m3s-1
at the node located at x = 9 m, y = 9 m, and z = 0.75 m. The water table is located 0.4
m below the sparging well (i.e., at z=0.35). Atmospheric pressure conditions were

Figure 2.B1 Hydraulic Head Profile at 0.4 and 0.8 days for Transient Infiltration in
an Unsaturated Vertical Soil Column.

Figure 2.B2 Hydraulic Head Profiles at .508d for Flow in an Unsaturated
Rectangular Soil Slab.

maintained on the surface boundary throughout the simulation and the bottom and lateral
boundaries were treated as no-flow conditions. The soil and fluid properties used in the
simulation were adopted from values given by Marely and Magee (1992) and are shown
below.
Saturated hydraulic conductivity, K = 3.92×10-8 m/s
Specific storage, Ss = 3.53×10-5
Water density, Dw = 1,000 kg m-3
Water viscosity, µw = 0.5 cP
Water compressibility, $w = 10-8 Pa-1
Air density, Da = 1.777 kg m-3
Air viscosity, µa = 0.1983@10-4 cP
Air compressibility, = 1.777@10-5 Pa-1
van Genuchten " = 0.145 m-1
van Genuchten $ = 2.7
Standard atmospheric pressure, Patm = 1.01×105 Pa
Gravitational acceleration, g = 9.8066 m/s2
The van Genuchten functions were used to describe the relative permeabilities and
capillary pressure of the air-water system.

Both transient and steady-state simulations were performed for this problem.
Transient simulations invoked the AT04 package for adaptive time stepping with an initial
time step size of 1 sec, a minimum )t=1×10-3s, maximum )t=1×105s, a time step
multiplier of 1.4 and a time step reduction factor of 2. The PCG4 package was selected
to solve the system of matrix equations with a maximum of 20 non-linear iterations for the
transient case and 100 non-linear equations for the steady-state simulation. Head closure
criteria were 0.1 and 0.21 m for the transient and steady-state simulations, respectively.
The simulated distributions of air pressure along a horizontal line through the well
are depicted in Figure 2.C1 for both simulations along with simulations using the finite-
element code of Huyakorn et al. (1994) at t=7 min, and at steady-state conditions. The
predicted and observed pressure profiles are in reasonable agreement with both codes. It
is interesting to note that the transient simulation reaches steady-state conditions in only
1.16 days showing the rapid equilibration process that occurs in the air phase. Absolute
air pressures for these simulations were computed from the potentiometric head by
equation (22) with Da calculated from equation (23).
2.4 APPLICATION EXAMPLES
Two example problems are provided to demonstrate the application of the variably
saturated flow option of the new BCF4 Package. In both examples, sufficient pumping
is allowed to occur to desaturate portions of an unconfined aquifer. The first example, is
a steady-state pumping simulation. The second example is a transient simulation in which
the wells are shut off after a four-year stress period to rejuvenate the aquifer. This
scenario is also modeled using the existing wetting option in MODFLOW (McDonald et
al., 1991) and the STAFF3D finite-element code (HydroGeoLogic, Inc., 1992). The
results obtained from different methods are compared and discussed below.

2.4.1 Problem 1—Steady-State Simulation of Pumping in an Unconfined
Aquifer
This test problem considers a 300-ft thick unconfined aquifer shown in Figure 2.2.
The modeled domain is a square of dimensions 75,000 ft x 75,000 ft. The top and bottom
of the aquifer are at elevations of 50 and –250 ft, respectively. The domain is subject to
a uniform and continuous vertical recharge of 3.0 x 10 –9 ft/s. The left (west) boundary
in the figure is a constant head boundary, and the remaining boundaries are no-flow
boundaries. Fifteen wells, screened over 100 ft of the bottom of the aquifer are each
pumped at a rate of 0.95 . The locations of wells are shown in Figure 2.2. Aquifer
parameters include:
Horizontal hydraulic conductivity, Kh (Kxx and Kyy) = ft/s
Vertical hydraulic conductivity, Kv (Kzz) = ft/s
Modeling Approach
The domain is uniformly discretized into 3 layers, 15 rows, and 15 columns of
grid blocks (i.e., )z = 100 ft, )x = )y = 5,000 ft). Thus, the value of VCONT (vertical
hydraulic conductivity divided by inter-layer distance between two adjacent nodal layers)
is calculated as . A constant head of zero is prescribed on the west side of top and
middle layers. The initial head distribution required by the code is uniformly set to zero
throughout the domain. The 15 wells are located in Layer 3. While these wells act as
sinks, surface recharge and constant head boundary on the west provide replenishment to
the aquifer system thus maintaining the desired steady flow conditions.
Two separate steady-state simulation runs are performed with: (a) input LAYCON
values of 3, and (b) input LAYCON values of 43 for all the three layers. Note that
LAYCON is a control parameter indicating the aquifer treatment option required for a
particular nodal layer. Case (a) uses the original MODFLOW formulation to compute
horizontal interblock transmissivity and conductance, whereas Case (b) activates the new
option included in the BCF4 Package. (See input instructions of the BCF4 Package.) For
both cases, the SIP Package is used to solve the matrix equation. Other than the difference
in input LAYCON values for the layers, the remaining input data for BCF4 Package and
the other Packages (Basic, Well, Recharge, SIP, and Output Packages) are identical. The
input data files used for the simulations are listed in Appendix B.
To maintain continuous recharge to the flow system, the prescribed recharge rate
is applied to the highest active cell in each vertical column. This is necessary for Case (a)
to allow continuous infiltration if any of the active cells receiving the recharge become
dry. The reason is that variable-head cells convert to no-flow cells if they are dry and do
not transmit recharge to the water table. This is not the case if the new variably saturated
flow option is used [Case (b)] and such a complication in data input is obviated.

Figure 2.C1 Steady-State and Transient Distribution of Air Pressure Near Sparging
Well for Air-Phase Flow Simulation.

Figure 2.2 A Schematic Diagram of the Unconfined Aquifer System.

Results and Discussions
Steady-state head distributions produced by the previous MODFLOW option [Case
(a)] and the new variably saturated flow option [Case (b)] are compared in Figures 2.3
through 2.5 for Layers 1, 2 and 3, respectively. Although all of the cells in the domain
are input as variable-head, approximately a quarter of the cells in Layer 1 are inactivated
when the previous BCF1 option in MODFLOW [Case (a)] is used for the simulation (see
Figure 2.3), because the simulated head values are below the bottom elevation of the layer
(–50 ft for Layer 1). Note that the variably saturated flow option [Case (b)] predicts a
head distribution inside the desaturated zones, thus allowing transmission of vertical
recharge to the water table as well as lateral flow components from neighboring cells.
Steady-state head distributions in Layers 2 and 3 (Figures 2.4 and 2.5, respectively) show
that the agreement between the two simulation options improves as the distance from the
desaturated zones increases. In the vicinity of the southeast corner of the domain, there
is a significant discrepancy between the two simulated head distributions. A close
examination of the results of Case (a) and Case (b) indicates that the horizontal interblock
conductance used in the variably saturated flow option of the BCF4 Package behaves
appropriately. Harmonic weighting of head-dependent transmissivities used in the original
scheme [Case (a)] causes larger drawdowns because transmissivities are weighted toward
lower values when drawdowns are large. From the physical standpoint, the harmonic
weighting should be applied on the saturated hydraulic conductivities and not the effective
transmissivities, which are head-dependent. (See Appendix A for demonstration.)
2.4.2 Problem 2—Transient Simulation of Pumping and Recovery in an

Unconfined Aquifer
The transient analysis presented in this example also considers a 300-ft thick
unconfined aquifer, shown in Figure 2.2. As in Problem 1, the modeled domain
corresponds to a square of dimensions 75,000 ft x 75,000 ft. The top and bottom of the
aquifer are at elevations of 50 and –250 ft, respectively. Initial groundwater levels in the
aquifer are at 0 ft throughout the domain. The domain is subject to a uniform and
continuous recharge of 3.0 x 10 –9 ft/s. The left (west) boundary in the figure is a constant
(zero) head boundary with the remaining boundaries being no-flow boundaries. There are
15 wells screened over 100 ft thickness of the bottom layer as shown in Figure 2.2.
However, in this problem, each well is pumped at a high rate of 3.85 . Pumping
from the wells is shut off after four years. (See Figure 2.6.) The aquifer is then allowed
to recuperate for 12 years. The aquifer parameters are:
Horizontal hydraulic conductivity, Kh (Kxx, Kyy) = 10 –4 ft/s

Vertical hydraulic conductivity, Kv (Kzz) = 10 –5 ft/s
Specific Yield, Sy = 0.03
Specific Storage, Ss = 10 –6 ft –1

Case (a) and Case (b).

Figure 2.4 Steady-State Head Distributions in Layer 2 - Comparison between Case
(a) and Case (b).

Figure 2.5 Steady-State Head Distributions in Layer 3 - Comparison between Case
(a) and Case (b).

Modeling Approach
As in Problem 1, the domain is uniformly discretized into 3 layers, 15 rows, and
15 columns. The value of VCONT (vertical hydraulic conductivity divided by inter-layer
distance between adjacent nodal layers) required by MODFLOW is calculated as 10 –7s –1.
A constant primary storage coefficient (SF1 defined as specific storage times block
thickness) of 10 – 4 is specified for each layer. A constant head of zero is prescribed on the
west side of top and middle discretized layers. Initially, the head in the domain is assumed
to be zero. Fifteen wells located in Layer 3 are allowed to pump at a rate of 3.85
over the first 4 years. Pumping is stopped during the subsequent period of 12 years to
allow aquifer recuperation.
Three separate transient simulation runs are performed. The first two simulations
utilize (1) the existing MODFLOW rewetting capability, and (2) the new variably saturated
flow (BCF4 Package) option of MODFLOW-SURFACT/MODHMS, respectively. These
two simulations are referred to as Cases (1) and (2), respectively. The third simulation
run [Case (3)] is performed independently using the finite-element STAFF3D code
(HydroGeoLogic, Inc., 1992). For the MODFLOW-based simulation runs, the SIP
Package is used to solve the matrix equation. Other than the BCF4 Package input, the rest
of the input data used for both cases are identical. Note that Case (1) requires additional
data to allow rewetting of dry cells.
To maintain continuous recharge in the MODFLOW rewetting package [Case (1)]
simulation option, surface recharge is applied to the highest active cell in each vertical
column (see Recharge Package input). This provision is necessary to allow continuous
infiltration, if any, of the active cells receiving the recharge becomes dry, since variable-
head cells convert to no-flow cells if they are dry, and do not transmit recharge to the
water table.

For the simulation runs performed using the original rewetting capability [Case
(1)] and the new variably saturated flow option [Case (2)], head distributions in Layers 1
and 3 (which represent two extreme saturation conditions) are presented in Figures 2.7
through 2.10. Figure 2.7 compares head distributions in Layer 1 at the end of the
pumping period (t=4 yr). For Case (1), approximately a quarter of the cells in Layer 1
are deactivated by 4 years of continuous pumping, but Layer 3 still remains fully active.
Note that a longer pumping period would lead to desaturation of Layer 3 and triggering
of instabilities in the numerical solution. Figure 2.8a and 2.8b compare head distributions
in Layer 3 at time value of 4 years. There is no inactivated cell in Layer 3. Significant
differences (by as much as 20 to 30 ft) between the head values of Cases (1) and (2) occur
in the vicinity of the individual wells. The two head distributions are in good agreement
away from the well field. Figures 2.9 and 2.10 present the head comparisons for Layers
1 and 3, respectively at the end of the recovery period (t=16 yr). In general, the previous
(rewetting) option of MODFLOW underpredicts the head values. The discrepancy
between the two head distributions is more pronounced in the southeast portion of the flow
domain.

Figure 2.6 Applied Pumping Stress at Each Well for Transient Analysis of the
Unconfined Aquifer System.

Figure 2.7 Head Distributions in Layer 1 at the End of Pumping (First) Period
(t=4 yr) Showing Comparison between Previous and New Simulation
Options.

Figure 2.8a Head Distribution in Layer 3 at the End of Pumping (First) Period (t=4
yr) Obtained Using the Previous MODFLOW Option.

Figure 2.8b Head Distribution in Layer 3 at the End of Pumping (First) Period (t=4
yr) Obtained Using the New Variably Saturated Flow Option in
MODFLOW-SURFACT/MODHMS.

Figure 2.9 Head Distribution in Layer 1 at the End of Recovery (Second) Period
(t=16 yr) Showing Comparison Between the Previous and New
Simulation Options.

Figure 2.10 Head Distributions in Layer 3 at the End of Recovery (Second) Period
(t=16 yr) Showing Comparison between the Previous and New
Simulation Options.

A comparison of head versus time relationships is made for three selected
observation points. Locations of these points with reference to Figure 2.2 are given
below.
Observation Point # Layer Row Column
1 1 7 11
2 2 11 11
3 3 11 10
Temporal variations of head at the three observation points (Figure 2.11) indicate
that both cases [Case (1) and (2)] are in good agreement until the scheme of Case (1)
converts variable-head cells into no-flow cells. After these conversions, the results of
Case (1) deviate from, but follow similar trends to those of Case (2). As stated earlier,
the ad hoc rewetting procedure of the BCF2 Package [Case (1)] may not produce reliable
numerical solutions. For example, if the pumping rate from the wells is increased from
3.85 to 3.90 in the first period of this transient simulation, the rewetting option of
MODFLOW fails to converge. However, the new variably saturated flow option produces
a convergent solution. The failure of the rewetting procedure used in the BCF2 Package
is due to the instability generated by cyclical drying and wetting (flip- flopping) of cells
between iterations.
The results obtained using the MODFLOW-SURFACT/MODHMS variably
saturated flow option [Case (2)] also are compared with those obtained using the
STAFF3D code, referred to as Case (3). Figure 2.12 shows a plot of temporal variation
of heads at the three observation points. There is good agreement between Cases (2) and
(3) through the entire simulation period.
2.5 OTHER EXTENSIONS TO THE BLOCK-CENTERED FLOW PACKAGE
Several minor enhancements have been made to the block-centered flow (BCF4)
package to allow even greater flexibility in simulations.
The first modification allows for input of vertical hydraulic conductivities of the
grid-blocks, as an option to input of leakance between gridblocks of adjacent layers. A
flag, IVHYC is incorporated in the first line of the BCF4 package input, to indicate this.
If IVHYC is set to one, the code expects input of vertical hydraulic conductivities for all
layers, and the leakance is computed internally based on a volume-weighted harmonic
mean of the vertical conductivity of each grid block and their separation distance. If
IVHYC is set to zero, the code expects input of leakance between layers, as in the original
MODFLOW.
The second modification allows for input of horizontal anisotropies on a cell-by-
cell basis, as an option to having a uniform anisotropy value for each model layer. A flag,
IANIXY is incorporated in the first line of the BCF4 package input, to indicate this. If
IANIXY is set to one, the code expects input of anisotropies on a cell-by-cell basis. If
IANIXY is set to zero, the code expects inputs of anisotropies on a layer-by-layer basis
as in the original MODFLOW.
The third modification allows for input of a curvilinear grid, as an option to using
a rectangular grid to discretize the domain in the areal plane. A flag, ICURVL is
incorporated in the first line of the BCF4 package input, to indicate this. If ICURVL is
set to one, a curvilinear grid is used, and the code expects input of DELR and DELC for
all nodes within a layer. Since layers are stacked vertically beneath each other, the DELR
and DELC values for all layers are computed internally from those of the first layer. If
ICURVL is set to zero, a rectangular grid is used as in the original MODFLOW. For this

Comparison between the New Simulation Option in MODFLOW-
MODHMS and STAFF3D.
MODHMS case, DELR is input only for the number of columns (NCOL), and DELC is
input only for the number of rows (NROW) in the grid. An additional flexibility provided
by MODHMS for unsaturated zone water flow simulation allows for deeper layers to be
represented by the confined/unconfined groundwater flow simulation capability provided
by the gravity-segregated vertical equilibrium (GSVE) pseudo-soil functions. This
capability is useful for unsaturated zone flows simulations in avoiding excessive vertical
discretization near the water table which would be needed to accurately represent the
unconfined water level conditions and associated aquifer flow. Thus, with IREALSL=1
or 2 (unsaturated/saturated water flow simulation) the top few layers could be set to solve
Richard’s Equation with a LAYCON value of 43 while thicker layers below could be used
to represent the unconfined/confined groundwater system. This is accomplished in
MODHMS by setting LAYCON=44 for these lower layers which will be treated as if
IREALSL=0 for those layers.
2.6 INPUT INSTRUCTIONS FOR THE BCF4 PACKAGE
2.6.1 General
The input data required by the BCF4 Package are similar to that of the previous
BCF Packages. The BCF4 Package incorporates the new variably saturated flow
formulation that provides a rigorous treatment of the unconfined flow problem with
desaturation and resaturation via the use of pseudo-soil functions. For comparative
purposes, we also include all of the previously documented schemes in the new package.
Thus, our BCF4 Package encompasses the BCF1 Unconfined Package (McDonald and
Harbaugh, 1988) which deactivates dry cells permanently, the BCF2 Package, which
converts no-flow cells to variable-head cells (McDonald et al., 1991), and the BCF3
Package, which simulates groundwater flow using different options for interblock
transmissivity relationships (Goode and Appel, 1992). It is suggested that users who are
interested in comparing the old and new schemes familiarize themselves with the
respective BCF Package documentations.
Input data for BCF4 primarily consists of property arrays to describe horizontal
hydraulic conductivity/transmissivity, vertical leakance, and top and bottom elevations of
cells and the storage coefficient and/or specific yield for a transient simulation. These
property arrays depend on input LAYCON values for each layer, as explained later. An
additional parameter array required by the BCF2 and BCF3 unconfined storage scheme
is WETDRY, which is used when the wetting flag of that option is active. The BCF4
memory allocation package reserves space for these arrays after reading certain flags.
These arrays are then initialized according to user input as discussed below. The input
instructions provided on the following pages are for the BCF4 Package. Note that BCF4
has been designed to maintain the overall compatibility with its predecessors.
To activate the new variably saturated flow option, the user specifies input
LAYCON for a layer as either 40 (if the layer is strictly confined) or 43 (otherwise). A
strictly confined layer is defined as the layer in which the piezometric head remains above
the top elevation of the layer throughout the simulation period. If no prior knowledge is
available for the head conditions in a layer, the user is advised to use an input LAYCON
of 43. The new unconfined flow formulation implemented in BCF4 eliminates major
awkwardness in input preparation as well as difficulties inherent in the previous cell
deactivation and rewetting schemes. We suggest that the user switch to this new approach
for more reliable converging simulations.
The LAYCON=40 or 43 schemes of the BCF4 package use weighted harmonic
averaging of hydraulic conductivities of adjacent grid-blocks to compute the conductance
between cells. An option of using weighted anthmetic averaging of hydraulic
conductivities of adjacent grid-blocks to compute the conductance between cells - similar

to how finite-element assembly is performed - is provided as an additional flexibility by
setting LAYCON=50 or 53.
2.6.2 BCF4 Input
Input for the BCF4 Package is read from the unit specified in IUNIT (1). A brief
explanation of the input variables is provided in the next subsection. The data are read
once at the beginning of the simulation. IUNIT array, which is read by the Basic Package,
determines the options or packages used in a MODFLOW simulation. (See Basic Package
of the original MODFLOW document for details.) Note that the extra input variable
required by BCF4 is shown in a shaded box.
1a. Data: ISS IBCFCB HDRY IWDFLG WETFCT IWETIT IHDWET IAXIS IREALSL ICNTRL IVHYC IANIXY ICURVL IOLDHDR
Format: I10 I10 F10.0 I10 F10.0 I10 I10 I10 I10 I10 I10 I10 I10 I10
*** Enter item 1b only if tabulated retention and permeability functions are used (IREALSL
= 5 or 6)
1b. Data: N_HSK

Format: I10
2. Data: LAYCON (NLAY) (Maximum of 200 layers)

Format: 40I2
(If there are 40 or fewer layers, use one record; otherwise, use more records as needed.)
**Enter item 3 only if air phase flow simulation is required (IREALSL=3 or 4).
3. Data: RHOWP RHOAP VISW VISG COMPWAT COMPAIR ATMGP GRAV INITW
Format: E10.3 E10.3 E10.3 E10.3 E10.3 E10.3 E10.3 E10.3 I10
**Enter item 4 only if the simulation is not axi-symmetric (IAXIS # 0).

**Enter item 4a only if IANIXY=0
4a. Data: TRPY (NLAY)

Utility Module: U1DREL. Note that U1DREL is the utility module that reads
the real one-dimensional (1-D) array (i.e., TRPY
for item 3).
**Enter item 4b only if IANIXY=1

4b. Data: TRPYNU (NCOL, NROW)
Utility Module: U2DREL. Note that item 4b is entered once for each layer of
the model grid.
**Enter item 5a only if ICURVL=0

5a. Data: DELR (NCOL)
Utility Module: U1DREL

**Enter item 5b only if ICURVL=1
5b Data: DELRCV (NCOL, NROW)
**Enter item 6a only if ICURVL=0
6a. Data: DELC (NROW)

**Enter item 6b only if ICURVL=1

6b. Data: DELCCV(NCOL, NROW)
**Enter item 7a only if INITW=0.
7a. Data: ZWT (NCOL, NROW)

**Enter item 7b only if INITW=1.

7b. Data: PWC (NCOL, NROW)
Note that item 7b consists of one array for each layer in the model grid. Arrays for
layers 1 through NLAY must be entered in order.
A subset of the following two-dimensional arrays is used to describe the properties

of each layer of the model grid. The arrays needed for each layer depend on the input
LAYCON value (see item 2) of each layer, whether the simulation is transient (ISS = 0) or
steady-state (ISS not 0), and if the wetting capability is active (IWDFLG not 0) for the
wetting schemes of BCF2 and BCF3. Note that IWDFLG should be set to zero for the new
variably saturated flow scheme introduced in the BCF4 Package (i.e., input LAYCON=43).
If an array is not needed, it must be omitted. There are no situations in which all arrays will
be required. The required arrays (items 8-15) for layer 1 are read first; then the arrays for
layer 2, etc., until the property arrays for all layers are input.
**Enter item 8 only if the simulation is transient (ISS = 0).
8. Data: SF1 (NCOL, NROW)


**Enter item 9 only if the input LAYCON value for this layer is one of the following:
0, 2, 10, 12, 20, 22
9. Data: TRAN (NCOL, NROW)

Note that for input LAYCON=40 or 43, the hydraulic conductivity, bottom and top
elevations are input in items 10, 11, and 14, respectively.
*Enter items 10, 11 and 12 only if the input LAYCON value for this layer is one of the
following:
1, 3, 11, 13, 21, 23, 31, 33, 40, 43
10. Data: HY (NCOL, NROW)

**Enter item 11 only if vertical hydraulic conductivities are to be entered. (i.e.

IVHYC > 0)
11. Data: VHY (NCOL, NROW)
12. Data: BOT (NCOL, NROW)

**Enter item 13 only if this layer is not the bottom most layer of the grid and if IVHYC =
0.
13. Data: VCONT (NCOL, NROW)

**Enter item 14 only if the simulation is transient (ISS = 0) and the input LAYCON value
for this layer is one of the following: 2, 3, 12, 13, 22, 23, 33, 43

**Enter item 15 only if the input LAYCON value for this layer is one of the following: 2,
3, 12, 13, 22, 23, 33, 40, 43
15. Data: TOP (NCOL, NROW)

**Enter item 16 only if the wetting capability is active (IWDFLG is non-zero) and input
LAYCON value for this layer is one of the following: 1, 3, 11, 13, 21, 23, 31, 33

16. Data: WETDRY (NCOL, NROW)
Note: The WETDRY array is irrelevant for input LAYCON = 40 or 43, and, therefore, item
14 is not needed for the new variably saturated flow scheme.
17. Data: VANAL (NCOL, NROW)

Note: Enter item 17 only if the input LAYCON value is 43 and IREALSL…0, 5, or 6.
18. Data: VANBT (NCOL, NROW)

Note: Enter item 18 only if the input LAYCON value is 43, and IREALSL…0, 5, or 6.
19. Data: VANSR (NCOL, NROW)

Note: Enter item 19 only if the input LAYCON value is 43, and IREALSL…0, 5, or 6.
20. Data: BROOK (NCOL, NROW)

Note: Enter item 20 only if the input LAYCON value is 43, and IREALSL=2 or 4.
***Enter item 21 only if IREALSL=7 and input LAYCON is 43.
21. Data: S_STAR (NCOL, NROW)

***Enter item 22 only if IREALSL=7 and input LAYCON is 43.
22. Data: D_STAR (NCOL, NROW)

**Enter items 23, 24 and 25 only if tabular retention input is required (IREALSL= 5
or 6) and input LAYCON is 43. Note that items 23, 24 and 25 are entered once for
each interpolation point, and then repeated for each of the remaining N_HSK
interpolation points of tabular input.
23. Data: H_VAL (NCOL, NROW)


24. Data: S_VAL (NCOL, NROW)
25. Data: K_VAL (NCOL, NROW)

2.6.3 Explanation of Fields Used in Input Instructions
ISS – is the steady-state flag

If ISS = 0, the simulation is transient.
If ISS is not 0, the simulation is steady-state.
IBCFCB – is a flag and a unit number.
If IBCFCB > 0, it is the unit number in which cell-by-cell flow terms
will be recorded whenever ICBCFL is set. (See Output
Control in McDonald and Harbaugh [1988].)
If IBCFCB = 0, cell-by-cell flow terms will not be printed or recorded.
If IBCFCB < 0, flow for each constant-head cell will be printed
whenever ICBCFL is set.
HDRY – is the head assigned to cells that are converted to dry during a simulation.
Although this value plays no role in the model calculations, it is useful as
an indicator when examining the simulated heads that are output from the
model. HDRY is thus similar to HNOFLO in the Basic (BAS) Package,
which is the value assigned to cells that are no-flow cells at the start of a
model simulation.
Note: When input LAYCON value for the layer is 40 or 43, the HDRY
value is not assigned to cells that dry during the simulation like in the other
LAYCON options 0, 1, 2, 3, 10, 11, 12, 13, 20, 21, 22, 23, 31, and 33.
IWDFLG – is a flag that determines if the wetting capability is active.
If IWDFLG = 0, the wetting capability is inactive.
If IWDFLG is not 0, the wetting capability is active.
Note: IWDFLG is not used if ANY input LAYCON value in item 2 is 40 or
43).
WETFCT – is a factor that is included in the calculation of the head that is initially
established at a cell when it is converted from dry to wet. (See IHDWET.)
Leave blank when wetting capability is inactive (i.e., IWDFLG = 0).
Note: This flag is irrelevant to input LAYCON = 40 or 43 conditions, and
may be left blank for these options.
IWETIT – is the iteration for attempting to wet cells. Cell wetting is attempted during
each of the IWETIT iterations. If IWETIT is 0, it is changed to 1. Leave
blank when wetting capability is inactive (i.e., IWDFLG = 0).
Note: This flag is irrelevant to input LAYCON = 40 or 43 conditions and
IHDWET – is a flag that determines which equation is used to calculate the initial head
at cells that become wet.
If IHDWET = 0, equation (3a) of McDonald et al. (1991) is used:
h = BOT + WETFCT (hn - BOT).

If IHDWET is not 0, equation (3b) of McDonald et al. (1991) is used:
h = BOT + WETFCT (THRESH).
IAXIS -- is a flag that determines if the simulation is axi-symmetric. If IAXIS

>0, the axi-symmetric approach is employed.
Note: For the axi-symmetric simulation approach (IAXIS >0) NROW

must be set equal to 1 in the Basic Package Input.
IREALSL -- is a flag indicating that real soil moisture functions are used to define
flow in the unsaturated zone above the water-table.
If IREALSL=0, the pseudo-soil relations are used to define the
water-table.
If IREALSL is greater than 0, then flow in the unsaturated zone is
treated rigorously.
If IREALSL=1, van Genuchten functions are used for retention
and relative permeability characteristics of the
unsaturated layers and water flow is simulated.
If IREALSL=2, van Genuchten function is used to characterize
moisture retention, and Brooks-Corey function is
used for relative permeability characteristics of the
unsaturated layers and air flow is simulated.
If IREALSL=4, van Genuchten functions are used to characterize
moisture retention and Brooks-Corey function is
used for relative permeability characteristics of the
If IREALSL=5, Tabular input of moisture retention and relative
permeability curves is used to characterize
unsaturated layers and flows of water is simulated.
If IREALSL=6, Tabular input of moisture retention and relative
permeability curves is used to characterize
unsaturated layers and flow of air is simulated.
If IREALSL=7, The bimodal hydraulic conductivity curve of
Mohanty (1997) is used to characterize unsaturated
zone water flow with preferential pathways as
discussed in Section 2.2B.
Note: This flag is required only for input LAYCON=43, 53, 44 or
54.

ICNTRL -- is a flag determining selection of mid-point or upstream weighting of
the relative permeability term.
If ICNTRL = 0, upstream weighting is used (suggested option).
If ICNTRL = 1, midpoint weighting is used.
Note: This flag is required only for input LAYCON=43, 53, 44 or
54.
IVHYC -- is the vertical hydraulic conductivity input flag.

If IVHYC = 0, vertical hydraulic conductivities are not read.
Leakances are read for multi-layer simulations.
If IVHYC = 1 , vertical hydraulic conductivities are read.
Leakances are computed from vertical conductivity
of each layer as per a volume weighted harmonic
mean.
IANIXY -- is the index for variable anisotropy on a cell-by-cell basis.
If IANIXY = 0, horizontal anisotropy is uniform within each layer
as in the original MODFLOW formulation.
If IANIXY=1, horizontal anisotropy is non-uniform within a layer
and is input on a cell-by-cell basis for all layers.
ICURVL -- is the index for using a curvilinear grid in the areal plane.
If ICURVL =1, a curvilinear grid is used with input of DELR and
DELC required for every node within a layer.
Note that layers are stacked vertically below each
other, hence, DELR and DELC for a node in lower
layers are the same as for the node vertically above
it.
If ICURVL =0, a rectangular grid is used, as in the original
MODFLOW.
IOLDHDR -- is a flag indicating if old header formats (MODFLOW-style) are to be

used for saving cell-by-cell flow or budget terms in the binary file.
If IOLDHDR =1, the old MODFLOW style headers are used for cell-
by-cell flow term output in the binary file.
If IOLDHDR =0, a new header format is used that includes the total
time value (TOTIM) and the stress-period time
value PERTIM in the header for cell-by-cell budget
term outputs.
N_HSK -- is the number of interpolation points used for tabulated input of

moisture retention curves and relative permeability functions for
unsaturated zone layers, when tabular input of these functions is
requested (via IREALSL=5 or 6).
LAYCON -- is the layer-type index array. Each element of LAYCON holds the
identification for the respective layer. Read one value for each layer.
There is a limit of 200 layers. If there are 40 or fewer layers, use one
record; otherwise, use two or more records. Leave unused elements in a
record blank. LAYCON is the parameter through which the new options
of the BCF4 Package may be implemented, as discussed below.
The BFC4 Package supports 20 different input LAYCON values. These values are
input in I2 format. From this value, the first digit (tens), if any, determines the application
of the type of interblock transmissivity function depending on the value of the digit. The
second digit (ones) of the input LAYCON value determines the layer characteristics as in the
original model. The meaning of the second digit is as follows (McDonald and Harbaugh,
1988):
0 - Strictly confined -
Transmissivity and storage coefficient of the layer are
kept constant for the entire simulation.
1 - Strictly unconfined - Transmissivity of the layer varies. It is calculated
from the saturated thickness and hydraulic
conductivity. The storage coefficient is constant in
time. This code is valid only for Layer 1.
2 - Confined/Unconfined - Transmissivity of the layer is kept constant. The
storage coefficient may alternate between confined and
unconfined values.
3 - Confined/Unconfined - Transmissivity of the layer is allowed to vary. It is
calculated from the saturated thickness and hydraulic
conductivity. The storage coefficient may alternate
between confined and unconfined values.
The above description can be summarized as follows (Goode and Appel, 1992):
The second digit of

the input LAYCON Layer Characteristics
value
0 Strictly confined; T and S constant in time
1 Strictly unconfined; T depends on head, S constant in time
2 Confined/Unconfined; T constant in time, S depends on
head because of storage conversion
3 Confined/Unconfined; T and S depend on head
Note that the new variably saturated flow option of the BCF4 Package are
implemented with input LAYCON = 40, 43, 50, 53, or 54 for all the layers in the model
grid. Furthermore, the values of LAYCON=44 or 54 may be used to represent a layer as
a unconfined/confined layer, in an otherwise unsaturated zone water flow simulation (i.e.,
IREALSL …0). Thus, when the second digit of the input LAYCON value is 4, its meaning
is the same as for 3 as discussed above. This new option requires a non-zero value of
IAPART. The flag IAPART, which is read from item 5 of the BAS Package, indicates
whether array BUFF is separate from array RHS.
The factor of ten represented by the first digit of the input LAYCON is stored in the
LAYAVG array. The second digit (ones) is stored in a LAYCON array. The stored values
of LAYCON and LAYAVG for an input LAYCON value are given in the following table:

Input LAYCON Value Stored LAYCON Value Stored LAYAVG Value
0 0 0
1 1 0
2 2 0
3 3 0
10 0 10
11 1 10
12 2 10
13 3 10
20 0 20
21 1 20
22 2 20
23 3 20
31 1 30
33 3 30
40 0 40
43 3 40
44 4 40
50 0 50
53 3 50
54 4 50
**Note that input LAYCON values 30, 32, 41, 42, 51, 52, are not allowed by MODHMS.
The meaning of the LAYAVG (tens digit with a factor of ten) is as follows:
Stored
LAYAVG Value Meaning
0 Harmonic mean interblock transmissivity (McDonald and
Harbaugh, 1988)
10 Arithmetic mean interblock transmissivity (Goode and Appel,
1992)
20 Logarithmic mean interblock transmissivity (Goode and
Appel, 1992)
30 Arithmetic mean saturated thickness times logarithmic mean
hydraulic conductivity (Goode and Appel, 1992)
40 Harmonic mean interblock hydraulic conductivity introduced
in BCF4 Package. Note that this LAYAVG value is to be
used when the rigorous unconfined storage schemes are
implemented.
50 Arithmetic mean interblock hydraulic conductivity. Except
for hydraulic conductivity calculation, LAYAVG=50 is
treated the same as LAYAVG=40.
Thus, the tens part of the input LAYCON value is stored in array LAYAVG, and
the ones part of the input LAYCON is stored in LAYCON array. The stored values of
LAYAVG determines the method of computing interblock transmissivity (when LAYAVG
= 10, 20, 30, 40, or 50).

RHOWP – Density of water.
RHOAP – Density of air at reference pressure.
VISW – Viscosity of water.
VISG – Viscosity of air.
COMPWAT – Compressibility of water.
COMPAIR – Compressibility of air.
ATMGP – Standard atmospheric pressure.
GRAV – Gravitational acceleration.
INITW -- Index for ambient water head input
= 0 when water is at hydrostatic equilibrium with heads computed
from a prescribed water table elevation
=1 when ambient head in water is input for all nodes of the
domain.
Values for RHOWP, RHOAP, VISW, VISG, COMPWAT, COMPAIR,

ATMGP, and GRAV depend on the units used in preparing input data sets. The
following table presents values for these parameters, at standard conditions, in different
sets of units commonly used.
Parameter SI Units cgs Units fps Units

RHOWP 1000 kg/m3 1 g/cm3 1.939 slugs/ft3
RHOAP 1.777 kg/m3 1.777×10-3 g/cm3 .003445 slugs/ft3
VISW 1×10-3 N.s/m2 1.0 cP 6.72×10-4 Rb/ft.s (Rbf.s/ft2)
VISG 1.983×10-5 N.s/m2 1.983×10-2 cP 1.3327×10-5 Rb/ft.s (Rbf.s/ft2)
COMPWAT 1-8 Pa-1 10-9 cm2/dyn 4.788×10-7 ft2/Rbf
COMPAIR 1.777×10-5 Pa-1 1.777×10-6 cm2/dyn 8.5083×10-4 ft2/Rbf
ATMGP 1.01×105 Pa 1.01×106 dyn/cm2 2.1094×103 Rbf/ft2
GRAV 9.8066 m/s2 980.66 cm/s2 32.2 ft/s2
TRPY – is a 1-D array containing an anisotropy factor for each layer. It is the ratio
of transmissivity or hydraulic conductivity (whichever is being used) along
a column to transmissivity or hydraulic conductivity along a row. Set to 1.0
for isotropic conditions. This is a single array with one value per layer. Do
not read an array for each layer; include only one array control record for
the entire array. If the value is the same for all layers, the entire array can
be specified by setting LOCAT to 0 and setting CNSTNT to the value that
applies to all layers. See below for explanation of LOCAT and CNSTNT.
Note: TRPY array is not needed for an axi-symmetric simulation

(IAXIS>0).

TRPYNU – is a 2-D array containing the cell-by-cell anisotropy factors within each
layer. It is the ratio of transmissivity or hydraulic conductivity
(whichever is being used) along a column to transmissivity or hydraulic
conductivity along a row. Set to 1.0 for isotropic conditions. One
array is read for each layer.
Note: TRPYNU array is not needed for an axi-symmetric simulation
(IAXIS>0).
LOCAT – is an index for the location of values to fill-in the array of concern.
<0, read an unformatted record containing the array values
= 0, set all the array values equal to constant (CNSTNT)
>0, read the formatted records containing the array values.
CNSTNT – is a constant to which all array values are set if LOCAT is equal to zero or
by which all array values are multiplied if LOCAT is not equal to zero.
DELR – is the cell width along rows. Read one value for each of the NCOL
columns. This is a single array with one value for each column. For an axi-
symmetric simulation (IAXIS>0), DELR represents radial width ()r) of a
cylindrical cell.
DELRCV – is the cell width along rows for a curvilinear grid. This is a 2-D array
containing one value for each NCOL column, for each of the NROW
rows of a layer.
DELC – is the cell width along columns. Read one value for each of the NROW
rows. This is a single array with one value for each row.
Note: DELC array is NOT read for an axi-symmetric simulation
(IAXIS>0).
DELCCV – is the cell width along columns for a curvilinear grid. This is a 2-D
array containing one value for each of the NCOL column, for each of
the NROW rows of a layer.
ZWT – is the elevation of water table in the domain used as a reference for air
flow calculation.
PWC – is the hydraulic head in the water phase used as a reference for air flow
calculations.
SF1 -- is the primary storage coefficient, the storability. Read only for a transient
simulation (steady-state flag, ISS, is 0). For input LAYCON equal to 1, 11,
21, or 31, SF1 will always be specific yield, while for input LAYCON equal
to 2, 12, 22, 3, 13, 23, 33, 40, or 43, SF1 will always be confined storage
coefficient. For input LAYCON equal to 0, 10, or 20, SF1 would normally
be confined storage coefficient. The primary storage coefficient divided by
the block thickness is equal to the specific storage (Ss) of equations (1) and
(2).
Note (1): An input LAYCON value of 0, 10, or 20 can also be used to
simulate water-table conditions where drawdowns are expected to remain
everywhere a small fraction of the saturated thickness and where there is no
layer above, or flow from above is negligible. In this case, specific yield
values would be entered for SF1.

TRAN – is the transmissivity along grid rows. TRAN is multiplied by TRPY to
obtain transmissivity along columns. TRAN is read only for a layer
represented by one of the following input LAYCON values: 0, 2, 10, 12,
20, 22.
HY – is the hydraulic conductivity along grid rows. HY is multiplied by TRPY
to obtain hydraulic conductivity along columns. HY is read only for a layer
represented by one of the following input LAYCON values: 1, 3, 11, 13, 21,
23, 31, 33, 40, 43.
VHY – is the vertical hydraulic conductivity of each cell. VHY is read only
for a layer represented by one of the following input LAYCON values:
1, 3, 11, 13, 21, 23, 31, 33, 40, 43.
BOT -- is the elevation of the aquifer bottom. Read only for a layer represented by
one of the following input LAYCON values: 1, 3, 11, 13, 21, 23, 31, 33,
40, 43.
VCONT -- is the vertical hydraulic conductivity divided by the thickness, from one
layer to the layer below. The value for a cell is the hydraulic conductivity
divided by thickness for the material between the node in that cell and the
node in the cell below. Because there is not a layer beneath the bottom
layer, VCONT is not specified for the bottom layer.
SF2 -- is the secondary storage coefficient, the specific yield. SF2 is read only if
the simulation is transient (steady-state flag, ISS, is 0) and input LAYCON
value for the layer is one of the following: 2, 3, 12, 13, 22, 23, 33, 43.
Note that SF2 is the porosity when solving for air flow or for Richard’s
equation in the unsaturated zone.
TOP -- is the elevation of the aquifer top. TOP is read only for a layer represented
by one of the following input LAYCON values: 2, 3, 12, 13, 22, 23 ,33,
40, 43.
WETDRY -- is a combination of the wetting threshold and a flag to indicate which
neighboring cells can cause a cell to become wet. If WETDRY < 0, only
the cell below a dry cell can cause the cell to become wet. If WETDRY >
0, the cell below a dry cell and the four horizontally adjacent cells can cause
a cell to become wet. If WETDRY is 0, the cell cannot be wetted. The
absolute value of WETDRY is the wetting threshold. When the sum of
BOT and the absolute value of WETDRY at a dry cell is equaled or
exceeded by the head at an adjacent cell, the cell is wetted. WETDRY is
read only if IWDFLG is not 0 and input LAYCON value is one of the
following: 1, 3, 11, 13, 21, 23, 31, 33.
VANAL -- is the van Genuchten parameter " for unsaturated soils. VANAL is
read only when the input LAYCON value for the layer is 43, and the
real soil moisture functions are requested with IREALSL…0.
VANBT -- is the van Genuchten parameter $ for unsaturated soils. VANBT is

read only when the input LAYCON value for the layer is 43, and the
VANSR -- is the residual saturation level for unsaturated soil. VANSR is read
only when the input LAYCON value for the layer is 43, and the real
soil moisture functions are requested with IREALSL…0.

BROOK -- is the Brooks-Corey exponent for computing relative permeability of
unsaturated soil. BROOK is read only when the input LAYCON value
for the layer is 43, and the real soil moisture functions, with Brooks-
Corey permeability relations are requested with IREALSL=2 or 4.
S_STAR – is the critical saturation parameter S* (see equation 5b of Section 2.2B)

for unsaturated soils with the bimodal conductivity function of Mohanty
et al (1997). S_STAR is read only when the input LAYCON value for
the layer is 43, and the bimodal conductivity function is requested with
IREALSL=7.
D_STAR – is the effective macroporosity fitting parameter (see equation 5b of

Section 2.2B) for unsaturated soils when using the biomodal
conductivity function of Mohanty et al (1997). D_STAR is read only
when the input LAYCON value for the layer is 43, and the bimodal
conductivity function is requested with IREALSL=7.
H_VAL – is the pressure head value for an interpolation point of tabular input for
moisture retention and relative permeability functions. H_VAL is read
only when the input LAYCON value for the layer is 43, and tabular
input of functions is requested with IREALSL=5 or 6.
S_VAL – is the corresponding water saturation value for interpolation point with
associated value of H_VAL. S_VAL is read only when the input
LAYCON value for the layer is 43, and tabular input of functions is
requested with IREALSL=5 or 6.
K_VAL – is the corresponding relative permeability value for the active flowing
phase, for interpolation point with associated value of H_VAL.
K_VAL is read only when the input LAYCON value for the layer is
43, and tabular input of functions is requested with IREALSL=5 or 6.

CHAPTER 3: THE FRACTURE-WELL (FWL4 AND FWL5) PACKAGES
CHAPTER 3: THE FRACTURE-WELL (FWL4 AND FWL5) PACKAGES
3.1 GENERAL
HydroGeoLogic, Inc., added to the U.S. Geological Survey Modular Groundwater
Flow Model, MODFLOW, new well packages that simulate a wellbore as a high-
conductivity fracture tube. These packages, which are an alternative to the existing well
package (WEL1 Package) in MODFLOW, are referred to as the Fracture-Well (FWL4 and
FWL5) Packages. Fracture-Well Packages are designed to overcome several problems
associated with the original WEL1 Package.
First, the FWL4 and FWL5 Packages emulate flow through multilayer wells. This
is because the fracture tube representation allows connection of aquifer cells (nodes) that are
associated with the well with 1-D finite diameter cylindrical well-cells. Additionally, the
total withdrawal rate prescribed for the well is rigorously incorporated. Volumetric fluxes
from each individual node associated with the fracture well are automatically computed by
the code, to sum to the total withdrawal from the well. On the other hand, the original
WEL1 Package implements only a source/sink term at an aquifer node representing a well.
Therefore, it requires the user to determine the contribution of each node to a physical well
that penetrates two or more nodes. These nodal flux contributions are typically estimated
such that the prescribed withdrawal rate is apportioned according to the transmissivities of
the contributing well nodes. Such an estimation might itself be a serious error thus
misrepresenting the physical system. Furthermore, for unconfined systems, prescribing the
nodal fluxes according to the WEL1 Package could cause computational difficulties. The
FWL4 and FWL5 Packages, however, automatically apportion the net well withdrawal
(input by the user) to the well nodes. For each time step, the computed nodal fluxes may
vary substantially, as is the case in the event of desaturation and resaturation of individual
layers.
Another feature of the new FWL4 and FWL5 Packages is that for an overpumped
unconfined system, the total well withdrawal is automatically adjusted when the water level
in the well has reached the well bottom. At this stage, the hydraulic head at the bottom well
node is fixed at well-bottom conditions, and the total withdrawal is computed to correspond
to the feasible quantity that the well can supply under these conditions. If a situation occurs
whereby the withdrawal demand is less than what the well can supply at bottom-hole
pressure conditions (i.e., say the pumping is reduced for a particular period or recharge is
increased), the hydraulic heads at the well nodes will bounce back to reflect these new
conditions, and the well will provide the full withdrawal demand. The previous WEL1
Package does not physically adjust the pumping rate when the water level drops below the
bottom of the well. Instead, it continues the computation into the non-physical realm, with
the head values at the well nodes continually dropping below the well-bottom elevation. On
other occasions, the WEL1 Package causes the system to become computationally unstable
by cyclically drying and wetting the pumping cells between iterations.
The FWL4 and FWL5 Packages also implement well-bore storage, which is
neglected in the previous versions of MODFLOW. The differences between the FWL4 and
FWL5 packages is that the FWL4 package assumes superposition to add the well-flow
equations to those of the porous matrix while the FWL5 package uses a logarithmic well
function to depict flow from the matrix block to the well node. Thus, a refined areal
discretization is still required in the vicinity of each well if the user is interested in accurate
predictions of local drawdowns with the FWL4 package.
Finally, the FWL4 and FWL5 Packages are intended for use only with the rigorous
variably saturated simulation option (i.e., input LAYCON = 40 or 43) of the BCF4 Package.
The following sections of this chapter provide further elaboration on the capabilities of the
FWL4 and FWL5 Packages and include a brief discussion of the formulation and
Chapter 3: The Fracture-Well (FWL4 and FWL5) Packages 3-1

demonstrative example problems. Numerical implementation of the fracture well
formulation is also verified against available analytical solutions in the literature, followed
by input instructions for the respective packages.
3.2 FORMULATION OF FRACTURE-WELL EQUATIONS
A fracture well is formulated for the FWL4 package by superposing 1-D line
elements representing the screened portion of a finite radius wellbore onto the 3-D cells that
represent the porous medium, as shown in Figure 3.1. Sudicky et al. (1995) detail the
derivation and use of fracture well elements and provide demonstrative examples. For the
sake of generality, we adapt their technique to our variably saturated flow formulation
implemented in the BCF4 Package. The formulation is further extended for the FWL5
package where well functions are used to send flow between porous matrix blocks and the
well nodes instead of using the superposition assumption. This is discussed later in the
section. Flow through the tubular fracture for both FWL4 and FWL5 packages is governed
by the 1-D variably saturated flow equation as follows:
(24)
where l is the distance along the well screen (L), is the hydraulic conductivity of the
well ( ), is the relative permeability within the well, is the dimensionless
saturated thickness of the well cell (ratio of saturated thickness to cell thickness), is the
well-bore specific storage (L–1), and is the hydraulic head in the well (L).
Pseudo-soil functions are used to define unsaturated flow behaviors in fracture wells,
as discussed in the BCF4 Package documentation (Chapter 2), such that the water-table
condition within a well cell is accurately represented by vertical equilibrium conditions. The
saturated hydraulic conductivity of the wellbore, which is estimated from laminar flow
theory by applying the Hagen-Poiseuille equation, may be expressed as (Sudicky et al.,
1995):
(25)
where D is the density of water (ML–3), g is the gravitational acceleration , : is the

viscosity of water (ML–1 T –1 ), and rw is the radius of the well (L). Typically, for large
diameter wells, the well hydraulic conductivity values are orders of magnitude larger than
interblock hydraulic conductivities of the porous matrix. Consequently, the simulation
results are insensitive to errors that may result from the use of equation (25).
As mentioned earlier, the FWL4 package uses superposition of the well node with
the porous matrix node. The FWL5 package on the other hand keeps them as separate nodes
with the Theim equation well function providing for flow between the matrix block and its
associated well node. Adapting the concept proposed by Bennett et al (1982) to the fracture-
well formulation for any layer k, the flux between the matrix block and its well node is given
by:
3-2 Chapter 3: The Fracture-Well (FWL4 and FWL5) Packages

Figure 3.1 Physical Situation (a) and Conceptualization (b) of a Fracture Well.
(26)
where Qmwk is the matrix to well flux in layer k; krk is the relative permeability, Kk and
Thickk are the horizontal conductance and thickness of the matrix block, hmk is the head in
the matrix block for layer k, hwk is the head in the well node at layer k, Sk is the skin factor
coefficient and re is the effective radius of the matrix block defined as (Halford and Hanson,
2002; Neville and Tonkin, 2004).
(27)
where )x and )y are the horizontal grid block dimensions of the cells containing the well.
The FWL5 package then seeks simultaneous solution for heads in the well and porous matrix
block nodes, which are distinct from each other; therefore, the porous matrix grid-block
does not require refinement around the wells for better accuracy of well flows, heads and
extraction conditions.
Information required from the user by the FWL4 and FWL5 Packages include: well
locations, identification of the top and bottommost layers in which a particular well is
screened, well bottom elevations, radii of the wells, withdrawal (–ve) or injection (+ve)
rates (L3 T –1 ) from the wells, and well-bore specific storage values (L–1). In addition, the
FWL5 package also requires input of a skin friction coefficient and of initial heads in the
well the instant any well is activated at any stress period. The FWL4 and FWL5 modules
organize the geometric information, compute K l using equation (25), and apply the discrete
fracture-well equation (24) to the respective well nodes. Withdrawals or injections are then
applied to the bottom most screened layer in the wells for every stress period. The
formulation appropriately distributes the stresses among the well nodes. Finally, when
necessary, a switching of boundary conditions at the well nodes (i.e., from prescribed fluxes
to prescribed heads) is performed automatically as described below.
Withdrawal continues at the prescribed rate in each well until the water level in the
well reaches the well bottom. This bottom elevation of a well is physically located anywhere
in the lowest nodal layer that is screened. When the piezometric head at the lowest well
node corresponds to the bottom elevation of a well, the head values are held fixed and the
well yield is allowed to fall below the demand and up to zero. If at a later stage, the aquifer
system recovers and the well yield increases to be larger than the demand, or if the demand
drops below the yield, the well reverts to withdrawing water at the prescribed pumping rate.
The total well yield and contributions of each layer to the well yield are reported in the
output file. In the following sections, illustrative examples are provided to verify and
demonstrate the applications of the FWL4 and FWL5 Packages.
3.3 VERIFICATION EXAMPLE
The Fracture-Well Package (FWL4) option is verified by comparing numerical
results with the analytical solutions obtained by Theis (1935) and Papadopulos and Cooper
(1967) for drawdown in a homogeneous aquifer with a constant discharge well. Unlike the
Theis solution, the solution of Papadopulos and Cooper accounts for well-bore storage.
Hence, two simulations are performed, one with and one without well-bore storage, to test
the numerical scheme used in FWL4.

A verification problem is described by Sudicky et al. (1995). In this problem, a 0.1
m radius fully-penetrating well is located at the center of a homogeneous, isotropic aquifer
of dimensions 1,000 m x 1,000 m x 5 m. The well withdraws water at a rate of 10 – 3 m 3/s.
A constant hydraulic head of 20 m is prescribed at the lateral boundaries, and the top and
bottom boundaries of the aquifer are assumed to be impermeable. The hydraulic
conductivity of the aquifer is 10– 4 m/s and specific storage is 10– 4 m– 1. Transient simulations
are performed using the FWL4 Package.
The domain is discretized into 5 layers of 1-ft thick with 35 rows and 35 columns
of variable grid spacings (from 111.1 ft to 0.1 ft near the well). The top and bottom
elevations of the aquifer are set to 0 and –5 m, respectively. The value of VCONT required
by MODFLOW (vertical hydraulic conductivity divided by inter-layer distance between any
two layers) is calculated as 10– 4 s–1. The primary storage coefficient of the aquifer (specific
storage times block thickness) is specified as 10 – 4. A constant head of 20 m is prescribed
on the lateral sides of the aquifer and a uniform initial head of 20 m is assumed. The 0.1
m radius well is located at the center of the domain (Row 18, Column 18) and is allowed to
pump at a rate of 10– 3 m 3/s for a period of 1000 seconds. As the well is fully penetrating,
the top and bottom of the well are located in Layers 1 and 5, respectively, and the well
bottom elevation is specified as –5 m. In order to compare the temporal variation of
drawdown at the well against the analytical solutions of Theis and Papadopulos, two
transient simulations are performed with the specific well-bore storage equal to zero and 0.1
m –1, respectively. For both simulations, the PCG4 Package is used to solve the matrix
equation. Drawdown versus time curves at the well face are plotted in Figure 3.2 along with
Theis and Papadopulos analytical solutions. As shown in the figure, MODFLOW-
SURFACT/MODHMS results obtained using FWL4 Package are in good agreement with
the respective analytical solutions, thus, verifying the fracture well formulation employed
in FWL4 Package.

Figure 3.2 Verification of the Numerical Solution (FWL4 Package) with the
Analytical Solutions of Theis (1935) and Papadopulos and Cooper (1967).
3.4 APPLICATION EXAMPLES
Two examples are provided to demonstrate application of the FWL4 Package and
a third example demonstrates verification and application of the FWL5 package. The first
example, which illustrates a fully 3-D case, considers a well pumping from an unconfined
aquifer. The second example, which illustrates a quasi 3-D case, considers a well pumping
from three aquifers separated by confining beds. The wells are purposely overpumped for
both cases, with rejuvenation of the aquifers in a second stress period. As an overall check
on the rigor of well boundary condition implementation, the response of the system is also
assessed using an independent simulation that do not invoke the FWL4 Package. For the
fully 3-D case, the head values at the well obtained from the fracture-well simulation are
imposed as time-dependent head conditions for the verification simulation, and the results
are compared with those obtained from original simulations in which the FWL4 Package was
used. The final example illustrates use of the FWL5 package with a quasi 3-D setup for a
well pumping from two aquifers separated by a confining bed. The confined case of this
problem is compared to analytical solutions for verification, and the unconfined case of this
problem demonstrates the drying capability of the package.
3.4.1 Problem 1—Fully Three-Dimensional Simulation of Withdrawal from an

Unconfined Aquifer
This problem considers a 300-ft thick homogeneous, isotropic unconfined aquifer
shown in Figure 3.3. The modeled domain is a square of 75,000 ft x 75,000 ft with top and
bottom elevations of the aquifer at +50 and –250 ft, respectively. Initial groundwater levels
in the aquifer are at zero throughout the domain. The left (west) boundary in the figure is
a constant head boundary and the remaining boundaries are no-flow boundaries. A fully-
penetrating well of diameter 0.4 ft and well-bore specific storage of 10 –6 ft –1 is located at the
center of the domain. Pumping from the well is stopped after operating a 10 ft 3/s capacity
pump for a period of 2.5 years. (See Figure 3.4.) The aquifer is then allowed to recuperate
for 4 years. The aquifer parameters include:
Hydraulic Conductivity (Kxx, Kyy, Kzz) = 10 – 4 ft/s
Specific Yield (Sy) = 0.03
Specific Storage (Ss) = 10 – 6 ft –1
Modeling Approach
The domain is discretized into 3 layers, 15 rows, and 15 columns ()x = )y = 5,000
ft and )z = 100 ft). The value of VCONT (vertical hydraulic conductivity divided by inter-
layer distance between two adjacent layers) is calculated as 10 – 6 s –1. A constant primary
storage coefficient (specific storage times block thickness) of 10–4 is specified for each layer.
A constant head of zero is prescribed on the west side of top and middle discretized layers.
The 0.4 ft diameter well is located at the center of the domain (Row 8, Column 8) and is
screened in all 3 layers to the bottom of the aquifer. The withdrawal demand is 10 ft 3/s for
the first 2.5 years. Note that the Fracture-Well (FWL4) Package automatically adjusts the
pumping rate when the aquifer can not supply the specified withdrawal rate (i.e., when water
level in the well reaches the well bottom). Pumping is stopped in the second stress period
for 4 years to allow aquifer recuperation. As the well is fully penetrating, the top and
bottom of the well are located in Layers 1 and 3, respectively, and the well bottom elevation
is specified as –250 ft.
A transient simulation is performed using the BCF4 (variably saturated flow option
of input LAYCON=43) and FWL4 Packages. The PCG4 Package is used to solve the matrix
equation. The input data files used for the simulation are listed in Appendix C.

Figure 3.3 A Schematic Diagram of the Unconfined Aquifer System for Fully 3-D
Simulation.
Figure 3.4 Pumping Operation in the Well for a Fully 3-D Simulation.
The temporal variations of water level in the well and total well flux are shown in
Figures 3.5 and 3.6, respectively. As shown in Figure 3.5, the water level in the well falls
gradually from its initial value (0) to the well bottom (-250 ft) in 2.15 years. The water level
remains at the well bottom until the pumping is stopped. During the second (recuperation)
period, the water level in the well rises from -250 ft to -17.13 ft. The corresponding
variation in total well flux can be observed in Figure 3.6. During the first stress period, the
total withdrawal from the well decreases from 10 to 7.63 ft 3/s. When pumping ceases, the
total well flux reduces to zero.
Employing the transient head distribution described in Figure 3.5, a prescribed well
head simulation is performed using identical hydrological conditions discussed above, but
without invoking the FWL4 Package. The objective of this simulation is to verify the
accuracy of results obtained in the previous simulation with the FWL4 Package. The
transient head values obtained from the previous simulation are specified for the cell at Layer
3, Row 8, and Column 8 (i.e., the bottom well node) of the unconfined aquifer (Figure 3.3).
To accommodate the prescribed transient head distribution, the cell is treated as a general
head boundary node with a large interface hydraulic conductivity (109 ft/s).
Figure 3.6 compares the simulation results obtained from the verification and
fracture well simulation runs. The two sets of results are in excellent agreement, thus
verifying the accuracy of the FWL4 Package for rigorous treatment of wells with prescribed
pumping rates.
3.4.2 Problem 2—Quasi Three-Dimensional Simulation of Withdrawal from a

Three-Aquifer System
This problem deals with three aquifers separated by two 10-ft thick zero permeability
confining beds (aquicludes), as shown in Figure 3.7. Note that leaky aquitards can readily
be handled using non-zero leakances. The modeled domain is a square of dimensions 75,000
ft. The top and bottom elevations of the three aquifers are as follows:
Aquifer Top Elevation (ft) Bottom Elevation (ft)
Aquifer 1 + 50 – 40
Aquifer 2 – 50 – 90
Aquifer 3 –100 –130
Initial piezometric heads in all aquifers are zero throughout the domain. The left (west)
boundary in the figure is a constant-head boundary with zero head and the remaining
boundaries are no-flow boundaries. A fully penetrating well of 0.4 ft diameter and well-bore
specific storage of 10 – 6 ft –1 is located at the center of the domain. Pumping from the well
is stopped after operating at 5 ft 3/s for a period of three years (Figure 3.8). The aquifers
are then allowed to recuperate for 9 years. The hydraulic parameters for the three aquifers
are as follows:
Hydraulic Conductivity (Kxx, Kyy, Kzz) = 10 – 4 ft/s
Specific Yield (Sy) = 0.1
Specific Storage (Ss) of Aquifer 1 = 1.11 x 10 – 5 ft – 1

Figure 3.5 Temporal Variation of Water Level in the Well for a Fully 3-D
Simulation.
Figure 3.6 Temporal Variation of Total Well Flux for a Fully 3-D Simulation:
Comparison Between Prescribed Flux (FWL4 Package) and Prescribed
Head (GHB Package) Simulation Results.
Figure 3.7 A Schematic Diagram of the Three-Aquifer System for a Quasi 3-D
Simulation.
Figure 3.8 Pumping Operation in the Well for a Quasi 3-D Simulation.
Modeling Approach
Each aquifer in the domain is treated as a nodal layer and the vertical leakage in the
confining beds (aquicludes) is zero. Each aquifer layer is uniformly divided into 15 rows
and 15 columns ()x = )y = 5,000 ft). Layers 1, 2, and 3 are 90, 40, and 30 ft thick,
respectively. The effects of the confining layers on flow are incorporated by specifying the
value of VCONT (vertical leakance between the aquifer layers) as zero. A constant primary
storage coefficient (specific storage times block thickness) of 0.001 is specified for each
aquifer layer. A constant head of zero is prescribed on the west side of the top and middle
discretized layers. A 0.4 ft diameter well is located at the center of the domain (Row 8,
Column 8) and is allowed to pump at a rate of 5 ft 3/s for the first 3 years. Note that the
Fractured-Well Package automatically apportions the total withdrawal to the individual
aquifers. In the event that the total demand is unfeasible (i.e., when the water level in the
well reaches the well bottom), the well withdrawal rate is adjusted to a value that the aquifer
system can supply. Pumping is stopped in the subsequent 9 years to allow aquifer
recuperation. As the well is fully penetrating, the top and bottom of the well are located in
Layers 1 and 3, respectively, and the well bottom elevation is specified as -130 ft. The well-
bore specific storage is 10 – 6 ft –1.
A transient simulation is performed with the BCF4 (variably saturated flow option
of input LAYCON=43) and FWL4 Packages. The PCG4 Package is used to solve the matrix
equation. The FWL4 Package prints the fluxes and head values at each well node at every
time step in addition to the total well flux.
Results and Discussion

Temporal variations of heads at the well nodes are presented in Figure 3.9. Flux
contributions of each aquifer to the well are shown in Figure 3.10. Initially, the contribution
of Aquifer 1 is more than 90% of the total withdrawal from the well (Figure 3.10). This
contribution declines when the head in its well node drops to the aquifer bottom elevation.
Aquifer 2 followed by Aquifer 3 then start contributing more to the applied pumping stress.
When the water level in the well reaches the well bottom (at t = 2.3 yr), the total yield from
the well begins to fall below the specified demand of 5 ft 3/s and decreases to 1.3 ft 3/s at 3
years. When the pumping is stopped in the second stress period (after 3 years), Aquifers
1 and 2 continue to supply water to the well, whereas Aquifer 3 gains water from the well.
Cross-sectional profiles of piezometric heads at the well nodes are presented in
Figures 3.11 through 3.14 for 2.0, 3.0, 4.5, and 7.5 years of simulation, respectively. The
piezometric heads in the three aquifers are presented for an east-west vertical cross-section
passing through the well along Row 8 (see Figure 3.7). As can be seen in Figure 3.11, at
t = 2 yr, the well piezometric levels are –39.98 ft, –89.97 ft, and –116.00 ft for the top,
middle, and bottom aquifers, respectively. Figure 3.12 represents the state of the aquifers
at the end of the first stress period. One and a half years after the pumping is stopped (at
t=4.5 yr), significant recovery of Aquifer 3 is observed because of the recharge it receives
from the two overlying aquifers through the wellbore (Figure 3.13). Figure 3.14, which
presents the piezometric heads at the middle of the second stress period (t=7.5 yr), indicates
that the water level in the well rises to –58.25 ft from the well bottom level of –130 ft. With
the increased piezometric levels, Aquifer 3 is converted to a fully confined aquifer. Aquifer
2 is recharged by incoming water from Aquifer 1 through the well and from the constant-
head boundary to the east. At the end of the second stress period (t=12 yr), water level in
the well rises to –23.21 ft and Aquifer 2 becomes fully confined (see also Figure 3.9).

Figure 3.9 Temporal Variation of Head at Well Nodes in a Quasi 3-D Simulation.

Figure 3.10 Temporal Variation of Well Fluxes of Aquifers for a Quasi 3-D
Simulation.
Figure 3.11 Vertical Profiles of Heads in the Aquifers in the First Stress Period at t=2
yr.
Figure 3.12 Vertical Profiles of Heads in the Aquifers at the End of First Stress
Period (t=3 yr).
Figure 3.13 Vertical Profiles of Heads in the Aquifers in the Second Stress Period at
t=4.5 yr.
Figure 3.14 Vertical Profiles of Heads in the Aquifers in the Second Stress Period at
t=7.5 yr.
3.4.3 Problem 3—Quasi Three-Dimensional Simulation of Withdrawal from a
Two-Aquifer System using the FWL5 package
The problem setup for this simulation is adapted from Neville and Tonkin (2004) and
considers a well that is open across two aquifers. The domain is aerially square with
dimensions of 47,000 ft to a side and with a pumping well in the middle extracting 62,840
ft3/d. No-flow boundary conditions are applied along the lateral edges of the domain. Two
cases are considered with this problem—the first is a confined simulation case which is
compared with the analytical solution of Papadopulos (1966) for verification and the second
is an unconfined simulation case that demonstrates application to well drying conditions.
The top and bottom elevations of the aquifers for the two cases are as follows.
Case 1: Confined Case 2: Unconfined

Top of Aquifer 1 -100 ft 10 ft
Bottom of Aquifer 1 -110 ft 0 ft
Top of Aquifer 2 -120 ft -10 ft
Bottom of Aquifer2 -130 ft -20 ft
Initial heads in the top aquifer are uniformly set at 10 ft. Initial heads in the bottom
aquifer are uniformly set at 30 ft. The hydraulic conductivity and specific storage values are
K=100 ft/d and S=0.0001 for the top aquifer, and K=400 ft/d and S=0.0001 for the
bottom aquifer. The well fully penetrates both aquifers and has a well-bore radius of 0.5 ft.
Modeling Approach
This problem was solved using both FWL4 and FWL5 packages. Each aquifer in
the domain is treated as a separate nodal layer with zero vertical leakage through the
confining bed. The FWL5 simulations use a uniformly spaced grid of 100 columns and 100
rows with a grid spacing of 470 ft. Initial heads in the well are the same as those in the
adjacent layer.
The FWL4 simulations use a variably spaced grid with a finer grid spacing
adjacent to the well. The first case (confined) simulation is conducted till 1.2 days to
exclude comparisons at later times when lateral boundary effects may interfere with
the analytical solution results. The second case (unconfined) simulation is conducted
till 160 days to allow the heads in the well to reach bottom-hole conditions.
The temporal variation of water level and flux at the well for the first (confined)
simulation case are presented in Figures 3.15 and 3.16 respectively. Results for the FWL4
and FWL5 cases match the analytical solution thus verifying the FWL4 and FWL5 packages.
Note that comparisons in the figures are for simulations with the MNW package which
compare very well with the analytical solutions (Neville and Tonkin, 2004). Further, the
FWL4 package provides correct heads and approportioning of fluxes between the two aquifer
layers with a refined grid around the well. Figures 3.17 and 3.18 present the temporal
variation of head and flux at the well for the second (unconfined) simulation case. Results
for the FWL4 and FWL5 package match closely indicating that both packages with different
well approximations provide similar results. Further, both packages can successfully handle
well drying conditions.

Figure 3.15 Temporal Variation of Heads for the Confined Case of Problem 3 to Test
FWL4 and FWL5 Packages.
Figure 3.16 Temporal Variation of Fluxes for the Confined Case of Problem 3 to Test
Figure 3.17 Temporal Variation of Heads for the Unconfined Case of Problem 3 to
Test FWL4 and FWL5 Packages.
Figure 3.18 Temporal Variation of Heads for the Confined case of Problem 3 to Test
3.5 INPUT INSTRUCTIONS FOR THE FWL4 AND FWL5 PACKAGES
3.5.1 General
The Fracture-Well (FWL4 and FWL5) Packages, which are an alternative to the
Well (WEL1) Package, simulate a well as a high-conductivity tubular fracture element. The
FWL4 and FWL5 Packages emulate a well by connecting all nodes that are associated with
the well with 1-D, finite diameter cylindrical well-cells of high conductivity. Conceptually,
a fracture well is formulated in the FWL4 package by superimposing these well elements
representing the screened portion of a wellbore onto the 3-D cells that represent the porous
matrix. Conceptually, the fracture well is formulated in the FWL5 package as separate well
elements representing the screened portion of a wellbore which receive flow according to
a well function (Thiem equation) from the associated 3-D cell that represents the porous
matrix. The advantages of FWL4 and FWL5 Packages over the WEL1 Package include the
following:
• Withdrawals from each individual node associated with the fracture well are
automatically computed by the code, to sum to the total withdrawal from the well
specified by the user. Therefore, there is no need for prior estimation of the
contribution of each node to a physical well that penetrates two or more nodes.
• For overpumped aquifer systems, the water level in the well is maintained at the
well bottom level and the total well withdrawal automatically adjusts to the
quantity that the well can supply at that stage. For multiple aquifers, each
aquifer’s contribution to the well is automatically adjusted to its yield for an
overstressed system. If a situation occurs whereby the withdrawal demand is less
than what the aquifer can yield at bottom-hole pressure conditions (i.e., the well
withdrawal is reduced for a particular pumping period or recharge is increased),
the water level in the well and thus piezometric heads at the well nodes will
bounce back to reflect the new stresses. This function of the FWL4 and FWL5
packages make then useful even for single layer systems or for systems where
wells pump entirely from a single layer.
• The FWL4 and FWL5 Packages account for the well-bore storage. It should be
noted that a finer areal discretization of the simulation domain is still required in
the region of pumping to accurately depict drawdowns in the vicinity of a well for
the FWL4 package, but this is not required for the FWL5 package since
logarithmic functions connecting grid blocks to well cells are not incorporated for
the FWL4 formulation which instead uses superposition.
The FWL4 Package can only be used with the new LAYCON options of the BCF4
Package (i.e., input LAYCON = 40 or 43).
3.5.2 FWL4 Input
Input for the Fracture-Well (FWL4) Package is read from the unit specified in
IUNIT (21). To invoke the FWL4 Package, an input LAYCON value of either 40 or 43 must
be used for all the layers in the model grid.
Input data for the FWL4 Package primarily consists of well locations, top and
bottom layers in which each well is screened, the bottom elevations and radii of the wells,
withdrawal (–ve) or injection (+ve) rates, and well-bore specific storage (L–1). Information
for each well is specified in one record in a given stress period. The number of records in
a stress period depends on the number wells in the domain.
1. Data: MXFWEL IFWLCB FCONST IFREWL4 MUTWL4
Format: I10 I10 F10.0 I10 I10

**Enter items 2 and 3 for each stress period
2. Data: ITMP
Format: I10
**Enter item 3 only if ITMP >0.

3. Data: LSTART LEND Row Col Elev Radius Q Storage
Format: I10 I10 I10 I10 F10.0 F10.0 F10.0 F10.0
(Input item 3 consists of one record for each well. If ITMP is negative or zero, item
3 is not read.)
3.5.3 Explanation of Fields Used in FWL4 Input Instructions

MXFWEL -- is the maximum number of wells in the simulation. These include
unplugged wells that may not be pumping but still act as a vertical
connection between layers.
IFWLCB -- is a flag and a unit number.
If IFWLCB > 0, it is the unit number on which cell-by-cell flow terms
for the well will be recorded whenever ICBCFL (see
Output Control) is set.
If IFWLCB = 0, cell-by-cell flow terms for the well will not be printed
or recorded.
If IFWLCB < 0, well injection rate will be printed whenever ICBCFL is
set.
(If IFBLCB=-1, flux-contribution of each FWL cell
will also be printed.)
FCONST -- is the multiplier used to compute hydraulic conductivity of the well.
FCONST is determined as follows:
KR = FCONST * (Radius)2
where FCONST = (see Equation 2 of the fracture well
formulation)
Radius = radius of the well screen [L];
ρ = density of water [M/L3];
g = acceleration due to gravity [L/T 2]; and
: = dynamic viscosity of water .
The value of FCONST depends on the length and time units used in preparing input
data sets. The following table presents FCONST values in different length and time units:
FCONT Value in dimensions

VALUE OF FCONST FOR VARIOUS LENGTH AND TIME UNITS
Length Time Units

Units Seconds Minutes Hours Days Years
Meters 1.23E06 7.36E07 4.41E09 1.06E11 3.87E13
Feet 3.74E05 2.24E07 1.35E09 3.23E10 1.18E13
If the length and time units used in the simulation are different from those given in the above
table, the equivalent value of FCONST must be computed using the relation given above.
IFREWL4 -- is a flag to read data items 2 and 3 of the FWL4 file in free format.
If IFREWL4 = 0, the FWL4 file is read in default format as noted in the
input instructions.
If IFREWL4 = 1, data items 2 and 3 of the FWL4 file are read in free
format.
MUTWL4 -- is a flag to suppress printout of FWL4 inputs in the output file.
If MUTWL4 = 0, FWL4 input read for every stress period is printed to the
output file.
If MUTWL4 = 1, printing of FWL4 input to the output file at every stress
period is suppressed, warnings would still be printed.
ITMP -- is the number of wells (including all unplugged wells) in a given stress
period.
Note: Unless either a new well is drilled or an old well is plugged, ITMP
remains constant and is equal to MXFWEL. In other words, an existing
well that is not discharging/recharging must be included in the data for each
stress period with its Q value set to zero.
LSTART -- is the layer number of the model cell that contains the top of the well
screen.
LEND -- is the layer number of the model cell that contains the well bottom.
Row -- is the row number of the model cell that contains the well.
Column -- is the column number of the model cell that contains the well.
Elevation -- is the elevation of the well bottom. Physically, this elevation should lie
within the bottom layer (LEND) of the well.
Radius -- is the radius of the well in the same units as any other length units used in
the simulation. Setting the radius of the well to zero effectively disconnects
the well fracture elements, since the fracture well conductivity (K R in
equations (1) and (2)) becomes zero. (The storage of the well also becomes
zero as a result of the well volume being zero).
Q -- is the volumetric flux (L3 T–1) into the well. A positive value indicates
injection of water in the well and a negative value indicates discharge from
the well.
Storage -- is the well-bore specific storage (Sfs ) given by the following formula: Sfs =
, assuming a saturated simulation for the aquifer, where L s

is the total screen length, rc is the radius of standing water pipe (or casing),
and r w is the screen radius. Note that the effect of well-bore storage is
considerable for large diameter wells, and particularly at earlier times of a
simulation. Well-bore storage effects may be neglected by setting the
storage term to zero. For unconfined simulations, the well-bore storage
resulting from desaturation of the well is automatically incorporated for the
well in equation (4).
3.5.4 Fracture Well (FWL5) Package Input
Input for the FWL5 Package is read from the unit specified in IUNIT (37), as
requested by the flow simulation (on file with default extension WL5). Input well properties
for all wells is first provided, consisting of well locations, top and bottom layers in which
each well is screened, bottom elevation, radius of the wells, and well-bore specific storage.
All wells are set to the inactive state at the start of the simulation. Then, for each stress
period, any number of wells may be activated or subsequently de-activated. For each well
that is activated at any stress period, additional information is sought for the pumping rate
of the well (negative for withdrawal and positive for injection), a flag to denote
inclusion/reading of skin-effect conductance, and a flag indicating input of initial head
conditions in a well. Additional respective input is requested if these flags are active. Note
that initial heads in the well need to be input whenever an inactive state is made active to
prevent unphysical values (HDRY) from expressing the pressure state of the well. The
FWL5 package performs these checks and will abort a simulation with the appropriate
message in the output listing file if it detects a well being activated without associated initial
head input. In addition, concentrations of injected fluid are specified for a transport
simulation, for each species. Note that the extra input variables required for transport are
shown in shaded boxes.
1. Data: MXFWEL5 IFWL5CB FCONST5 IFWL5HD IFWL5CC INITH5 ICNTF IFREWL5 MUTWL5
Format: I10 I10 F10.0 I10 I10 I10 I10 I10 I10
Note: ICNTF is not used if ICNTNU = 0 in the ATO package.
2. Data: LSTART LEND Row Column Elevation Radius Storage

Format: I10 I10 I10 I10 F10.0 F10.0 F10.0
** Enter the remaining items (3, 4, 5 and 6) for each stress period.
3. Data: ITMP
Format: I10
** Enter items 4, 5 and 6 only if ITMP >0, for each of the ITMP = NCHANGE number of
wells that are activated or de-activated.
4. Data: LWELL IBOUNDF5 SKINI INHD FQ Conc(I), I=1, NSPECI

Format: I10 I10 I10 I10 F10.0 200F10
** Enter item 5 only if SKINI < 0 (also ITMP and IBOUNDF5 for this well have to be greater
than zero).
5. Data: SKNF (K), K=LSTART, LEND

Format: 8F10.3
Note that LSTART and LEND are starting and ending layers of well LWELL and are defined
for all wells via data item 2.

** Enter item 6 only if INHD = 1 (also ITMP and IBOUNDF5 for this well have to be greater
than zero).
6. Data: HDFWL5 (K), K=LSTART, LEND

Format: 8F10.3
** Enter item 7 only if INHD = 1 and also if ITRAN ≠ 0. (also ITMP and IBOUNDF5 for
this well have to be greater than zero).
7. Data: CONFWL5 (K, LSPEC), K=LSTART, LEND

Format: 8F10.3
Note: Item 7 is to be repeated (for each species) for the LSPEC = 1, NSPECI contaminant
species of simulation.
3.5.5 Explanation of Fields Used in FWL5 Input Instructions
MXFWEL5 -- is the maximum number of wells in the simulation. These include wells
that may not be active (sink or source) but still act as a vertical connection
between layers.
IFWL5CB -- is a flag and a unit number for flow simulation.
If IFWL5CB > 0, it is the unit number on which cell-by-cell flow terms
for the well will be recorded whenever ICBCFL (see
Output Control) is set. Default extension of file is
CW5.
If IFWL5CB = 0, cell-by-cell flow terms for the well will not be printed
or recorded.
If IFWL5CB < 0, well injection rate will be printed whenever ICBCFL
is set.
(If IFWL5CB=-1, flux-contribution of each FWL
cell will also be printed.)
FCONST5 -- is the multiplying factor used to compute hydraulic conductivity of the well
given by:
KR = FCONST5 * (Radius)2
where FCONST =
: = dynamic viscosity of water .
The value of FCONST5 depends on the length and time units used in preparing
input data sets. The following table presents FCONST5 values in different length and time
units:
FCONT5 Value in dimensions

VALUE OF FCONST5 FOR VARIOUS LENGTH AND TIME UNITS
Time Units
Length
Units
Seconds Minutes Hours Days Years
Meters 1.23E06 7.36E07 4.41E09 1.06E11 3.87E13
Feet 3.74E05 2.24E07 1.35E09 3.23E10 1.18E13
If the length and time units used in the simulation are other than those given in the above
table, equivalent value of FCONST5 must be computed using the relation given above.
IFWL5HD -- is a flag and unit number for saving FWL5 heads (and
concentrations for transport) to a separate file .
If IFWL5HD>0 it is the unit number on which a separate file with

default extension FH5 is opened for saving heads (and
concentrations) at all FWL5 nodes.
If IFWL5HD = 0 the FWL5 heads (and concentrations) are saved to

the HDS (and CON) files.
If IFWL5HD < 0 Heads (and concentrations) of FWL5 nodes are

not saved.
IFWLCC -- is a flag and a unit number for transport simulation.

If IFWL5CC > 0, it is the unit number on which mass-flux-terms
of species from wells (Sink/Source) will be recorded whenever
ICCCFL (see Output Control) is set. Default extension of file is
TW5.
If IFWL5CC = 0, mass-flux-terms will not be printed or
recorded.
If IFWL5CC < 0, mass-flux-terms of species from wells
(Sink/Source) will be printed whenever ICCCFL is set.
INITH5 -- is a flag and a unit number for reading initial heads for all FWL5
nodes.
If INITH5 = 0, initial heads are read and assigned based on the
flag INHD in data item 4.
If INITH5 >1, it is the Fortran unit number from which the
FWL5 initial heads are read in binary format. The file and
formats for this case are the same as what is produced from a
previous simulation which may represent steady-state conditions
from which to begin a simulation. With this flag, the rest of the
file need not be changed. Heads read based on the INHD flag of
data item 4, are overwritten by heads read from file on unit
number INITH5 with a default extension of 2DF, or 2DR
depending on the state of the flag IFWL5HD of the previous
simulation.

ICNTF -- is a flag and unit number for reading restart heads (and
concentrations when a transport simulation is performed) from an
external file created by a previous simulation.
If ICNTF = 0, do not look for restart heads or restart
concentrations in a file created by a previous simulation.
If ICNTF > 0, it is the Fortran unit number from which to read
FWL5 heads (and concentrations) in binary format for the restart
simulation. If ICNTF = ICNTNU (restart option flag in ATO
package), then heads (and concentrations) are read from the file
with extension HDC which contains heads from the previous
simulation. Otherwise heads (and concentrations) are read from file
with default extension FHC which contains FWL5 heads (and
concentrations) from the previous simulation. Concentrations are
present in the restart file only if a transport simulation is performed.
IFREWL5 -- is a flag to read data items 2 to 7 of the FWL5 file in free format.
If IFREWL5= 0, the FWL5 file is read in default format as noted
in the input instructions.
If IFREWL5=1, data items 2 to 7 of the FWL5 file are read in
free format.
MUTWL5 -- is a flag to suppress printout of FWL5 inputs in the output file.

If MUTWL5= 0, FWL5 input read for every stress period is
printed to the output file.
If MUTWL5=1, printing of FWL5 input to the output file at
every stress period is suppressed, warnings would still be printed.
LSTART -- is the layer number of the model cell that contains the top of the
well screen.
Elevation -- is the elevation of the well bottom. Physically, this elevation should
lie within the bottom layer (LEND) of the well.
Radius -- is the radius of the well in the same units as any other length units
used in the simulation. Setting the radius of the well to zero,
effectively disconnects the well fracture elements, since the fracture
well conductivity (KR in equations (1) and (2)) becomes zero. (The
storage of the well also becomes zero because of the well volume
being zero). The well action of preventing unphysical
overpumping, however, is still active.
Storage -- is the well-bore specific storage (Sfs ) given by the following
formula: Sfs = (1/Ls) , for a saturated simulation assumption
for the aquifer, where L s is the total screen length, rc is the radius
of standing water pipe (or casing), and is the screen radius.
Note that the effect of well-bore storage is considerable for large

diameter wells, and particularly at earlier times of a simulation.
Well-bore storage effects may be neglected by setting the Storage
term to zero. For unconfined simulations, the well bore storage
resulting from desaturation of the well is automatically incorporated
for the well in equation (4) the FWL4 Package (Volume I) and
Storage represents any compressibility effects of the wellbore.
ITMP -- is a flag and a counter.
If ITMP < 0, Fracture well data from the last stress period will
be reused.
If ITMP $ 0, ITMP is the number of fracture wells (including
abandoned but unplugged) that require input for the
current stress period.
LWELL -- is the well number (entered sequentially in identifying the wells).
IBOUNDF5 -- is the ‘IBOUND’ value of the well indicating if the well is active or
inactive for this stress period.
If IBOUNDF5 > 0, the well is active.
If IBOUNDF5 = 0, the well is inactive.
SKINI -- is an index or value of skin factor for a well.
If SKINI = 0 then all skin factors of the well are set to zero.
If SKINI > 0 then the value of SKINI is used as the skin factor for
all layers of the well.
If SKINI < 0 then the skin factors will be read separately for each
layer of the well.
INHD -- is the index for reading initial heads (or concentrations for
transport) of the well.
If INHD = 0 then do not read initial heads. The well is probably
already active with a change of pumping rate (or concentrations for
transport)or is being deactivated.
If INHD = 1 then read initial head array for all layers of this well
(useful for restart simulation).
If INHD = 2 then the initial head array (or concentrations for
transport) for this well is set to the value of the adjacent porous
matrix block, for all well layers.
If INHD = 3 then the initial head array (or concentrations for
transport) for this well is set to the value of the bottom most porous
matrix block, for all well layers.
Note: INHD should not be zero when an inactive cell is made
active.
FQ -- is the volumetric flux (L3 T–1) into the well. A positive value
indicates injection of water in the well and a negative value
indicates discharge from the well.

Conc(I) -- is the concentration of Species I. The input value is used only if
it is an injection well (i.e., when Q is positive) and denotes the
concentration of injected water. Note that NSPECI is the total
number of species in the transport simulation.
SKNF(K) -- is the value of the skin factor for all layers (from LSTART to
LEND) of the well.
Note: Setting a very high value for the skin factor (1020) effectively disconnects that layer’s
porous matrix block from the well thereby simulating casing effects within the layer.
HDFWL5(K) -- is the initial head value for all layers (from LSTART to LEND) of
the well
CONFWL5
(K,LSPEC) -- is the initial concentration value for all layers (from LSTART to
LEND) of species LSPEC, of the well.

CHAPTER 4: THE RECHARGE-SEEPAGE FACE BOUNDARY (RSF4)
CONDITION PACKAGE
CHAPTER 4: THE RECHARGE-SEEPAGE FACE BOUNDARY (RSF4)
CONDITION PACKAGE
4.1 GENERAL
The MODFLOW code accommodates groundwater recharge boundary conditions
using the Recharge (RCH1) package. However, the RCH1 package has certain limitations
in representing unconfined systems. If an unconfined aquifer saturates to the ground
surface, the capacity of the aquifer to absorb the supplied recharge is reduced, with the
remaining water being shed as surface runoff. The RCH1 Package is incapable of
handling such a situation. Instead, it continues to provide the aquifer with the supplied
recharge (as in confined systems) with heads continually rising well above the ground
surface. The new RSF4 Package of MODFLOW-SURFACT is designed to eliminate this
physically unrealistic condition. The RSF4 Package allows the supplied recharge into the
groundwater system if the water table is below a user prescribed pool (ponding) elevation.
The pool elevation corresponds the ground surface elevation when there is no ponding.
If, at anytime, the water table reaches the pool elevation, the simulation allows only as
much recharge to occur to maintain the prescribed pool conditions. The remaining
recharge is not accepted into the simulation domain. The output of the simulation reflects
this reduction in volumetric recharge flux (L3/T) to the system resulting from the system
saturating up to the ponding height. Note that the RSF4 Package is uncoupled from the
Streamflow-Routing Packages (STR1 or SFR1), and, therefore, this surplus recharge is
not routed into prescribed streams in any manner.
Further, for a MODHMS simulation, the unconfined seepage option conflicts with
the overland flow domain where the domain is active and hence a very high ponding
elevation should be set in these locations to prevent interference of this package with
overland flow, if the unconfined seepage seepage opion is used in the RSF4 package.
An additional capability of the new RSF4 package allows for convenient input of
transient recharge/precipitation independent of the ‘stress period’ conceptualization of
MODFLOW. When this option is used, the recharge time-series may be input via a
separate file that determines how recharge over the domain changes with time regardless
of the stress-period setup, to provide convenience in input and generally smaller input file
sizes, when recharge changes rapidly and often within a simulation period.
In addition to surface recharge-drainage conditions, the RSF4 Package also can
be used to simulate a seepage face boundary condition by prescribing the elevation of a
seepage face boundary node. Thus, zero flux occurs at the boundary, until the water table
builds up to the seepage face elevation. If the water table reaches the seepage face
elevation, the aquifer drains to maintain seepage face elevation (i.e., atmospheric pressure)
conditions. If the water table drops below the seepage face elevation conditions during
the simulation, zero flux is maintained at the seepage face boundary.
The unconfined seepage option of the RSF4 package is intended for use only with
the rigorous variably saturated simulation option (i.e., input LAYCON = 40 or 43) of the
BCF4 package. The capabilities of the RSF4 package are introduced below, including a
brief discussion of the formulation and demonstrative example problems.
4.2 FORMULATION OF THE RSF4 BOUNDARY CONDITION PACKAGE
The RSF4 Package may be applied to any situation involving recharge and
atmospheric seepage face boundary conditions. Input data required by the package
includes the recharge rate, which is stored in a two-dimensional (2-D) array (RECH) with
one element for each horizontal cell location, as in the previous MODFLOW Recharge
Package. In addition, to allow unconfined recharge conditions, the RSF4 Package requires
pool elevations (PNDEL) as input. As long as the water-table elevation (HNEW) of a cell
remains below its pool elevation, the entire recharge rate (RECH) is applied to the
Chapter 4: The Recharge-Seepage Face Boundary (RSF4) Condition Package 4-1

boundary. If, during a nonlinear iteration, the water-table level exceeds the pool level,
the pool level is maintained and the recharge to the system decreases to the maximum
recharge rate acceptable by the system under pool elevation conditions. If, at anytime
during the simulation, the water table falls below the prescribed pool elevation, the full
recharge rate is applied to the system. Further, if the pool elevation is above land surface,
the storage term is augmented by the additional volume of ponded water, using the
modified Picard method (Celia et al., 1990). Ponding is assumed to occur over the full
area of a cell unless the flag IPNDPOR is turned on. In that case, small lakes and ponds
within a grid-block may have their volume accounted (in transient cases) as a fraction of
the grid-block size which is additionally read in the array PNDPOR. The cell-by-cell flow
file for the recharge module further includes budget values for the ponded storage term for
transient simulations. It should further be noted that negative values of ‘recharge’ listed
in the cell-by-cell- flow file or mass budget terms indicate a discharge occurring under
seepage-face conditions, from the ‘RECHARGE’ boundary nodes.
The RSF4 Package also simulates lateral seepage face boundary conditions. The
code provides zero flux to (or from) the system until the water table elevation is above the
seepage face elevation, at which point water discharges from the system at seepage face
elevation conditions, for any node within the domain delineated as a possible seepage-face
node.
4.3 RAPID, TRANSIENT RECHARGE CONDITIONS
Recharge and precipitation are taken to be constant within a stress period of
MODFLOW. For rapid rainfall events, it would then become tedious to break up a short
time period into several stress periods during a simulation. An option has therefore been
provided, to vary recharge or precipitation within a simulation, independent of the stress
periods. The transient recharge is applied zonally, and an index, MXZRCH, is read by
the RSF4 package indicating the maximum number of zones on which transient recharge
is applied. If MXZRCH is zero, the transient recharge package is not invoked by the
simulation. However, if transient recharge exists, an additional integer array is read by
the RSF4 package, identifying the areal location of each zone. A zonal value of zero
indicates that no transient recharge zone occurs over that location, and the default value
of recharge is applied to that location, as read in for the stress period in the RECH array.
At locations where the zonal value is not zero, the default recharge for the stress period
in the RECH array is added to the appropriate value for that zone, from the recharge time
series data supplied in a separate file. This file, with default extension .RTS, contains the
recharge time series tabulated for all zones. Each row of the file contains the starting time
of an event (TSTART), the ending time of an event (TEND), a multiplying factor (for unit
conversions), followed by one column of recharge/precipitation rates for each of the
zones. If a subsequent TSTART value does not match with a TEND value from the
previous row, zero recharge is applied to each of the zones in the interim period. Thus,
long hiatus periods need not be included as a separate row in the .RTS file.
The time periods between events, or within an event need not be fixed, as this
module is designed for use with adaptive time stepping schemes. The adaptive time
marching scheme ensures that each TSTART and TEND is hit exactly by the simulation,
to ensure that the events are accurately represented. Within an event or between events,
the adaptive time stepping schemes try to maximize computational efficiency.
It should be noted that the recharge provided in the RECH array is added to values
supplied in the RTS file to provide a total recharge at any node. Hence, applied irrigation
or other human sources of recharge may be supplied on a stress-period basis via the RECH
array of the RSF package while rapidly varying precipitation events may be supplied via
the RTS file.
4-2 Chapter 4: The Recharge-Seepage Face Boundary (RSF4) Condition Package

4.4 VERIFICATION AND APPLICATION EXAMPLES
Two example problems are provided to demonstrate verification and application
of the RSF4 package. The first problem considers an unconfined aquifer that receives
water from surficial recharge, and discharges into long parallel drains. The second
problem involves an embankment draining along a vertical seepage face. These problems
are compared with analytical and numerical solutions available for the respective
situations.
4.4.1 Problem 1—Flow to Parallel Drains in an Unconfined Aquifer
This problem concerns flow to parallel drains in an isotropic and homogeneous
unconfined aquifer subject to vertical recharge from precipitation as shown in Figure 4.1.
As a result of the 2-D nature of the problem, only one vertical slice may be analyzed. The
distance between drains is 1,000 ft and the depth of the aquifer is 75 ft. The initial
hydraulic head in the aquifer is 50 ft. The domain is subject to a constant precipitation
recharge of 0.01 ft/d. The drains maintain a constant head of 50 ft. Aquifer parameters
include:
Hydraulic conductivity (K) = 1 ft/d
Specific yield, S y = 0.2
Modeling Approach
As described earlier, the problem is 2-D, and only one slice (NROW = 1) of unit
width (DELC = 1 ft) is selected for analysis. Furthermore, a line of symmetry exists
midway between the drains (at x = 500 ft), hence only the first half section of the domain
is discretized as shown in Figure 4.2. The domain is discretized into 10 layers of
thickness 7.5 ft each, and 20 columns of length 25 ft each. The base of the aquifer is at
the datum (0 ft) and the top is at 75 ft. The value of VCONT for each of the layer
(vertical conductivity divided by interlayer distance between two adjacent layers) is 0.1333
d–1. The initial head value and boundary head at the drain node is at 50 ft. Input
LAYCON values of 43 were used for all of the layers, to make use of the new variably
saturated flow formulation, and the SIP package was selected to solve the equations. Two
cases were simulated: Case 1 with a recharge of 0.01 ft/d applied to the domain, and Case
2 with a recharge of 0.02 ft/d applied to the domain.
Results and Discussion

The results of the simulations are shown in Figures 4.3 and 4.4 for Cases 1 and
2, respectively, for different time values. For the Case 1 simulation, the water table
remains below the ground surface and the entire recharge is applied to the domain
throughout the simulation. The steady-state solution for this case is compared with the
analytical solution of Bear (1979) in Figure 4.3 and shows an excellent match. For Case
2, the water table reaches the ground surface in certain parts of the flow domain. Therein,
the water table is then maintained at the pool elevation (ground surface) and the recharge
to the groundwater system decreases, as is reflected in the simulation output.
4.4.2 Problem 2—Seepage Through a Square Embankment
This problem concerns steady seepage through a square embankment. It is chosen
to test the RSF4 Package’s ability to simulate seepage faces. A rigorous (variably
saturated flow) numerical solution to this problem is provided by Cooley(1983).
Geometry, boundary conditions, and isotropic soil parameters are shown in Figure 4.5.
A 10 m x 10 m isotropic embankment with isotropic conductivity properties of K = 0.0l
m/d holds 10 m of water on its left, with a head of 2 m being maintained on the right.
A seepage face, however, may develop along this boundary. Owing to the 2-D nature of
this problem, only a vertical cross-section of the problem is analyzed.

Figure 4.1 Schematic Diagram Depicting Flow to Parallel Drains in an Unconfined
Aquifer.

Figure 4.2 Discretization Scheme Used for the Simulation of Flow through Parallel
Drains.

Figure 4.3 Comparison of Numerical and Analytical Solutions for the Parallel
Drain Problem (Recharge = 0.01 ft/d).

Figure 4.4 Transient Movement of the Water Table for the Parallel Drain Problem
(Recharge = 0.02 ft/d).

Figure 4.5 Steady Seepage through a Square Embankment.

Modeling Approach
The problem configuration is 2-D. Therefore, only one slice (NROW=1) of unit
width (DELC=1 m) is selected for analysis. Vertically, the domain was discretized into
12 columns and 10 layers of 1 m thick. At the left and right boundaries, a column width
(DELR) of 0.1 m was used to capture boundary effects accurately. A head of 10 m is
applied along the left boundary, and a head of 2 m is applied at the bottom two cells on
the right-hand-side boundary. The remaining cells along the right-hand-side boundary are
possible seepage faces and are incorporated into the simulation with the RSF4 Package.
The value of VCONT required by MODFLOW is 0.01 d–1. Input LAYCON values of 43
were used for each layer to use the variably saturated flow formulation of the BCF4
Package, and the SIP package was used to solve the matrix equations. The input data files
used for the simulation are listed in Appendix D.

A steady-state simulation of the problem was performed starting with initial heads
of 10 m. Figure 4.6 shows the simulated water-table elevation and piezometric head along
the base of the domain. These results compare very well with results obtained by using
the finite-difference and finite-element models of Cooley (1983) and Huyakorn et al.
(1986).
4.5 INPUT INSTRUCTIONS FOR THE RSF4 PACKAGE
4.5.1 RSF4 Input
Input to the Recharge Seepage Face (RSF4) Package is read from the unit specified
in IUNIT(8). The additional input variables required by RSF4 are shown in a shaded box.
Note that the new RSF4 Package encompasses the Recharge (RCH1) Package of the
original MODFLOW (McDonald and Harbaugh, 1988).
Input for transient recharge conditions occurs via two files. The first is the
recharge-seepage face package input file on IUNIT (8), which indicates the presence of
transient recharge to the model and number of zones of recharge. The second file contains
the recharge time series (default extension .RTS) for each of the zones. This file is format
free and opened internally when transient recharge is applied.
1. Data: NRCHOP IRCHCB IUNCNF MXSEEP MXZRCH IPNDPOR
Format: I10 I10 I10 I10 I10 I10
**Enter items 2 through 6 for each stress period.
2. Data: INRECH INIRCH INPNDEL INSEEP

Format: I10 I10 I10 I10
**Enter item 3a only if INRECH $ 0.
3a. Data: RECH(NCOL, NROW)
Utility Module: U2DREL. Note that U2DREL is the utility module that
reads two-dimensional array (i.e., RECH for item 3a).
**Enter item 3b only if INRECH > 0 and MXZRCH > 0. The tabulated, zonal recharge
values are added to the default recharge values read above in item 3a.
3b. Data: IZNRCH (NCOL, NROW)

Utility Module: U2DINT

*Enter item 4 only if the recharge option (NRCHOP) is equal to 2 and INIRCH $ 0.
4. Data: IRCH(NCOL, NROW)

**Enter item 5 (both 5a and 5b) only if the unconfined option is used (IUNCNF> 0) and
INPNDEL $ 0.
5a. Data: PNDEL(NCOL, NROW)

**Enter item 5b only if ponding porosity flag is on (IPNDPOR=1)
5b. Data: PNDOR(NCOL, NROW)

**Enter item 6 only if INSEEP > 0.
6. Data: Layer Row Column Elevation

Format: I10 I10 I10 F10.0
(Input item 6 normally consists of one record for each seepage face boundary cell. If
INSEEP is negative or zero, item 6 is not read.)
Explanation of Fields Used in Input Instructions
NRCHOP -- is the recharge option code (flag). Recharge rates are defined in a two-
dimensional array, RECH, with one value for each vertical column.
Accordingly, recharge is applied to one cell in each vertical column, and
the option code determines which cell in the column is selected for
recharge.
1 - Recharge is only to the top grid layer.
2 - Vertical distribution of recharge is specified in array IRCH.
3 - Recharge is applied to the highest active cell in each vertical
column. A constant-head node intercepts recharge and prevents
deeper infiltration.
IRCHCB -- is a flag and a unit number for flow simulation.
If IRCHCB > 0, it is the unit number on which cell-by-cell flow
terms will be recorded whenever ICBCFL (see
If IRCHCB # 0, cell-by-cell flow terms will not be printed or
recorded.
IUNCNF -- is the unconfined option code (flag).
If IUNCNF > 0, recharge seepage face boundary condition is employed
and recharge (RECH) will be applied until water table
reaches the ponding elevations described in PNDEL
array.
If IUNCNF # 0, the specified recharge (RECH) is applied always like
in a confined case.
MXSEEP -- is the maximum number of seepage boundary cells active at one time.

MXZRCH -- is a zonal time-series flag for input of recharge
If MXZRCH = n, recharge varies as a time series, and n is the total
number of recharge zones used in the simulation.
If MXZRCH = 0, MODFLOW’s input structure for recharge is used with
variations only at stress periods.
Note 1: MXZRCH > 0 indicates that recharge varies at a different (possibly

much smaller) time scale than the stress period of MODFLOW, and that
an independent time series for recharge will be provided (in file *.RTS)
for each zone as detailed in Section 4.5.2.
Note 2: A zonal index array identifies the grid blocks which lie within a particular
zone. An index value of zero allows for the default application of
MODFLOW’s recharge for that stress period (as read into the RECH
array).
Note 3: Zonal locations for the recharges in the time series may be varied between
stress periods. Thus, the zones themselves may be varied on a stress
period basis. This is controlled by the flag INRECH, which also controls
reading of recharge rates of each stress period.
IPNDOR -- is a flag for reading the ponding porosity
If IPNDPOR = 0, ponding porosity is not read and the full cell area is
assumed to be ponded.
If IPNDPOR = 1, an array of ponding porosity is read. The ponding
porosity is the fraction of the total cell area that is
wetted by lakes/ponds, and is used to determine the
total ponded storage, if ponding occurs.
INRECH -- is the RECH read flag.
If INRECH $ 0, an array of recharge rates, (RECH) is read.
If INRECH < 0, recharge rates from the preceding stress period are
used.
Note: If MXZRCH > 0, INRECH also controls input array for zonal indexes
(IZNRCH). Hence, zonal locations may be changed at every stress
period.
INIRCH -- is the IRCH read flag. When NRCHOP is two,
If INIRCH $ 0, an array of layer numbers (IRCH) is read.
If INIRCH < 0, the array (IRCH) used in the preceding stress period
is reused.
Note: When NRCHOP is one or three, INIRCH is ignored.
INPNDEL -- is the PNDEL read flag, when the unconfined option is used (IUNCNF
> 0).
If INPNDEL $ 0, an array of ponding elevations (PNDEL) is read.
If INPNDEL < 0, ponding elevations from the preceding stress period
are used.
INSEEP -- is a flag and a counter of seepage boundary cells.
If INSEEP < 0, seepage face boundary data from the preceding stress
period will be reused.
If INSEEP $ 0, INSEEP is the number of seepage face boundaries
during the current stress period.

RECH -- is the recharge rate (L/T). Read only if INRECH is greater than or equal
to zero.
IZNRCH -- is an integrater array identifying the zone value of each node.

Note: The default value of recharge as read into the array RECH at each
stress period will be applied to the cells with a zone value of zero.
IRCH -- is the layer number array that defines the layer in each vertical column
where recharge is applied. Read only if NRCHOP is two and if INIRCH
is greater than or equal to zero.
PNDEL -- is the ponding elevation array that defines the ponding elevation in each
vertical column. Typically, the ponding elevation coincides with ground-
surface elevation if there is no water pooling above the ground surface.
Read only if IUNCNF is greater than zero and INPNDEL is greater than
or equal to zero.
PNDPOR -- is the ponding porosity array that defines the fraction of cell area that is
occupied by ponds/lakes, when water level is above the top of the domain,
up to the ponding elevation PNDEL. This fraction is used in transient
simulations to track volumes/rates of ponded water.
Layer -- is the layer number of the cell representing seepage face boundary.
Row -- is the row number of the cell representing seepage face boundary.
Column -- is the column number of the cell representing seepage face boundary.
Elevation -- is the elevation of the seepage face.
4.5.2 Transient Recharge Time-Series File

The transient recharge time-series file is opened internally on FORTRAN Unit 96,
when transient recharge conditions are applied. The file is in free format and contains one
row for each event within the simulation. Data entries on each row are as follows:
Data: TSTARTR TENDR TMULTR RZ(1) RZ(2)...... RZ(MXZRCH)

Format: Free
Explanation of fields used in input.
TSTARTR - Starting time of a rainfall/precipitation event

TENDR - Ending time of a rainfall/precipitation event
Note that TSTARTR and TENDR are referenced to the total simulation time and not to
the stress period time, when multiple stress periods occur in a simulation.
TMULTR - Multiplying factor applied to recharge rates of all the zones for the given
event. This scaling is convenient for unit conversions.
RZ(1) - Recharge rate in Zone 1 for this event.
RZ(2) - Recharge rate in Zone 2 for this event.
•
•
•

RZ(MXZRCH) - Recharge rate in Zone MXZRCH for this event.
A sample RTS file is shown in Table 4.1. In this example, the domain is divided into 8
zones of recharge. No rainfall or precipitation occurs for the first 2.375 days of
simulation, so the simulation would proceed rapidly till the first recharge event, adapting
its time-step size depending on convergence behavior. At 2.375 days, a rainfall event
occurs over some of the zones, for a one hour period, till 2.417 days. At 2.417 days, the
intensity (and/or zone) of the recharge event changes, with these entries occurring in the
second row of the file, and so on. Note that no rainfall event occurs between 2.5 and 4
days, and an entry is not required in the file for this situation.
Table 4.1 Example input file for recharge time series
TSTARTR TENDR RZ(1) RZ (2) RZ RZ RZ RZ RZ RZ

(d) (d) TMULTR ft/d ft/d (3) (4) (5) (6) (7) (8)
2.375 2.417 1.0 0 0 0.225 0.177 .033 0 0 0
2.417 2.458 1.0 0 0 0 0.13 0.13 0.022 0 0
2.458 2.500 1.0 0 0 0 0 0.088 0 0 0
4.000 4.042 1.0 0.169 0.551 0.551 0.337 0.178 0 0 0
4.042 4.083 1.0 0.151 0.447 0.333 0.106 0.08 0 0.691 0.633
4.083 4.125 1.0 0.100 0.050 0 0 0 0.150 0.379 0.300
4.125 4.167 1.0 0 0 0 0 0 0 0.083 0.022
Note: MXZRCH = maximum number of recharge zones = 8

CHAPTER 5: THE ADAPTIVE TIME-STEPPING AND OUTPUT
CONTROL (ATO4) PACKAGE
CHAPTER 5: THE ADAPTIVE TIME-STEPPING AND OUTPUT
CONTROL (ATO4) PACKAGE
5.1 GENERAL
In the MODFLOW code, the time domain (for transient flow simulations) is
discretized using a backward-difference formulation, with time-step values predetermined
by specifying the following parameters: the number of time steps (NSTP), the length of
the stress period (PERLEN), and a multiplying factor (TSMULT) which increments the
time-step size using a geometric series. Limitations of using such a scheme are listed
below.
• If the matrix solver or the non-linear iterative scheme (for unconfined situations)
fails to converge in a given time step, the computation is aborted.
• The pre-selected time step sizes may be inefficient to the overall solution, even
when converged numerical results are attained.
• The geometric progression of time-step sizes must be maintained. As a result, the
user lacks flexibility in assigning specific time values to examine the flow system.
Therefore, output can only be requested at time values that lie on the time
progression series, which needs to be determined by the user in advance.
To overcome the above difficulties and enhance efficiency of solution, an adaptive
time-stepping and output control package (ATO4) has been developed. The adaptive time-
stepping scheme selects a time-step size depending on the anticipated non-linearities of the
system for a given calculation. If the anticipated non-linearities are not significant, a
larger time-step size is selected to aggressively move forward with the simulation. If
anticipated non-linearities are severe, a smaller time-step size is selected to ensure
convergence for that time step. In the event that the solution fails to converge for a given
time step, the time-step size is further reduced, and the solution is repeated.
Time values of simulation results are not known apriori for an automatic time-
stepping scheme; therefore, the ATO4 Package also includes new modules for output
control, which determine the time values at which output is desired. With the ATO4
Package invoked, the use of the previous Output Control (OC) Package of MODFLOW
becomes unnecessary. Note that the output control of the original MODFLOW is tedious
and cumbersome to implement. Output control for the new (ATO4) scheme, on the other
hand, is simple and straightforward. The user only needs to input the time values (in time
units used previously) at which output is desired for any given stress period, and the detail
of the requested output (i.e., whether one or all of heads, drawdowns, mass budgets, and
cell-be-cell flow terms are to be output or not). The ATO4 Package automatically
recognizes the selected time values for output and adjusts the time-step size to perform
computations at these print times. Thus, for example, a user can examine the results after
10, 11, and 15 months of a 15 month stress period simulation. The ATO4 Package will
not only provide efficient time stepping throughout the simulation, but it also will ensure
that computations are performed at 10, 11, and 15 months, the time values for which the
output is requested.
Finally, the ATO package provides options for restart of a simulation from a non-
zero starting time value. The OC package of MODFLOW does not allow for restarts and
it can get extremely tedious to develop new input files for performing restarts specially
when multiple stress-periods are involved.
Chapter 5: The Adaptive Time-Stepping and Output Control (ATO4) Package 5-1
5.2 FORMULATION OF ADAPTIVE TIME-STEPPING USED IN THE ATO4
PACKAGE
The adaptive time-step size selection scheme implemented in the ATO4 Package
is simple and effective. The user supplies an initial time-step size (DELT), a maximum
time-step size (TMAX), a minimum time-step size (TMIN), a time-step enlargement factor
(TSMULT), and a time-step reduction factor (TSDIV) for each stress period of length
PERLEN. The initial time-step size, DELT, is used for the first time-step of the stress
period. If convergence is achieved within 35% of the maximum number of iterations
(MXITER), the time-step size is increased by the multiplying factor TSMULT. For a
convergence rate between 35% and 65% of MXITER, the time-step size is unaltered from
the previous DELT, and the convergence behavior of the system is considered optimal for
this value of DELT. If, however, the number of iterations required to achieve
convergence exceeds 65% of MXITER, the time-step size is reduced by a dividing factor
TSDIV. If convergence is not achieved for a particular time-step, the time-step size is
reduced by a factor 5.0 and computations are redone for the smaller time-step size. In
addition, if a print time (i.e., the target time value at which the output printing/saving is
required) is reached, the time-step size is selected to include the print time value. If the
value of DELT is reduced to be smaller than TMIN, the simulation is aborted. This
technique provides a means of exiting the program simulation if severe difficulties are
encountered. Finally, the value of DELT is constrained to be less than or equal to TMAX.
This constraint is significant for accuracy of time discretization, even when the non-
linearities are mild or insignificant. The user is provided with the flexibility of printing
either at specified times using the TMVEC array or after a specified number of time-steps
(NPSTP), or both.
A transient simulation of pumping in an unconfined aquifer is performed to verify
the ATO4 Package and demonstrate its implementation. The selected problem is depicted
in the BCF4 Package writeup, wherein the simulation is performed using the original
MODFLOW’s time-stepping and output control. A comparison of these two simulation
runs is made in terms of computational efficiency and robustness.
The transient analysis presented in this example considers a 300-ft thick
unconfined aquifer, shown in Figure 5.1. The modeled domain corresponds to a square
of dimensions 75,000 ft x 75,000 ft. The top and bottom of the aquifer are at elevations
of 50 and -250 ft, respectively. Initial groundwater levels in the aquifer are at zero
throughout the domain. The domain is subject to a uniform and continuous recharge of
3.0 x 10–9 ft/s. The left (west) boundary in the figure is a constant (zero) head boundary
with the remaining boundaries being no-flow boundaries. Fifteen wells are screened over
100 ft of the aquifer bottom and located as shown in Figure 5.1. Each well is pumped at
a high rate of 3.85 ft 3/s. Pumping from the wells is stopped after four years. (See Figure
5.2.) The aquifer is then allowed to recuperate for 12 years. The aquifer parameters are:
Horizontal hydraulic conductivity, K xx. Kyy = 10–4 ft/s
Vertical hydraulic conductivity, K zz = 10–5 ft/s
Specific Yield, S y = 0.03
Specific Storage, S s = 10–6 ft–1
5-2 Chapter 5: The Adaptive Time-Stepping and Output Control (ATO4) Package
Figure 5.1 A Schematic Diagram of the Unconfined Aquifer System.
Figure 5.2 Applied Pumping Stress at Each Well for Transient Analysis of the
Unconfined Aquifer System.
Modeling Approach
The domain is discretized into 3 layers, each 100 ft thick with 15 rows and 15
columns of a uniform block of 5,000 ft areally. Thus, the value of VCONT required by
MODFLOW (vertical hydraulic conductivity divided by inter-layer distance between any
two layers) is calculated as 10–7 s–1. A constant primary storage coefficient (specific
storage times block thickness) of 10–4 is specified for each layer. A constant head of zero
is prescribed on the west side of top and middle discretized layers. Initially, the head in
the domain is assumed to be zero. Fifteen wells located in Layer 3 are allowed to pump
at a rate of 3.85 ft 3/s over the first 4 years. Pumping is stopped during the subsequent
period of 12 years, to allow aquifer recuperation.
Two separate transient simulation runs are performed, using (1) the existing
MODFLOW time-stepping and output control, and (2) using the new automatic time-
stepping and output control (ATO4) package. These two simulations are referred as Case
(1) and Case (2), respectively. The unconfined flow option of input LAYCON = 43 was
selected for both cases. The SIP Package is used to solve the matrix equations for both
cases. Time-step control parameters for Case (1) include: MXITER = 50, DELT =
7.875 x 10 6 s, and TSMULT = 1. Time-step control parameters for Case (2) include:
MXITER = 50, DELT = 7.875 x 10 6 s, TSMULT = 1.7, TSDIV = 2.0, TMAX = 2
x 10 7 s, and TMIN = 8 x 10 5 s. The input data files using the ATO4 Package [Case (2)]
are listed in Appendix E.
This problem has been examined previously, in the context of the BCF4 package.
This section focuses on comparison of the old and new time-stepping schemes, as applied
to the example problem. Figures 5.3 and 5.4 compare head distributions in layers 1 and
3 respectively, at the end of the pumping period (t=4 yr) for both cases, and Figures 5.5
and 5.6 compare head distributions in layers 1 and 3 respectively, at the end of the
simulation period (t=16 yr). Comparison of the results is excellent, and the ATO4
package performance is verified.
The number of nonlinear iterations required to complete the simulation were 544
iterations for Case 1, and 412 iterations for Case 2. The CPU times for the Case 1 and
Case 2 simulations were 33 seconds and 19 seconds, respectively. The ATO4 package
automatically selects the optimal time-step size for Case 2, thereby making the simulation
more efficient than for Case 1. Note that the simulation for Case 1 was aborted (due to
failure of convergence) if a TSMULT value of 1.7 was used to increment the time-step
size, as was done in Case 2. Further, if the maximum iteration count were more
restrictive (say MXITER=25 instead of 50), Case 1 simulations would abort, while Case
2 simulations would cut the time-step size for unsuccessful iterations, and proceed until
the simulation is completed.
Finally, note that the ATO4 Package used in Case 2 allows for flexibility in
output, and the user may demand output at 3.2 years and 4 years of simulation as is noted
in the output file. The simulation will perform optimally (using large time-step values
possibly exceeding 0.8 years) until it comes to 3.2 years, upon which the simulation solves
the system at both 3.2 and 4 years. The previous MODFLOW versions do not allow this,
and predetermining the time-step at which one wishes to have output, can be tedious.
Figure 5.3 Head Distributions in Layer 1 at the End of Pumping (First) Period
(t=4 yr) Showing Comparison between the Previous and New Time-
Stepping Schemes.
Figure 5.4 Head Distributions in Layer 3 at the End of (First) Stress Period (t=4
yr) Showing Comparison between the Previous and New Time Stepping
Schemes.
Stepping Schemes.
Stepping Schemes.
5.4 INPUT INSTRUCTIONS FOR THE ATO4 PACKAGE
5.4.1 General
Input to the ATO4 Output Control is read from the unit specified in IUNIT (22).
The previous output option, which is specified in IUNIT (12), is ignored if the new ATO4
Package is activated. If both IUNIT (12) and IUNIT (22) are zero, no output control data
are read, and default output control is used. Under the default, head and total budget are
printed at the end of every stress period (McDonald and Harbaugh, 1988). Similar to the
old output control package, all printer output goes to unit 6 as specified in the main
program. If necessary, the unit number for printer output can be changed to meet the
requirements of a particular computer.
The ATO package further includes flags and parameters to restart (or continue)
a simulation from a non-zero time value. This capability is active when the variable
TCNTNU is non-zero and represents the time value (in simulation units of time) from
which the simulation is restart. The first time-step size value of the restart simulation is
also then input in variable DTCNTNU (in simulation units of time). When this is the case,
the simulation will wind itself up to time TCNTNU (appropriately winding stress period
information on boundary conditions and time values of the recharge or ET time-series
files) and proceed from that point onwards with the first time-step size being DTCNTNU.
Initial heads (and concentrations for transport) as well as values of initial storage in
interception in the IPT package and initial values of the ON/OFF switch of regulated
structures in the CHF package (for MODHMS) are obtained from the respective initial
conditions as input for the simulation. Further flags are available for direct input of
starting heads from the binary HDS file (and/or starting concentrations from the CON file)
of a previous MODHMS simulation for the same problem. When the flag (and unit
number) ICNTNU is set to a positive number, the file containing starting heads (all binary
head outputs of a previous run) is opened on the specified unit number (and with default
extension of HDC—which may be obtained by renaming the previous simulation’s HDS
file to extension HDC) and the file is wound up till the specified time (TCNTNU) is
reached. The head values (in all simulated domains) available in the file at this time value
(or greater, if data is not available for time TCNTNU) are used for starting the current
restart simulation. Similarly, when the flag and unit number, ICNTNUC is set to a
positive value, and when a transport simulation is performed, the file containing starting
concentrations is opened on the specified unit number (and containing default extension
of COC—which may be obtained by renaming the previous simulation’s CON file to
extension COC). This file is then wound up to the time value of TCNTNU (or greater)
to obtain initial concentrations for all simulation domains (which include immobile and
mobile domains in the subsurface for dual porosity cases, and, the OLF surface and the
CHF domain as appropriate) and all species of the restart simulation.
5.4.2 ATO4 Input
1. Data: IHEDFM IDDNFM IHEDUN IDDNUN ICONFM ICONUN

Format: I10 I10 I10 I10 I10 I10
TCNTNU DTCNTNU ICNTNU ICNTNUC IUNITMB

F10.3 F10.3 I10 I10 I10
2. Data: INCODE IHDDFL IBUDFL ICBCFL ICONFL ICONBD ICCCFL
Format: I10 I10 I10 I10 I10 I10 I10
3. Data: DELT TMIN TMAX TSMULT TSDIV NPRTS NPSTP
Format: F10.0 F10.0 F10.0 F10.0 F10.0 I10 I10
**Enter item 4 only if NPRTS > 0.
4. Data: TMVEC(NPRTS)
Format: 8F10.0 (number of records = NPRTS/8 + 0 or 1)
**Enter item 5 only if INCODE $ 0.
5. Data: Hdpr Ddpr Hdsv Ddsv Conpr Consv
Format: I10 I10 I10 I10 I10 I10
(Record 5 is read 0, 1, or NLAY times, depending on the value of INCODE.
IHEDFM – is a code for the format in which heads will be printed.
IDDNFM – is a code for the format in which drawdowns will be printed. Format
codes have the same meaning for both head and drawdown. A positive
format code indicates that each row of data is printed completely before
starting the next row. This means that when there are more columns in
a row than will fit on one line, additional lines are used as required to
complete the row. This format is called the wrap format. A negative
format code indicates that the printout is broken into strips where only
that number of columns that will fit across one line are printed in a strip.
As many strips are used as are required to print the entire model width.
This format is called the strip format. The absolute value of the format
code specifies the printout format as follows.
0 - (10G11.4) 3 - (15F7.1) 6 - (15F7.4) 9 - (20F5.2) 12 - (10G11.4)

1 - (11G10.3) 4 - (15F7.2) 7 - (20F5.0) 10 - (20F5.3)
2 - (9G13.6) 5 - (15F7.3) 8 - (20F5.1) 11 - (20F5.4)
Note if IREALSL…0, the drawdown is a meaningless parameter and this switch

is associated with printout of saturations of the active phase.
IHEDUN -- is the unit number to which heads will be written if they are saved on
disk.
IDDNUN -- is the unit number to which drawdowns will be written if they are saved
on disk.
Note if IREALSL…0, the drawdown is a meaningless parameter and this switch
is associated with printout of saturations of the active phase. Thus, active phase
saturations are printed to the DDN files when IREALSL…0.
ICONFM - is a code for the format in which concentrations are to be printed.

- The format codes are listed above (see the explanation of
IDDNFM).
ICONUN – is the unit number to which concentrations of each species are to be

written if they are saved on disk.
TCNTNU – Flag and time value for continuation of a previous simulation. If
TCNTNU = 0, the simulation is NOT a continuation of a previous
simulation. If TCNTNU > 0, the simulation continues from a previous
simulation, from the time value provided for this input.
DTCNTNU - time-step size value used for the continuation simulation from time value
of TCNTNU. If the simulation is not a continuation simulation, this
variable is not required/used.
ICNTNU – Flag and unit number for reading restart heads from an external file
created by a previous simulation.
If ICNTNU = 0, do not look for restart heads in a file created by a
previous simulation.
If ICNTNU > 0, it is the Fortram unit number from which to read heads
for the restart simulation (on file with extension HDC). The file will be
wound to the appropriate time value (TCNTNU) to obtain the required
starting heads.
ICNTNUC – Flag and unit number for reading restart concentrations from an external
file created by a previous simulation (when a transport simulation is
performed).
If ICNTNUC = 0, do not look for restart concentrations in a file created
by a previous simulation.
If ICNTNUC > 0, it is the Fortram unit number from which to read
concentrations for the restart simulation (on file with extension COC).
The file will be wound to the appropriate time value (TCNTNU) to obtain
the required starting concentrations.
IUNITMB -- Flag and unit number for saving nodal mass balance terms to a binary file.
If IUNITMB # 0 Node-by-node flux balances are not saved.
If IUNITMB > 0 Node-by-node flux balances are saved to a binary file
with default extension NFB and the value of IUNITMB is the Fortran Unit
number on which the file is written.
INCODE -- is the head/drawdown output code. It determines the number of records
in input item 5.
If INCODE < 0, layer-by-layer specifications from the last time steps
are used. Input item 5 is not read
If INCODE = 0, all layers are treated the same way. Input item 5 will
consist of one record.
If INCODE > 0, input item 5 will consist of one record for each layer.
IHDDFL – is a head and drawdown output flag.
If IHDDFL = 0, neither heads nor drawdowns will be printed or saved
on disk.
If IHDDFL … 0, head and drawdowns (or saturations) will be printed
or saved according to the flags for each layer
specified in input item 5.
IBUDFL -- is a budget print flag for flow simulation.
If IBUDFL < 0, overall volumetric budget will be printed every “n”
timesteps where n=abs (IBUDFL)
If IBUDFL = 0, overall volumetric budget will not be printed.
If IBUDFL > 0, overall volumetric budget will be printed as specified
by either NPRT or NPSTP in Item 5 input.
(Note that the overall volumetric budget will always be printed at the end
of a stress period, even if the value of IBUDFL is zero.)
ICBCFL -- is a cell-by-cell flow-term flag.

If ICBCFL = 0, cell-by-cell flow terms are not printed or saved.
If ICBCFL … 0, cell-by cell flow terms are printed or recorded on
disk depending on flags set in the component of flow
packages, i.e., IWELCB, IRCHCB, etc.
ICONFL – is a concentration output flag.

If ICONFL = 0, concentrations are not printed or saved on
disk.
If ICONFL … 0, concentrations of each species are printed or

saved according to the flags for each layer
specified in input item 3.
ICONBD – is a budget print flag for transport simulation.

If ICONBD = 0, transport budget is not printed.
If ICONBD … 0, transport budget is printed
(Note that the transport budget is always printed at the end of a

stress period, even if the value of ICONBD is zero.)
ICCCFL – is a cell-by-cell mass-flux-term flag for transport simulation.

If ICCCFL = 0, transport mass-flux-terms are not saved or
printed.
If ICCCFL … 0, transport mass-flux-terms are printed or

recorded on disk depending on flags set in
other packages, i.e., IWELCC, IRCHCC,
etc.
DELT -- is the first time step size of the stress period.
TMIN -- is the minimum value of time step size allowed in the stress period.
TMAX -- is the maximum value of time step size allowed in the stress period.
TSMULT – is the time step multiplier. TSMULT should NOT be less than 1.0. This
value overwrites the TSMULT value read in item 9 of Basic Package
Input.
TSDIV -- is the reduction factor of time step size. TSDIV must be greater than 1.0.
NPRTS – is the number of print times specified in item 4. If no print time is
specified (i.e., NPRTS=0), overall volumetric budget and heads/
drawdowns will be printed or saved at the end of the stress period
according to the flags specified in input item 5. If NPRTS=0, item 4
must be skipped.
NPSTP – is an output printout control parameter. NPSTP=n means that the overall
volumetric budget, heads/drawdowns, and cell-by-cell flow terms will be
printed or saved at every nth time step of the stress period according to
the flags specified in input items 2 and 5. Set NPSTP equal to zero if the
step wise printing/saving is NOT required.
TMVEC -- is a 1-D array containing NPRTS number of print time values at which
printing or saving of volumetric budget, heads/drawdowns, or cell-by-cell
flow terms is desired in a particular stress period. The printing/saving is
performed according to the flags specified in input items 2 and 5. Time
is assumed to be zero at the beginning of a stress period. The print times
must be specified in ascending order of magnitude and the difference
between two successive values must be at least TMIN.
Note: This item must be skipped if NPRTS=0.
Hdpr -- is the output flag for head printout.
If Hdpr = 0, head is not printed for the corresponding layer.
If Hdpr … 0, head is printed for the corresponding layer.
Ddpr -- is the output flag for drawdown printout.
If Ddpr = 0, drawdown is not printed for the corresponding layer.
If Ddpr … 0, drawdown is printed for the corresponding layer.
Hdsv -- is the output flag for head save.
If Hdsv = 0, head is not saved for the corresponding layer.
If Hdsv … 0, head is saved for the corresponding layer.
Ddsv -- is the output flag for drawdown save.
If Ddsv = 0, drawdown is not saved for the corresponding layer.
If Ddsv … 0, drawdown is saved for the corresponding layer.
Conpr -- is the output flag for concentration printout.
If Conpr = 0, concentration is not printed for the corresponding layer.
If Conpr … 0, concentration is printed for the corresponding layer.
Consv -- is the output flag for concentration save.
If Consv = 0, concentration is not saved for the corresponding layer.
If Consv … 0, concentration is saved for the corresponding layer.
CHAPTER 6: THE PRECONDITIONED CONJUGATE GRADIENT (PCG4)
PACKAGE
CHAPTER 6: THE PRECONDITIONED CONJUGATE GRADIENT (PCG4)
PACKAGE
6.1 GENERAL
The MODFLOW code generates a system of equations describing the groundwater
flow system in a finite-difference form. This set of algebraic equations is normally solved
by using either a Slice Successive Over-Relaxation (SSOR) or a Strongly Implicit
Procedure (SIP) Package. For complex field simulations, efficiency and robustness may
be lacking in these solution routines. A Preconditioned Conjugate Gradient (PCG2)
Package is provided by Hill (1994) to provide MODFLOW with an alternative solver.
The preconditioning schemes in PCG2 are the least squares polynomial preconditioner
presented by Saad (1985) or the optimal Chebyshev polynomial preconditioner of Meyer
et al. (1989). These schemes are chosen primarily on computer storage considerations and
often perform poorly on large-scale field studies. Recognizing this point, we incorporate
the PCG4 Package which contains a PCG solver based on partial LU decomposition as a
preconditioner. Computer storage requirements are more than those required by the
previous solvers in use with MODFLOW. However, the PCG4 solver is simple, robust
and efficient. The solver has been used on many groundwater modeling projects of
various sizes and complexities. The remaining portions of this chapter are devoted to
formulation of the PCG4 package and input instructions for its use.
6.2 FORMULATION OF STANDARD PCG SCHEME
The Preconditioned Conjugate Gradient (PCG) method (see Kershaw, 1978) has
emerged as one of the most promising iterative methods for solving large sparse symmetric
matrix equations generated by finite element or finite difference approximations of multi-
dimensional field problems. A number of recent papers describe successful conjunctive
applications of PCG schemes in conjunction with finite-element and finite-difference
methods to groundwater flow problems. These include publications by Gambolati (1980),
Kuiper (1981, 1987), Gambolati et al. (1986), Kalf (1988), and Meyer et al. (1989).
The symmetric PCG solver implemented into the PCG4 Package of MODFLOW
is based on a two-step procedure developed by Meijerink and van der Vorst (1977) and
extended and implemented by Kershaw (1978) and Anderson (1983). It is used to solve
the system of algebraic equations resulting from the Picard scheme of the finite element
flow formulation. The first step of this PCG algorithm involves preconditioning the finite
difference coefficient matrix A of a general symmetric matrix system Ax = y by
computing an incomplete triangular Cholesky decomposition (LDLT) of A. The second
step involves Conjugate Gradient iterations of an initial guess x0 of the vector x. The CG
recursive equations are as follows:
(26a)
(26b)
Chapter 6: The Preconditioned Conjugate Gradient (PCG4) Package 6-1

Then, for i = 0,1,2,....iter1, we have
(26c)
(26d)
(26e)
(26f)
This scheme is repeated until the L 2 norm or the L 4 norm of the head error is less than
the prescribed tolerance criteria. A general sparse matrix storage scheme is employed for
which the matrix structure is n×R where n is the number of nodes in the system and R is
the global connectivity structure of the discretized system of equations (R is 7 for a 3-D
finite difference structure and 5 for a 2-D finite difference structure).
6.3 INPUT INSTRUCTIONS FOR THE PCG4 PACKAGE
6.3.1 PCG4 Input
Input required by the PCG4 Package is similar to that of the PCG2 Package
implemented in MODFLOW by Hill (1994). The PCG4 iterative solver is a simple
solution scheme requiring only closure tolerance limits and a maximum limit of solver
iterations as input. There are no iteration parameters to be estimated, nor other choices
to be made by the user. A solver-print-control-flag and maximum outer iterations (>1
for unconfined systems) are also requested as input.
Input for the PCG4 Package is read from a unit specified in IUNIT (13) of the
Basic Package input file. The input is as follows:
1. Data: MXITER ITERI INEWT BFACT RESRED IDMPBOT

Format I10 I10 I10 F10.0 F10.0 I10
2. Data: HCLOSE MUTPCG MNORTH RCLOSE

Format F10.3 I10 I10 F10.3
Note that input parameters in shaded boxes are required only if Newton Raphson
linearization is used as discussed in Chapter 7.
MXITER -- is the maximum number of outer iterations—that is, calls to the solution
routine. For a linear problem MXITER should be 1, unless more than 50
inner iterations are required, when MXITER could be as large as 10. A
larger number (generally less than 100) is required for a nonlinear
problem.
ITERI -- is the maximum number of inner iterations. For nonlinear problems,
ITERI usually ranges from 6 to 600, depending on the problem, and may
be set at 600 with closure controlled by the tolerance limits.
6-2 Chapter 6: The Preconditioned Conjugate Gradient (PCG4) Package

INEWT -- is the Newton-Raphson linearization index. If INEWT = 0 the old
modified Picard MODFLOW schemes are used. If INEWT = 1
then Newton-Raphson linearization is performed for unconfined or
unsaturated cases.
BFACT -- Factor by which backtracking is performed if residual reduction

criterion is not met. BFACT must be greater than zero and less than
one.
RESRED -- Residual reduction factor to force backtracking. RESRED must be

greater than zero. If RESRED is less than one, the backtracking
algorithm forces a residual reduction at every Newton iteration. For
RESRED greater than unity, increases in residual are allowed by the
prescribed factor to allow the Newton algorithm to move easily out
of local minima. As RESRED increases, there are less chances for
backtracking. The user should determine which value of this
parameter gives optimal convergence for the particular problem.
See the NRB1 package documentation in Chapter 7 for suggested
usage.
Note that BFACT and RESRED are required only when INEWT = 1.
IDMPBOT -- is an index for using aquifer geometry to dampen head oscillations

between iterations.
If IDMPBOT=1, The dampening is applied.
If IDMPBOT=0, The dampening is not applied.
Note: This dampening may be effective in achieving convergence for
aquifer systems with highly irregular layering geometries that
otherwise have difficulties converging, as a result of heads within
any iteration falling significantly below the bottom of the modeled
system.
HCLOSE -- is the head change criterion for convergence. When the maximum
absolute value of the head change at all nodes during an iteration is less
than or equal to HCLOSE the nonlinear iterations are terminated.
Commonly, HCLOSE is set to equal 0.01.
MUTPCG -- is a flag which controls printing from the solver. If MUTPCG = 0, no
printing from the solver is suppressed. If MUTPCG = 1, the solver
iteration summary is suppressed. If MUTPCG = 2, the solver iteration
summary, as well as convergence behavior at each iteration, is
suppressed. Usually this flag should be set to 1 or 2 to prevent lengthy
outputs. Solver convergence behavior at each iteration needs to be
examined only if the solver fails and the convergence is stalling.
Chapter 6: The Preconditioned Conjugate Gradient (PCG4) Package 6-3

MNORTH -- is the maximum number of orthogonalizations allowed for the
Orthomin solver to solve the transport matrix equation. A
suggested value of MNORTH ranges from 5 to 10, but lower
values may be used to conserve array space. Smaller MNORTH
may, however, weaken the solver’s performance. MNORTH
larger than 10 may help the solver to handle difficult problems
but, in general, diminishing returns are expected for MNORTH
much larger than 10.
RCLOSE -- is the relative closure criterion for convergence. This parameter

is used for convergence of inner iterations, when the relative head
change is less than the prescribed value. Note that HCLOSE is
also used to check for inner iteration convergence, and
convergence is achieved when either RCLOSE or HCLOSE
values have been met by the respective changes at any inner
iteration.
Note: If RCLOSE is set to zero, its value is taken as HCLOSE x 10-3.
6-4 Chapter 6: The Preconditioned Conjugate Gradient (PCG4) Package

CHAPTER 7: THE NEWTON-RAPHSON LINEARIZATION WITH
BACKTRACKING NRB1 PACKAGE
CHAPTER 7: THE NEWTON-RAPHSON LINEARIZATION WITH
BACKTRACKING NRB1 PACKAGE
7.1 GENERAL
Rigorous unsaturated zone flow simulations with water or air as the active phase
are typically highly nonlinear requiring small time-steps and several iterations for solution
to practical situations. For several cases, one-step steady state simulations may not
converge with the Picard iteration scheme of MODFLOW for unconfined situations, or
with the modified Picard scheme of MODFLOW-SURFACT/MODHMS (Chapter 2) for
unsaturated flow (of air or water) situations, or with pseudo-soil functions determining
water table elevations in unconfined aquifers where rigorous unsaturated flow simulation
is not required. A Newton-Raphson scheme providing quadratic convergence can greatly
alleviate convergence difficulties. The NRB1 package discussed here integrates the
Newton-Raphson scheme with a backtracking algorithm to control step-size, into
MODFLOW for use with non-linear situations which can benefit especially when Picard
iterations are excessive or fail entirely.
The NRB1 package includes a Newton-Raphson linearization and a backtracking
scheme to stabilize the Newton iterations. The formulation is designed to be compatible
with the current MODFLOW scheme of solving for heads at all nodes at every iteration.
Further, this formulation requires no special treatment for any of the boundary conditions
currently simulated by MODFLOW. Finally a backtracking scheme limits the increase in
residuals at any iteration while an underrelaxation technique assists with oscillatory
behavior of the solution between iterations. These issues are discussed in the rest of this
chapter.
7.2 FORMULATION OF NEWTON RAPHSON SCHEME
MODFLOW, using implicit time discretization, expresses the equations for fluid
flow in a porous medium by the matrix equation
(27)
where superscripts k+1 an k denote current and previous time levels respectively, )t is
the time-step size, B is a diagonal storage matrix, A is the flux matrix, h the potentiometric
head, and F is an external flux. In MODFLOW code terminology hk+1 is HNEW, which
is solved for in equation (27); Bk+1/)t is HCOEF, the storage term added to the diagonal
of the left hand side matrix; Ak+1 contains the flow term comprising of CC, CR, and CV
in their respective diagonal an upper diagonal locations; and Fk+1+Bhk/)t is the RHS
vector. For the new schemes in the BCF4 package discussed in Chapter 2, the RHS vector
is assembled using the modified Picard scheme of Cilea et al (1990). The modified Picard
scheme, in effect, applies a Newton type linearization to the storage term so additional
consideration is only required for Newton-Raphson linearization of the flow terms, Ak+1,
of equation (27). This can be expresses as (Huyakorn an Pinder ,1983)
Chapter 7: The Newton-Raphson Linearization with Backtracking NRB1 Package 7-1

(28)
where r and r+1 denote previous and present iteration levels at the current time value and
)hr+1 is the iteration displacement vector or update vector
(29)
Substituting equation (29) in (28) and rearranging yields the Newton expansion for the
flow term as
(30)
where
(31)
which may be expanded for each row of the matrix as
(32)
The Aij contribution to the left-hand-side matrix from the flux term is already filled into
the matrix (CC, CR, and CV) during assembly of the system of equations by MODFLOW.
Therefore, the NRB1 package only needs to add the last term in equation (32) to the left
hand side of the system of equations to provide a Newton-Raphson linearization, along
with adding the right-hand-side of equation (30) to the RHS matrix assembled by
MODFLOW.
The left-hand-side matrix of MODFLOW is symmetric. The modified Picard
method applies a Newton type linearization on the storage term which is only on the
diagonal, therefore the matrix remains symmetric. The additional terms of equation (32),
however, create an unsymmetric system of equations which are not compatible with the
solvers that exist in MODFLOW. Therefore, solution to hk+1 is sought for this asymmetric
system using an Orthomin solution scheme (Anderson, 1983a,b; Behie and Vinsome,
1982; Panday et al., 1993). This solver is also used later by the transport routines of
MODFLOW-SURFACT/MODHMS, and is available through use of the PCG4 package
discussed in Chapter 6.
Boundary conditions of MODFLOW do not require any additional treatment for
this Newton-Raphson scheme. Prescribed head boundaries in equation (30) are
7-2 Chapter 7: The Newton-Raphson Linearization with Backtracking NRB1 Package

manipulated in the same manner as for the Picard scheme. Source/sink terms are added
to the RHS vector as in the Picard scheme, and therefore also do not require additional
treatment. The remaining boundaries of MODFLOW (general head boundary, drains,
streams, rivers, evapotranspiration, recharge-seepage face) may be generally lumped into
one category for this discussion, as being expressed by an equation of the form
(33)
where q is the boundary flux, B is a leakance factor and ho is a known head on the outside
of the boundary. Expressing equation (33) in Newton-Raphson form gives
(34)
with dq/dh obtained from equation (33) as -B. Substituting equations (33) and (29) into
(34) and rearranging gives
(35)
which is the exact form of the respective boundaries assembled by the Picard iteration
scheme of MODFLOW. No additional treatment is therefore required for these
boundaries.
Full Newton iterations may diverge for many situations, and techniques are applied
to stabilize the solution. Oscillatory behavior in solution is dampened by using the
underrelaxation formula of Cooley (1983). The solution is further stabilized by using a
backtracking formula. After each solution for hk+1 by MODFLOW, the norm residual
of equation (27) is calculated. If this residual is larger than a factor of RESRED from the
value at the previous iteration, the update vector for the current iteration is shortened by
a factor of BFACT (note that 0<BFACT<1). This process is repeated until the residual
of the current iteration is within RESRED times its value at the previous iteration, before
proceeding to the next Newton iteration. Strict residual reduction may be enforced by
setting RESRED to less than one. Although additional computational effort is required per
iteration in implementing Newton-Raphson linearization with backtracking, experience has
indicated its robustness to prove more computationally efficient in solving highly nonlinear
problems.
7.3 INPUT INSTRUCTIONS FOR THE NRB1 PACKAGE
Three parameters are required as input for implementing the NRB1 package. The
switch INEWT is input to instruct the code to use the Newton-Raphson formulation. The
asymmetric matrix generated by the Newton scheme renders the symmetric solvers of
MODFLOW useless. Therefore, the switch INEWT is associated with the PCG4 package
discussed in Volume 1, Chapter 5 of the MODFLOW-SURFACT/MODHMS
documentation. When INEWT=1, the Newton Raphson solution scheme with the
Orthomin asymmetric solver is chosen; when INEWT=0, the Picard scheme of
MODFLOW with the symmetric conjugate gradient solver of PCG4 is chosen. Further,
when INEWT is one, the two backtracking parameters, BFACT and RESRED are required
Chapter 7: The Newton-Raphson Linearization with Backtracking NRB1 Package 7-3

as input. BFACT needs to be greater than zero and less than one. RESRED needs to be
greater than zero. A value greater than 1015 or zero turns off backtracking. A value of
RESRED less than one forces residual reduction at every Newton iteration, while values
larger than one allow for the residual to increase at any Newton iteration, by the
prescribed factor. The backtracking count and ‘accepted residual’ are printed to the output
file at each Newton iteration if the switch MUTPCG = 0 or 1 (see PCG4 input
instructions). See Volume 1 Chapter 6 for input instructions to the PCG4 package which
details the use of the switch INEWT and the factors BFACT and RESRED.
General suggestions for use follow. Advanced users may use their own
experience.
1. If Picard iterations are excessive, or fail to converge, or if the simulation is
progressing only with prohibitively small time-steps sizes, Newton Raphson
linearization should be employed to overcome difficulties (set INEWT=I).
2. As a first attempt, perform no backtracking (leave BFACT and RESRED
blank, or set RESRED $ 1015).
3. If convergence difficulties persist, perform severe step cuts (BFACT = 0.1
to 0.2) allowing for only a slight increase in the residual value (RESRED
= 1 to 2).
4. If convergence difficulties still persist, perform severe step cuts (BFACT =
0.1 to 0.2) but allow for a considerable increase in the residual (RESRED
= 10 to 50) to escape local minima in the error surface.
Converged solutions to extremely difficult problems have been obtained with
reasonable computational effort, by following the above suggestions. Simulation
performance may then be fine-tuned by adjusting BFACT, RESRED, MXITER, and
ITER1 in the PCG4 input package, and TSMULT, TSDIV, and TMAX in the ATO4
package for transient simulations.
7-4 Chapter 7: The Newton-Raphson Linearization with Backtracking NRB1 Package

8.1 GENERAL
An observation node package has been included in MODFLOW-
SURFACT/MODHMS to simplify post-processing of breakthrough curves for transient
simulations, at any node(s) within the domain. Flow simulation output shows the pressure-
time relationship at the requested node(s) of the simulation, and transport output includes
the concentration-time relationship for each contaminant simulated. A separate output file
is created consisting the breakthrough outputs, which may be easily post-processed to
provide breakthrough curves.
8.2 INPUT INSTRUCTIONS FOR THE OBS1 PACKAGE
8.2.1 OBS1 Input
Input for the OBS1 package is read from the unit specified in IUNIT(6) of the
basic file. The required input consists of the number of observation nodes, their locations,
and a unit number on which the breakthrough information will be output. By default, the
extension of the input file on IUNIT(6) is .OBS, and the extension of the output file which
contains the breakthrough curve is .OBW.
1. Data: MXOBSW IOBSUN
Format: I10 I10
**Enter items 2 and 3 for each stress period.

2. Data: ITMP
Format: I10
**Enter item 3 only if ITMP >0

3. Data: IWELL K I J
MXOBSW -- is the maximum number of observation nodes requested for the

simulation.
IOBSUN -- is a flag and unit number for recording breakthrough information.
If IOBSUN >0, it is the unit number on which the breakthrough
information is recorded in binary mode.
If IOBSUN =0, observation node information is not printed or
recorded.
If IOBSUN <0, indicates that breakthrough information will be
recorded in ASCII mode, on unit š OBSUN š.
ITMP -- is a flg and a counter.
If ITMP <0, observation node data from the last stress period will be
reused.
Chapter 8: Observation Node (OBS1) Package 8-1

If ITMP $, ITMP is the number of observation nodes (NOBS) for the
IWELL -- is the well’s index number.
K -- is the layer number within the finite difference grid of the observation
node.
I -- is the row number within the finite difference grid of the observation
node.
J -- is the column number within the finite difference grid of the
observation node.
8.3 OUTPUT OF OBS1 PACKAGE
For each stress period, the output in the .OBW file (created on unit IOBSUN)
begins as follows:
NUMBER OF OBSERVATION WELLS = ________ FOR STRESS PERIOD _______
The next line depends on the simulation option ITRAN.
If ITRAN = 0, (flow only) the output is as follows:
LIST OF : WELL NO., TIME, AND HEAD
followed by lines containing the breakthrough at each well.
If ITRAN = 1, (flow and transport) the output is as follows:
LIST OF: WELL NO., TIME, HEAD, AND CONCS FOR THE ___ SPECIES
OF SIMULATION
If ITRAN = 2, (transport only) the output is as follows:
LIST OF: WELL NO., TIME AND CONCS FOR THE ___ SPECIES OF
SIMULATION
8-2 Chapter 8: Observation Node (OBS1) Package

Anderson, D.V., 1983. ILUCG3: Subprograms for the solution of a linear asymmetric
matrix equation arising from a 7, 15, 19, or 27 point 3D discretization, Computer
Physics Communications, 30:43-49.
Anderson, D.V., 1983. ICG3: Subprograms for the solution of a linear symmetric
matrix equation arising from a 7, 15, 19, or 27 point 3D discretization, Computer
Physics Communications, 30:51-57.
Banta, E.R., 2000. MODFLOW-2000, the U.S. Geological Survey Modular
Groundwater Model - Documentation of Packages for Simulating
Evapotranspiration with a Segmented Function (ETS1) and Drains with Return
Flow (DRT1), U.S. Geological Survey Open -File Report 00-466, 127 p.
Bear, J., 1979. Hydraulics of Groundwater, McGraw-Hill, New York, 569 pp.
Bennett, G.D., A.L. Kontis and S.P. Larson, 1982. Representation of Multiaquifer Well
Effects in Three-Dimensional Ground-Water Flow Simulation, Groundwater,
20(3):334-341.
Brooks, R.H. and A.T. Corey, 1966. Properties of porous media affecting fluid flow.
ASCE J. Irrig. Drain. Div., 92 (IR2): 61-88.
Celia, M.A., Bouloutas, E.T., and Zarba, R.L., 1990. A general mass-conservative
numerical solution for the unsaturated flow equation. Water Resour. Res., 27(7):
1483-1496.
Cooley, R.L., 1983. Some new procedures for numerical solution of variably saturated
flow problems, Water Resour. Res., 19(5), 1271-1285.
Gambolati, G., 1980. Fast solution to finite element flow equations by Newton iteration
and modified conjugate gradient method, Int. J. Numer. Meth. Eng., 15:661-675.
Gambolati, G., Sartoretto, F., and Uliana, F., 1986. A conjugate gradient finite element
model of flow for large multiaquifer systems, Water Resour. Res.,
22(7):1003-1015.
Goode, D.J. and Appel, C.A., 1992. Finite-difference interblock transmissivity for
unconfined aquifers and for aquifers having smoothly varying transmissivity: U.S.
Geological Survey Water-Resources Investigations Report 92-4124.
Halford, K.J., and R.T. Hanson, 2002. User Guide for the Drawdown-Limited, Multi-
Node Well (MNW) Package for the U.S. Geological Survey’s Modular Three-
Dimensional Finite-Difference Ground-Water Flow Model, versions MODFLOW-
96 and MODFLOW-2000. U.S. Geological Survey Open-File Report, 02-293,
33 pp.
Hill, M.C., 1994. Preconditioned Conjugate-Gradient 2 (PCG2), a Computer Program
for Solving Groundwater Flow Equations, U.S. Geological Survey Water
Resources Investigations Report 90-4048.
Huang, K., B.P. Mohanty, and M.Th.van Genuchten, 1996. A new convergence
criterion for the modified Picard iteration method to solve the variably-saturated
flow equation. Journal of Hydrology, 178:69-91.
Huyakorn, P.S., S. Panday, and Y.S. Wu, 1994. A three-dimensional multiphase flow
model for assessing NAPL contamination in porous and fractured media, 1.
Formulation. Journal of Contaminant Hydrology, 16:109-130.
Huyakorn, P.S., Springer, E.P., Guvanasen, V., and Wadsorth, T.D., 1986. A three-
dimensional finite element model for simulating water flow in variably saturated
porous media. Water Resour. Res., 22(12): 1790-1808.
Chapter 9: References 9-1

HydroGeoLogic, Inc., 1992. STAFF3D: A Three-Dimensional Finite Element Code for
Simulating Fluid Flow and Transport of Radionuclides in Fractured Porous Media.
Code Documentation Report.
Kalf, F.R.P., 1988. A variably saturated finite element model for three- dimensional
seepage face problems, PhD. thesis, Univ. of New South Wales, Sidney,
Australia, 339pp.
Kershaw, D.S., 1978. The incomplete Cholesky-conjugate gradient method for the
iterative solution of systems of linear equations, J. Comp. Physics, 26:43-65.
Kuiper, L.K., 1981. A comparison of the incomplete Cholesky-conjugate gradient
method with the strongly implicit method as applied to the solution of
two-dimensional groundwater flow equations, Water Resour. Res.,
17(4):1082-1086.
Kuiper, L.K., 1987. A comparison of iterative methods as applied to the solution of the
nonlinear three-dimensional groundwater flow equation, SIAM J. Sci. Stat.
Comput., 8(4):521-528.
McDonald, M.G. and Harbaugh, A.W., 1988. A modular three-dimensional finite-
difference groundwater flow model: U.S. Geological Survey Techniques of Water-
Resources Investigations Book 6, chapter A1.
McDonald, M.G., Harbaugh, A.W., Orr, B.R., and Ackerman, D.J., 1991. A method
of converting no-flow cells to variable-head cells for the U.S. Geological Survey
modular finite-difference ground-water flow model: U.S. Geological Survey
Open-File Report 91-536.
Marley, M.C., Li, F. and Magee, S., 1992. The application of a 3-D model in the design
of air sparging systems. Proc. Conf. on 1992 Petroleum Hydrocarbons and
Organic Chemicals in Ground Water: Prevention, Detection and Restoration, Nov.
4-6, 1992, Houston, TX, pp. 337-392.
Meijerink, J.A., and van der Vorst, H.A., 1977. An iterative solution method for linear
systems of which the coefficient matrix is a symmetric M-Matrix, Math. of
Comp., 31(137):148-162.
Meyer, P.D., Valocchi, A.J., and Ashby, S.F., and Saylor, P.E., 1989. A numerical
investigation of the conjugate gradient method as applied to three-dimensional
groundwater flow problems in randomly heterogeneous porous media, Water
Resources Research, 25(6):1440-1446.
Mohanty, B.P., R.S. Bowman, J.M.H. Hendrickx, and M.T. van Genuchten, 1997. New
piecewise-continuous hydraulic functions for modeling preferential flow in an
intermittent-flood-irrigated field, Water Resources Research, 33(9):2049-2063.
Mualem, Y., 1976. A new model for predicting the hydraulic conductivity of unsaturated
porous media, Water Resour. Res., 12(3):513-522.
Neuman, S.P., 1972, Theory of flow in unconfined aquifers considering delayed response
of the water table, Water Resour. Res., 8, 4, pp. 1031-1045.
Neville, C.J., and M.J. Tonkin, 2004. Modeling Multiaquifer Wells with MODFLOW,
Groundwater, 42(6):910-919.
Panday, S., Y.S. Wu, P.S. Huyakorn, and E.P. Springer, 1994. A three-dimensional
multiphase flow model for assessing NAPL contamination in porous and fractured
media, 2. Porous medium simulation examples. Journal of Contaminant
Hydrology, 16:131-156.
Papadopulos, I.S., 1966. Nonsteady flows to multiaquifer wells. Journal of Geophysical
Research, 71(20):4791-4797.
9-2 Chapter 9: References

Papadopolus, I.S., and Cooper Jr., H.H. (1967). Drawdown in a well of large diameter,
Water Resources Research, 3(1), 241-244.
Romeu, R.H., and Noetinger, B., 1995. Calculation of inter nodal transmissivities in
finite difference models of flow in heterogeneous porous media. Water Resour.
Res., 31:(4) 943-959.
Saad, Y., 1985. Practical use of polynomial preconditionings for the conjugate gradient
method, SIAM Journal of Scientific and Statistical Computing, 6(4):685-881.
Shan, C. (1995). Analytical solutions for determining vertical air permeability in
unsaturated soils. Water Resour. Res., 31: (9) 2193-2200.
Sudicky, E.A., Unger, A.J.A., and Lacombe, S. (1995). A noniterative technique for the
direct implementation of well bore boundary conditions in three-dimensional
heterogeneous formations, Water Resources Research, 31(2), 411-415.
Theis, C.V. (1935). The relation between the lowering of the piezometric surface and the
rate and duration of discharge of a well using groundwater storage, EOS Trans.
AGU, 16, 519-524.
van Genuchten, M. Th., G.F. Pinder, and W.P. Saukin, 1977. Modeling of leachate and
soil interactions in an aquifer, Management of Gas and Leachate in Lanfills, U.S.
EPA, Report. EPA-600/9-77-026, pp. 95-103.
van Genuchten (1980). A closed-form equation for predicting the hydraulic conductivity
of unsaturated soils, Soil Science of American Journal 44, 892-898.

APPENDICES
Appendix A: Harmonic Averaging of Transmissivities
Appendix B: Input Data for the Simulation of Steady-State Pumping in an Unconfined

Aquifer
Appendix C: Input Data for the Fully 3-D Simulation of Withdrawal from an Unconfined
Aquifer
Appendix D: Input Data for the Simulation of Steady-State Seepage Through a Square
Embankment
Appendix E: Input Data for the Transient Simulation of Pumping and Recovery in an
Unconfined Aquifer Using the ATO4 Package
Appendix F: MODPATH Documentation For MODFLOW-SURFACT/MODHMS
Appendix G: Using ET Time-Series with the Evapotranspiration (EVT) Package or with

the Segmented Evapotranspiration (ET2) Package
Appendix H: Time Varying Flow and Head Boundary Package with Step Function Option
Appendix I: Addendum for PCG5 Solver

APPENDIX A:
HARMONIC AVERAGING OF TRANSMISSIVITIES
Natural porous medium systems exhibit a wide range of heterogeneities which must
be accounted for in groundwater flow simulations. The heterogeneities of material
properties are provided to a finite difference model as nodal values for the individual grid
blocks (cells). The original MODFLOW Block-Centered Flow Packages BCF1 and BCF2
use a harmonic mean of the effective transmissivities of adjacent blocks to calculate the
interblock transmissivity. For unconfined systems, the effective transmissivity of a block
is computed as the effective hydraulic conductivity multiplied by the saturated block
thickness.
Harmonic averaging of interblock transmissivities can cause misrepresentations in an
aquifer system, as illustrated in one extreme example depicted in Figure A.1. Consider
A and B to be adjacent grid blocks in an unconfined aquifer being flooded from the left.
The situation at a given instant is such that grid block A has an average water-table
elevation of h a which is above the bottom of the layer, and block B has an average water-
table elevation of h b, below the bottom of the layer. The transmissivity of grid-block A
is some positive value, since h a is above the bottom. The transmissivity of grid-block B
is zero, since its saturated thickness is zero. The Harmonic mean transmissivity is
therefore computed as zero, and grid-block B never saturates, contrary to what is
physically expected.
A more appropriate and physically consistent procedure is to use the Harmonic mean
of the saturated hydraulic conductivities (or fully saturated transmissivities), with a
midpoint (or upstream) weighting on the relative permeabilities (which are related to the
saturated grid-block thickness by the pseudo-soil functions for our variably saturated
scheme implementation). It is noted (Romeu and Noetinger, 1995) that the Harmonic
mean of block conductivities leads to an underestimation of the equivalent conductivity,
and refined grids may be used to overcome this problem (in addition to inaccuracies
associated with grid-averaging of the results). However, the harmonic averaging of head-
dependent transmissivities leads to a further underestimation of aquifer transmissive
property, as is evidenced by the larger drawdowns near wells predicted by the previous
package of MODFLOW (see Problems 1 and 2 of the BCF4). Further, this problem may
not be overcome by grid refinement.
Appendix A A-1
M:\Projects\MODFLOW Manuals\SURFACT and HMS Ver.3\Vol I\Appendices_A-E.wpd
Figure A.1 A Case of Adjacent Grid Blocks Where Harmonic Averaging of
Head-Dependent Transmissivities Is Inadequate.
A-2 Appendix A
APPENDIX B:
INPUT DATA FOR THE SIMULATION OF STEADY-STATE
PUMPING IN AN UNCONFINED AQUIFER
The input data sets used for the steady-state simulation of an unconfined aquifer
system are listed in this Appendix. Two simulations were performed using: Case(a) the
original MODFLOW option and Case(b) the variably-saturated flow option of the BCF4
Package. The problem description is presented in Section 2.4.1 of the BCF4 Package.
The input data files are listed in the following order:
1. The Basic Flow (BAS) Package Input
2a. The Block-Centered Flow (BCF4) Package Input for Case (a)
2b. The Block-Centered Flow (BCF4) Package Input for Case (b)
3. The Well (WEL) Package Input
4. The Recharge (RCH) Package Input
5. The Strongly Implicit Procedure (SIP) Package Input
6. The Output Control (OC) Package Input.
Appendix B B-1
1 2 3 4 5 6 7 8
123456789012345678901234567890123456789012345678901234567890123456
78901234567890
SAMPLE PROBLEM FOR STEADY-STATE SIMULATION
BASIC MODEL WITH RECHARGE AND WELLS
3 15 15 1 1
11 12 0 0 0 0 0 18 19 0 0 22
1 1 IAPART,ISTRT
1 1(15I3) 3 IBOUND-1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
1 1(15I3) 3 IBOUND-2
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
B-2 Appendix B
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
0 1 IBOUND-3
999.99
0 0. HEAD-1
0 0. HEAD-2
0 0. HEAD-3
1.26E08 1 1.0 PERLEN,NSTP,TSMULT
2a. The Block-Centered Flow (BCF4) Package Input for Case (a)
1 2 3 4 5 6 7 8
123456789012345678901234567890123456789012345678901234567890123456
78901234567890
1 0 777.7 0 0.0 0 0
333
0 1. TRPY
0 5000. DELR
0 5000. DELC
0 0.0001 HY-1
0 -50. BOT-1
0 1.0E-7 VHY/THICK-1
0 50. top-1
Appendix B B-3
0 0.0001 HY-2
0 -150. BOT-2
0 -50. top-2
0 0.0001 HY-3
0 -250. BOT-3
0 -150. top-3
2b. The Block-Centered Flow (BCF4) Package Input for Case (b)
1 2 3 4 5 6 7 8
123456789012345678901234567890123456789012345678901234567890123456
78901234567890
1 0 777.7 0 0.0 0 0
434343
0 1. TRPY
0 5000. DELR
0 5000. DELC
0 0.0001 HY-1
0 -50. BOT-1
0 50. top-1
0 0.0001 HY-2
0 -150. BOT-2
0 -50. top-2
0 0.0001 HY-3
0 -250. BOT-3
0 -150. top-3
B-4 Appendix B
Appendix B B-5
1 2 3 4 5 6 7 8
123456789012345678901234567890123456789012345678901234567890123456
78901234567890
15 0 MXWELL,IWELBD
15 NWELL
3 5 11 -0.95
3 4 6 -0.95
3 6 12 -0.95
3 9 8 -0.95
3 9 10 -0.95
3 9 12 -0.95
3 9 14 -0.95
3 11 8 -0.95
3 11 10 -0.95
3 11 12 -0.95
3 11 14 -0.95
3 13 8 -0.95
3 13 10 -0.95
3 13 12 -0.95
3 13 14 -0.95

1 2 3 4 5 6 7 8
123456789012345678901234567890123456789012345678901234567890123456
78901234567890
3 0 NRCHOP,IRCHBD
1 INRECH
0 3.E-9 RECH-1
B-6 Appendix B
1 2 3 4 5 6 7 8
123456789012345678901234567890123456789012345678901234567890123456
78901234567890
500 5 MXITER,NPARM
1. .0001 0 .001 1 ACCL,ERR,IPCALC,WSEE
6. The Output Control (OC) Package Input
1 2 3 4 5 6 7 8
123456789012345678901234567890123456789012345678901234567890123456
78901234567890
12 9 0 0
0 1 1 1
1 0 0 0
Appendix B B-7
APPENDIX C:
INPUT DATA FOR THE FULLY 3-D SIMULATION OF
WITHDRAWAL FROM AN UNCONFINED AQUIFER
The input data used for the fully 3-D simulation of withdrawal from an unconfined
aquifer are presented in this Appendix. The problem description is presented in Section
3.4.1 of the FWL4 Package. The input data files are listed in the following order:
2. The Block-Centered Flow (BCF4) Package Input
3. The Fractured-Well (FWL4) Package Input
4. The Preconditioned Conjugate Gradient (PCG4) Package Input
Appendix C C-1
1 2 3 4 5 6 7 8
123456789012345678901234567890123456789012345678901234567890123456
78901234567890
SAMPLE PROBLEM FOR TRANSIENT SIMULATION
WITH A FRACTURE-WELL
3 15 15 2 1
11 00 0 0 0 0 0 00 00 0 0 22 13 21
1 1 IAPART,ISTRT
1 1(15I3) 3 IBOUND-1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
1 1(15I3) 3 IBOUND-2
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
C-2 Appendix C
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
0 1 IBOUND-3
999.99
0 0. HEAD-1
0 0. HEAD-2
0 0. HEAD-3
Appendix C C-3
1 2 3 4 5 6 7 8
123456789012345678901234567890123456789012345678901234567890123456
78901234567890
0 0 777.7 0 0.0 0 0
434343
0 1. TRPY
0 5000. DELR
0 5000. DELC
0 0.0001 SF1-1
0 0.0001 HY-1
0 -50. BOT-1
0 0.03 SF2-1
0 50. top-1
0 0.0001 SF1-2
0 0.0001 HY-2
0 -150. BOT-2
0 0.03 SF2-2
0 -50. top-2
0 0.0001 SF1-3
0 0.0001 HY-3
0 -250. BOT-3
0 0.03 SF2-3
0 -150. top-3
3. The Fractured-Well (FWL4) Package Input
1 2 3 4 5 6 7 8
C-4 Appendix C
123456789012345678901234567890123456789012345678901234567890123456
78901234567890
1 0 3.74E05 MXFWEL,IFWLBD,FCONST(1/(ft-s))
1 NWELL
1 3 8 8 -250. 0.2 -10. 1.0E-6
1 NWELL
1 3 8 8 -250. 0.2 0. 1.0E-6
4. The Preconditioned Conjugate Gradient (PCG4) Package Input
1 2 3 4 5 6 7 8
123456789012345678901234567890123456789012345678901234567890123456
78901234567890
50 300 MXITER,NPARM
0.0001 .001 2
Appendix C C-5
1 2 3 4 5 6 7 8
123456789012345678901234567890123456789012345678901234567890123456
78901234567890
12 9 0 0
0 1 1 1
1 0 0 0
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
C-6 Appendix C
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
Appendix C C-7
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
C-8 Appendix C
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
Appendix C C-9
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
C-10 Appendix C
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
Appendix C C-11
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
C-12 Appendix C
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
Appendix C C-13
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
-1 1 1 1
C-14 Appendix C
APPENDIX D:
INPUT DATA FOR THE SIMULATION OF STEADY-STATE
SEEPAGE THROUGH A SQUARE EMBANKMENT
The input data used for the simulation of steady seepage through a square
embankment are listed in this Appendix. The problem description is presented in Section
4.3.2 of the RSF4 Package. The input data files are listed in the following order:
3. The Recharge-Seepage Face Boundary (RSF4) Package Input
Appendix D D-1
1 2 3 4 5 6 7 8
123456789012345678901234567890123456789012345678901234567890123456
78901234567890
STEADY-STATE SEEPAGE THROUGH A SQUARE EMBANKMENT
VERIFICATION WITH COOLEY RESULTS
10 1 12 1 4
11 0 0 0 0 0 0 18 19 0 0 0
1 1 IAPART,ISTRT
1 1(12I3) 3 IBOUND-1
-1 1 1 1 1 1 1 1 1 1 1 1
1 1(12I3) 3 IBOUND-2
-1 1 1 1 1 1 1 1 1 1 1 1
1 1(12I3) 3 IBOUND-3
-1 1 1 1 1 1 1 1 1 1 1 1
1 1(12I3) 3 IBOUND-4
-1 1 1 1 1 1 1 1 1 1 1 1
1 1(12I3) 3 IBOUND-5
-1 1 1 1 1 1 1 1 1 1 1 1
1 1(12I3) 3 IBOUND-6
-1 1 1 1 1 1 1 1 1 1 1 1
1 1(12I3) 3 IBOUND-7
-1 1 1 1 1 1 1 1 1 1 1 1
1 1(12I3) 3 IBOUND-8
-1 1 1 1 1 1 1 1 1 1 1 1
1 1(12I3) 3 IBOUND-9
-1 1 1 1 1 1 1 1 1 1 1 -1
1 1(12I3) 3 IBOUND-10
-1 1 1 1 1 1 1 1 1 1 1 -1
999.99
0 10. HEAD-1
0 10. HEAD-2
D-2 Appendix D
0 10. HEAD-3
0 10. HEAD-4
0 10. HEAD-5
0 10. HEAD-6
0 10. HEAD-7
0 10. HEAD-8
1 1.0 (12F3.0) 0 HEAD-9
10 10 10 10 10 10 10 10 10 10 10 2
1 1.0 (12F3.0) 0 HEAD-10
10 10 10 10 10 10 10 10 10 10 10 2
Appendix D D-3
1 2 3 4 5 6 7 8
123456789012345678901234567890123456789012345678901234567890123456
78901234567890
9 0 777.7 0 0.0 0 0
43434343434343434343
0 1. TRPY
11 1.(9F8.3) 5 DELR
0.10 0.90 1.00 1.00 1.00 1.00 1.00 1.00 1.00
1.00 0.90 0.10
0 1. DELC
0 0.01 HY-1
0 9. BOT-1
0 0.01 VHY/THICK-1
0 10. top-1
0 0.01 HY-2
0 8. BOT-2
0 0.01 VHY/THICK-2
0 9. top-2
0 0.01 HY-3
0 7. BOT-3
0 0.01 VHY/THICK-3
0 8. top-3
0 0.01 HY-4
0 6. BOT-4
0 0.01 VHY/THICK-4
0 7. top-4
0 0.01 HY-5
0 5. BOT-5
0 0.01 VHY/THICK-5
0 6. top-5
0 0.01 HY-6
D-4 Appendix D
0 4. BOT-6
0 0.01 VHY/THICK-6
0 5. top-6
0 0.01 HY-7
0 3. BOT-7
0 0.01 VHY/THICK-7
0 4. top-7
0 0.01 HY-8
0 2. BOT-8
0 0.01 VHY/THICK-8
0 3. top-8
0 0.01 HY-9
0 1. BOT-9
0 0.01 VHY/THICK-9
0 2. top-9
0 0.01 HY-10
0 0. BOT-10
0 1. top-10
3. The Recharge-Seepage Face Boundary (RSF4) Package Input
1 2 3 4 5 6 7 8
123456789012345678901234567890123456789012345678901234567890123456
78901234567890
1 0 0 8 NRCHOP,IRCHBD,IUNCNF,MXSEEP
1 0 0 8 INRECH,INIRCH,INPNDEL,INSEEP
0 0.0 RECH-1
1 1 12 9.5
2 1 12 8.5
3 1 12 7.5
4 1 12 6.5
5 1 12 5.5
6 1 12 4.5
Appendix D D-5
7 1 12 3.5
8 1 12 2.5
1 2 3 4 5 6 7 8
123456789012345678901234567890123456789012345678901234567890123456
78901234567890
500 5 MXITER,NPARM
1. .0001 0 .001 1 ACCL,ERR,IPCALC,WSEED
D-6 Appendix D
APPENDIX E:
INPUT DATA FOR THE TRANSIENT SIMULATION OF PUMPING
AND RECOVERY IN AN UNCONFINED AQUIFER USING THE
ATO4 PACKAGE
The input data used for the transient simulation of pumping from an unconfined
aquifer system using the Adaptive Time-Stepping and Output Control (ATO4) Package are
listed in this Appendix. The problem description is presented in Section 5.3 of the ATO4
Package. The input files are listed in the following order:
6. The Adaptive Time-Stepping and Output Control (ATO4) Package Input.
Appendix E E-1
1 2 3 4 5 6 7 8
123456789012345678901234567890123456789012345678901234567890123456
78901234567890
Transient Simulation of Pumping and Recovery
in an Unconfined Aquifer
3 15 15 2 1
11 12 0 0 0 0 0 18 19 0 0 00 22
1 1 IAPART,ISTRT
1 1(15I3) 3 IBOUND-1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
1 1(15I3) 3 IBOUND-2
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
E-2 Appendix E
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
0 1 IBOUND-3
999.99
0 0. HEAD-1
0 0. HEAD-2
0 0. HEAD-3
1 2 3 4 5 6 7 8
123456789012345678901234567890123456789012345678901234567890123456
78901234567890
0 0 777.7 0 0.0 0 0
434343
0 1. TRPY
0 5000. DELR
0 5000. DELC
0 0.0001 SF1-1
0 0.0001 HY-1
0 -50. BOT-1
Appendix E E-3
0 0.03 SF2-1
0 50. top-1
0 0.0001 SF1-2
0 0.0001 HY-2
0 -150. BOT-2
0 0.03 SF2-2
0 -50. top-2
0 0.0001 SF1-3
0 0.0001 HY-3
0 -250. BOT-3
0 0.03 SF2-3
0 -150. top-3
1 2 3 4 5 6 7 8
123456789012345678901234567890123456789012345678901234567890123456
78901234567890
15 0 MXWELL,IWELBD
15 NWELL
3 5 11 -3.85
3 4 6 -3.85
3 6 12 -3.85
3 9 8 -3.85
3 9 10 -3.85
3 9 12 -3.85
3 9 14 -3.85
3 11 8 -3.85
3 11 10 -3.85
3 11 12 -3.85
3 11 14 -3.85
E-4 Appendix E
3 13 8 -3.85
3 13 10 -3.85
3 13 12 -3.85
3 13 14 -3.85
0 NWELL
1 2 3 4 5 6 7 8
123456789012345678901234567890123456789012345678901234567890123456
78901234567890
3 0
1
0 3.E-9 RECH-1
-1
1 2 3 4 5 6 7 8
123456789012345678901234567890123456789012345678901234567890123456
78901234567890
50 5 MXITER,NPARM
1. .0001 0 .001 1 ACCL,ERR,IPCALC,WSEED
6. The Adaptive Time-Stepping and Output Control (ATO4) Package Input
Appendix E E-5
1 2 3 4 5 6 7 8
123456789012345678901234567890123456789012345678901234567890123456
78901234567890
12 9 0 0
0 1 1 1
0.7875E07 0.08E07 0.20E08 1.7 2.0 2 0
1.0e08 1.26144e8
1 0 0 0
0 1 1 1
0.23625e8 0.23E07 0.60E08 1.7 2.0 1 0
3.78432e8
1 0 0 0
E-6 Appendix E
APPENDIX F: MODPATH DOCUMENTATION FOR MODFLOW-
SURFACT/MODHMS
APPENDIX F: MODPATH DOCUMENTATION FOR MODFLOW-
SURFACT/MODHMS
F.1.0 INTRODUCTION
MODPATH is a particle tracking post-processing program designed to work with
MODFLOW. MODPATH is documented in User's Guide for
MODPATH/MODPATH-PLOT Version 3 (Pollock, 1994), which provides the particle
tracking methodology, details of input and output files, time concepts, and examples. The
document is available online at:
http://water.usgs.gov/software/code/ground_water/modpath/doc/ofr94464.pdf
The version of MODPATH documented here (henceforth referred to as
MODPATH-SURFACT) is modified from MODPATH Version 3 (Pollock, 1994) to
make it compatible with MODFLOW-SURFACT. This documentation for MODPATH-
SURFACT, is consistent with the original MODPATH documentation. Additional input
variables required by this version of MODPATH-SURFACT are shown in a shaded box.
MODPATH-SURFACT requires these additional variables because of the following
enhancements made to MODFLOW-SURFACT.
1. MODFLOW-SURFACT uses the BCF4 package that provides a rigorous
treatment of unconfined flow using a variably saturated formulation with pseudo-soil water
retention functions to enhance convergence for desaturation/resaturation conditions. The
original MODFLOW code converts a desaturated variable-head cell into a no-flow cell.
2. The BCF4 package also includes real soil options to solve variably saturated
flow situations in a more rigorous fashion by providing moisture retention and
permeability characteristics of the unsaturated layers.
3. The adaptive time-stepping and output control (ATO) package used by
MODFLOW-SURFACT uses a time-stepping scheme with automatic generation and
control of time steps that are not known aprori. In the MODFLOW code, the time domain
(for transient flow simulations) is discretized, with predetermined time-step values.
4. MODFLOW-SURFACT uses a fracture-well (FWL4 and FWL5) packages
that simulate a wellbore as a high conductivity fracture tube. The FWL4 and FWL5
packages truly emulate a multilayer well and automatically apportion the net well
withdrawal to well nodes. Also, with FWL4 and FWL5 packages, the total well
withdrawal is automatically adjusted when the water level in the well reaches the well
bottom.
Before MODPATH-SURFACT can be run, MODFLOW-SURFACT must be run with the
following settings.
1. Flag MPHFLAG in Basic Flow (BAS) package should be equal to 1. If
MPHFLAG =1, then a binary file *.mph is generated by MODFLOW-SURFACT. This
MPH file is used by MODPATH-SURFACT to determine time step information. As
described above, MODFLOW-SURFACT uses adaptive time-stepping (ATO) package
which dynamically determines time stepping depending on the anticipated non-linearities
of the system for a given calculation.
2. All cell-by-cell flow components should be recorded in a single, common
file (CBB file). Use the same unit number for cell-by-cell outputs from all
packages(IBCFCB in BCF file, IFWLCB in FWL file, IDRNCB in DRN file, IRIVCB in
RIV file, IGHBCB in GHB file, IRCHCB in RCH file and so on).
Appendix F: MODPATH Documentation for MODFLOW-SURFACT/MODHMS F-1

M:\Projects\MODFLOW Manuals\SURFACT and HMS Ver.3\Vol I\Appendix F.wpd
3. The cell-by-cell flow information should be written to the CBB file at every
time step during the simulation. This is achieved by setting appropriate flags in the
Adaptive Time-stepping and Output Control (ATO) package (Flags ICBCFL=1,
NPRTS=0 and NPSTP=1), and by providing a unit number in Block Centered Flow
(BCF) package (IBCFCB) as well as all other boundary condition packages (FWL, GHB,
RIV, RCH, EVT etc.) used in the simulation.
F.2.0 MODPATH FILES

MODPATH obtains data from a combination of file and interactive keyboard
input. Files that contain basic information about the flow system, such as geometry,
boundary conditions, and cell-by-cell budget terms, are referred to as Flow System Files.
Interactive keyboard input is used primarily to supply information about modeling options
that change frequently from run to run. Keyboard input is recorded in a special data file
called a Response File. Output from MODPATH consists of a series of files containing
information about particle coordinates and time of travel. Detailed descriptions about
particle tracking methodology, time concepts, input and output files, different options for
running MODPATH are presented in MODPATH Version 3 (Pollock, 1994).
MODPATH always requires the following set of data files:
2.1. Flow System Files - MODFLOW-SURFACT and MODPATH data files
that contain information about the physical dimensions of the ground-water
system, boundary conditions, hydrologic properties, cell-by-cell budget output,
and head. MODPATH-SURFACT requires the following flow system files:
2.1.1. MPH file is a binary file containing time stepping information
generated as output by MODFLOW-SURFACT.
2.1.2. MODPATH’s main data file.
2.1.3. MODFLOW-SURFACT stress package data files (i.e., all
boundary conditions).
2.1.4. Ancillary data files that are referenced by the main data file or any
of the stress package data files.
2.1.5. Cell-by-cell budget file generated as binary output by
MODFLOW-SURFACT.
2.1.6. Heads generated as output file by MODFLOW-SURFACT.
2.2. Name File - A data file containing a list of input and output files required
by MODPATH-SURFACT, FORTRAN unit numbers, and file-type descriptions.
This information is used by MODPATH-SURFACT to open the files.
F.2.1.1 MPH File
MPH file is a binary file produced by MODFLOW-SURFACT. This file is
produced only if the flag MPHFLAG in the Basic Flow (BAS) package is equal to 1. The
MPH file contains time step information used for the simulation. This file is produced for
steady-state as well as transient simulations. The MPH file is a 'sequential' access,
'binary' file with time stepping information written in the following format. This set of
data is written for every time step for all stress periods.
F-2 Appendix F: MODPATH Documentation for MODFLOW-SURFACT/MODHMS

KSTP KPER PERTIM
KSTP -- Time step number. Format: Integer*4.
KPER -- Stress period number. Format: Integer*4.
PERTIM -- Time elapsed in the stress period. Format: Real*4.
F.2.1.2 Main Data File
The main data file is used by MODPATH to define most of the basic elements of the
MODPATH simulation. MODPATH is designed to read data from a single main data file.
The MODPATH main data file is opened in the Name file as type MAIN.
The main data file consists of 10 data items:
• Item 1 -- Grid Dimensions and Specifications
• Item 2 -- Options
• Item 3 -- Layer Type Codes
• Item 4 -- Confining Bed Codes
• Item 5 -- Horizontal Grid Spacing
• Item 6 -- Model Layer and Confining Layer Thickness
• Item 7 -- Top and Bottom Elevations for Model Layers
• Item 8 -- Boundary Array
• Item 9 -- Porosity
• Item 10 -- Stress Period and Time Step Data
Data items 6 and 7 provide alternate ways of specifying vertical discretization. Each
MODPATH simulation uses either data item 6 or data item 7, but not both. Data item 10
is omitted for steady-state simulations. Detailed instructions for these data items are
provided on the pages that follow.

Item 1 -- Grid Dimensions and Specifications
1. NCOL NROW NLAY NCBL IGRID NPER MAXSIZ HNOFLO HDRY NPART
NCOL -- is the number of columns
NROW -- is the number of rows
NLAY -- is the number of layers
NCBL -- is the number of quasi-3D confining layers
IGRID -- is the grid geometry code
0 = grid with variable thickness and (or) nonhorizontal layers.
1 = true rectangular 3-D grid. (Layer 1 can be unconfined for
IGRID=1 provided that the bottom elevation of layer 1 is equal to a
constant elevation.)
NPER -- is the number of stress periods. NPER also indicates whether the
simulation is steady-state or transient. To specify a steady-state
simulation, set NPER=0.
MAXSIZ -- is the maximum allowed size (in bytes) of the Composite Budget File. If
MAXSIZ = 0, the program uses a default value that is set in the
MODPATH main program. The default value can be changed by resetting
the MAXSIZ value in the MODPATH main program and then
recompiling the source code. If a transient MODPATH run will generate
a composite budget file larger than MAXSIZ, a warning message is issued
and an option is provided to either quit or proceed.
HNOFLO -- is the value specified in MODFLOW to represent head in inactive cells.
HDRY -- is the head assigned by MODFLOW to cells that have gone dry. If the
original version of MODFLOW was used with the BCF1 package, HDRY
should be set to "1.0E+30". If the BCF2 package was used in the
MODFLOW simulation, HDRY must equal the value specified in the
MODFLOW simulation.
NPART -- is the maximum number of particles allowed for a MODPATH run. The
value of NPART has no effect in MODPATH-PLOT. If NPART is
omitted or set equal to 0, MODPATH automatically resets NPART to a
default value that is defined in the MODPATH main program. For the
standard version of the MODPATH source code, there is no advantage to
setting NPART to a value other than 0 because array memory is not
allocated dynamically. However, providing the variable NPART makes
it relatively easy to take advantage of FORTRAN compilers that allow
arrays to be dimensioned dynamically at runtime. The use of dynamic
memory allocation requires that MODPATH’s main source program be
modified as appropriate for each specific FORTRAN compiler.

Item 2 -- Options and MPH File Name
2A. Options
Options -- is a 1-line character record that may contain one or more keyword codes
that control options in MODPATH. A blank line must be included for this
data item even if none of the keyword options is used.
XSECTION indicates that the model is a 1-row cross-section for which IBOUND and the
Grid Unit Array (GUA) should each be read as a single, 2-dimensional array with
dimensions of NCOL and NLAY.
COMPACT indicates that MODPATH should generate endpoint, pathline, and time series
files as text files using the global node number to indicate cell location. If COMPACT is
omitted, the cell location is designated using the row-column-layer grid indices (as in
previous versions of MODPATH).
BINARY indicates that the endpoint, pathline, and time series files will be generated by
MODPATH in binary form.
METERS indicates to MODPATH-PLOT that distances are in meters. The METERS
parameter only affects the label on the distance scale produced by MODPATH-PLOT. It
has no affect on computations produced by MODPATH. If the option METERS is not
specified, MODPATH assumes that length units are in feet. If a length unit other than feet
or meters is used in MODFLOW, the scaling performed by MODPATH-PLOT will not
be correct.
WT(OFF) indicates that the water table profile will not be drawn on true cross-section
plots. If none of the three "WT" options is specified, MODPATH-PLOT draws the
water-table profile as a solid line with the same color used for drawing contour lines.
WT(DOTTED) indicates that the water-table profile will be drawn as a dotted line on true
cross-section plots.
WT(DASHED) indicates that the water-table profile will be drawn as a dotted line on true
cross-section plots.
ATO indicates that Adaptive Time-Stepping and Output Control (ATO) package is used
in the MODFLOW-SURFACT/MODHMS simulation.
FWL indicates that Fracture Well (FWL4) package is used in the MODFLOW-
SURFACT/MODHMS simulation.
FWL5 indicates that Multinode Fracture Well (FWL5) package is used in the
MODFLOW-SURFACT/MODHMS simulation.
REALSL indicates that the flag IREALSL is greater than 0 in the Block-Centered Flow
(BCF) package in the MODFLOW-SURFACT/MODHMS simulation. IREALSL is a flag
indicating real soil moisture functions are used to define flow in the unsaturated zone
above the water-table.

2B. MPH File Name
***** INCLUDE ITEM 4 ONLY IF KEYWORD ‘ATO’ IS USED IN ITEM 2A.
MPH File Name is the complete MPH file name generated as a result of the MODFLOW-
SURFACT/MODHMS simulation. The MPH file is generated only when the flag
MPHFLAG is greater than 0 in Basic Flow (BAS) package.
Item 3 -- Layer Type Codes

3. LAYCON(NLAY)
LAYCON values used in the BCF4 package of MODFLOW-SURFACT/MODHMS

should be used here. A LAYCON value must be provided for all layers. All values in
this file are read in free format. LAYCON values should be separated by at least 1 space.
For details on LAYCON, refer to the Input Instructions Section of the BCF4 Package.
Item 4 -- Confining Bed Codes

***** INCLUDE ITEM 4 ONLY IF NCBL > 0
4. NCON(NLAY)
NCON -- is the array containing the confining layer code for each model layer.
0 = no quasi-3D confining layer associated with the model layer.
1 = a quasi-3D confining layer is associated with the model layer.
Item 5 -- Horizontal Grid Spacing
5A. DELR(NCOL)
FORMAT: 1-D real array reader (U1DREL)
5B. DELC(NROW)
DELR -- is the grid spacing along rows (x-direction in MODPATH).
DELC -- is the grid spacing along columns (y-direction in MODPATH).

Item 6 -- Model Layer and Confining Layer Thickness
***** INCLUDE ITEM 6 ONLY IF IGRID=1
***** When ITEM 6 is used, ITEM 7 must be omitted
6A. DELZ(NLAY)
***** INCLUDE 6B ONLY IF NCBL > 0
6B. DELZCB(NLAY)
6C. ZBL1
DELZ -- is the grid spacing (cell thickness) in the vertical direction (z-direction
in MODPATH). The value input for DELZ of the top layer is ignored
if the top layer is a water-table layer (LAYCON = 1).
DELZCB -- is the thickness of quasi-3D confining layers. NLAY values must be
read even if only some of the model layers have underlying quasi-3D
confining layers. A value of zero should be entered for model layers
that are not underlain by quasi-3D confining layers. The bottom layer
cannot have an underlying quasi-3D confining layer, so the value of
DELZCB for the bottom layer must be zero.
ZBL1 -- is the bottom elevation of layer 1. This elevation datum is used to
calculate the elevations for all layers from layer thicknesses.
Item 7 -- Top and Bottom Elevations for Model Layers

***** INCLUDE ITEM 7 ONLY IF IGRID=0
***** When ITEM 7 is used, ITEM 6 must be omitted
REPEAT THE FOLLOWING DATA IN SEQUENCE FOR EACH LAYER:
**** DO NOT INCLUDE 7A FOR LAYER 1 IF LAYCON=1 ****
7A. ZTOP(NCOL,NROW)
7B. ZBOT(NCOL,NROW)
ZTOP -- is the array containing the elevation of the top of the model layer. The
ZTOP array is not specified for layer 1 when layer 1 is unconfined
(LAYCON=1).
ZBOT -- is the array containing the elevation of the bottom of the model layer. If the
layer contains an underlying quasi-3D confining layer, ZBOT corresponds
to the top of the confining layer. The thickness of that confining layer is
calculated by taking the difference in elevation between ZBOT and the
elevation of the top of the underlying model layer.

Item 8 -- Boundary Array
In the current version of MODFLOW, the IBOUND array is read one layer at a time by
specifying a 2-D array of data for each layer in the grid. In this version of MODPATH,
if the model is a 1-row cross-section and the keyword XSECTION is specified in the
options data record (Main data file item 2), the IBOUND array will be read as a single 2-D
array for the entire cross-section. Future versions of MODFLOW may support the
XSECTION option for IBOUND data in cross-sectional models.
If this IS NOT a 1-row cross-section, read a 2-D array of data for each layer by repeating
the following data sequence for each layer:
8. IBOUND(NCOL,NROW)
FORMAT: 2-D integer array reader (U2DINT)
Or, if this IS a 1-row cross-section with XSECTION specified, read a single 2-D array
for the cross-section:
8. IBOUND(NCOL,NLAY)
FORMAT: 2-D integer array reader (U2DINT)
IBOUND -- is the boundary array containing cell type codes.
IBOUND < 0 : specified head cell
IBOUND = 0 : inactive cell
IBOUND > 0 : active cell
The absolute value of IBOUND is the zone code, which is used by MODPATH-PLOT to
determine the color of pathlines and particle points. MODPATH requires that the absolute
value of IBOUND be less than 1000.
Item 9 -- Porosity
REPEAT THE FOLLOWING DATA IN SEQUENCE FOR EACH LAYER:
9A. POR(NCOL,NROW)
If the model layer has a quasi-3D confining layer, then
9B. PorCB(NCOL,NROW)
POR -- is the array containing porosity values for a model layer.
PorCB -- is the array containing porosity values for a quasi-3D confining layer
associated with the model layer. If a model layer includes a quasi-3D
confining layer, the PorCB array immediately follows the POR array for
that model layer.

Item 10 -- Stress Period and Time Step Data
***** ITEM 10 IS NOT REQUIRED FOR STEADY STATE SIMULATIONS
10A. TBEGIN
TBEGIN -- is the time value assigned to the beginning of the MODFLOW simulation.
Any convenient value may be specified, including values less than zero.
Item 10B should be repeated NPER times, once for each stress period.
10B. Perlen Nstp Tsmult
Perlen -- is the stress period length.
Nstp -- is the number of time steps in the stress period.
Tsmult -- is the time step multiplication factor.
Note that if keyword ‘ATO’ is used in Item 2A, time stepping information in each stress
period is read from MPH file specified in Item 2B. Item 10 B is however, still required
as input.
10C. BeginPeriod BeginStep MaxPeriod MaxStep

Data item 10C marks the range of time steps that will be processed from the MODFLOW
cell-by-cell budget file and used to generate the MODPATH composite budget file. The
MODFLOW cell-by-cell budget file must contain data for all of the time steps within the
range specified in data item 10C. The MODFLOW cell-by-cell budget file also may
contain data for time steps that are outside of the specified range.
BeginPeriod, BeginStep -- the stress period and time step of the first time step in the
block of time steps that will be processed.
MaxPeriod, MaxStep -- the last stress period and maximum number of time steps
in any stress period in the block of time steps that will be
processed.
Item 10C is not read if the MODPATH analysis uses an existing composite budget file.

F.2.2 Name File
The name file contains one line for each input data file. This file provides the
information needed to manage the input data files. MODPATH and MODPATH-PLOT
prompt the user to specify the name file. The file then is opened internally within the
program. Each line consists of 3 items: (1) a character string signifying the type of data
file, (2) an integer file unit number, and (3) the file name. The data file is read using free
format input. Any valid unit number for a given operating system may be specified.
However, MODPATH reserves several numbers in the range 80 to 99 for internal use.
Files can be specified in any order. For a typical simulation, the name file might look like:
main 11 main.dat
rch 12 recharge.dat
wel 13 well.dat
data 31 ibound.1
budget 50 budget.out
head(binary) 60 head.out
The first data item on each line is a string of up to 15 characters that identifies the
type of data file. The character string identifier is not case sensitive. MODPATH and
MODPATH-PLOT reserve several special character strings to indicate specific data files:
MAIN = main data file
RCH = recharge package
WEL = well package
RIV = river package
STR = stream package
DRN = drain package
GHB = general head boundary package
EVT = evapotranspiration package
FWL = FWL4 fracture well package
WL5 = multinode fracture well package (FWL5)
CBF = composite budget file
DCF = drawing commands file
ENDPOINT = endpoint file
PATHLINE = pathline file
TIME-SERIES = time series file
TIME = time data file
LOCATIONS = starting locations file
BUDGET = binary budget file produced by MODFLOW
HEAD(BINARY) = binary head file produced by MODFLOW
HEAD = text head file produced by MODFLOW
SAT(BINARY) = binary saturation file produced by MODFLOW-
SURFACT
SAT = text saturation file produced by MODFLOW-
SURFACT
DRAWDOWN(BINARY) = binary drawdown file produced by MODFLOW
DRAWDOWN = text drawdown file produced by MODFLOW
CONTOUR-DATA = text file containing 2-D data to contour
CONTOUR-LEVEL = text file containing contour levels
DATA = ancillary text input data files
LIST = summary output file
GUA = grid unit array file

Ancillary data files usually contain large arrays that are referenced by array control
records in other data files. Ancillary data files always must be declared as type DATA.
MODPATH’s data file requirements are summarized in Table F-1.
Table F-1 Summary of Data File Requirements
Needed for
Needed for MODPATH Include in
Description File Type Keyword MODPATH? PLOT? Name file?
Main data file MAIN yes yes yes
Recharge package RCH yes1 no yes
Well package WEL yes yes yes
River package RIV yes1 yes yes

1
Stream package STR yes yes yes
Drain package DRN yes1 yes yes
GHB package GHB yes1 yes yes
ET package EVT yes1 no yes
FWL4 package FWL yes1 yes yes
FWL5 package WL5 yes yes yes
Composite budget CBF transient only transient cross- optional3

file sections only
Drawing DCF no user’s option optional3

commands file for map view
plots
Endpoint file ENDPOINT yes yes optional3
Pathline file PATHLINE pathline mode pathline mode optional3
Time series file TIME-SERIES time series time series optional3

mode mode
Time data file TIME user’s option no optional3
Starting locations LOCATIONS user’s option no optional3

file
Binary budget file BUDGET yes1 no yes
Head file (text or HEAD yes1 yes2 yes

binary) HEAD(BINARY)
Saturation file SAT yes4 no yes4

(text or binary) SAT(BINARY)
Drawdown file DRAWDOWN no user’s option yes

(text or binary) DRAWDOWN (BINARY) for map view
plots
Contour data file CONTOUR-DATA no user’s option optional3

for map view
plots

Needed for
Needed for MODPATH Include in
Description File Type Keyword MODPATH? PLOT? Name file?
Contour level file CONTOUR-LEVEL no user’s option optional3
for map view
plots
Ancillary text DATA user’s option user’s option yes

data file
Summary output LIST yes yes optional3

file
Grid unit array GUA no user’s option yes

file
Notes: 1. MODFLOW stress package, budget, and head files are not used by MODPATH in transient simulations
that read data directly from a composite budget file.
2. The MODFLOW head file is not required by MODPATH-PLOT for some plot types. See section Plot
Types in Chapter 4.
3. Users have the option of specifying these files in the name file or allowing MODPATH to prompt for file
names or, in some cases, assign default file names.
4. Saturation file is needed only if real soil moisture functions are used to define flow in the unsaturated zone
above the water table (IREALSL >0 in BCF file), in a MODFLOW-SURFACT/MODHMS simulation.
MODFLOW-SURFACT/MODHMS writes saturation to DDN file if real soil moisture functions are used.

F.3.0 EXAMPLES
Three sample problems are presented to illustrate application of MODPATH-
SURFACT. All the problems are based on the same hypothetical flow system shown in
Figure 1. There are 9 fracture wells performing extraction pumping at a rate of 50,000
ft3/day/well. These wells extract water from all 5 layers. All the layers use LAYCON =
43 option of the BCF4 package used by MODFLOW-SURFACT. Recharge occurs at the
western portion of the model at a rate of 0.1 ft/day as shown in Figure 1.
F.3.1 Example 1
In this example the system is at steady-state. MODFLOW-SURFACT was used
to simulate flow with pseudo-soil water retention functions. This example demonstrates
the use of LAYCON option 43 in BCF4 package, and the use of fracture wells via the
FWL4 package when using MODPATH-SURFACT.
The resulting head distribution in Layer 3 is shown in Figure 2. Particle tracking
was performed using this version of MODPATH. Particles were introduced in layer 1 of
the model. Figure 2 also shows particle pathlines in plan view as well as the
cross-sectional view.
F.3.2 Example 2
Example 2 is a transient flow simulation performed on the problem described in
Example 1. The simulation was performed for a 100 day period. This example
demonstrates how to use MODPATH when a MODFLOW-SURFACT simulation uses the
ATO package.
Results for this simulation are presented in Figure 3. This figure presents head
distribution in layer 3 as well as particle pathlines. These particles were introduced in layer
1 of the model.
F.3.3 Example 3
This example is an extension of Example 2. Example 3 uses real-soil moisture
retention functions for a transient MODFLOW-SURFACT simulation. This illustrates the
use of MODPATH with MODFLOW-SURFACT when REALSL option is used in the
BCF4 package.
Results of this simulation are presented in Figure 4.
F.4.0 REFERENCES
Pollock, David W., 1994, User's Guide for MODPATH/MODPATH-PLOT Version 3:
A particle tracking post-processing package for MODFLOW, the U. S. Geological
Survey finite-difference ground-water flow model, U. S. Geological Survey
Open-File Report 94-464.

Recharge
0.1 ft/day
100 ft
80 ft
60 ft
Cross-Section
40 ft
20 ft
0 ft
Areal View
(100 cells x 50ft)

5000 ft
Fracture Wells
5000 ft
Constant Head (100 cells x 50ft) Constant Head
50 ft 30 ft
Figure 1. Groundwater grid and boundary conditions applied to the example
problem.

Figure 2. Head distribution and particle pathlines in model layer 3 at
steady-state condition for Example 1. Particles were introduced in
model layer 1.

time=100 days for a transient simulation in Example 2. Particles
were introduced in model layer 1.

time=100 days for a transient simulation in Example 3. This
example is simulated with real-soil moisture retention functions
used by MODFLOW-SURFACT. Particles were introduced in
model layer 1.

APPENDIX G: USING ET TIME-SERIES WITH THE
EVAPOTRANSPIRATION (EVT) PACKAGE OR WITH THE
SEGMENTED EVAPOTRANSPIRATION (ET2) PACKAGE
APPENDIX G: USING ET TIME-SERIES WITH THE
EVAPOTRANSPIRATION (EVT) PACKAGE OR WITH THE
SEGMENTED EVAPOTRANSPIRATION (ET2) PACKAGE
G.1 GENERAL
ETmax, the maximum rate of evapotranspiration, is taken to be constant within a
stress period of MODFLOW. For rapid changes in ETmax, it would then become tedious
to break up a short time period into several stress periods during a simulation. An option
has therefore been provided, to vary ETmax within a simulation, independent of the stress
periods. The transient ETmax is applied zonally, and an index, MXZETS, is read by the
EVT or ET2 packages indicating the maximum number of zones on which transient ETmax
is applied. If MXZETS is zero, the transient ETmax package is not invoked by the
simulation. However, if transient ETmax exists, an additional integer array is read by the
EVT or ET2 packages, identifying the areal location of each zone. A zonal value of zero
indicates that no transient ETmax zone occurs over that location, and the default value of
ETmax is applied to that location, as read in for the stress period in the EVTR or ETSR
arrays. At locations where the value is not zero, the EVTR or ETSR array is updated with
the appropriate value for that zone, from the ETmax time series data which is supplied in
a separate file. This file, with default extension .ETS, contains the ETmax time series
tabulated for all zones. Each row of the file contains the starting time (TSTART), the
ending time (TEND), a multiplying factor (for unit conversions), followed by one column
of ETmax rates for each of the zones. If a subsequent TSTART value does not match with
a TEND value from the previous row, zero ETmax is applied to each of the zones in the
interim period. Thus, long periods without ET need not be included as a separate row in
the .ETS file.
The time periods between events, or within an event need not be fixed, as this
module is designed for use with adaptive time stepping schemes. The adaptive time
marching scheme ensures that each TSTART and TEND is hit exactly by the simulation,
to ensure that the temporal changes are accurately represented. Within a period or
between periods, the adaptive time stepping schemes try to maximize computational
efficiency.
Details on the EVT and ET2 packages are available in the respective USGS reports
documenting these packages. Note that the ET2 package is denoted as ETS1 in the
original MODFLOW documentation (Banta, 2000). The current documentation describes
extension of these packages to include time-series input of ETmax.
G.2 EVAPOTRANSPIRATION (EVT) PACKAGE INPUT

Input to the Evapotranspiration (EVT) Package is read from the unit specified in
IUNIT (5).
FOR EACH SIMULATION
1. Data: NEVTOP IEVTCB MXZETS2
Format: I10 I10 I10
FOR EACH STRESS PERIOD
Appendix G: Using ET Time-Series with the EVT and ET2 Packages G-1
M:\Projects\MODFLOW Manuals\SURFACT and HMS Ver.3\Vol I\Appendix G.wpd
EVT1RP
2. Data: INSURF INEVTR INEXDP INIEVT
**Enter item 3 only if INSURF $ 0
3. Data: SURF
Module: U2DREL
**Enter item 4a only if INEVTR$0 and MXZETS2=0
4a. Data: EVTR
Module: U2DREL
**Enter item 4b only if INEVTR$0 and MXZETS2…0
4b Data: IZNETS2
Module: U2DREL
Note that either item 4a or 4b is read depending on whether a time-series for maximum ET
flux (EVTR) is requested or not
**Enter item 5 only if INEXDP$0
5. Data: EXDP
Module: U2DREL
**Enter item 6 only if NEVTOP=2 and INIEVT$0
6. Data: IEVT
Module: U2DINT

NEVTOP -- is the evapotranspiration (ET) option code. ET parameters (ET
surface, maximum ET rate, and extinction depth) are specified in
two-dimensional arrays, SURF, EVTR, and EXDP, with one value
for each vertical column. Accordingly, ET is calculated for one
cell in each vertical column. The option codes determine for which
cell in the column ET will be calculated.
1 - ET is calculated only for cells in the top grid layer.
2 - The cell for each vertical column is specified by the user in
array IEVT.
3 - ET is calculated for highest active cell.
G-2 Appendix G: Using ET Time-Series with the EVT and ET2 Packages
IEVTCB -- is a flag and a unit number.
If IEVTCB > 0, it is the unit number on which cell-by-cell flow
If IEVTCB # 0, cell-by-cell flow terms will not be printed or
recorded.
MXZETS2 -- Index for input of zonal time-series of the maximum ET flux

(EVTR) via a separate ETS file.
= 0 if no, maximum ET flux (EVTR) is read as per stress
periods.
= n if yes, a separate file is used to read time-series of
maximum ET flux (EVTR), and n is the total number of
ET zones used in the simulation.
Note: The ET time-series (ETS) associated with zonal input of
ETmax is read as per instructions provided in Section G.4.
INSURF -- is the ET surface (SURF) read flag.
If INSURF $ 0, an array containing the ET surface elevation

will be read.
If INSURF < 0, the ET surface from the preceding stress period
will be reused.
INEVTR -- is the maximum ET rate (EVTR) read flag.
If INEVTR $ 0, an array containing the maximum ET rate will

be read.
IF INEVTR < 0, the maximum ET rate from the preceding
stress period will be reused.
IZNETS2 -- is the integer array defining the zone value of each node, for zonal
input of the ET time series.
Note that the applied ET will be zero in cells with a zone value of
zero.
INEXDP -- is the extinction depth (EXDP) read flag.
If INEXDP $ 0, an array containing the extinction depth (EXDP)

will be read.
If INEXDP < 0, the extinction depth from the preceding stress
INIEVT -- is the layer indicator (IEVT) read flag. It is used only if the ET
option (NEVTOP) is equal to two.
If INIEVT $ 0, an array containing the layer indicators (IEVT)

will be read.
If INIEVT < 0, layer indicators used during the preceding
SURF -- is the elevation of the ET surface.
EVTR -- is the maximum ET rate (volume of water per unit area (LT-1)).
EXDP -- is the ET extinction depth.
IEVT -- is the layer indicator array. For each horizontal location, it
indicates the layer from which ET is removed. It is needed only if
the ET option is equal to two.
G.3. EVAPOTRANSPIRATION SEGMENTS (ET2) PACKAGE

Input to the ET2 Package is read from the file that is type “ET2” in the name file or
from the unit specified in IUNIT (41). All single-valued variables are free format if the
option “FREE” is specified in the Basic Package input file; otherwise, the variables have
10-character fields. Arrays are read by an array-reading utility module, either U2DREL
or U2DINT (Harbaugh and others, 2000), as indicated. Optional variables are shown in
brackets.
FOR EACH SIMULATION
0. #Text
Item 0 is optional - “#” must be in column 1. Item 0 can be repeated as many
times a desired.
1. NETSOP IETSCB NPETS NETSEG MXZETS1
2. INETSS INETSR INETSX [INIETS [INSGDF] ]
3. ETSS (NCOL, NROW) - U2DREL - If INETSS $0
4a. ESTR (NCOL, NROW) - U2DREL - If INETSR $0 and MXZETS1=0
4b. IZNETS1 (NCOL, NROW) - U2DREL - If INETSR $0 and MXZETS1=0.
Note that either item 4a or 4b is read depending on whether a time series for
maximum ET flux (ETSR) is requested or not.
5. ETSX (NCOL, NROW) - U2DREL - If INETSX $0
6. IETS (NCOL, NROW) - U2DINT - If NETSOP = 2 and if INIETS $0
7. PXDP (NCOL, NROW) - U2DREL If NETSEG > 1 and INSGDF $0
8. PETM (NCOL, NROW) - U2DREL If NETSEG > 1 and INSGDF $0
If NETSEG >1, (NETSEG - 1) repetitions of Items 7 and 8 are read. If NETSEG >2,
Items 7 and 8 are read for the uppermost segment intersection, followed by repetitions of
Items 7 and 8 for successfully lower intersections.
Explanation of Variables Read by the Evapotranspiration Segments Package
Text - is a character variable (79 characters) that starts in column 2. Any
characters can be included in Text. The “#” character must be in column
1. Lines beginning with “#” are restricted to these first lines of the input
file. Text is written to the LIST output file when the input file is read
and provides an opportunity for the user to include information about the
model both in the input file and the associated output file.
NETSOP - is the evapotranspiration (ET) option code. ET variables (ET surface,
maximum ET rate, and extinction depth) are specified in layer variables,
ETSS, ETSR, and ETSX, with one value for each vertical column of cells
in the model grid. Accordingly, ET is calculated for one cell in each
vertical column. The option codes determine the cell within a column for
which ET will be calculated. If NETSOP = 1, ET is calculated only for
cells in the top grid layer.
If NETSOP - 2, the cell for each vertical column is specified by the user
in variable IETS.
If NETSOP = 3, ET is calculated for the top most active grid block.
IETSCB - is a flag and unit number.
If IETSCB > 0, it is the unit number to which ETS1-Package cell-by-cell
flow terms will be written when “SAVE BUDGET” or a non-zero value
for ICBCFL is specified in Output Control (Harbaugh and others, 2000).
IETSCB must be a unit number associated with a file listed with type
“DATA (BINARY)” or “DATAGLO (BINARY)” in the name file.
If IETSCB #0, ETS1-Package cell-by-cell flow terms will not be written.
NPETS - is the number of evapotranspiration-segments parameters. No used. Set
equal to zero.
NETSEG - is the number of segments used to define the relation of evapotranspiration
rate to hydraulic head in the interval where the evapotranspiration rate is
variable.
MXZETS - Index for input of zonal time-series of the maximum ET flux

(ETSR) via a separate ETS file
= 0 if no, maximum ET flux (ETSR) is read as per stress-
periods.
= n if yes, a separate file is used to read time-series of maximum
ET flux (ETSR), and n is the total number of ET zones used in
the simulation.
Note: The ET time-series file (ETS) associated with zonal input of
ETmax is read as per instructions provided in Section G.4.
INETSS - is the ET surface (ETSS) read flag.
If INETSS $ 0, a layer variable containing the ET surface (ETSS) will be

read from Item 5 in the ETS1 input file.
If INETSS < 0, the ET surface from the preceding stress period will be
reused.
INETSR - is the ESTR read flag. Its function depends on whether or not parameters
are being used.
If INETSR $0, a layer variable containing the maximum ET rate (ETSR)

will be read from Item 6 in the ETS1 input file.
If INETSR <0, the maximum ET rate from the preceding stress period
will be reused.
IZNETS1 - Integer array defining the zone value of each node, for zonal input of the
ET time series.
Note that the applied ET will be zero in cells with a zone value of zero.
INETSX - is the extinction depth (ETSX) read flag.
If INETSX $0, a layer variable containing the extinction depth (ETSX)

If INETSX <0, the extinction depth from the preceding stress period will
be reused.
INIETS - is the layer indicator (IETS) read flag. It is read if the option (NETSOP)
is equal to two or if NETSEG >1. If NETSEG >1 and NETSOP is not
equal to two, INIETS is ignored and IETS is not read.
If INIETS $0, a layer variable containing the layer indicator (IETS) will
be read from Item 9 in the ETS1 input file.
If INIETS < 0, layer indicators used during the preceding stress period
will be reused.
INSGDF - is the segment definition read flag. It is read only if NETSGEG > 1.
If INSGDF $0, two layer variables to define PXDP and PETM for each
fo (NETSEG - 1) segment intersections are read from Items 10 and 11,
respectively, of the ETS1 input file.
If INSGDF <0, PXDP and PETM from the preceding stress period will
be reused.
ETSS - is the elevation of the ET surface (L).
ETSR - is the maximum ET flux (volumetric flow rate per unit area, LT1)
Pname - is the name of a parameter that will be used to define the ETSR the ETSR
variable in the current stress period.
Iname - is an instance name that is ready only if Pname is a time-varying
parameter. Multiple instances of the same time-varying parameter are not
allowed in a stress period.
IETSPF - is an optional format code for printing the ETSR variable after it has been
defined by parameters. The format codes are the same as those used in
the U2DREL array reading utility module (Harbaugh and others, 2000).
ETSX - is the ET extinction depth (L). This variable is read only if INETSX $0.
IETS - is the layer indicator variable. For each horizontal location, IETS
indicates the layer from which ET is removed. It is read only if the ET
option (NETSOP) is equal to two and if INIETS $0.
PXDP - is a proportion of the extinction depth (dimensionless), measured
downward from the ET surface, which, with PETM, defines the shape of
the relation between the evapotranspiration rate and head. The value of
PXDP must be between 0.0 and 1.0, inclusive. Repetitions of PXDP and
PETM are read in sequence such that the first occurrence represents the
bottom of the first segment, and subsequent repetitions represent the
bottom of successively lower segments. Accordingly, PXDP values for
later repetitions (representing lower segments) should be greater than
PXDP values for earlier repetitions.
PETM - is a proportion of the maximum evapotranspiration rate (dimensionless)
which, with PXDP, defines the shape of the relation between the
evapotranspiration rate and head. The value of PETM should be between
0.0 and 1.0, inclusive. Repetitions of PXDP and first segment, and
subsequent repetitions represent the bottoms of successively lower
segments. Accordingly, PETM values for later repetitions (representing
lower segments) generally would be less than PETM values for earlier
repetitions.
G.4 TRANSIENT POTENTIAL EVAPOTRANSPIRATION TIME-SERIES FILE

The transient ETmax time series file is used with either the EVT or the ET2
package, when the flag MXZETS is not zero in their respective inputs. This file contains
the time series information for each of the MXZETS zones of ETmax. Note that the
zonation is defined in the respective EVT or ET2 input files.
The transient potential evapotranspiration time-series file is opened on FORTRAN
Unit 97, when transient ET conditions are applied. The file is in free format and contains
one row for each ET period within the simulation. The default filename for the file is
fname.ETS where fname is the filename for the simulation. Data entries on each row are
as follows:
Data: TSTARTE TENDE TMULTE EZ(1) EZ(2)...... EZ(MXZETS)
Format: Free
Explanation of fields used in input.

TSTARTE - Starting time of input record for time-varying PET
TENDE - Ending time of input record for time-varying PET
Note that TSTARTE and TENDE are referenced to the total simulation
time and not to the stress period time, when multiple stress periods
occur in a simulation.
TMULTE - Multiplying factor applied to ET rates of all the zones for the given event.
This scaling is convenient for unit conversions.
EZ(1) - ET rate in Zone 1 for this event.
EZ(2) - ET rate in Zone 2 for this event.
•
•
•
EZ(MXZETS) - ET rate in Zone MXZETS for this event.
A sample ETS file is shown in Table G.1. In this example, the domain is divided into 8
zones of ET. No ET occurs for the first 2.375 days of simulation, so the simulation would
proceed rapidly till the first record, adapting its time-step size depending on convergence
behavior. At 2.375 days, ET occurs over some of the zones, for a one hour period, till
2.417 days. At 2.417 days, the intensity (and/or zone) of the ET changes, with these
entries occurring in the second row of the file, and so on. Note that no ET occurs between
2.5 and 4 days, and an entry is not required in the file for this situation.
Table G.1 Example input file for ET time series
TSTARTE TENDE EZ(1) EZ (2) EZ EZ EZ EZ EZ EZ

(d) (d) TMULTE ft/d ft/d (3) (4) (5) (6) (7) (8)
2.375 2.417 1.0 0 0 0.225 0.177 .033 0 0 0
2.417 2.568 1.0 0 0 0 0.13 0.13 0.022 0 0
2.458 2.500 1.0 0 0 0 0 0.088 0 0 0
4.000 4.042 1.0 0.169 0.551 0.551 0.337 0.178 0 0 0
4.042 4.083 1.0 0.151 0.447 0.333 0.106 0.08 0 0.691 0.633
4.083 4.125 1.0 0.100 0.050 0 0 0 0.150 0.379 0.300
4.125 4.167 1.0 0 0 0 0 0 0 0.083 0.022
Note: MXZETS = maximum number of ET zones = 8
APPENDIX H: TIME VARYING FLOW AND HEAD BOUNDARY PACKAGE
WITH STEP FUNCTION OPTION
APPENDIX H: TIME VARYING FLOW AND HEAD BOUNDARY PACKAGE
WITH STEP FUNCTION OPTION
H.1 GENERAL
The Time-Variant Flow and Head Boundary Package was developed at the USGS
(Leake and Lilly, 1997) to allow prescribed flux and prescribed head cells to take on
different values for each time step. It provides a means of simulating flux variations and
head variations with time that can be provided in a tabular fashion. This package is
adapted here to extend the time-varying prescribed flux and prescribed heads to include
step-functions between tabulated time values as an additional option to interpolation.
Details of the FHB package are provided in the USGS documentation of the FHB
package. The step function implementation of this package is done via an input flag
ITHTFHB. When this flag is set to zero, the original MODFLOW implementation fo this
package is used with interpolation of head or flux values from the tabular input. Note that
adaptive time-stepping may be used with this package and the computational time values
need not honor the tabulated time values of the FHB package. This is also the case for use
of MODFLOW’s time stepping scheme. However, if adaptive time stepping is used and
if the flag ITHTFHB is set to 1 then adaptive time-stepping will honor the time values
provided in the tabular time array for FHB input. Head and flux values for intermediate
times are still interpolated from the tabular input values. If, however, the flag ITHTFHB
is set to 2, the head and flux values are implement as step functions instead of
interpolation. Note that this package has also been extended for the surface-water domains
of MODHMS.
H.2 INPUT INSTRUCTIONS FOR FHB PACKAGE
Input instructions for the FHB package are an extension to input instructions for
the FHB package specified by Leake and Lilly (1997). Input for FHB is read from
IUNIT(30), specified in the Basic Package input (McDonald and Harbaugh, 1988, chap.
4, p. 9–11). All input is free format, which requires each of the numbered data groups to
start on a new input record. More than one record can be used for any data group and
numbers within data groups must be separated by at least one space or a comma. Integer
data types cannot include a decimal point. Blank spaces are not treated as zeros. The FHB
package has been extended to include a step function capability in addition to interpolation
of the prescribed time-series of head or flux at any node. For a transport simulation, the
FHB package further includes possible temporal variations of concentrations of species at
the flow and head boundary nodes. For flow boundary nodes, the parameter IAUX of
each node determines if the concentrations of species are input or whether a zero value is
applied by default. For head boundary nodes, the parameter IAUX further indicates if
concentrations from the HCN package are to be used.
FOR EACH SIMULATION

1. Data: NBDTIM NGWFLW NGWHED IFHBSS IFHBCB NFHBX1 NFHBX2
Type: Integer Integer Integer Integer Integer Integer Integer
NOLFLW NOLHED NCHFLW NCHHED ITHTFHB IFHBCC

Integer Integer Integer Integer Integer Integer
Note that the shaded items, NOLFLW, NOLHED, NCHFLW and NCHHED are used
only with MODHMS and may be set to zero for a SURFACT simulation.
Appendix H: Flow and Head Boundary Package with Step Function Option H-1
M:\Projects\MODFLOW Manuals\SURFACT and HMS Ver.3\Vol I\Appendix H.wpd
Omit data item 2 if NFHBX1=0. Input item 2 consists of one record for each of NFHBX1
auxiliary variables.
2. Data: VarName Weight
Type: Character Real
Omit data item 3 if NFHBX2=0. Input item 3 consists of one record for each of NFHBX2
auxiliary variables.
Data items 4a and 4b are required for all simulations. Include NBDTIM times in data item
4b.
4a. Data: IFHBUN CNSTM IFHBPT
Type: Integer Real Integer
4b. Data: BDTIM(NBDTIM)
Type: Real
Omit data item 5a if all NGWFLW, NOLFLW and NCHFLW=0.
Omit data item 5b if NGWFLW=0. Input item 5b consists of one set of numbers for each
of NGWFLW cells. Each set of numbers includes layer, row and column indices, an
integer auxiliary variable, and NBDTIM values of specified flow.
5b. Data: Layer Row Column IAUX FLWRAT(NBDTIM)
Type: Integer Integer Integer Integer Real
Omit data item 5bii if ITRAN = 0 or IAUX = 0
5bii. Data: FBCONC (NSPECI, NBDTIM)
Type: Real
Note that data item 5bii consists of one entry for each of the NBDTIM interpolation time
values for a species. Data item 5bii is repeated for each species associated with transport.
Also note that if IAUX (of data item 5b) = 0, the concentration of all species is taken as
zero for this node.
Omit data item 5c if NOLFLW=0. Input item 5c consists of one set of numbers for each of
NOLFLW cells. Each set of numbers includes overland row and column indices, an integer
auxiliary variable, and NBDTIM values of specified flow.
5c. Data: Row Column IAUX FLWRAT(NBDTIM)

Type: Integer Integer Integer Real
Omit data item 5cii if ITRAN = 0 or IAUX = 0

5cii. Data: FBCONC (NSPECI, NBDTIM)
Type: Real
Note that data item 5cii consists of one entry for each of the NBDTIM interpolation
time values for a species. Data item 5cii is repeated for each species associated with
transport. Also note that if IAUX (of data item 5c) = 0, the concentration of all
species is taken as zero for this node.
H-2 Appendix H: Flow and Head Boundary Package with Step Function Option
Omit data item 5d if NCHFLW=0. Input item 5d consists of one set of numbers for
each of NCHFLW cells. Each set of numbers includes channel segment number, an
5d. Data: Segment IAUX FLWRAT(NBDTIM)

Type: Integer Integer Real
Omit data item 5dii if ITRAN = 0 or IAUX = 0

5dii. Data: FBCONC (NSPECI, NBDTIM)
Type: Real
Note that data item 5dii consists of one entry for each of the NBDTIM interpolation
time values for a species. Data item 5dii is repeated for each species associated with
transport. Also note that if IAUX (of data item 5d)=0, the concentration of all
Note that items 5c and 5d are only required for a MODHMS simulation.
Omit data items 6a and 6b if NFHBX1=0 or if all NGWFLW, NOLFLOW and

NCHFLOW=0. Include one set of data items 6a and 6b for each of NFHBX1 auxiliary
variables. Input item 6b consists of one set of numbers for each of NGWFLW, NOLFLW
and NCHFLW cells. Each set includes NBDTIM values of the variable.
6b. Data: AuxVar(NBDTIM)
Type: Real
Omit data item 7a if all NGWHED, NOLHED and NCHHED=0.

Omit data item 7b if NGWHED=0. Input item 7b consists of one set of numbers for each
of NGWHED cells. Each set of numbers includes layer, row, and column indices, an
integer auxiliary variable, and NBDTIM values of specified head.
7b. Data: Layer Row Column IAUX SBHED(NBDTIM)
Omit data item 7bii if ITRAN = 0 or IAUX # 0

7bii. Data: HBCONC (NSPECI, NBDTIM)
Type: Real
Note that data item 7bii consists of one entry for each of the NBDTIM interpolation time
values for a species. Data item 7bii is repeated for each species associated with transport.
Also note that if IAUX (of data item 7b) = 0, the concentration of all species is taken as
zero for this node. Further, if IAUX (of data item 7b)<0, the concentrations input in the
HCN package are used for the simulation.
Omit data item 7c if NOLHED=0. Input item 7c consists of one set of numbers for
each of NOLHED cells. Each set of numbers includes overland, row, and column
indices, an integer auxiliary variable, and NBDTIM values of specified head.
7c. Data: Row Column IAUX SBHED(NBDTIM)

Omit data item 7cii if ITRAN = 0 or IAUX # 0

7cii. Data: HBCONC (NSPECI, NBDTIM)
Type: Real
Note that data item 7cii consists of one entry for each of the NBDTIM interpolation
time values for a species. Data item 7cii is repeated for each species associated with
transport. Also note that if IAUX (of data item 7c) = 0, the concentration of all
species is taken as zero for this node. Further, if IAUX (of data item 7c)<0, the
concentrations input in the HCN package are used for the simulation.
Omit data item 7d if NCHHED=0. Input item 7d consists of one set of numbers for
each of NCHHED cells. Each set of numbers includes channel segment number, an
integer auxiliary variable, and NBDTIM values of specified head.
7d. Data: Segment IAUX SBHED(NBDTIM)

Omit data item 7dii if ITRAN = 0 or IAUX # 0

7dii. Data: HBCONC (NSPECI, NBDTIM)
Type: Real
Note that data item 7dii consists of one entry for each of the NBDTIM interpolation
time values for a species. Data item 7dii is repeated for each species associated with
transport. Also note that if IAUX (of data item 7d)=0, the concentration of all
species is taken as zero for this node. Further, if IAUX (of data item 7d)<0, the
Omit data items 8a and 8b if NFHBX2=0 or if NGWHED, NOLHED and NCHHED all
=0. Include one set of data items 8a and 8b for each of NFHBX2 auxiliary variables.
Input item 8b consists of one set of numbers for each of NGWHED, NOLHED, NCHEED
cells. Each set includes NBDTIM values of the variable.

Type: Real
NBDTIM -- is the number of times at which flow and head will be specified
for all selected cells.
If NBDTIM = 1, specified flow and head values will
remain constant for the entire simulation.
If NBDTIM > 1, specified flow and head values will be
computed for each time step using linear
interpolation.
NGWFLW -- is the number of groundwater cells at which flows will be
specified.
NGWHED -- is the number of groundwater cells at which head will be specified.
IFHBSS -- is the FHB steady-state option flag. If the simulation is transient,
the flag is read but not used. For steady-state simulations, the
flag controls how specified-flow, specified-head, and
auxiliary-variable values will be computed for each steady-state
solution.
If IFHBSS = 0, values of flow, head, and auxiliary
variables will be taken at the starting
time of the simulation. This results in use
of the first value in arrays FLWRAT,
SBHED, and AuxVar for each respective
boundary cell.
If IFHBCB … 0, values of flow, head, and auxiliary
variables will be interpolated in the same
way that values are computed for
transient simulations.
IFHBCB -- is a flag and unit number.
If IFHBCB > 0, it is the unit number on which
cell-by-cell flow terms will be recorded
whenever ICBCFL is set (see McDonald
and Harbaugh, 1988, chap. 4, p. 14–15).
If IFHBCB # 0, cell-by-cell flow terms will not be
recorded.
NFHBX1 -- is the number of auxiliary variables whose values will be
computed for each time step for each specified-flow cell.
computed for each time step for each specified-head cell.
NOLFLW -- is the number of overland cells at which flows will be specified.
NOLHED -- is the number of overland cells at which heads will be specified.
NCHFLW -- is the number of channel cells at which flows will be specified.
NCHHED -- is the number of channel cells at which heads will be specified.
ITHTFHB -- is an index for controlling time-stepping such that tabular entries
for the time inputs are exactly met by the adaptive time-stepping
routines.
IFITHTFHB = 0, the adaptive time-stepping is independent of
tabular time inputs for FHB.
If ITHTFHB=1, the time steps are adapted to honor the tabular
time value entries of the FHB input.
If ITHTFHB=2, the head or flux values are provided as step
functions instead of interpolating between tabular time values
entries. Thus, the head or flux value associated with a time entry
is maintained till the next time entry at which stage the head or
flux values associated with the new time entry are implemented.
Note that time steps are adapted to honor the tabular time values
when ITHTFHB=2 to accurately capture each step change in
head or flux.
IFHBCC -- is a flag and unit number for printing mass flux output of a
transport simulation.
If IFHBCC>0, it is the unit number on which cell-by-cell mass
flux terms will be recorded wherever ICCCFL is set in the ATO
or OC packages.
If IFHBCC#0, cell-by-cell mass flux terms will not be recorded.
VarName -- is the name of an auxiliary variable. Name can include up to 16

characters with no embedded blank characters.
Weight -- is the time-weighting factor for an auxiliary variable specifying
the fraction of each time step at which the value of the variable
will be interpolated. Value must be in the range from 0.0 to 1.0.
IFHBUN -- is the unit number on which data lists will be read. The same or
different unit numbers can be used to read lists in data items 4b,
5b, 5c and 5d, 6b, 7b, 7c and 7d, and 8b.
CNSTM -- is a constant multiplier for data list BDTIM (data item 4b),
FLWRAT (part of data item 5b, 5c and 5d), SBHED (part of data
item 7b, 7c and 7d), and auxiliary variables in data items 6b and
8b.
IFHBPT -- is a flag for printing values of data lists in items 4b, 5b, 5c and
5d, 6b, 7b, 7c and 7d, and 8b.
If IFHBPT > 0 data list read at the beginning of the
simulation will be printed.
If IFHBCB < 0 data list read at the beginning of the
simulation will not be printed.
BDTIM -- is simulation time at which values of specified flow and (or)
values of specified head will be read. NBDTIM values are
required.
Layer -- is the layer index of specified-flow cell (data item 5b, 5c and 5d)
or specified-head cell (data item 7b, 7c, and 7d).
Row -- is the row index of specified-flow cell (data item 5b, 5c and 5d)
or specified-head cell (data item 7b, 7c and 7d).
Column -- is the column index of specified-flow cell (data item 5b and 5c)
or specified-head cell (data item 7b and 7c).
Segment -- is the channel segment number of specified-flow cell (data item 5d
and 7d).
IAUX -- is a dummy parameter.
FLWRAT -- is volumetric rate of flow at specified-flow cells. A list of
NBDTIM values must be specified for each of NFLW
specified-flow cells.
FBCONC -- is the concentration of species for fluid entering the domain via
the specified-flow cells. If a value is provided for a node during
outflow conditions, the value is ignored and the outflow
concentration is determined by the concentration of water leaving
the domain at the specified-flux cell.
AuxVar -- is value of real auxiliary variable at specified-flow and
specified-head cells. A list of NBDTIM values must be specified
for each of NFLW specified-flow cells and for each of NHED
specified-head cells.
SBHED -- is an array containing NBDTIM values of the head for each
specified-head cell.
HBCONC -- is the concentration of species for fluid entering the domain via
the specified-head cells. If a value is provided for a node during
outflow conditions, the value is ignored and the outflow
concentration is determined by the concentration of water leaving
the domain at the specified-head cell.
APPENDIX I: ADDENDUM FOR PCG5 SOLVER
APPENDIX I: ADDENDUM FOR PCG5 SOLVER
I.1 GENERAL
The PCG5 package is a proprietary solver that may be included with MODFLOW-
SURFACT. This package implements the MATIS matrix solver developed by Dr. Peter
Forsyth at the University of Waterloo. The PCG schemes are the same as those discussed
in Chapter 6 of Volume I of this document. However, the PCG5 solver provides for
additional levels of fill during ILU decomposition (the PCG4 solver only accommodates
a fill level of zero). It further includes red-black ordering schemes and a variety of
acceleration techniques (only the Orthomin acceleration scheme is currently released with
MODFLOW-SURFACT) to make this package more robust and efficient than the PCG4
package. Experience has shown this package to be up to 20 times faster than the PCG4
package in solving typical flow and transport problems.
I.2 ITERATIVE SPARSE MATRIX SOLVER (PCG5) PACKAGE INPUT
Input required by the PCG5 package is similar to that of the PCG4 package
discussed in Chapter 6 of Volume I of the MODFLOW-SURFACT/MODHMS document.
The PCG5 iterative solver is a simple solution scheme requiring only closure tolerance
limits and a maximum limit of solver iterations as input. There are no iteration parameters
to be estimated, nor other choices to be made by the user. A solver-point-control-flag and
maximum outer iterations (> 1 for unconfined systems) are also requested as input, to be
consistent with the other solver packages of MODFLOW. The PCG5 package may be used
in place of PCG4. This solver handles matrices for both flow and transport. For a
transport simulation, either of PCG4, PCG5, or LMG solvers may be used.
Basic Package input file, as requested by the flow simulation. The input is as follows:
Format: I10 I10 I10 F10.0 F10.0 I10
2. Data: HCLOSE MUTPCG MNORTH LEVELS IACCELS NATURL ISTOR1 ISTOR2

Format: F10.3 I10 I10 I10 I10 I10 I10 I10

routine. For a linear problem MXITER should be 1, unless more than
50 inner iterations are required, when MXITER could be as large as 10.
A larger number (generally less than 100) is required for a nonlinear
problem.
ITERI usually ranges from 6 to 600, depending on the problem, and
may be set at 600 with closure controlled by the tolerance limits.

unsaturated cases.
Appendix I: Addendum for PCG5 Solver I-1

M:\Projects\MODFLOW Manuals\SURFACT and HMS Ver.3\Vol I\Appendix I.wpd
criterion is not met. BFACT must be greater than zero and less
than one.

algorithm forces a residual reduction at every Newton iteration.
For RESRED greater than unity, increases in residual are allowed
by the prescribed factor to allow the Newton algorithm to move
easily out of local minima. As RESRED increases, there are less
chances for backtracking. The user should determine which value
of this parameter gives optimal convergence for the particular
problem. See the NRBI package documentation in Chapter 7 Vol.
1, for suggested usage.

between iterations.
Note: This dampening may be effective in achieving convergence
for aquifer systems with highly irregular layering geometries that
system.
Commonly, HCLOSE is set equal 0.01.
iteration summary is suppressed, if MUTPCG = 2, the solver iteration

Orthomin solver to solve the transport matrix equation. A suggested
value of MNORTH ranges from 5 to 10, but lower values may be used
to conserve array space. Smaller MNORTH may, however, weaken
the solver’s performance. MNORTH larger than 10 may help the
solver to handle difficult problems but, in general, diminishing returns
are expected for MNORTH much larger than 10.
LEVELS -- Level of ILU factorization. If NATURL=0, LEVELS should normally

be > 0. LEVELS = 0 is a no-fill ILU.
I-2 Appendix I: Addendum for PCG5 Solver

IACCELS -- Type of acceleration method requested.
IACCELS =0 Conjugate gradient for symmetric positive
definite matrix.
=1 ORTHOMIN acceleration.
=2 CGS Conjugate gradient squared.
=3 CGSTAB (stabilized CGS).
=4 GMRES.
Note that IACCELS=1 (the Orthomin scheme) is the only option provided with
NATURL -- If NATURL=0 a red/black reduced system is used (default),

NATURL=1 does NOT attempt to produce a reduced system. Note
that LEVEL controls the level of the ILU.
ISTOR1 -- is a factor controlling the amount of storage allocated in the workspace
array used by the package to store pointers needed in the factorize and
solve phases. Value of 20 is recommended for ISTOR1.
ISTOR2 -- is a factor controlling the amount of storage allocated in the workspace

array used by the package to store intermediate results during the
factorize and solve phases. Value of 19 is recommended for ISTOR2.
Appendix I: Addendum for PCG5 Solver I-3

DOCUMENTATION
VOLUME II: TRANSPORT MODULES
BY:
HYDROGEOLOGIC, INC.
TEL: (703) 478-5186
FAX: (703) 471-4180

HYDROGEOLOGIC, INC.
DOCUMENTATION
VOLUME III: SURFACE WATER FLOW MODULES
BY:
HYDROGEOLOGIC, INC.
TEL: (703) 478-5186
FAX: (703) 471-4180

HYDROGEOLOGIC, INC.
DOCUMENTATION
BY:
HYDROGEOLOGIC, INC.
TEL: (703) 478-5186
FAX: (703) 471-4180

HYDROGEOLOGIC, INC.
CONTENTS - VOLUME II: TRANSPORT MODULES
Page
1.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-1
1.2 Attributes of Transport Modules in MODFLOW-SURFACT/
MODHMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-2
1.3 MODFLOW-SURFACT/MODHMS Operation and Input Options 1-3
1.4 Document Organization and Usage Guide . . . . . . . . . . . . . . . . 1-3
CHAPTER 2: ANALYSIS OF CONTAMINANT TRANSPORT (ACT) MODULES

2.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-1
2.2 Formulation of the Multispecies Transport Problem . . . . . . . . . 2-1
2.2.1 Governing Equation for Transport in the Active Phase . . 2-1
2.2.2 Governing Equation for Transport in Both Air and Water
Phases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-3
2.2.3 Governing Equation for Transport in Air and Water Phases
with a Residual or Immobile NAPL Phase . . . . . . . . . . 2-5
2.2.4 Retardation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-8
2.2.5 Hydrodynamic Dispersion . . . . . . . . . . . . . . . . . . . . . 2-8
2.2.6 Radioactive Decay or Biodegradation . . . . . . . . . . . . . 2-9
2.2.7 Compressibility Effects . . . . . . . . . . . . . . . . . . . . . . . 2-11
2.2.8 Initial and Boundary Conditions . . . . . . . . . . . . . . . . . 2-11
2.3 Finite-Difference Approximation . . . . . . . . . . . . . . . . . . . . . . 2-12
2.3.1 Discretized Equations . . . . . . . . . . . . . . . . . . . . . . . . 2-12
2.3.2 Implementation of Boundary Conditions . . . . . . . . . . . . 2-16
2.3.3 Solution Procedures . . . . . . . . . . . . . . . . . . . . . . . . . 2-16
2.3.4 Concentration Output . . . . . . . . . . . . . . . . . . . . . . . . 2-18
2.4 Mass Budget Computation . . . . . . . . . . . . . . . . . . . . . . . . . . 2-18
2.5 Mass Budget Output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-19
CHAPTER 3: USER’S GUIDE FOR THE ACT MODULES

3.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-1
3.2 Code Structure and Operation . . . . . . . . . . . . . . . . . . . . . . . . 3-1
3.3 Basic Procedure for Input Preparation . . . . . . . . . . . . . . . . . . 3-5
3.4 Input Instructions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-5
3.4.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-5
3.4.2 Basic Flow (BAS) Package Input . . . . . . . . . . . . . . . . 3-7
3.4.3 Basic Transport (BTN1) Package Input . . . . . . . . . . . . 3-12
3.4.4 Prescribed-Head-Concentration Boundary (HCN1)
Package Input . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-22
3.4.5 Prescribed Concentration Boundary (PCN1) Package
Input . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-23
3.4.6 Well (WEL) Package Input . . . . . . . . . . . . . . . . . . . . 3-24
3.4.7 Drain (DRN) Package Input . . . . . . . . . . . . . . . . . . . 3-26
3.4.8 River (RIV) Package Input . . . . . . . . . . . . . . . . . . . . 3-27
3.4.9 General-Head Boundary (GHB) Package Input . . . . . . . 3-28
3.4.10 Recharge-Seepage Face (RSF4) Package Input . . . . . . . . 3-30
3.4.11 Fracture Well (FWL4) Package Input . . . . . . . . . . . . . 3-33
3.4.12 Fracture Well (FWL5) Package Input . . . . . . . . . . . . . 3-35
Contents i
M:\Projects\MODFLOW Manuals\SURFACT and HMS Ver.3\Vol II\WordPerfect Files\R03-05.360_1-5.wpd
CONTENTS - VOLUME II: TRANSPORT MODULES (CONTINUED)
Page
3.4.13 Preconditioned Conjugate Gradient (PCG4) Package

Input . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-39
3.4.14 Output Control (OC) Package Input . . . . . . . . . . . . . . 3-41
3.4.15 Adaptive Time-Stepping and Output Control (ATO4)
Package . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-44
3.4.16 Evapotranspiration (EVT) Package Input . . . . . . . . . . . 3-50
3.4.17 Evapotranspiration Segments (ET2) Package . . . . . . . . 3-53
3.4.18 Input Instructions for FHB Package . . . . . . . . . . . . . . 3-56
CHAPTER 4: VERIFICATION AND APPLICATION

4.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-1
4.2 Level I: Problems with Analytical Solutions . . . . . . . . . . . . . . 4-1
4.2.1 One-Dimensional, Single-Species Transport . . . . . . . . . . 4-1
4.2.2 One-Dimensional Transport of a Three-Member
Radioactive Decay Chain . . . . . . . . . . . . . . . . . . . . . . 4-6
4.2.3 Transport Due to an Injection/Withdrawal Well . . . . . . . 4-6
4.2.4 Two-Dimensional Transport from a Point Source in a
Steady Uniform Flow Field . . . . . . . . . . . . . . . . . . . 4-12
4.2.5 Transport Due to an Injection-Withdrawal Well Pair . . . 4-12
4.3 Level II: Practical Problems with Complexities That Preclude
Analytical Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-18
4.3.1 Two-Dimensional Transport in a Heterogeneous
Confined Aquifer with a Steady-State Flow Field . . . . . 4-21
4.3.2 One-Dimensional Transport in Transient Unconfined
Flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-25
4.3.3 Two-Dimensional Transport in a Heterogeneous
Unconfined Aquifer with a Transient Flow Field . . . . . . 4-30
4.3.4 Two-Dimensional Transport in the Water Phase
of an Unsaturated Rectangular Soil Slab . . . . . . . . . . . 4-40
4.3.5 Two-Dimensional Transport in an Active Water Phase
with Mass Transfer and Diffusion in the Inactive Air
Phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-40
4.3.6 Dissolution and Volatization of an Immobile Multi-
Component NAPL with Advective and Dispersive
Transport . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-46
4.3.6.1 NAPL Dissolution—Active Water Phase . . . . . 4-46
4.3.6.2 NAPL Volatization—Active Air Phase . . . . . . 4-52
4.3.7 Three-Dimensional Study of Air Sparging at a NAPL
Contaminated Site . . . . . . . . . . . . . . . . . . . . . . . . . 4-52
4.4 Level III: Actual Field Simulation Problems . . . . . . . . . . . . . . 4-59
4.4.1 Cross-Sectional Simulation of Chloride Transport at the
Borden Site . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-59
4.4.2 Three-dimensional Transport of Radionuclide at a Waste
Disposal Site . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-68
ii Contents
Page
CHAPTER 6: ADDENDUM FOR DUAL DOMAIN SIMULATIONS

6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-1
6.2 Formulation of The Dual Domain Equations . . . . . . . . . . . . . . 6-3
6.3 Numerical Implementation . . . . . . . . . . . . . . . . . . . . . . . . . . 6-5
6.4 Input/output Considerations . . . . . . . . . . . . . . . . . . . . . . . . . 6-7
6.5 Verification Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-8
6.6 Input Instructions For Dual Domain Simulations . . . . . . . . . . . 6-10
6.6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-10
6.6.2 Basic Flow (BAS Package Input) . . . . . . . . . . . . . . . . 6-10
6.6.3 BCF4 Input . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-17
6.6.4 Basic Transport (BTN1) Package Input . . . . . . . . . . . . 6-28
6.7 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-39
CHAPTER 7: ADDENDUM FOR DENSITY-DEPENDENT FLOW/TRANSPORT

SIMULATIONS
7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-1
7.2 Formulation of the Density-Dependent Flow/Transport
Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-1
7.3 Numerical Implementation and Solution Options . . . . . . . . . . . 7-2
7.4 Input/Output Considerations . . . . . . . . . . . . . . . . . . . . . . . . . 7-5
7.5 Verification Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-6
7.6 Input Instructions for Density-Dependent Simulations . . . . . . . . 7-6
7.6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-6
7.6.2 Basic Transport Package Input (.BTN) . . . . . . . . . . . . . 7-10
Input . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-22
7.7 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-23
CHAPTER 8: REACTION (RCT) PACKAGE

8.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-1
8.2 Formulation of the Reaction Module . . . . . . . . . . . . . . . . . . . 8-1
8.2.1 Module 1: Instantaneous Aerobic Decay of BTEX . . . . . 8-2
8.2.2 Module 2:Instantaneous Degradation of BTEX using
Multiple Electron Acceptors . . . . . . . . . . . . . . . . . . . 8-4
8.2.3 Module 3: Kinetic-limited degradation of BTEX using
multiple electron acceptors . . . . . . . . . . . . . . . . . . . . 8-6
8.2.4 Module 4: Double Monod Model . . . . . . . . . . . . . . . . 8-10
8.2.5 Module 5: Sequential Decay Reactions . . . . . . . . . . . . . 8-12
8.2.6 Module Aerobic/Anaerobic Model for PCE/TCE
Degradation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-13
8.2.7 User Defined Module . . . . . . . . . . . . . . . . . . . . . . . . 8-16
8.3 Reaction Package Input . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-20
8.4 Model Verification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-22
8.4.1. One-Dimensional Transport of a Three-Member
Radioactive Decay Chain . . . . . . . . . . . . . . . . . . . . . 8-22
8.4.2. Chain Decay of NAPL in a Single Cell . . . . . . . . . . . . 8-23
8.5 Model Simulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-25
Contents iii
Page
8.5.1. Module 1: Instantaneous Aerobic Decay of BTEX . . . . . 8-25

8.5.2. Module 2: Instantaneous Degradation of BTEX using
8.5.3 Module 3: Kinetic-limited Degradation of BTEX Using
8.5.4 Module 4: Double Monod Model . . . . . . . . . . . . . . . . 8-31
8.5.5 Module 5: Sequential Decay Reactions . . . . . . . . . . . . 8-33
8.6 Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-37
8.7 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-38
iv Contents
FIGURES
Page
Figure 2.1 Spatial Discretization of an Aquifer System (After McDonald and

Harbaugh, 1988). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-13
Figure 2.2 Cell I, j, k and Indices for the Six Adjacent Cells. . . . . . . . . . . 2-14
Figure 3.1 Flow Chart Depicting Functional Operation of the Integrated
MODFLOW-SURFACT/MODHMS Code. . . . . . . . . . . . . . . . 3-2
Figure 3.2 Flow Chart Depicting the Computational Procedures for Transient
Transport Analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-3
Figure 4.1 One-Dimensional Transport of a Conservative Species. . . . . . . . 4-2
Figure 4.2 Concentration Profile at 50 Days of Simulation of Single Species
One-Dimensional Transport "L=5 m (Pe=2). . . . . . . . . . . . . . 4-3
Figure 4.5 One-Dimensional Transport of a Three-Member Radioactive Decay
Chain. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-7
Figure 4.6 Species Concentration Distributions Showing a Comparison of
Analytical and MODFLOW-SURFACT/MODHMS Solutions for
One-Dimensional Transport of a Three-Member Radioactive
Decay Chain. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-9
Figure 4.7 Transport Due to an Injection/Withdrawal Well. . . . . . . . . . . . 4-10
Figure 4.8 Breakthrough Curve from Simulation of Transport Due to an
Injection/Withdrawal Well. . . . . . . . . . . . . . . . . . . . . . . . . . 4-13
Figure 4.9 Two-Dimensional Transport from a Point Source in a Uniform
Flow Field: (a) Problem Description, and (b) Model
Discretization. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-15
Figure 4.10 Concentration Profiles Along the Center Line of the Plume
Obtained from Simulation of Two-Dimensional Transport from
a Point Source in a Steady Uniform Flow Field. . . . . . . . . . . . . 4-16
Figure 4.11 Problem Description of Transport Due to an Injection-Withdrawal Well
Pair. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-17
Figure 4.12 Breakthrough Concentrations at Withdrawal Well Obtained from
Simulation of Transport with an Injection-Withdrawal Well Pair. . 4-20
Figure 4.13 Problem Description for Two-Dimensional Transport in a
Heterogeneous Confined Aquifer. . . . . . . . . . . . . . . . . . . . . . 4-22
Figure 4.14 Pumping Well Concentrations Obtained from Simulation of Two-
Dimensional Transport in a Heterogeneous Aquifer. . . . . . . . . . 4-23
Figure 4.15 Concentration Distribution at One Year of Simulation of Two-
Dimensional Transport in a Heterogeneous Confined Aquifer, (a)
MODFLOW-SURFACT/MODHMS, and (b) MT3D. . . . . . . . . 4-24
Figure 4.16 Mass Budget from Simulation of Two-Dimensional Transport in a
Heterogeneous Confined Aquifer. . . . . . . . . . . . . . . . . . . . . . 4-26
Figure 4.17 IN-OUT Mass Budget from Simulation of Two-Dimensional
Transport in a Heterogeneous Confined Aquifer. . . . . . . . . . . . 4-27
Figure 4.18 Mass Balance Errors from Simulation of Two-Dimensional
Transport in a Heterogeneous Confined Aquifer, (a) Dimensional
Error, and (b) Normalized Error. . . . . . . . . . . . . . . . . . . . . . 4-28
Contents v
FIGURES (CONTINUED)
Page
Figure 4.19 One-Dimensional Transport in Transient Unconfined Flow,

(a) Problem Description, and (b) Gallery Operation. . . . . . . . . . 4-29
Figure 4.20 Water-Table Profiles Obtained from One-Dimensional
Unconfined Flow Simulation. . . . . . . . . . . . . . . . . . . . . . . . 4-31
Figure 4.21 Concentration Profiles from Simulation of One-Dimensional
Transport in Transient Unconfined Flow. . . . . . . . . . . . . . . . . 4-32
Figure 4.22 Mass Budget from Simulation of One-Dimensional Transport in
Transient Unconfined Flow. . . . . . . . . . . . . . . . . . . . . . . . . 4-33
Figure 4.23 Normalized Mass Balance Error from Simulation of One-
Dimensional Transport in Transient Unconfined Flow. . . . . . . . 4-34
Heterogeneous Unconfined Aquifer, (a) Model Domain, and (b)
Pumping Schedule. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-35
Figure 4.25 Temporal Distribution of Hydraulic Head at the Pumping Well
from Simulation of Transport in a Heterogeneous Unconfined
Aquifer. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-36
Figure 4.26 Pumping Well Breakthrough Curve from Simulation of Transport
in a Heterogeneous Unconfined Aquifer. . . . . . . . . . . . . . . . . 4-37
Figure 4.27 Mass Budget from Simulation of Transport in a Heterogeneous
Unconfined Aquifer. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-38
Figure 4.28 Normalized Mass Balance Error from Simulation of Transport
in a Heterogeneous Unconfined Aquifer. . . . . . . . . . . . . . . . . 4-39
Figure 4.29 Geometry and Boundary Conditions for Simulation of Solute
Transport in an Unsaturated Rectangular Soil Slab. . . . . . . . . . 4-41
Figure 4.30 Simulated Contaminant Concentrations in an Unsaturated
Rectangular Soil Slab at 0.508d. . . . . . . . . . . . . . . . . . . . . . 4-43
Figure 4.31 Problem Description for Leaching of Volatile Pesticide from the
Soil Surface with Air Phase Diffusion. . . . . . . . . . . . . . . . . . 4-44
Figure 4.32 Simulated Pesticide Concentrations for Problem 4.3.5 (a) 1 d,
(b) 60 days, and (c) 20 yrs. . . . . . . . . . . . . . . . . . . . . . . . . . 4-47
Figure 4.33 Simulated Pesticide Concentrations for Problem 4.3.5 with No
Mass Transfer to the Air Phase (a) 1 d, (b) 60 d, and (c) 20 yrs. . 4-48
Figure 4.34 Hydraulic Head and Water Saturation Profiles for NAPL
Dissolution Example 4.3.6.1. . . . . . . . . . . . . . . . . . . . . . . . 4-49
Figure 4.35 Mass distribution of PCE, TCE, DCE, and VC as a function of
dissolution time. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-50
Figure 4.36 Concentration contours of (a) PCE, (b) TCE, (c) DCE, and
(d) VC at 1,000 days. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-54
Figure 4.37 (a) Equivalent hydraulic head contours and (b) water saturations
of the active gas phase flow problem. The black box indicates
the perimeter of the NAPL source zone. . . . . . . . . . . . . . . . . 4-55
vaporization time. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-57
(d) VC at 0.1 days. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-58
Figure 4.40 Gasoline Saturations Along a Horizontal Plane Through the Well
at 10 Years. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-61
vi Contents
FIGURES (CONTINUED)
Page
Figure 4.41 Mass Loss of Gasoline from the Various Phases in the Subsurface Over
the 10 year Simulation Period. . . . . . . . . . . . . . . . . . . . . . . . 4-62
Figure 4.42 Location of Borden Landfill Showing Monitoring Network.
Cross-section A-AN is along Longitudinal Plume Axis (from
Frind and Hokkanen, 1987). . . . . . . . . . . . . . . . . . . . . . . . . 4-63
Figure 4.43 Cross-section A-AN Showing Aquifer Depth and Location of
Monitoring Devices (from Frind et al., 1985). . . . . . . . . . . . . . 4-64
Figure 4.44 Finite-Difference Grid Used in Borden Flow and Transport
Simulation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-65
Figure 4.45 Recharge Distribution Used for the Flow Simulation. . . . . . . . . 4-66
Figure 4.46 Comparison of Simulated Hydraulic Heads Produced by
(a) Finite-Element Model of Frind and Hokkanen (1987), and
(b) MODFLOW- Model. . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-69
Figure 4.47 Comparison Between (a) Observed Chloride Plume in 1979, and
(b) Simulated Plume at 40 Years. . . . . . . . . . . . . . . . . . . . . . 4-70
Figure 4.48 Domain Discretization, Geometry and Horizontal Conductivity
Distribution for Simulation of Three-Dimensional Transport of
Radionuclide at a Waste Disposal Site. . . . . . . . . . . . . . . . . . . 4-71
Figure 4.49 Schematic Cross-Section of the Modeled Aquifer System for Simulation
of Three-dimensional Transport of Radionuclide at
a Waste Disposal Site. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-72
Figure 4.50 Steady-State Water-Table Profiles Obtained from Simulation
of 3-D Groundwater Flow in the Aquifer System Underlying
Waste Disposal Site. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-73
Figure 4.51 Particle Pathlines Obtained from Simulation of 3-D Groundwater
Flow in the Aquifer System Underlying Waste Disposal Site. . . . 4-74
Figure 4.52 Concentration Distribution at 27 Years of Simulation of Transport
of Radionuclide at a Waste Disposal Site. . . . . . . . . . . . . . . . . 4-76
Figure 4.53 Concentration Distribution after 78 Years of Simulation of
Transport of Radionuclide at a Waste Disposal Site. . . . . . . . . . 4-77
Figure 6.1 Examples and Schematic of Dual Domain System . . . . . . . . . . . 6-2
Figure 6.2 Typical Matrix Block Fracture Units: (a) Slab Matrix Block
Representation, and (b) Spherical Matrix Block Representation . . 6-6
Figure 6.3 Effluent Curves for a Sorbing Medium Showing a Co9mparison of
Numerical and Analytical Solutions for Mass Transfer Coefficient
(") Values to 0.0, 0.15, and 4 . . . . . . . . . . . . . . . . . . . . . . . 6-11
Figure 7.1 Problem Description of Seawater Intrusion in a Coastal Confined
Aquifer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-8
Figure 7.2 Chloride Concentrations for Steady-State Simulation of Saltwater
Intrusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-9
Figure 7.3 Chloride Concentrations for Transient Analyses of Saltwater
Intrusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-10
Figure 8.1 Conceptual Model for Chlorinated Solvent Biodegradation (after
Clement, 1997) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-13
Figure 8.2 Comparison of Results for BTN and RCT Modules for Three
Member Chain Decay with Decay in Water and Soil Phases . . . . 8-23
Figure 8.3 Comparison of Results for BTN and RCT Modules for Three
Member Chain Decay with Decay in Water Phase Only . . . . . . . 8-23
Contents vii
FIGURES (CONTINUED)
Page
Figure 8.4 Comparison of Straight Chain Decay of PCE with Presence of

NAPL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-24
Figure 8.5 BTEX and Oxygen Concentrations at 1000 Days Obtained Using
Module 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-25
Figure 8.6 BTEX, Oxygen, Nitrate, Ferrous Iron, Sulfate and Methane
Concentrations at 1000 Days Obtained Using Module 2 . . . . . . 8-27
Figure 8.7 BTEX, Oxygen, Nitrate, Ferrous Iron, Sulfate, Methane
Concentrations at 1000 Days Obtained Using Module 3 . . . . . . 8-30
Figure 8.8 Electron Donor, Electron Acceptor, and Aqueous Phase Bacteria
Concentration at 1000 Days Obtained Using Module 4 . . . . . . . 8-32
Figure 8.9 PCE, TCE, DCE, and VC Concentrations at 1000 Days for
Module 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-34
Figure 8.10 PCE, TCE, DCE, VC, ETH, and Cl Concentrations at 1000 Days
Obtained Using Module 6 . . . . . . . . . . . . . . . . . . . . . . . . . . 8-37
viii Contents
TABLES
Page
Table 2.1 Mass Balance Reporting for Transport Calculations . . . . . . . . . 2-20

Table 2.2 Mass Balance Terms that Can be Printed from the
MODFLOW-SURFACT/MODHMS Transport Routines . . . . . . 2-21
Table 3.1 Packages Relevant to Transport Simulation . . . . . . . . . . . . . . . 3-6
Table 4.1 Parameter Values and Initial and Boundary Conditions for
Simulation of One-Dimensional Transport of a Three-Member
Radioactive Decay Chain. . . . . . . . . . . . . . . . . . . . . . . . . . . 4-8
Table 4.2 Model Parameters for Simulations of Transport From an
Injection/Extraction Well. . . . . . . . . . . . . . . . . . . . . . . . . . . 4-11
Table 4.3 Model Parameters For Simulation of Two-Dimensional Transport From
a Point Source in a Steady, Uniform Flow Field. . . . . . . . . . . . 4-14
Table 4.4 Model Parameters for Simulation of Transport with an Injection-
Withdrawal Well Pair. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-19
Table 4.5 Physical Parameter Values for Simulation of Transport in an
Unsaturated Rectangular Soil Slab . . . . . . . . . . . . . . . . . . . . . 4-42
Table 4.6 Material Parameter Values for Simulation of Transport in an
Active Water Phase with Volatization and Diffusion in the
Air Phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-45
Table 4.7 Physical Properties of Soil . . . . . . . . . . . . . . . . . . . . . . . . . . 4-51
Table 4.8 Chemical Properties of Contaminants. . . . . . . . . . . . . . . . . . . 4-53
Table 4.9 Physical Properties Controlling Active Gas Phase Flow. . . . . . . 4-56
Table 4.10 Material Parameter Values for Air Sparging Simulation of
Example 4.3.7. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-60
Table 4.11 Parameters for the Borden Site Groundwater Flow and Chloride
Transport Simulation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-67
Table 4.12 Transport Parameters for Simulation of Three-Dimensional
Transport of Radionuclides at a Waste Disposal Site. . . . . . . . . 4-75
Table 6.1 Material Properties for Dual Domain Column Simulations . . . . . 6-9
Table 8.1 Reaction Parameters Used in 8.5.3 . . . . . . . . . . . . . . . . . . . . 8-28
Contents ix
x Contents
MODFLOW-SURFACT/MODHMS
ABSTRACT
MODFLOW-SURFACT/MODHMS is a fully integrated groundwater flow and
solute transport code developed by HydroGeoLogic, Inc., based on the widely popular
U.S. Geological Survey modular three-dimensional (3-D) groundwater flow modeling
code, MODFLOW. This volume (Volume II of the code documentation) describes solute
transport modules incorporated into MODFLOW-SURFACT/MODHMS to allow the
simulation of single-species and multi-component contaminant transport. The physical and
mathematical concepts underlying the transport computational modules are detailed, and
implementation of these concepts into the modular structure of MODFLOW-
SURFACT/MODHMS is discussed. Also provided are instructions for the preparation
of input files for transport simulations.
In order to verify numerical techniques and demonstrate field applications of
MODFLOW-SURFACT/MODHMS, three levels of code testing are presented. For Level
I, five problems with available analytical solutions are considered. These problems range
from a simple problem of one-dimensional (1-D) transport in a semi-infinite column to a
relatively complex problem of transport in a non-uniform flow field created by a pair of
recharging and pumping wells. A variety of cases were used to assess the accuracy of
numerical solutions and demonstrate certain advantages of the Total Variation Diminishing
(TVD) schemes used in MODFLOW-SURFACT/MODHMS. For Level II testing, three
practical problems with complexities (e.g., material heterogeneity and transient water table
and pumping conditions) that preclude analytical solutions are considered. These example
problems were designed to test the code’s utility and flexibility and highlight pertinent
model features (including the code’s ability to produce mass-conservative and numerically
stable and accurate solutions) which are necessary for realistic field simulations. Four
additional test problems are subject to Level II testing, to verify the multiphase transport
formulations incorporated in Version 2 of MODFLOW-SURFACT/MODHMS. Finally,
Level III code testing focuses on actual field applications. Two field simulation examples
are provided. The first example concerns a cross-sectional simulation of chloride transport
in a shallow unconfined aquifer system underneath the Canadian Forces Base (CFB)
Borden Landfill site. The second example concerns a 3-D simulation of radionuclide
transport from a waste disposal site in the USA. Results obtained from the present study
compared well with field observations and previous modeling studies.
1.1 GENERAL
MODFLOW-SURFACT/MODHMS is a fully integrated groundwater flow and
solute transport code based on the U.S. Geological Survey modular groundwater flow
model, MODFLOW. Since its development, MODFLOW has been widely used and
enhanced by incorporation of additional modules to expand its capabilities. Recent
enhancements include a streamflow routing package, a horizontal-flow barrier package,
an interbed-storage package, and a time-variant specified-head package. The basic finite-
difference package was also modified to circumvent difficulties arising from cell
desaturation/resaturation. Included in the modification are additional options for
interblock transmissivity averaging. All these enhancements, which appear in previous
versions of MODFLOW, have been incorporated into MODFLOW-
SURFACT/MODHMS. HydroGeoLogic, Inc., has further extended MODFLOW’s
simulation capabilities by including a rigorous variably saturated treatment of unconfined
systems, robust and accurate solution schemes, allowances for prescribed ponding recharge
and seepage face conditions, and a rigorous treatment of multi-layered wells. Our new
flow packages are generally referred to as “SURF packages” and are presented in Volume
I of the code documentation.
Recognizing that few contaminant transport codes consistent with MODFLOW are
available, HydroGeoLogic, Inc., has developed compatible transport modules and
combined these with the original MODFLOW modules and SURF packages to produce the
fully integrated MODFLOW-SURFACT/MODHMS code. This report, which constitutes
Volume II of the code documentation, describes the formulation, capability, and use of the
new Analysis of Contaminant Transport (ACT) modules. Key simulation features
provided by the ACT modules include the following:
• Advective-dispersive transport of multiple chemical species in steady-state or
transient flow fields
• Linear or nonlinear retardation for each species
• First-order decay and biochemical degradation of contaminants in water and soil
• Generation of transformation products.
Version 2 has additional modeling capabilities including
• Mass partitioning of a single or multicomponent contaminant with diffusive mass
movement in the inactive phase
• Mass partitioning of a single or multicomponent contaminant from an immobile
NAPL phase with advective and dispersive transport in the active phase and
diffusive transport in the inactive phase.
MODFLOW-SURFACT/MODHMS uses state-of-the-art numerical schemes for
solving the transport equation. Included in the code are Total Variation Diminishing
(TVD) flux limiting schemes designed to provide accurate, physically correct, and strictly
mass-conserved numerical solutions. An adaptive implicit scheme is used to minimize
temporal discretization errors. The user may also control the degree of implicitness of the
solution, from fully implicit to explicit. Upstream, central, and combined upstream-
central difference schemes are also provided as options for quicker solutions (but usually
less accurate than the TVD solution). The matrix equations resulting from the finite-
difference approximations are solved using an efficient Orthomin matrix solver. The

primitive mass-conservative form of the transport equation is used for all schemes,
providing strictly mass-conserved numerical solutions. The various schemes mentioned
are intended to provide quality results to novice modelers, as well as full control of
solution options to experienced modelers.
The Input/Output structure of MODFLOW-SURFACT/MODHMS is an extension
of the MODFLOW format. MODFLOW-SURFACT/MODHMS maintains complete
compatibility between the flow and transport analyses with respect to numerical
procedures, data structures, spatial and temporal discretizations, and boundary conditions.
MODFLOW-SURFACT/MODHMS conducts a transport simulation using all of the
MODFLOW data sets, and duplication of data input for flow and transport simulations is
eliminated. Information on transport parameters and first-type transport boundary
conditions is supplied to the code via additional input files. Second-type and third-type
transport boundary conditions are supplied by simple modifications of the MODFLOW
boundary data files.
1.2 ATTRIBUTES OF TRANSPORT MODULES IN MODFLOW-

SURFACT/MODHMS
MODFLOW-SURFACT/MODHMS provides its users with the following key benefits for
groundwater flow and contaminant transport modeling:
• Comprehensive capability to perform groundwater flow and contaminant transport
simulations using dynamic allocation of computer memory. Furthermore, the use
of MODFLOW as a basis provides quick learning for MODFLOW users who
require contaminant transport simulation capabilities.
• Rigorous solution of the primitive form of transport equation while maintaining
full compatibility with MODFLOW and all of its simulation options. Other
MODFLOW-based transport codes neglect transient flow storage effects caused
by factors such as compressibility effects and transient water-table and pumping
conditions that lead to drainage and rewetting of grid cells. Such assumptions may
result in unrealistic transport simulations.
• Handling of transport of contaminants with transformation (daughter) products,
taking into account first-order biochemical degradation or radioactive chain decay
with linear or non-linear equilibrium sorption.
• Handling of multiphase transport of contaminants including an immobile,
depleting NAPL phase.
• Availability of axi-symmetric flow and transport analysis options. These options
allow the user to obtain highly efficient analyses of aquifer tests, single-well tracer
tests, and contaminant transport scenarios involving well injection and pumping.
• Accurate, robust, and efficient mass-conservative transport solution schemes.
These schemes are seamlessly integrated into MODFLOW-SURFACT/MODHMS
and are suitable for simulations of site-specific field characterization and
remediation problems on microcomputers.
• Consistent transport numerical solution options. One option uses a standard
upstream weighting scheme that is highly efficient but only first-order accurate.
The second option uses the more computationally demanding, highly accurate
TVD scheme with the van Leer flux limiter. Given certain requirements for
numerical accuracy and computation speed, the user may readily choose a suitable
option.

1.3 MODFLOW-SURFACT/MODHMS OPERATION AND INPUT
OPTIONS
MODFLOW-SURFACT/MODHMS may be operated in a stand-alone mode or
as part of a complete package that includes a graphical user interface. The first option is
relevant if the user acquires just MODFLOW-SURFACT/MODHMS or wishes to perform
simulations outside of the user interface system.
This section focuses on the first option (i.e., running MODFLOW-
SURFACT/MODHMS in the stand-alone mode). Before running the code, the user will
need to create input data files for a given simulation problem.
For each simulation run, MODFLOW-SURFACT/MODHMS performs both
groundwater flow and solute transport analyses. In view of this, the user should take the
following steps in setting up input files:
1. Refer to the original MODFLOW user’s manual (McDonald and Harbaugh, 1988)
and set up input files for the relevant flow packages following the instructions in
the manual.
2. If usage of the new SURF packages is desirable, proceed to the relevant chapters
of Volume I of the MODFLOW-SURFACT/MODHMS code documentation and
follow the input instructions for the selected flow packages.
3. Refer to Chapter 3 of this document (Volume II) and follow the input instructions
for the relevant packages required in running the ACT modules.
Note that steps 1 and/or 2 may be skipped by the user who already has existing
MODFLOW input files for groundwater flow simulation.
1.4 DOCUMENT ORGANIZATION AND USAGE GUIDE

This document is organized into eight chapters. The remaining chapters and their
purposes are outlined below.
Chapter 2 presents the basic ACT modules. This chapter should be read by the
user who is interested in relevant information pertaining to the transport model
formulation, numerical techniques, and mass balance computational schemes used by
Chapter 3 presents input instructions for the basic ACT modules. Included in this
chapter is a brief discussion of the MODFLOW-SURFACT/MODHMS code functional
operation. Basic procedures for setting up input files and specific input instructions for
the individual packages that need to be invoked for the transport analysis are provided.
Relevant sections of Chapter 3 need to be read by the user in order to gain insight into the
code operation and transport input preparation.
Chapter 4 presents model verification and application to groundwater
contamination problems. The user should read this chapter to obtain information
pertaining to testing of MODFLOW-SURFACT/MODHMS against analytical solutions,
comparison with other available numerical models, such as MT3D and FTWORK, and
field applications.
Chapter 5 provides a list of references cited in the earlier chapters of this
document.
Chapter 6 is an addendum section that provides the formulation, relevant input
instructions, example applications and references related to the dual domain simulation

capabilities of MODFLOW-SURFACT/MODHMS. This chapter should be read by the
user who is interested in incorporating complexities of fractured conditions in transport
simulations.
instructions, example applications and references related to the density-dependent flow and
transport simulation capabilities of MODFLOW-SURFACT/MODHMS. This chapter
should be read by the user who is interested in saltwater intrusion or related problems
where the density of contaminants affects the flow field.
instructions, example applications and references related to the general reaction package
of MODFLOW-SURFACT/MODHMS. This package includes several specific reactions
of interest for biodegradation of industrial contaminants as well as the capability of
incorporating user-defined reactions among several mobile and immobile chemical
component species. It has been developed as an extension to the RT3D (Clement, 1977)
reaction modules which are seamlessly integrated into the contaminant transport routines
of MODFLOW-SURFACT/MODHMS.

CHAPTER 2: ANALYSIS OF CONTAMINANT TRANSPORT (ACT)
MODULES
CHAPTER 2: ANALYSIS OF CONTAMINANT TRANSPORT (ACT) MODULES
2.1 GENERAL
The ACT modules are integrated into MODFLOW-SURFACT/MODHMS to
perform a numerical solution of the advective-dispersive transport equation in transient or
steady-state flow fields for up to five solute species. The numerical schemes used in a
transport analysis are strictly mass-conservative. Linear or non-linear (Freundlich)
equilibrium adsorption is allowed for each species. First-order biochemical degradation
is allowed in the fluid phases and on soil. A straight-chain-degradation (or radioactive
decay) with generation of transformation products is also implemented. Options are
provided to allow transport in the active phase only; transport in both air and water phases
(one of them being inactive to the flow equation, in which diffusion is the transport
mechanism); and transport in both air and water phases of contaminants which dissolve
and volatize from an immobile depleting NAPL phase. The block-centered
finite-difference spatial discretization fully compatible with the MODFLOW formulation
is used with state-of-the-art TVD flux limiting schemes to solve the transport equations.
The formulation and implementation of the transport simulator is discussed in this chapter.
2.2 FORMULATION OF THE MULTISPECIES TRANSPORT PROBLEM

2.2.1 Governing Equation for Transport in the Active Phase
The partial differential equation governing 3-D transport of a contaminant species,
k, in a variably saturated porous medium may be written in the primitive form as follows
(Bear, 1979):
∂ ⎛ k ∂c k ⎞ ∂ ∂ ∂
∂xi
⎜ Dij
⎜
⎟−
⎟
∂x j ⎠ ∂xi
vi c k =
∂t
(
φSα c k + )
∂t
ρ Β c sk ( ) ( )
⎝
+ λαk φSα c k + λks ρ Β c sk − qc *k + Γ k (1)
NPAR NPAR
− ∑ξ
j =1
kj λα φSα c −
j j
∑ξ
j =1
kj λ sj ρ Β c sj , i or j = 1,2,3
where Dij is the apparent hydrodynamic dispersion tensor, ck is solute concentration of

component k in the active phase, csk is the concentration of component k adsorbed to the
soil, vi is the Darcy velocity, N is the effective porosity, S" is the saturation of the active
fluid phase ", 8"k and 8sk are the first-order decay coefficients for component k in the
active fluid phase " and soil respectively, q is the volumetric flow rate via sources (or
sinks) per unit volume of the porous medium, c*k is the solute concentration of the sources
(or sinks) and ' is the mass transfer rate of component k from the active phase to the
inactive phase (which is zero for contaminant transport in the active phase only, i.e., it is
displayed here, for later generalization of the equation to include mass transfer).
Component (j) is the parent component for daughter product k when transformation
products occur, and the last two terms in equation (1) represent generation of component
k as a result of decay of all parents j = 1 to NPAR (where NPAR is the number of parents
of k), in water and on soil respectively. >kj is the fraction of parent component j
transforming into component k (which is unity for a straight chain decay where one unit
Chapter 2: Analysis of Contaminant Transport (ACT) Modules 2-1

of parent produces one unit of daughter.). Branched chain decay details are provided later
in Section 2.2.6. The active phase " is considered as the phase for which the flow
equation is solved, as discussed in Chapter 2 of Volume 1. Equation (1) expresses the
transport of a decaying contaminant in a sorbing porous medium because of advection and
Fickian dispersion in the active phase.
The governing equation for transport of contaminant in a variably saturated porous
medium, Equation (1), reduces to the saturated transport equation when the active fluid
phase " is water (w), and Sw=1 (i.e., for fully confined aquifers or saturated cells). Recall
that the index IREALSL in the BCF4 input package determines the flow simulation’s
active phase. If IREALSL=0, then pseudo-constitutive relations determine the water table
in a water-active system. If IREALSL= 1 or 2, then water in the active phase in a
variably saturated system and if IREALSL = 3 or 4, then air (a) is the active phase of a
variably saturated system. For unconfined conditions (i.e., using pseudo-constitutive
relations), grid blocks above the water table have a saturation value equal to zero (residual)
and transport occurs in the dry cells under steady state conditions, with mass flux into it
equal to the mass flux out of it from the bottom, thus keeping consistency with GSVE
considerations of the flow equation for this case. For 0<Sw<1, flow and contaminant
transport occurs under unconfined conditions, recalling that Sw represents the saturated
grid-block thickness, when pseudo-soil relations are used to determine the water table
(i.e., when IREALSL=0).
The storage term for contaminant concentration in the active phase (first term on
the right hand side of equation(1)) is expanded for the case of transport in a transient
unsaturated flow field wherein the specific yield (Sy) values used for flow are different
(smaller) than the porosity (N) values of equation (1). Specifically, the storage of
contaminant occurs within two regions of the pore space in any grid block. The first
region includes the space occupied by the term Sy and the second region includes the
remaining pore space (N!Sy). For example, the unconfined flow solution using pseudo-
soil functions considers Sy as the drainable porosity, the remaining pore space being filled
by water that is immobile. Thus, expanding the storage term gives:
∂
∂t
(
φSα c k =
∂
∂t
) [
S y Sα c k + (φ − S y )S active c k ] (1b)
where Sactive is the saturation of the active phase in the remaining pore space (N!Sy). This
expansion is needed because the change of saturation of fluid is calculated as per space Sy,
occurring due to net inflow minus outflow (which is used for the Darcy velocity in the
transport equation). The remaining space is not involved in the advective term, and it may
optionally be assumed to be filled by the active phase (Sactive=1) or by the inactive phase
(Sactive=0). A flag IPHSFLG is read with the basic transport input to indicate whether the
remaining space is filled by the active (IPHSFLG=1) phase or the inactive (IPHSFLG=0)
phase. The mass conservative nature of the primitive form of the transport equation which
uses volumetric fluid fluxes in the advective term to move contaminant (rather than
velocities as used by the reduced form of the transport equation), is acute to this physical
situation. A similar expansion is not performed for the decay or gain-from-parent terms
of equation (1) since they do not pose a mass conservation issue when N is different from
Sy.
2-2 Chapter 2: Analysis of Contaminant Transport (ACT) Modules

2.2.2 Governing Equation for Transport in Both Air and Water Phases
Equation (1) assumes that contaminant resides in the active phase only, with
partitioning to the soil grains due to adsorption. A contaminant may, however, transfer
mass across phase boundaries into the inactive phase, with diffusive transport occurring
in the inactive phase. Thus, for water being the active phase, a volatile contaminant would
transfer mass into the air phase and diffuse in the air phase as well (advection in the air
phase being neglected). Similarly, a contaminant in an active air phase may dissolve in
the water phase and be transported by diffusion therein, neglecting water phase advection.
The diffusive mass transport equation for contaminant k in the inactive phase may be
expressed as:
∂ ⎡ k ∂c kp ⎤ ∂
( ) ( )
NPAR
⎢ Dij ⎥= φS p c p + λ pφS p c p − ∑ ξ kj λ pj φS p c pj − Γ k
k k k
(2)
∂xi ⎢⎣
p
∂x j ⎥⎦ ∂t j =1
where the subscript p denotes the inactive (or passive) phase in the flow system solved by
the flow modules.
For the case of transport in a transient unsaturated flow-field wherein the specific
yield (Sy) values used for flow are different (smaller) than the porosity (N) values of
equation (2), the storage term for contaminant in the inactive phase is expanded in a
similar fashion to equation (1b) as:
∂
∂t
(
ϕS p c kp =
∂
∂t
) [
S y S p c kp + (ϕ − S y )S active c kp ] (2b)
when the remaining pore space (N!Sy) is filled by the active (Sactive=1) or inactive (Sactive
=0) phase as determined by the user via input flag, IPHSFLG.
Equilibrium mass partitioning may be assumed between air and water phases (one
of which is the active phase, the other being inactive with respect to the flow equation) to
express the concentration of contaminant k in the air phase, cak, as a function of its
concentration in the water phase, cwk as
c ak = K aw
k
c wk (3)
and summing Equations (1) and (2) gives the governing equation for transport in both air
and water phases as:
∂ ⎡ k
{ ( ) }∂∂cx
⎤ ∂
( ∂
) [( ∂
) ] ( )
k
⎢ Dij + K pα Dij
k k
⎥− vi c k = ϕ Sα + K pkα S p c k + ρ Β c sk
∂xi ⎢⎣
p
j ⎥⎦ ∂xi ∂t ∂t
+ (λ ϕS )c ∑ ξ (λα ϕSα + λ ϕS )
NPAR
α
k
α + λ ϕS p K
k
p
k
pα
k
+ λ ρ c − qc −
k
s
k
Β s
*k
kj
j j
p p K pjα c j (4)
j =1
NPAR
− ∑ξ
j =1
kj λsj ρ Β csj , i or j = 1,2,3

where subscripts " and p refer to the active and passive phases respectively; K pkα is the
partition coefficient for contaminant k from the active phase, " to the passive phase, p; and
K pkα = K aw
k
if water is the active phase or K pkα = 1 / K aw
k
if air is the active phase.
Equation (4) can be put back into the form of Equation (1) by defining effective
parameters, so that the numerical solution to Equation (4) is the same as that of Equation
(1), with effective parameters replacing the respective parameter values of Equation (1).
These effective parameters are defined as:
(D ) k
ij eff = Dijk + K pkα Dijk ( ) p
(5)
S eff = Sα + K pkα S p (6)
(λ S )
k
eff = λαk Sα + λkp S p K pkα (7)
(λ S )
j
eff = λαj Sα + λ pj S p K pjα (8)
where the subscript eff denotes the effective parameter which replaces the respective
parameter value in Equation (1) during numerical solution to the transport equation.
MODFLOW-SURFACT/MODHMS invokes the multiphase transport equation via
an index, IEQPART. If this index is zero (or not input), only active phase transport
simulation is performed. If IEQPART=1, transport occurs with advection and dispersion
in the active phase, and diffusion in the inactive phase. For this case, the code requires
the air-water partition coefficient, and the diffusion coefficient in the passive phase for
each component, as input, in addition to the input required when IEQPART=0. Also
recall that the index IREALSL in the BCF4 package determines which phase will be
treated as the active phase. If IREALSL is 1 or 2, then water is the active phase, and if
IREALSL is 3 or 4, then air is the active phase. The user is cautioned into fully
understanding the conceptual system when IREALSL=0 is used with the multiphase
transport option.
It should be noted that the inactive phase saturation is computed such that
Sp+S"=1. If the flow-field is transient, the saturation of the inactive phase changes along
with active phase saturation changes, as a result of honoring the above constraint. While
saturation changes in the active phase are balanced by the advective term in equation (1),
the saturation changes in the passive phase are not. The instantaneous equilibration
assumption for passive phase flow does not provide the fluxes of this phase which cause
phase saturation to change and hence neglecting this advective term in equation (2) omits
a mass transport mechanism that would cause Sp to change in the storage term at any time
step. Therefore, for charging Sp resulting due to active phase flow, equation (2) is not a
mass-conserved equation. Options could be provided for keeping Sp fixed during transport
computations (to an initial value or a user-defined value), however, these would mis-
represent the physics to maintain mass conservation. Therefore, the mass balance errors
for any species should be noted for transport simulations with inactive phase partitioning
in a transient unsaturated flow-field and if large, its impact or the simulation should be
examined in detail. Analysis for that species would probably be superior if conducted with
the passive phase activated instead.

2.2.3 Governing Equation for Transport in Air and Water Phases with a
Residual or Immobile NAPL Phase
A further extension of Equation (4) may be obtained by also considering an
immobile NAPL phase from which contaminant components may dissolve into water and
volatize into air within the subsurface for subsequent transport, with the mass transfer
process depleting the NAPL phase saturations. Contaminant components may also diffuse
within a continuous NAPL phase, with degradation and generation of components within
the NAPL phase. A mass balance equation for contaminant k in the NAPL phase may be
expressed as:
∂ ⎡ k ∂cnk ⎤ ∂
( ) ( )
NPAR
⎢ Dij ⎥= φS n cn + λnφS n cn − ∑ ξ kj λnj φS n cnj − Γnk
k k k
(9)
∂xi ⎢⎣
n
∂x j ⎥⎦ ∂t j =1
where the subscript n denotes the NAPL phase and the term 'nk is the mass transfer term
for component k from the NAPL phase to the other fluid phases present, which must also
be added to the right-hand-side of Equation (4) for completeness. Note that when
Equation (9) is added to Equation (4), this term (i.e., 'nk) cancels out, and equilibrium
partitioning may be used to express NAPL phase concentrations of component k, in terms
of the water phase concentration as:
c nk = K nw
k
c wk (10)
k
where K nw is the equilibrium partitioning coefficient for contaminant k from water to
NAPL. Note that due to equilibrium considerations, Equation (3) may be divided by
Equation (10) to express NAPL phase concentration of component k in terms of its
concentration in the air phase as:
k
K nw
c = K c = k c ak
k
n
k k
na a (11)
K aw
where K nak is the partitioning coefficient for component k from air to NAPL. Equation
(9) may be added to Equation (4) and the result may be put back in the form of Equation
(1) by defining effective parameters, as was done in section 2.2.2, as:
(D ) k
ij eff = Dijk + K pkα Dijk ( ) p
+ K nkα Dijk ( )
n
(12)
S eff = Sα + K pkα S p + K nkα S n (13)
(λ S )
k
eff = λαk Sα + λkp S p K pkα + λkn S n K nkα (14)
and

(λ S )
j
eff = λαj Sα + λ pj S p K pjα + λ nj S n K njα (15)
where Knkα is the partitioning coefficient of component k from the active phase, ", to the
NAPL phase, n. K nkα = K nw

k
[of Equation (10)] when water is the active phase, and
K nkα = K nak = K nw
k k
/ K aw [of Equation (11)] when air is the active phase.
The index IEQPART=2 invokes the multiphase transport equation with immobile
NAPL in MODFLOW-SURFACT/MODHMS. For this case, MODFLOW-
SURFACT/MODHMS also requires the NAPL-water partition coefficient and diffusion
coefficient in NAPL for each component, as input, in addition to all input required when
IEQAPART=1 (see section 2.2.2). Furthermore, the initial NAPL saturation and
composition of this immobile NAPL is also required. As contaminants dissolve and/or
volatize from the NAPL phase, the NAPL saturations reduce until no NAPL phase resides
in the simulation domain. Alternatively, if NAPL-soluble components are generated, a
NAPL phase may appear. The flash package described below handles
appearance/disappearance of the NAPL phase by consideration of contaminant component
mass conservation, and a compositional framework for the NAPL phase.
It is noted that NAPL saturation changes in equation (9) result from phase transfer
mechanisms only and therefore equation (9) is a mass conserved equation. In addition,
solution of multiphase transport within the unsaturated, single-phase flow-field of
MODFLOW-SURFACT/MODHMS assumes that NAPL occupies the pore space N,
without consideration of its presence in the flow equation.
The effective transport equation in a multiphase porous medium discussed above,
may be solved for each contaminant introduced to the subsurface via the NAPL phase.
Thus, each contaminant is redistributed spatially within the domain due to multiphase
advection, dispersion, decay and generation, processes. The phase-state of the system,
however, will be thrown into disequilibrium in regions where a NAPL phase exists, or
where it is likely to form. A flash calculation is then performed at each grid-block in the
region (where NAPL exists) to set-back phase equilibrium conditions. Thus, a
compositional consideration is required for the NAPL phase such that all NAPL
contaminant components must be included in the problem definition. Hence, the mass of
the NAPL phase is the sum of the mass of its components giving
∑Ck
k
n = ρn (16)
or alternatively, defining a mass fraction of component k in the NAPL phase, Xnk, as the
mass of k divided by the mass of the NAPL phase, we have from Equation (16)
∑Xk
k
n =1 (17)
Note that by definition, we have
C nk = ρ n X nk (18)

The net mass of component k within a computational grid block is conserved for
the flash calculation, when mass is redistributed among phases. Thus
∑ (ρα Sα X α ) k
N
(
= ∑ ρα Sα X αk )
O (19)
α α
where subscript N refers to new (after flash) and O refers to old (before flash) values.
Equation (19) is a set of k equations for each NAPL component k. Equations (17) and (19)
therefore form a set of k+1 equations which may be solved for the k+1 unknowns, Sn and
Xnk for each component k, to reset equilibrium. Note that density for air and water
required by equation (19) are obtained as discussed for the respective flow equations, and
NAPL phase density may be obtained as
ρ n = ∑ X nk ρ nk (20)
k
where Dnk is the density of pure component k in its liquid state [note that X"k for " other
than NAPL in equation (19) may be obtained from the equilibrium partitioning relations
of (3), (10), and (11)]. This set of non-linear equations may be solved by Newton
iteration, at each grid block. Solution to this set of equations is called the flash package
which takes the dis-equilibrium concentrations obtained after transporting all contaminants,
and re-assigns contaminant masses to the various phases present such that their
concentrations are in equilibrium with an existing NAPL phase. Mass conservation is
maintained, therefore, the NAPL saturation changes due to contaminant mass transfer in
or out of it are also computed.
Two states of disequilibrium may occur in any grid block after the transport
equations are solved for all components. First, where a NAPL phase is already present,
dissolution and volatization of its components will reduce NAPL concentrations from their
equilibrium constraint value of (17). The flash package remedies this by decreasing NAPL
saturations. Second, where a NAPL phase is absent, generation terms in the transport
equations may create NAPL-soluble components whereby the constraints of (16) or (17)
are exceeded. This is an indication of NAPL phase appearance which will be quantified
by the flash package. Where no NAPL phase exists at any time, at any location in the
system, the sum of concentrations of all contaminants, as calculated for a NAPL phase will
be less than the constraint (17), and the flash routine may be skipped.
Due to the mass equilibration considerations of the flash package, simulations
using IEQPART=2 require further user input of the densities of each species k in its pure
liquid state. In addition, options are provided for interpreting the input species
concentrations via the flag IINIT. If IINIT is zero, the code takes the input concentration
to represent active phase concentrations in equilibrium with immobile NAPL. This option
is also useful for restarting a simulation to continue the transients further in time. When
IINIT is one, the code interprets input concentrations as resident NAPL phase
concentrations which are in equilibrium with the other phases present. When IINIT is
two, the input concentrations are interpreted as concentration of species in the initial
inventory of the insitu NAPL phase. This mass is first flashed into the other phases
present (assumed initially contaminant-free), before proceeding with the transport
simulation. For all cases of IINIT, the user should ensure that concentration input is
consistent when a NAPL phase exists. Thus, when Sn>0 (i.e., NAPL phase exists)
constraint of equation (17) cannot be violated and the user should ensure that the mass
fractions of all contaminants in the NAPL sum to unity. Equation (20) may then be used
to determine NAPL phase density, and equation (18) to then determine the consistent set
of input values for species concentrations. Alternatively, the user should ensure that sum
of the mass fraction is less than unity at a node where NAPL saturation is zero. If the set

of species concentrations is inconsistent with the above requirements, it is likely to be a
data input error. Benefit of the doubt is, however, given to the user and only a warning
message is issued regarding this inconsistency at any node.
2.2.4 Retardation
The sorbed and solute concentrations of component k are related by the non-linear
equilibrium Freundlich isotherm expressed as
( )
csk = κ k c k * *η k (21)
where 6 and 0 are the Freundlich coefficient and exponent respectively, for component
k k
k. Note that setting 0k to unity provides a linear isotherm for adsorption of component
k.
2.2.5 Hydrodynamic Dispersion
The hydrodynamic dispersion tensorial components are computed using the
following constitutive relations for isotropic porous media (Scheidegger, 1961):
vi v j
Dij = α T v δ ij + (α L − α T ) + βDoδ ij (22)
v
where v is the magnitude of the velocity vector, (

v = v12 + v 22 + v32 ) 1
2
, "L and "T are
longitudinal and transverse dispersivities, respectively, δij is the Kronecker delta, Do is the
free-water molecular diffusion coefficient, and $ is the tortuosity given by the Millington-
β = Sw φ
10 4
Quirk (1961) equation as 3 3
. Equation (22) typically is used to provide
dispersivities for 3-D isotropic porous media. Alternatively, the hydrodynamic dispersion
may be computed from the relations provided by Burnett and Frind (1987) for stratified
porous media:
v x2 v y2 v z2
D xx = α L + αT +αv + βDo (23a)
v v v
v x2 v y2 v z2
D yy = α T +αL +αv + βD0 (23b)
v v v
v x2 v y2 v2
Dzz = α v + α v + α L z + βD0 (23c)
v v v
D xy = D yx = (α L − α T ) v x v y v (23d)
D zx = D xz = (α L − α v ) v z v x v (23e)
where subscripts L, T, and v are the indices for the horizontal longitudinal, horizontal
transverse, and vertical transverse directions, respectively. Note that equations (23a)

through (23e) collapse to equation (22) when "v = "T. As a third alternative, the
hydrodynamic dispersion maybe computed from relations provided by Guvanasen (1998)
for quasi-3D flow situations or cases with dominant vertical flow components, as
v x2 v y2 v z2
D xx = α Lh + α Th + α Tv + βDo (24a)
v v v
v x2 v y2 v z2
D yy = α Th + α Lh + α Tv + βDo (24b)
v v v
v x2 v y2 v z2
D zz = α Tv + α Tv + α Lv + βDo (24c)
v v v
vx v y
Dxy = D yx = (α L − α T ) (24d)
v
⎛ α + α Lv ⎞v v
D xz = D zx = ⎜ Lh − α Tv ⎟ x z (24e)
⎝ z ⎠ v
⎛ α + α Lv ⎞ v y vz
D yz = D zy = ⎜ Lh − α Tv ⎟ (24f)
⎝ z ⎠ v
where the subscripts Lh, Th, Lv, and Tv are indices for horizontal longitudinal, horizontal
transverse, vertical longitudinal and vertical transverse directions respectively. Note that
equations (24) collapse to equations (23) when Lv=Lh. Equations 22, 23, and 24 are thus
the 2, 3, and 4 component dispersivity equations respectively, which are successive
generalizations of Scheidegger’s isotropic equation. The set of equations (24) is typically
used to calculate dispersivities in a three dimensional system where vertical flow
components are significant. In such cases, the vertical and horizontal components of
dispersion can be an order of magnitude apart for their respective longitudinal and
transverse components. The set of equations (23) typically is used to calculate
dispersivities in 3-D non-isotropic systems with mainly horizontal flows. In such cases,
the vertical transverse dispersivity is typically an order of magnitude less than the
horizontal transverse dispersivity for areally extensive systems.
2.2.6 Radioactive Decay or Biodegradation
The third and fourth terms on the right-hand-side of equation (1) represent a first-
order degradation (i.e., radioactive or biodegradation) in the active phase and on soil,
respectively, while the fifth and sixth terms represent the gain in mass of contaminant
species because of first-order degradation of a parent species in the active phase and on
soil, respectively. For the case of multi-phase transport, decay can occur in both water
and air phases, and, if residual NAPL is included in the simulation, decay is modeled in

the NAPL phase as well. The radioactive decay rate is typically dependent on the
radionuclide species, while the biodegradation rate is dependent on the chemical species
as well as the soil type or zone (which would determine microbial activity). MODFLOW-
SURFACT/MODHMS allows for both types of degradation. Furthermore, radioactive
decay occurs at the same rate in all phases in which the radionuclide species exists (i.e.,
water and soil), whereas biodegradation rates may be different for different phases. The
degradation rate constant (8) may be computed from the half-life (t1/2) of a radionuclide
component from 8=ln2/t½, where ln is the natural logarithm.
For the case of chain decay or chain chemical reactions, MODFLOW-
SURFACT/MODHMS treats five components. These components may be part of a
straight chain with one unit of parent component decaying to one unit of the daughter, or
may be treated with a more generalized, decay/generation model that includes branched
chains. The first case is indicated by switch ICHAIN = 1 while the second general model
is invoked by setting ICHAIN = 2. For the first case, a straight chain sequence may be
expressed as:
where component decays to component with decay rates 81",s (in fluids
" and on soil), and so on to component . If a flag for chain decay is on
(ICHAIN=1, see Chapter 3), the code can treat up to five components in a linear chain
decay sequence, whereby a loss of mass (or moles or curies) of component because
of decay causes an equal gain in mass (or moles or curies) of component which is
thereby generated. Any or all of the decay rate coefficients may be set to zero to model
sets of sub-chains. Note that component does not have any generation terms since it
is the ultimate parent component. Also note that the daughter product of component
is not modeled. If the flag for chain decay is off (ICHAIN=0, see Chapter 3), the code
can model up to five independent decaying, retarded species in one single simulation.
For the second case, a more generalized decay model is applied by setting ICHAIN = 2.
A branched chain decay may be simulated in the most general terms, as exemplified by:
ξ 21 ξ 42
1 λ1 2 λ2 4 λ4
ξ 31 ξ 43
ξ 53
3 λ3 5 λ5
Where component decays to and in ratios >21 and >31 respectively,

i.e., decay of one unit of species produces >21 units of component , and >31 units
of species . Component decays to with decay of one unit of producing
>42 units of . Components also decays to component with decay of one unit
of producing >43 units of . Finally, for this example, species also decays
to with one unit of decay of producing >53 units of .
The general chain decay formulation is also useful for straight chains where one
unit of parent j decays to produce >jk units of daughter k. Thus, the factor >jk also allows

for scaling of daughter concentrations to perform decaying transport calculations in any
concentration units for the various components being modeled.
2.2.7 Compressibility Effects
Flow compressibility effects may significantly affect contaminant transport
behavior when large changes occur in the water pressure, typical of nuclear waste
repository situations or deep well injection problems. To incorporate compressibility
effects appropriately, the effective porosity is related to the specific storage by neglecting
water compressibility effects and extending Terzaghi’s effective stress concept to include
air in the voids (Brutsaert, 1964; Bear and Pinder, 1978; Bear et al., 1984; Panday and
Corapcioglu, 1991) to give
dφ
S w S s dh = (25)
1−φ
where h is the hydraulic head.
Equation (25) enables a user to compare computed concentrations with field
values, without needing to convert from reservoir pressure conditions to atmospheric
conditions, since most of the saturated storage of the transient flow field occurs through
matrix compressibility.
2.2.8 Initial and Boundary Conditions
The initial condition of the transport problem is represented by
c k ( x1 , x 2 , x3 , t = 0) = c ok (26)
where ck are the concentrations in the active phase. For the case where NAPL is included
in the simulation, these concentrations may be provided within the residual NAPL phase
which are subsequently flashed to the other phases under equilibrium conditions. When
NAPL is included in the simulation, the NAPL saturations within the domain are required
as initial conditions as well.
The boundary conditions around the model domain are expressed as
First Type : c k ( x1 , x2 , x3 , t ) = c~ k on B1′ (27a)
∂c k
Second Type : Dij ni = qcD on B2′ (27b)
∂x j
∂c k
Third Type : Dij ni − vi ni c k = qcT on B3′ (27c)
∂x j
where B1′ is the boundary portion where concentration in the active phase is prescribed
~k
as c , B2′ , and B3′ are portions of the boundary where the dispersive and the total solute
mass fluxes are prescribed as q cD and q cT , respectively; and ni is the outward unit normal
vector. Note that an outflow boundary condition is a special case of equation (27c) with q cT = vi vi c k
thus leading to the condition of zero dispersive flux.

Sources or sinks of the flow system are incorporated into the transport equation
via the source/sink term. They may be classified as areally distributed or point sources
or sinks. Examples of sources include recharge and injection wells. Unconfined recharge
boundaries, constant-head, and general-head boundaries, rivers, and streams may be
sources or sinks to the system depending on potentiometric conditions. It is necessary to
specify the concentration of a source within the active phase for all species, at inflow
boundaries of the system, to pose a well-defined problem. Examples of sinks include
seepage faces, drains, pumping wells, and evapotranspiration boundaries. For sinks, the
concentration of withdrawn water (when water is the active phase) is generally equal to
the concentration of groundwater in the aquifer (i.e., c*=c). There is one exception,
however, when dealing with evapotranspiration. In this case, pure water is assumed to be
extracted and thus the concentration of the evapotranspiration flux is zero (i.e., c*=0).
When air is the active phase, the concentration of contaminants in the extracted air is equal
to its concentration in the computational grid block which extracts air. Finally, prescribed
concentration conditions may occur anywhere in the domain, including at source/sink
nodes to the flow system. Examples include large bodies of fluid influencing the modeled
flow and transport system, or leaching of contaminant from a solid or immobile phase
within the domain, into the ambient water.
2.3 FINITE-DIFFERENCE APPROXIMATION

2.3.1 Discretized Equations
The transport modules of MODFLOW-SURFACT/MODHMS follow the same
convention for spatial and temporal discretizations as that used by the flow modules.
Shown in Figure 2.1 is the right-handed Cartesian coordinate system employed for spatial
discretization of an aquifer system. An assumption is made that x, y, and z correspond
to x1, x2, and x3, respectively. With the depicted coordinate system, ∆rj along the row
direction is equivalent to ∆xj along the x-axis; ∆ci along the column direction is equivalent
to ∆yi along the y-axis; and )vk is equivalent to )zk along the vertical z-axis.
Using the cell indices illustrated in Figure 2.2, the finite-difference form of the
governing transport equation (1) for cell i, j, k may be written as:
( ) ( ) ( ) (
∇ x τ xx ∇ x c k + ∇ y τ yy ∇ y c k + ∇ z τ zz ∇ z c k − ∇ x Q x c k − ∇ y Q y c k − ∇ z Q z c k ) ( ) ( )
{ ( ) ( ) (
= (∆x j ∆y i ∆z k ) ∇ t φS w c k + ∇ t ρ B c sk + λ kwφS w c k + λ ks ρ B c sk − λ ks −1 ρ B c sk −1 − λ kw−1φS w c k −1 )n +1
}− (Qc *) n +1
(28)
(
+ ∇x τ ∇ yc k
xy
k
) + ∇ (τ x
k
xz ∇zc k
) + ∇ (τ ∇
y
k
yx x c k
) + ∇ (τy
k
yz ∇zc k
) + ∇ (τ
z zx ∇ xc k
) + ∇ (τz
k
zy ∇ yc k
)
k
where c is the time-weighted concentration for component k in the active phase; Lx, Ly,
and Lz are finite-difference gradient operators in the x, y, and z directions, respectively;
Lt is the temporal difference operator; n+1 denotes the current time level; τ xx ,τ yy ,τ zz
are the discretized dispersive terms; and Qx, Qy, Qz are the volumetric fluxes in the
principal coordinate directions. The discretized cross dispersion terms,
τ xy ,τ xz ,τ yx ,τ yz ,τ zx and ττ zy , are optionally taken into account by the code if
requested. Calculation of all dispersion coefficients at interblock locations are performed
by accounting for variable cell thickness as discussed by Zheng, 1990. Note that

Figure 2.1 Spatial Discretization of an Aquifer System (After McDonald
and Harbaugh, 1988).

Figure 2.2 Cell I, j, k and Indices for the Six Adjacent Cells.

superscript k denotes component k and superscript k-1 denotes the parent component of k,
under chain decay conditions. Also, if multiphase transport (with or without immobile
NAPL) is considered, the equivalent form of the multi-phase transport equation is used
with effective parameters defined by equations (5) through (8) or equations (12) through
(15). The time averaged concentration may be written as:
c k = θ ck ( ) n +1
+ (1 − θ ) c k ( ) n
(29)
where n is the previous time level, n+1 is the current time step, and 2 is the Crank-
Nicolson factor which is unity for fully implicit, 0.5 for Crank-Nicolson, and 0 for fully
explicit time discretization.
The interblock concentrations in the advection terms on the left-hand-side of
equation (28) are calculated using a mass conservative second-order TVD scheme (Roe,
1981; Harten, 1983; Sweby, 1984; Yee, 1987; Cox and Nishikawa, 1991; Blunt and
Rubin, 1992; Arminjon, and Dervieux, 1994; Forsyth, 1994; Liu et al., 1994; and Unger,
et al. 1996) with the van Leer flux limiter (van Leer, 1977, 1979). Expansion of the
interblock advective flux term may be written as
( )
∇ i Qi c k = Qi + 12 cik+ 12 − Qi − 12 cik− 12 (30a)
where I is any coordinate direction (i=1,2,3), i+½ is the forward face of the grid block,
and i-½ is the backward face of the gridblock in the ith coordinate direction. The
concentration at the face c~i k+ 12 may be expressed in a TVD form as
σ (γ ij )
cik+ 12 = cups
k
+
2
[c k
dwn − cups
k
] (30b)
where ups is the upstream point and dwn is the downstream point of the nodal pair i and
i+1, and F is the flux limiter term, which is a function of (ij, the smoothness sensor.
Concentration at the face cik− 12 is expressed in a similar fashion. The TVD scheme
minimizes numerical dispersion and preserves monotonicity properties of the exact solution
and appears in equation (30b) as a correction to the upstream solution. The smoothness
sensor is calculated from concentrations at the downstream, upstream, and second point
upstream nodes and detects rapid changes in concentration. The van Leer flux limiter
weights the solution from fully upstream to fully downstream. When the degree of
implicitness of the solution varies, the Crank-Nicolson factor is chosen as close to ½ as
possible, without violating the TVD property of the solution (Blunt and Rubin, 1992).
The TVD property, when applied to the advective term, ensures physically correct
solutions without spurious oscillations even for totally advective transport. As stated by
Harten (1983), the TVD property guarantees that for a nonlinear, scalar equation or a
linear system of equations, the total variation (TV) of the solution will not increase as the
solution advances in time, i.e.,
( )
TV c n +1 ≤ TV c n ( ) (31a)

where n+1 and n denote the current and previous time levels, and the total variation of
a quantity c is defined as
∞
TV (c ) = ∑c i − ci −1 (31b)
i = −∞
in which i and i–1 denote adjacent grid points. The van Leer limiter used in MODFLOW-
SURFACT/MODHMS is capable of capturing a shock within three nodes, under optimal
time stepping conditions, thus providing accurate, physically correct mass conserved
solutions. For TVD schemes with automatic control over implicitness, the time weighting
factor is adaptively adjusted for optimal accuracy of temporal discretization, while still
maintaining the TVD property. Options for upstream, midstream, and mixed spatial
weighting of the advective terms are also provided for quick initial simulations toward
calibration.
For transport through a fracture well (see the FWL4 Package described in Volume
I of the MODFLOW-SURFACT/MODHMS code documentation), it is assumed that
dispersion and storage in the wellbore are negligible since advection is the dominant
transport mechanism.
2.3.2 Implementation of Boundary Conditions
All external and internal flow boundaries of the domain (including
injection/withdrawal wells) are treated as third-type boundaries for transport. The
concentrations of contaminant species, input by the user at all possible inflow boundaries,
are multiplied by the fluid flux and implemented into the right-hand side vector. Outflow
boundaries to the system do not require user supplied concentration values, which are
ignored if provided. For outflow boundaries, the fluid flux is added to the diagonal term
of the left-hand-side matrix and outflow concentrations are computed by the code. All
third-type (fluid/contaminant flux) boundaries act in an additive fashion, so a user may
implement, for example, recharge, evapotranspiration, and drain conditions, all on one
node. For instance, for an active water phase, recharge water with prescribed
concentrations enters the cell, fresh water (without contaminants) leaves due to
evapotranspiration thus further affecting the cell concentration, and water with calculated
resultant concentration for that cell leaves it through the drain boundary. Note that
evapotranspiration is treated by allowing only water but no contaminant efflux. Prescribed
concentration (first-type) conditions may be implemented anywhere in the domain,
including at flow boundaries. Note that this prescribed concentration is supplied for the
active fluid phase in the system. Numerical implementation is performed by multiplying
the diagonal by a large number and replacing the right-hand-side by the large number
times the prescribed concentration value for the respective node. Note that first-type
boundary conditions supersede all other third-type conditions prescribed at the same node.
2.3.3 Solution Procedures
Nonlinearities in adsorption or of implicit TVD schemes used to solve the
transport equation are treated using Picard iterations. The second term on the right hand
side of equation (1) is expanded using the modified Picard method due to the non-linearity
of the Freundlich isotherm of equation (21). This term is written in finite difference form
as
∂ ρ
∂t
( ) (
ρ Β csk = Β csk n+1,m+1 − csk n
∆t
) (32)

Where superscript n refers to the old time step, n+1 denotes the current time
step, m refers to the latest iterate value of cn+1 available, and m+1 is the next iteration.
A Taylor series expansion of the first term on the right hand side gives
n +1, m
⎡ dc sk ⎤
c k n +1, m +1
s =c k n +1, m
s +⎢ k ⎥ {c k n +1, m +1
− c k n +1,m } (33)
⎣ dc ⎦
where
dcsk
dc k
( ) (
= η kκ k c k * * η k − 1 ) (34)
is obtained from differentiating equation (21). The coefficient matrix therefore gets the
diagonal term ρ β ∆t dc s dc
k k
( )
and the right hand side vector gets
[ρ ck n
Β s (
/ ∆t − ρ Β c sk n +1,m / ∆t + ρ Β / ∆t dc sk / dc k n +1,m )] as the discretized
adsorbed contaminant storage term per unit volume of the system. Iterations are
performed on each component k, till convergence before proceeding to the next component
of a chain decay situation, or to the next time step of a transient simulation.
The Cooley (1983) under-relaxation scheme is used to dampen flip-flop behavior
during iterations. The system of matrix equations is solved using the Orthomin scheme
(Anderson, 1983a, b; Behie and Vinsome, 1982; Panday et al., 1993). Cross-dispersion
terms are included for the mixed and upstream weighted schemes, if desired, and are
implemented through the right-hand-side vector. For Courant numbers not much greater
than unity, the user may choose to apply the cross-dispersion term or the TVD flux limiter
as a flux correction with two iterations, as is done in Flux Corrected Transport (FCT)
schemes (Boris and Book, 1973). Time-lagging these terms is common practice by
modelers, and this option may be invoked by requesting only one iteration. However,
iterations are relatively inexpensive, and time-lagging is not advised.
Non-linearities also occur in solution to the transport equation, when an immobile
NAPL phase is simulated. For this case, the transport equations for all contaminant
species are solved first. Then, the flash package is invoked on grid blocks which contain
NAPL. The presence of NAPL is indicted if its saturation is greater than zero.
Alternatively, the appearance of NAPL within a grid block is indicated when the constraint
(16) or (17) is exceeded. The flash package is bypassed for grid blocks that contain no
NAPL saturations, and where concentrations are less than the constraint (16) or (17).
After contaminants have been flashed to equilibrium phase conditions throughout the
domain, an iterative procedure repeats the computations until convergence is achieved.
The flash package itself solves the system of equilibrium equations (17) and (19)
simultaneously using a Newton-Raphson scheme.
When an implicit TVD solution or a nonlinear transport problem does not
converge within the prescribed number of iterations for a particular flow time step, the
time-step size is reduced by a factor of 2 and the solution is reattempted for the reduced
time-step size. This process is repeated until convergence is achieved or a user-defined
number of sub-steps is exceeded, whereby the computations are aborted. Successful
computations for all sub-steps are performed before proceeding to the next flow time step.
When an explicit TVD scheme encounters stability problems, the target time-step is

divided into a number of stable sub-steps for which computations are performed before
proceeding to the next flow time step, if a user defined maximum is not exceeded.
2.3.4 Concentration Output
The concentration distribution of the active phase for every chemical species is
reported in the output file, if print options are on for the given time step. Concentrations
during sub-steps (i.e., when a time-step size entering the ACT modules is reduced due to
non-convergence or stability considerations) are not reported. The formats of these
outputs are similar to the respective formats which are used by MODFLOW to output
heads. Iteration summaries for nonlinear cases or TVD solutions and convergence
histories are reported for every time step in the output file. This information is important
in adjusting time-step sizes, automatic time-stepping parameters, and maximum iteration
counts for subsequent simulations in order to maximize speed and accuracy in
computation.
2.4 MASS BUDGET COMPUTATION

A transport mass budget is computed at the end of each time step and accumulated
to provide summarized information on the total mass of contaminant into or out of the
groundwater flow system, as well as, the rates and the individual contributions. The
integral form of (1) is obtained with application of Green’s theorem to the dispersion and
advection terms.
⎡∂ ∂
⎛ k ∂c k
⎜ Dij
⎞
− vi c k ⎟ni dΒ − ∫
( k
) k
(
⎢ ∂t ϕS w c + ∂t ρ Β c s + λwϕS w c
k k
) ⎤
⎥ d∀ + qc*k d∀ = 0 (35)
∫ ⎜ ∂x j ⎟ ⎢ k ⎥ ∫
⎝ ⎠
⎣⎢+ λs ρ Β c s − λw ϕS w c − λs ρ Β c s ⎦⎥
Β ∀ k k −1 k −1 k −1 k −1 ∀
where œ is the solution domain with boundary B, and the source/sink function is
represented by
nw
∫ qc
*k
d∀ = ∑ QI ′ c I*′k δ ( xi − xiI ′ ) (36)
∀ I ′ =1
with QIN being the net volumetric flow rate of the well, or any other internal flow boundary
node, and * being the Dirac delta function at xi=xiIN (i=1,2,3). All terms have been
defined in Section 2.2, and the superscript k denotes the component index for a species,
while k&1 is the component index of the parent of component k, in a chain decay situation.
Also note that coefficients in equation (35) represent effective values of equation (5)
through (8) or (12) through (15) when multiphase transport is activated with
IEQPART>0.
Upon combining (35) and (36) and discretizing the terms using the block-centered
finite-difference approach, the following expression for the rate of material balance error
for a particular time step is obtained:
ε&M = ∑ (τ xx ∇ x c k − Qx c k )I * − ∑
nΒ nr nc nL
I *=1 i =1
∑ ∑ (∆x ∆y ∆z )⋅
j =1 k =1
j i k
[∇ (ϕS c )+ ∇ (ρ c )+ (λ ϕS c )
t w
k
t
k
Β s
k
w w
k n +1
+ λks ρ Β csk − λkw−1ϕS w c k −1 (37)
] ( )
nw
− λks −1 ρ Β c sk −1 + ∑ QI ′ c I*′k
n +1
I ′=1

where nB is the number of nodes on the flow boundary; nr , nc , and nL are the numbers
of rows, columns, and layers, respectively; and the remaining symbols are as defined
previously.
Equation (37) can be expressed in the form:
nΒ n n n
ε& n +1
M = − ∑ FIΒ* − ∑ FIWS − ∑ FISS − ∑ FIWD
I *=1 I =1 I =1 I =1
(38)
n n n nw
−∑ F + ∑ F D
IS
P
IW +∑F +∑F P
IS
w
I′
I =1 I =1 I =1 I ′ =1
where n is the total number of nodes in the grid (n=nr×nc×nL); the first term is the net
advective-dispersive flux along all boundaries of the domain, the second term is the net
rate of mass storage in the fluid phases, the third term is the net rate of mass storage on
the soil, the fourth term is the net rate of decay in the fluid phases, the fifth term is the net
rate of decay on soil, the sixth term is the net rate of gain of component by decay of its
parent in fluid phases, the seventh term is the net rate of gain of component by decay of
its parent on soil, and the final term is the net rate of mass injection-withdrawal,
respectively.
2.5 MASS BUDGET OUTPUT

The individual components of the mass budget are reported by MODFLOW-
SURFACT/MODHMS in the output file in the same format as that used by MODFLOW.
Thus the fluxes are separated into categories IN and OUT, and the mass flow rate budget
for the current time step, as well as the cumulative mass budget since the start of
simulation, are presented (Table 2.1). If transport occurs in more than one fluid phase
(IEQPART>0.), the storage, decay, and mass generation of contaminants in the other
phases are printed accordingly. Mass fluxes across all boundaries are categorized
individually for the respective boundary conditions (i.e., river, stream, drain, recharge-
seepage face, evapotranspiration, prescribed head boundaries, and prescribed
concentration boundaries) and are also saved to a binary file, if requested, for post-
processing calculations. Note that the storage gain and degradation loss terms are reported
under the OUT category (Items 2a, 2b, 2c, and 2d respectively), and the storage loss
(Items 1a and 1b) and gain-from-parent terms are reported under the IN category as is
done in MODFLOW to report positive numbers. Note that related output will occur for
the air and NAPL phases, if the situation requires the code to do so. Absolute and relative
errors of simulation are also reported in the output file, as (IN-OUT), and 2* (IN-
OUT)/(IN+OUT) *100% (Items 2h and 2i respectively).
A more conventional way of computing mass balance errors (Domenico and
Schwartz, 1990) is to individually sum the net boundary influx (I) (where I = 1c + 1d of
Table 2.1), the boundary outflux (O) (where O = 2c + 2f of Table 2.1), and the net
change of storage within the domain ()S, where )S includes gains in storage, gain from
parent, and loss from decay, i.e., )S = 2a + 2b +2c +2d – 1a – 1b for the example in
Table 2.1), and use the following formula:
E = I − O − ∆S
where + is the error, which should be zero if the solution satisfies mass conservation
perfectly. It may be noted that + is also equal to IN-OUT as expressed in Table 2.1. The
mass balance computation provides a global measure of the behavior of the numerical

solution. If the solution is well behaved, the mass balance errors will be small. The
schemes used for the ACT modules should provide normalized mass balance errors that
are only a fraction of a percent.
Table 2.1 Mass Balance Reporting for Transport Calculations
Item Cumulative Masses, M Rates for this time step, M/T
1 IN IN
a Storage Loss in Water =0.45621E+09 Storage Loss in Water =0.15969
b Storage Loss on Soil =0.00000E+00 Storage Loss on Soil =0.00000E+00
c Well Boundaries =0.47304E+09 Well Boundaries =0.00000E+00
d Const Head Bound. =0.11712E-21 Const Head Bound. =0.00000E+00
e Total In =0.92925E+09 Total In =0.15969
2 OUT OUT
a Storage Gain in Water =0.46746E+09 Storage Gain in Water =0.35344E-02
b Storage Gain on Soil =0.00000E+00 Storage Gain on Water =0.00000E+00
c Decay in Water =0.00000E+00 Decay in Water =0.00000E+00
d Decay on Soil =0.00000E+00 Decay on Soil =0.00000E+00
e Well Boundaries =0.37618E+09 Well Boundaries =0.15616
f Const Head Bound. =0.85605E+08 Const Head Bound. =0.00000E+00
g Total Out =0.92925E+09 Total Out =0.15969
h In - Out =10.070 In - Out =0.38820E–07
i % error =0.10836E–05 % error =0.24309E–04
Individual components of all boundary fluxes, cell-by-cell storage terms in all fluid
phases and on the soil, as well as the cell-by-cell decay terms in all fluid phases and on the
soil are saved to a binary file if requested, for post-processing calculations to examine
accumulation and decay behavior in subregions of interest within the domain. The utility
module UBUDSV of MODFLOW is used for binary output. Components of the storage,
degradation (or decay), and generation terms in all fluid phases and on soil for each node
are written to unit IBCFCC defined in the basic transport data file. The text in the heading
of each component of printout includes, in sequential order,
NSTORAGE I FLUIDSN
NSTORAGE ON SOILN
NDECAY I FLUIDSN
NDECAY ON SOILN
NGENERAT. I FLUIDSN
NGENERAT. ON SOILN, for each species.
The user may post-process the results provided for all fluid phases, to provide
storage, decay, or generation of contaminant species in any or all of the individual fluid
phases (air, water, and NAPL), if required.

Components of boundary fluxes are written to files requested by the individual
boundary types. The utility module UBUDSV of MODFLOW is used here as well.
UBUDSV of MODFLOW is used here as well. Table 2.2 lists all mass balance
transport terms that can be printed from the MODFLOW-SURFACT/MODHMS transport
routines and the TEXT header written by UBUDSV to the corresponding cell-by-cell
binary file for each term.
Table 2.2 Mass Balance Terms that Can be Printed from the MODFLOW-
SURFACT/MODHMS Transport Routines
TEXT written to Cell- Binary
Mass Balance by-Cell Binary Output Output File
Term Description Files Extension
CSWLOS loss of storage from water 'STORAGE I FLUIDS' *.TBT
CSALOS loss of storage from air 'STORAGE I FLUIDS' *.TBT
CSNLOS loss of storage from NAPL 'STORAGE I FLUIDS' *.TBT
CSWGAN gain of storage from water 'STORAGE I FLUIDS' *.TBT
CSAGAN gain of storage from air 'STORAGE I FLUIDS' *.TBT
CSNGAN gain of storage from NAPL 'STORAGE I FLUIDS' *.TBT
CSSLOS loss of storage from soil ' STORAGE ON SOIL' *.TBT
CSSGAN gain of storage of soil ' STORAGE ON SOIL' *.TBT
CDWLOS decay loss in water ' DECAY IN FLUIDS' *.TBT
CDALOS decay loss in air ' DECAY IN FLUIDS' *.TBT
CDNLOS decay loss in NAPL ' DECAY IN FLUIDS' *.TBT
CDSLOS decay loss in soil ' DECAY ON SOIL' *.TBT
CPWGAN gain from parent in water 'GENERAT.I FLUIDS' *.TBT
CPAGAN gain from parent in air 'GENERAT.I FLUIDS' *.TBT
CPNGAN gain from parent in NAPL 'GENERAT.I FLUIDS' *.TBT
CPSGAN gain from parent on soil 'GENERAT. ON SOIL' *.TBT
CFWLLOS loss to fracture well FWL4 ' FWELS' *.TFW
CFWLGAN gain from fracture well FWL4 ' FWELS' *.TFW
CFWL5LOS loss to fracture well FWL5 ' FWL5' TW5
CFWL5GAN gain from fracture well FWL5 ' FWL5' TW5
CWELLOS loss to well ' WELLS' *.TBW
CWELGAN gain from well ' WELLS' *.TBW
CHCNLOS loss from constant head boundaries ' CONSTANT HEAD' *.THC
CHCNGAN gain from constant head boundaries ' CONSTANT HEAD' *.THC
CDRNLOS loss to drain boundaries ' DRAINS' *.TBD
CRSFLOS loss to RSF boundaries ' RECHARGE' *.TRC
CRSFGAN gain from RSF boundaries ' RECHARGE' *.TRC
CP_SWLOS loss of ponded water storage ' PONDED STORAGE' *.TRC
CP_SWGAN gain of ponded water storage ' PONDED STORAGE' *.TRC
CP_DWLOS decay loss in ponded water ' PONDED DECAY' *.TRC
CP_PWGAN gain from parent in ponded water ' PONDED GENERAT.' *.TRC
CEVTLOS loss to EVT boundaries ' ET' *.TE1

TEXT written to Cell- Binary
Mass Balance by-Cell Binary Output Output File
Term Description Files Extension
CETSLOS loss to ETS boundaries ' ET2' *.TE2
CRIVLOS loss to river boundaries ' RIVER LEAKAGE' *.TRV
CRIVGAN gain from river boundaries ' RIVER LEAKAGE' *.TRV
CGHBLOS loss to GHB boundaries ' HEAD DEP BOUNDS' *.TBG
CGHBGAN gain from GHB boundaries ' HEAD DEP BOUNDS' *.TBG
CPCNLOS loss to prescribed concentration boundaries ' CONSTANT CONC' *.TPC
CPCNGAN gain from prescribed concentration boundaries ' CONSTANT CONC' *.THC
CSWIMLOS* loss of storage from water (immobile region) --
CSWIMGAN* gain of storage from water (immobile region) --
CDWIMLOS* decay loss in water (immobile region) --
CPWIMGAN* gain from parent in water (immobile region) --
CSSIMLOS* loss of storage from soil (immobile region) --
CSSIMGAN* gain of storage from soil (immobile region) --
CDSIMLOS* decay loss in soil (immobile region) --
CPSIMGAN* gain from parent in soil (immobile region) --
CDUALLOS* dualrate term loss --
CDUALGAN* dualrate term gain --
* Currently storage terms for the dual porosity immobile phase are added to the respective terms in the mobile
phase and reported together.

3.1 GENERAL
This chapter presents the MODFLOW-SURFACT/MODHMS code structure and
operation, and provides instructions for preparation of input files of the ACT modules and
relevant packages. For each computer run, MODFLOW-SURFACT/MODHMS provides
a dual analysis of flow and associated transport problems. As mentioned earlier, input
preparation for MODFLOW-SURFACT/MODHMS is straightforward and follows the
MODFLOW format structure. Transport simulations make use of all MODFLOW data
sets and duplication of this data is eliminated.
The input preparation guide in this chapter is presented, assuming that the user has
existing data files for the groundwater flow analysis and is ready to proceed to edit these
files as well as create additional input files for the contaminant transport analysis. If this
assumption is incorrect, the user will need to refer to the MODFLOW User’s Manual and
Volume I of MODFLOW-SURFACT/MODHMS document to create the flow input files.
3.2 CODE STRUCTURE AND OPERATION

MODFLOW-SURFACT/MODHMS retains the modular structure of the original
MODFLOW code. The similar design of code structure is intended to facilitate the use
of flow and transport modules in a consistent manner and minimize the user’s learning
curve.
MODFLOW-SURFACT/MODHMS provides complete analysis of flow and
associated transport problems. Figure 3.1 shows a general flow chart depicting the code’s
functional operation. Computer memory is first allocated for the arrays of the flow and
transport modules. This is followed by routines to read and process data for flow and
transport. Boundary data for flow and transport are then read for each stress period, and
the time-stepping loop is entered. Solution to the groundwater flow equation is then
performed, followed by solution to the solute transport equation for each time-step. If
flow is steady-state, the flow equation is solved only for the first time-step. The transport
modules are entered once for each flow time-step as seen in Figure 3.1. Shown in Figure
3.2 are the general procedures for transient transport analysis. The transport routine
assembles and solves the transport equation for each species considered in the simulation.
For non-linear cases, iterations are performed until convergence is obtained for each
species.
Three types of transport simulations are possible with MODFLOW-
SURFACT/MODHMS as described below.
1. Transient transport simulation in a transient flow field.
2. Transient transport simulation in a steady-state flow field.
3. Steady-state transport simulation in a steady-state flow field.
The first scenario is the most general case of transient transport in a transient flow
field, where the transport equation is solved for each flow time-step, for each stress
period. The second scenario is the situation where a steady-state flow simulation is
performed by setting the steady-state flag (i.e., ISS … 0) in MODFLOW. The associated
transport simulation is performed subsequently, for all time-steps, in this steady-state
velocity field, for each stress period. If the flow field changes between stress periods, it
Chapter 3: User’s Guide for the ACT Modules 3-1

Figure 3.1 Flow Chart Depicting Functional Operation of the Integrated
MODFLOW-SURFACT/MODHMS Code.
3-2 Chapter 3: User’s Guide for the ACT Modules

Figure 3.2 Flow Chart Depicting the Computational Procedures for Transient
Transport Analysis.

field (Scenario 3) by setting the steady-state flag (i.e., ISS … 0) of MODFLOW and using
a very large (1010) time-step size.
If, in dealing with Scenarios 2 and 3, the steady-state flow does not converge, then
the problem needs to be solved using a transient simulation approach for a long time
period until storage terms diminish to negligible values and steady-state conditions are
achieved. Then, the final set of head values may be used as a starting condition for the
code to complete the analysis in a subsequent combined flow/transport simulation run
using the fully transient approach. The flow solution should converge quickly since the
flow field is near or at steady-state.
Transport solution options in MODFLOW-SURFACT/MODHMS may be
specified via two flags set in the Basic Transport (BTN1) Package (described in Section
3.4.2). The flag IACLVL controls spatial accuracy and the flag THETRD controls
temporal discretization accuracy. The flag IACLVL provides the following spatial
approximation options:
1. full upstream weighting (IACLVL=0);
2. mixed (combined upstream-central difference) weighting with upstream factor
automatically determined by the code (IACLVL=1);
3. user-defined spatial weighting (IACLVL=–1); or
4. the TVD scheme with the van Leer flux limiter (IACLVL–2).
The first three schemes are linear schemes, while the last scheme is nonlinear, for
non-explicit time stepping. Novice transport modelers are advised to set IACLVL = 0 or
1 for early calibration simulations, which may be switched to IACLVL = –2 for final
simulations requiring greater accuracy.
The flag THETRD provides the following temporal discretization options:
1. automatic selection of implicitness (THETRD=0);
2. fully implicit time weighting (THETRD=1);
3. Crank-Nicolson time weighting (THETRD=0.5); and
4. explicit time weighting (THETRD=0.0001).
Novice transport modelers are advised to always use THETRD=0. For non-TVD
schemes, this will provide a Crank-Nicolson time-weighting with second-order accuracy
in time. For TVD schemes, the implicitness is chosen as close to Crank-Nicolson as
possible without violating the TVD property. Experienced modelers may select any value
of THETRD, from fully implicit to fully explicit. Stability conditions are checked
automatically by MODFLOW-SURFACT/MODHMS, when explicit schemes are used.
The Flux Corrected Transport (FCT) scheme may be implemented for linear transport
problems by setting IACLVL=–2 and the maximum number of transport iterations
(MXITERC) to 2. The user should ensure that Courant Numbers in the simulation are not
much greater than unity when the FCT scheme is used, to preserve solution accuracy.
Note that the Courant number is defined as V)t/)x, where V is the contaminant or solute
velocity, )t is the time-step, and )x is the grid nodal spacing.
For non-linear transport simulations (due to non-linear retardation or the use of
the non-linear TVD schemes), the maximum number of transport iterations (MXITERC)
should be greater than one. If convergence is not achieved within the specified number
of iterations (suggested number is between 5 and 25) the time-step size is halved, and the
transport solution is repeated for that species. This process is repeated until convergence

is achieved or a maximum number of time-step reductions (NNOTCV) is reached whereby
the simulation is aborted. Converged solutions are obtained for all sub-time-steps before
proceeding to the next target time-step. Time lagging this term may be achieved by setting
MXITERC=1, but this is not advised. An FCT type of update may be performed by
setting MXITERC to two.
For transport simulations that include a residual, dissolving NAPL phase (i.e., if
IEQPART=2) the index MXOUTIT determines the maximum number of nonlinear
transport iterations (suggested value between 5 and 25), and the index NITFLASH
determines the maximum number of Newton iterations to be performed during the flash
calculation (suggested value between 10 and 30). If MXOUTIT is set to one, the
saturation update for the redistributing NAPL phase is time lagged, but this is not
recommended. A correction iteration is performed if MXOUTIT=2, before moving onto
the next time step, as in FCT schemes. For MXOUTIT $3, the transport solution
proceeds until convergence. If convergence is not achieved within the prescribed
MXOUTIT iterations, the time step size is halved, and the transport solution for all species
is repeated along with the flash calculations, for the smaller time value. Converged
solutions are obtained for all subtime-steps before proceeding to the next target time step.
Note that automatic time-stepping is not affected by the convergence behavior of nonlinear
transport simulations. Automatic time-stepping, when used for steady-state flow
conditions, behaves like the regular time-step-size incrementing scheme of MODFLOW
with a maximum limit on the time-step size which cannot be exceeded. The maximum
time-step size is an important consideration for all simulations involving transport, since
large Courant numbers may produce significant numerical dispersion because of temporal
discretization errors. Good judgement on the user’s part with respect to spatial and
temporal discretizations is not easily replaceable.
3.3 BASIC PROCEDURE FOR INPUT PREPARATION

As mentioned earlier, MODFLOW-SURFACT/MODHMS is designed to conduct
a dual analysis of groundwater flow and contaminant transport in a single computer run.
Thus it is assumed at this stage that the user has created or obtained input files for
groundwater flow simulation. With this assumption, the basic procedures for input
preparation necessary for involving transport simulation options are straightforward.
In essence, the user needs to modify the relevant flow input files and create up to
three new input files containing additional information required for transport modeling.
Shown in Table 3.1 are the identification and input functions of the relevant flow packages
and three transport packages for which detailed instructions for input preparation are
presented in the next section.
3.4 INPUT INSTRUCTIONS

3.4.1 General
This section provides general and specific instructions for preparation of the
additional input data required for the code to conduct transport simulation concurrently
with groundwater flow simulation in the same computer run.
The user specifies the requirement for transport simulation via a control variable
(ITRAN) of the BAS Package, described in Section 3.4.2. The BAS Package has been
modified to incorporate ITRAN and additional input file units needed for the transport
analysis. These input file units are used by the BTN1, HCN1, and PCN1 Packages to

Table 3.1 Packages Relevant to Transport Simulation
Package Relevant Input Function

BAS Reads values of certain control variables (flags) for
(Basic Flow) transport simulation and opens additional files needed.
*BTN1 Reads control variables and arrays containing aquifer

(Basic Transport) and contaminant transport parameters.
*HCN1 Reads prescribed concentration values for model cells

(Head-Concentration) that are being treated as constant-head or specified-head
cells.
*PCN1 Reads prescribed concentration values for model cells
(Prescribed that receive first-type transport boundary conditions.
Concentration)
WEL Reads a mass-flux-terms flag and concentration values
(Well) specified for the injection well cells (Q>0).
DRN Reads a mass-flux-terms flag for the drain cells.

(Drain)
RIV Reads a mass-flux-terms flag and prescribed
(River) concentration values for the river cells.
GHB Reads a mass-flux-terms flag and prescribed

(General-Head- concentration values for the GHB cells.
Boundary)
RSF4 Reads a mass-flux-terms flag and concentration values
(Recharge-Seepage for recharge water to be applied to recharge cells.
Face)
FWL4/FWL5 Reads a mass-flux-terms flag and concentration values
(Fracture Well) for the injection wells.
PCG4 Reads an Orthomin solver flag.

(PCG Solver)
OC Reads output control variables.
(Output Control)
ATO4 Reads output control variables.
(Adaptive Time-
Stepping)
EVT/ET2 Reads a mass-flux-terms flag and fraction of
(Evapotranspiration) contaminant that leaves with ET flux.
* = New Transport Package

read data pertaining to transport parameters and first-type boundary conditions as
described in Sections 3.4.3 through 3.4.5. Section 3.4.6 through 3.4.11 provide input
instructions for the WEL, DRN, RIV, GHB, RSF4, and FWL4 Packages, all of which
have been modified to incorporate additional sink/source data for the transport simulation.
Note that the sink/source packages are invoked only when they are needed (i.e., when the
relevant sinks/sources are present in the flow domain).
In order to solve the transport matrix equation, MODFLOW-SURFACT/MODHMS
employs an iterative method with matrix preconditioning and orthomin accelerations,
referred to as the Orthomin technique and included in the PCG4 Package. Section 3.4.12
provides the input instructions for the PCG4 solver. Finally, Sections 3.4.13 and 3.4.14
detail the input instructions for the OC and ATO4 Packages. The entire flow and transport
simulation printouts are controlled by the package selected from these two. Note that use
of adaptive-time-stepping of the ATO4 Package is advocated especially for unconfined
flow simulations.
3.4.2 Basic Flow (BAS) Package Input
Input for the BAS package is read from Unit 1 as specified in the main program
of MODFLOW. The BAS package has been modified to require additional flags
indicating that a transport simulation is requested and to open three additional files that are
needed. These flags correspond to the unit numbers on IUNIT(17), IUNIT(23), and
IUNIT(24). The flag ITRAN indicates the transport simulation requirements. If the flag
is zero, no transport simulation is required and only a flow simulation is performed. If
the flag is unity, a dual flow-and-transport simulation is required. Note that the user may
turn off the flag ITRAN at any stage of the modeling study to perform the flow simulation
only. No other data need be disturbed. The capability for performing steady-state flow
analysis for some stress periods and transient analysis for others, has been included from
a similar concept used in MODFLOW 2000. The entire input data structure is presented
here for the BAS package for completeness. The additional variables required for
transport analysis are shown in shaded boxes.
1. Data: HEADNG(32)
Format: 20A4
2. Data: HEADNG (continued)
Format: 12A4
3. Data: NLAY NROW NCOL NPER ITMUNI ITRAN IDUAL MPHFLAG

Format: I10 I10 I10 I10 I10 I10 I10 I10
4. Data: IUNIT(60)
Format: 60I3
(BCF WEL DRN RIV EVT TLK GHB RSF SIP DE4 SOR OC
STR PG2 HFB LAK2 IBS CHD FWL4 ATO OLF
CHF PRL OBS GFD FHB IPT RCT SLC LMG PCG5 RES FWL5 SFR1
LAK3 GAGE ETS1)

5. Data: IAPART ISTRT
Format: I10 I10
**Item 6 consists of one array for each layer in the model grid. Arrays for 1 through
NLAY layers must be entered in order.
6. Data: IBOUND(NCOL, NROW)

Utility Module: U2DINT. Note that U2DINT is the utility module that reads
a two-dimensional integer array (i.e., IBOUND for item 6).
7. Data: HNOFLO DATUM

Format: F10.3 F10.3
8. Data: SHEAD(NCOL, NROW)

**Enter item 9 for each stress period.
9. Data: PERLEN NSTP TSMULT SS/TR

Format: F10.0 I10 F10.0 8X, A2

HEADNG -- is the simulation title that is printed on the printout. It may be up to 128
characters long; 80 in the first record and 48 in the second. Both
records must be included even if they are blank.
NLAY -- is the number of model layers.
NROW -- is the number of model rows.
NCOL -- is the number of model columns.
NPER -- is the number of stress periods in the simulation.
ITMUNI -- indicates the time unit of model data. (It is used only for printout of
elapsed simulation time. It does not affect model calculations.)
0- undefined 3 - hours
1- seconds 4 - days
2- minutes 5 - years
The unit of time must be consistent for all data values that involve time.
For example, if years is the chosen time unit, stress-period length, time-
step length, transmissivity, etc., must all be expressed using years for
their time units. Likewise, the length unit must also be consistent.
ITRAN -- is a flag that determines if the simulation is either flow only or flow and
transport.

If ITRAN = 0, the simulation is for flow only.
If ITRAN = 1, the simulation is for both flow and transport.
IDUAL -- For use with dual porosity module.

MPHFLAG -- When MPHFLAG > 0 and ISS = 0, the ATO4 package is used.
MODFLOW-SURFACT/MODHMS writes out a binary file
*.mph, which is used by MODPATH to determine time step
information.
IUNIT -- is a 60-element table of input units for use by all major options. The
additional packages required for flow and transport simulation are
shown in shaded boxes. Note that the Preconditioned Conjugate
Gradient (PCG) solver specified in IUNIT(13) must be used for the
transport simulation (i.e., ITRAN=1). For the flow simulation, any
solver (SIP, SSOR, or PCG), may be used.
IUNIT LOCATION MAJOR OPTION

1 Block-Centered Flow (BCF4) Package
2 Well (WEL) Package
3 Drain (DRN) Package
4 River (RIV) Package
5 Evapotranspiration (EVT) Package
6 Transient Leakage Package (TLK)
7 General-Head-Boundary (GHB) Package
8 Recharge-Seepage Face Boundary (RSF4) Package
9 Strongly Implicit Procedure (SIP) Package
10 Direct Solver Package (DE4)
11 Slice Successive Over-Relaxation (SSOR) Package
12 Output Control (OC) Package
13 Preconditioned Conjugate Gradient (PCG4) Package

14 Streamflow-Routing (STR) Package
15 Preconditioned Conjugate Gradient (PCG2) Package
16 Horizontal-Flow Barrier (HFB) Package
17 Basic Transport (BTN1) Package

18 Lake (LAK2) Package
19 Interbed-Storage (IBS) Package
20 Time-Variant Specified-Head (CHD) Package
21 Fractured-Well (FWL4) Package

22 Adaptive Time-Stepping and Output Control (ATO4)
Package
23 Prescribed-Concentration Boundary (PCN1) Package
24 Specified-head Concentration Boundary (HCN1)

Package
25 Overland Flow Package (OLF)1
26 Channel Flow Package (CHF)1
27 Analytical Infiltration Package (PRL)1
28 Observation Node Package (OBS)
29 General Finite Differences (GFD)
30 Flow and Head (time-varying) Boundary Package (FHB)
31 Interception of Precipitation and Evapotranspiration
Package (IPT)1
32 Reactive Transport Package (RCT)
33 Saltchem Module (SLC)
34 Link - AMG Package (LMG)2
35 Preconditioned Conjugate Gradient 5 Package (PCG5)
36 Reservoir Package (RES)
37 Multi-Node Fracture Well Package (FWL5)
38 Steam-Flow Routing Package (SFR1)
39 Lake 3 Package (LAK3)
40 Gage Package (GAGE)
41 Evapotranspiration Segments Package (ETS1)
1. Note: OLF, CHF, PRL and IPT packages are not available with MODFLOW-
SURFACT/MODHMS.
2. Note: LMG Package links are provided, however the AMG solver is not available.
If IUNIT(n) # 0, the corresponding major option is not being used.

If IUNIT(n) > 0, the corresponding major option is being used and
data for that option will be read from the unit
number contained in IUNIT(n). The unit numbers
in IUNIT should be integers from 1 to 98. It is
recommended that a different number be used for
each major option. Printer output is assigned to
unit 6 (unless it is changed to meet computer
requirements). That unit number should not be
used for any other input or output. The user is also

permitted to assign unit numbers for output. Those
numbers should be different from those assigned to
input. The Basic Package reads input from unit 1
(unless it is changed to meet computer
requirements). It is permissible but unwise to use
that unit for other major options. Note that if both
IUNIT(12) and IUNIT(22) are greater than zero,
IUNIT(12) is set equal to zero. (Refer to OC
Package input or ATO4 Package input for details.)
IAPART -- indicates whether array BUFF is separate from array RHS.
If IAPART = 0, the arrays BUFF and RHS occupy the same space.
This option conserves space. This option should be
used unless some other package explicitly says
otherwise.
If IAPART … 0, the arrays BUFF and RHS occupy different space.

This option must be used if input LAYCON value of
any model layer is either 31, 33, 40, or 43. Note
that LAYCON values for layers are input in item 2
of the BCF4 Package.
ISTRT -- indicates whether starting heads are to be saved. If they are saved, they
will be stored in array STRT. They must be saved if drawdown is
calculated.
If ISTRT = 0, starting heads are not saved.
If ISTRT … 0, starting heads are saved.
IBOUND -- is the boundary array.
If IBOUND(J,I,K) < 0, cell J,I,K has a constant head where
J,I,K are the column, row, layer
numbers of the cell, respectively.
If IBOUND(J,I,K) = 0, cell J,I,K is inactive.
If IBOUND(J,I,K) > 0, cell J,I,K is active.
HNOFLO -- is the value of head to be assigned to all inactive cells (IBOUND = 0)
throughout the simulation. Since heads at inactive cells are unused, this
does not affect model results but serves to identify inactive cells when
head is printed. This value is also used as drawdown at inactive cells
if the drawdown option is used. Even if the user does not anticipate
having inactive cells, a value for HNOFLO must be submitted.
DATUM -- is a shift value applied to the datum of all head and elevation
calculations. This amount is subtracted from all head and
elevation values internally, to minimize computational round-off
errors. Note that heads and elevations of a simulation are reported
with respect to the original datum and this entry is only used
internally for scaling purposes.
SHEAD -- is head at start of the simulation. Regardless of whether starting head

is saved, these values must be input to initialize the solution.
PERLEN -- is the length of a stress period. It is specified for each stress period.

NSTP -- is the number of time steps in a stress period. Note that if ATO4
Package is used (i.e., IUNIT(22) > 2), the number of time steps in a
stress period are determined by the code.
TSMULT -- is the multiplier for the length of successive time steps. The length of
the first time step DELT(1) is related to PERLEN, NSTP, and
TSMULT by the relation
DELT(1) = PERLEN(1-TSMULT) / (1-TSMULT**NSTP).
Note that if ATO4 Package is used (i.e., IUNIT(22) > 0), TSMULT
value read from BAS input will be overwritten by that read from ATO4
Package.
SS/TR -- is a character variable that indicates if the stress period is transient or
stead state. The only allowed options are ‘SS’ and ‘TR’. If left blank,
the entire simulation proceeds in steady or transient mode according to
flag ISS in the BCF file. Note that if this option is used, the flag ISS
should be set to zero in the BCF file.
3.4.3 Basic Transport (BTN1) Package Input
Parameter input for the transport simulation is read by the BTN1 Package from
a file specified in IUNIT(17). Recall that the IUNIT array, which is read by the Basic
Flow (BAS) Package, determines the options or packages required by the simulation and
hence IUNIT (17) must not be set to zero. This package is always required with a
transport simulation. Note that ITRAN must be set equal to 1 in the BAS input file to
activate a transport simulation. Input parameters for the new BTN1 Package are listed
below along with their respective formats. Following this list is an explanation of the
input parameters.
1. Data: LINR IDCYTP ILAMWS IACLVL NSPECI ICHAIN IDISP IEQPART NDENS IUNCAD
Format: I10 I10 I10 I10 I10 I10 I10 I10 I10 I10
**Enter item 2a only if IEQPART = 2.
2a. Data: COMPNAPL NITFLASH MXOUTIT IINIT SNCLOSE

Format: F10.0 I10 I10 I10 F10.0
**Enter item 2b only if IEQPART = 2 and if IREALSL = 1 or 2 (i.e., when water is the
active phase).
2b. Data: RHOWP RHOAP COMPWAT COMPAIR ATMGP GRAV

Format: F10.0 F10.0 F10.0 F10.0 F10.0 F10.0
**Enter item 3 only if IACLVL = –1.
3. Data: ALFX ALFY ALFZ

Format: F10.0 F10.0 F10.0
4. Data: CNOFLO CCLOSE MXITERC NNOTCV THETRD IBCFCC ICROSS NOMATRIX IPHSFLG
Format: F10.0 F10.0 I10 I10 F10.0 I10 I10 I10 I10
** Each of the following items (5 through 21) consists of NLAY two-dimensional arrays,
one for each layer in the model grid. Thus, NLAY number of arrays must be specified

in item 5 (sequentially, 1 through NLAY for each of the layers) before entering the arrays
of item 6. NLAY is the number of layers in the model grid. If an item is not needed, it
must be omitted.
**Enter items 5 and 6 only if input LAYCON value of any model layer is neither 40 nor
43. Input LAYCON values for layers are read from item 2 of the BCF4 Package. Note
that for input LAYCON = 40 or 43 (i.e., for the variably saturated flow option) the bottom
and top elevations of all the layers are read by the BCF4 Package and, therefore, are not
requested below.
5. Data: BOT(NCOL, NROW)

Utility Module: U2DREL.
Note that item 5 consists of one array for each layer in the model grid. Arrays for layers
1 through NLAY must be entered in order.
6. Data: TOP(NCOL, NROW)

7. Data: LDISP (NCOL, NROW)

**Enter item 8 only if IDISP > 0.
8. Data: TDISP (NCOL, NROW)

**Enter item 9 only if IDISP >1.
9. Data: VTDISP (NCOL, NROW)

10. Data: VLDISP (NCOL, NROW)


11. Data: PHI (NCOL, NROW)
**Enter item 12 only if LINR >0.
12. Data: RHOB (NCOL, NROW)

13a. Data: DIFF (I), I=1, NSPECI

Format: 5F10.0
**Enter item 13b only if IEQPART >0.
13b. Data: DIFF2 (I), I=1, NSPECI

Format: 5F10.0
**Enter item 13c only if IEQPART=2.
13c. Data: DIFF3 (I), I=1, NSPECI

Format: 5F10.0
**Enter item 14a only if IEQPART >0 or if IREALSL >2.
14a. Data: EQPARTAW (I), I=1, NSPECI

Format: 5F10.0
**Enter item 14b only if IEQPART=2.
14b. Data: EQPARTNW (I), I=1, NSPECI

Format: 5F10.0
**Enter item 15 only if IEQPART=2.
15. Data: RHOK (I), I=1,NSPECI

Format: 5F10.0
16. Data: SNAPL (NCOL, NROW)


**The following items are used to input chemical properties of each species participating
in the simulation. Items 17 through 21 are sequentially input for each of the chemicals,
i.e., all of the arrays (items 17 through 21) for Species 1 are read first; then all of the
arrays for Species 2, etc. If an array is not needed, it must be omitted.
**Enter item 17 only if LINR > 0.
17. Data: KD (NCOL, NROW)

**Enter item 18 only if LINR = 2.
18. Data: ETA (NCOL, NROW)

19. Data: ALAM1 (NCOL, NROW)

** Enter item 20a only if ILAMWS = 0.
20a. Data: ALAM2 (NCOL, NROW)

Note that item 20a consists of one array for each layer in the model grid. Arrays for
** Enter item 20b only if ILAMWS = 0.
20b. Data: ALAM3 (NCOL, NROW)

Note that 20b consists of one array for each layer in the model grid. Arrays for layers 1
through NLAY must be entered in order.
** Enter item 20c only if ILAMWS = 0.
20c. Data: ALAM4 (NCOL, NROW)

Note that 20c consists of one arrary for each layer in the model grid. Arrays for layers

21. Data: SCONC (NCOL, NROW)
***Enter Items 22a, 22b and 22c only if ICHAIN=2
22a) Data: NPAREN(I)
Format: I10
22b) Data: KPAREN (J), J=1, NPAREN
Format: 5I10
22c) Data: SFACTR(J), J=1, NPAREN
Format 5F10.3
LINR -- is the retardation index.
If LINR = 0, retardation is not applied.
If LINR = 1, retardation is linear.
If LINR = 2, retardation is nonlinear.
IDCYTP -- is the degradation type index.
If IDCYTP = 1, the degradation is only species dependent (e.g.,
radioactive decay).
If IDCYTP = 0, the degradation is both species and location
dependent (e.g., biodegradation).
ILAMWS -- is an index for the type of degradation on soil.
If ILAMWS = 1, the rate of degradation is same in all phases
including soil.
If ILAMWS = 0, the rate of degradation is different in each
phase including soil.
IACLVL -- is the index for the choice of numerical scheme.
If IACLVL = –2, adaptive TVD scheme with the van Leer flux
limiter.
If IACLVL = –1, upstream weighting scheme with user-specified
upstream factors.
If IACLVL = 0, fully upstream weighting scheme.
If IACLVL = 1, upstream weighting with upstream factors
automatically determined by the code.
Note that the TVD scheme is highly accurate but more computationally
demanding than the other upstream weighting schemes. If steep
concentration fronts are expected, the TVD scheme should be selected.
On the other hand, the fully upstream weighting scheme is the most

efficient but only first-order accurate. If the grid is sufficiently refined
and numerical dispersion is not expected to be significant, the fully
upstream weighting scheme should be selected. The user who is
familiar with upstream weighting may prefer partial to full-upstream
weighting to control numerical oscillations while avoiding too much
numerical dispersion. This is achieved with IACLVL=–1 or 1. With
IACVL=–1, the scheme becomes a central difference scheme if zero
upstream factors are specified.
NSPECI -- is the total number of species to be considered in the transport
simulation.
ICHAIN -- is the index for chain-decay conditions.
If ICHAIN = 2, Chain decay or transformation with daughter
products is accounted for with a branched chain
where one parent can have several daughter
products, or one daughter product may have
several parent components.
If ICHAIN = 1, chain-decay or transformation with daughter
products is taken into account , with one unit of
parent decaying to one unit of daughter in a
straight chain.
If ICHAIN = 0, transformation of one species to another is not
allowed (i.e., single-species transport).
IDISP -- is an index to determine dispersivity data input requirements.
If IDISP = 0, only longitudinal dispersivity data is read. Items
7, 8, and 9 must be omitted.
If IDISP = 1, longitudinal and transverse dispersivity data are
read. Items 8 and 9 must be omitted.
If IDISP = 2, longitudinal, transverse, and vertical transverse
dispersivity data are read. Item 9 must be
omitted.
If IDISP = 3, longitudinal, transverse, vertical transverse, and
vertical longitudinal dispersivity data are read
from items 6, 7, 8, and 9, respectively.
IEQPART -- Index to determine if transport occurs in phases other than the
active phase.
If IEQPART = 0, transport occurs in active phase only.
If IEQPART = 1, transport occurs in air and water phases.
If IEQPART = 2, NAPL phase is also present, and contaminants
are flashed into air and water phases for
transport.
Note that if IEQPART …0, then transport occurs by advection +
dispersion in the active phase, and by diffusion in the passive
phases. The index IREALSL in the BCF4 input package determines
whether water or air is the active phase. Equilibrium partitioning
is used to distribute contaminants among all phases present.

NDENS -- Flag for use with density-dependent module.
IUNCAD -- Index for treatment of adsorption in unconfined grid-blocks.
For IUNCAD = 1, adsorbed mass on solid phase is multiplied
by saturation as:
Cs = Kd * Cw * Saturation
Note: the IUNCAD=1 option may be used only when IREALSL=0
(i.e., when vertical equilibrium is considered using the pseudo-soil
option, or traditional MODFLOW options for the flow simulation).
This option includes consideration of adsorption occurring only in
the water-saturated portion of the grid block.
For IUNCAD = 0, adsorbed mass on solid phase is calculated
according to the traditional multiphase
approach as:
Cs = Kd * Cw
Note: IUNCAD=0 for IREALSL …0.

COMPNAPL -- is the compressibility of NAPL. COMPNAPL may typically be set
to zero due to negligible effects.
NITFLASH -- Number of Newton iterations used to equilibrate the flash package.
MXOUTIT -- Maximum number of outer iterations for convergence of transport
equation when in equilibrium with an immobile NAPL phase.
IINIT -- is the index for concentration initialization when NAPL phase is
present.
If IINIT=0, contaminant concentrations are read in for the
active phase.
NAPL phase in equilibrium with the active phase.
NAPL phase with other phases being assumed
contaminant free for these initial conditions. Mass
is first redistributed by the flash package to
equilibrate contaminant mass among all phases.
SNCLOSE -- Closure tolerance on NAPL saturation.
RHOWP -- Density of water.
RHOAP -- Density of air at reference pressure.
COMPWAT -- Compressibility of water.
COMPAIR -- Compressibility of air.
ATMGP -- Absolute gas pressure of reference conditions.
GRAV -- Gravitational acceleration.
Note: Input instructions for the BCF4 package supply standard values
for RHOWP, RHOAP, COMPAIR, ATMGP, and GRAV in several
commonly used sets of units.

ALFX -- is the user-specified upstream weighting factor in x-direction (i.e.,
along the Row Direction). Read only if IACLVL = –1.
ALFY -- is the user-specified upstream weighting factor in y-direction (i.e.,
along the Column Direction). Read only if IACLVL = –1.
ALFZ -- is the user-specified upstream weighting factor in z-direction (i.e.,
along the Vertical Direction). Read only if IACLVL = –1.
CNOFLO -- is the value of concentration to be assigned to all inactive cells (i.e.,
cells for which IBOUND = 0). Although this value plays no role
in the model calculations, it serves to identify inactive cells while
examining the concentration output. Thus, CNOFLO is similar to
HNOFLO of the Basic Flow (BAS) Package.
CCLOSE -- is the concentration change criterion for convergence in the case of
nonlinear iterations. When the maximum absolute value of the
concentration change at all nodes during an iteration is less than or
equal to CCLOSE the nonlinear iterations are terminated. A value
on the order of 10 –3 to 10–5 of the maximum concentration is
suggested for CCLOSE.
MXITERC -- is the maximum number of outer iterations (i.e., calls to the
solution routine) in the case of nonlinear iterations. A value within
the range of 10 to 50 is suggested for MXITERC.
NNOTCV -- is the maximum number of time-step cuts allowed in solving
transport equations before aborting the simulation. A value within
the range of 3 to 15 is suggested for NNOTCV.
THETRD -- is a control parameter for time weighting factor.
If THETRD = 1.0, fully implicit scheme is used.
If THETRD = 0.5, Crank-Nicolson scheme is used.
If THETRD = 0.0001, explicit scheme is used.
If THETRD = 0.0, Crank-Nicolson factor is selected automatically.
IBCFCC -- is a flag and a unit number on which storage and decay terms for all
phases are saved if requested. The corresponding terms
representing generation of species are also saved if multi-species
chain decay simulation is opted.
If IBCFCC > 0, it is the unit number on which cell-by-cell
storage and decay terms will be recorded
whenever ICCCFL (see Output Control) is set.
If IBCFCC # 0, cell-by-cell storage and decay terms will not be
printed or recorded.
ICROSS -- is a flag indicating if cross-dispersion terms are to be included for
the non-TVD schemes (i.e., for –1 # IACLVL #1, where IACLVL
is entered through item 1).
If ICROSS = 0, cross-dispersion terms are neglected.
If ICROSS = 1, cross-dispersion terms are included.
NOMATRIX -- is the index for not writing coefficient matrix.

If NOMATRIX = 1, coefficient matrix is not written, and is
computed whenever it is required.
If NOMATRIX = 0, coefficient matrix is written to be used
subsequently, if possible.
Note that one may be faster than the other depending on CPU speed
and disk writing speed of the computational environment.
IPHSFLG -- is the index for the phase that occupies space between porosity and
specific yield, when transient flow simulation has a different
specific yield from porosity of transport simulation.
If IPHSFLG =0, active phase occupies space.
If IPHSFLG =1, inactive phase occupies space.
BOT -- is the elevation of the layer bottom. Read only if the variably
saturated flow modeling option is not used in the BCF4 Package.
TOP -- is the elevation of the layer top. Read only if the variably saturated
flow modeling option is not used in the BCF4 Package.
LDISP -- is longitudinal dispersivity (L).
TDISP -- is transverse dispersivity (L). Read only if IDISP > 0.
VTDISP -- is vertical transverse dispersivity (L). Read only if IDISP > 1.
VLDISP -- is vertical longitudinal dispersivity (L). Read only if IDISP > 2.
PHI -- is effective porosity. For dual domain simulations, PHI is the
porosity within the mobile domain, Nm, defined as the void space in
the mobile domain per unit volume of mobile domain.
RHOB -- is bulk density of soil ( ).

M
L3
Read only if LINR > 0. For dual
domain simulations, RHOB is the bulk density of the mobile
domain, if the immobile domain has distinct adsorption
characteristics from the mobile domain (IDUAL=3). If IDUAL=2
then both domains have similar adsorption characteristics and this
parameter defines the adsorption for both domains. If IDUAL=1,
then adsorption is assumed to occur only in the mobile domain, and
this parameter defines the mobile domain adsorption characteristics.
DIFF(I) -- is the molecular diffusion coefficient (L2/T) of species I in the active
phase. Thus, for IREALSL=0, 1, 2, water is the active phase and
DIFF(I) is the free-water molecular diffusion coefficient for species
I. For IREALSL=3, 4, air is the active phase and DIFF(I) is the
molecular diffusion coefficient of species I in air.
DIFF2(I) -- is the molecular diffusion coefficient (L2/T) in the inactive phase.
Note that inactive phase transport via diffusion occurs when
IEQPART…0.
DIFF3(I) -- is the molecular diffusion coefficient (L2/T) in the NAPL phase.
Note that NAPL phase contaminant partitioning and diffusion occur
when IEQPART=2.
EQPARTAW(I) -- is the gas-water partitioning coefficient for species I, defined as
conc_gas=EQPARTAW*conc_water. Note that EQPARTAW is

required as input if air is the active phase, even though the
simulation does not opt for multiphase transport (i.e., even if
IEQPART=0). This is because adsorption is assumed to be
mediated through the water phase, and contaminant needs to be
partitioned into the water phase for subsequent adsorption. For this
case, EQPARTAW may be set equal to 1 if adsorption is assumed
to occur directly from the air phase.
EQPARTNW(I) -- is the NAPL-water partitioning coefficient for species I, defined as
conc_NAPL=EQPARTNW*conc_water.
RHOK(I) -- is the density of pure component species I in its liquid state.
SNAPL -- is the NAPL saturation at the start of the simulation.
KD -- is soil partition coefficient ( ). Read only if LINR > 0.

L3
M
ETA -- is the Freurdlich isotherm exponent (0) in equation (2) of Chapter

2. Read only if LINR = 2.
ALAM1 -- is rate of contaminant degradation in the active fluid phase (1/T).
ALAM1 is also the degradation rate of contaminant on soil, if all
degradation rates are specified as equal.
ALAM2 -- is rate of contaminant degradation on soil (1/T). Read only if
ILAMWS = 0. Note that ALAM2 is not needed if the decay rate
on soil is specified to be same as that in the fluids (i.e., if ILAMWS
= 1 in item 1).
ALAM3 -- is rate of contaminant degradation in the inactive phase (1/T). Note
that inactive phase transport is considered only when IEQPART…0.
Further, ALAM3 is not needed if all decay rates are considered as
equal (i.e., if ILAMWS=1).
ALAM4 -- is the rate of contaminant degradation in the NAPL phase (1/T).
Note that NAPL phase is considered only when IEQPART=2.
SCONC -- is concentration at start of the simulation (M/L3).
Note that KD, ETA, ALAM1, ALAM2, and SCONC are required for each species
in the simulation depending on the indexes set in item 1. Further note that if
IEQPART=2 then the parameter IINIT defines the condition for which the initial
concentration is read into SCONC.
NPAREN(I) -- is the number of parent species of component I
KPAREN (J) -- Component index of the J th parent that decays to component I (for
each J = 1, NPAREN (I)).
SFACTR (J) -- Fraction of parent component J that decays to component I. (for
each J = 1, NPAREN (I)). Decay of one unit of J produces
SFACTR units of component I, therefore SFACTR may also be
used to scale units in a decay process.

3.4.4 Prescribed-Head-Concentration Boundary (HCN1) Package Input
The new HCN1 Package is used to specify concentrations of different species at
the following model cells: (1) model cells that are treated as constant-head cells in the
Basic Flow (BAS) Package (i.e., when IBOUND< 0), and (2) cells that are treated as
time-variant specified-head boundaries in the CHD1 Package of MODFLOW. (See Leake
and Prudic, 1988, for the documentation for the CHD1 package.) This input file is not
required if the simulated domain does not contain a specified-head boundary (i.e., if
IBOUND $ 0 for all the model cells) and the CHD1 Package is not used in the flow
simulation. A file, however, is requested on IUNIT(24) for a transport simulation, if the
flow problem contains prescribed-head cells. This file is necessary if prescribed-head
conditions occur in the groundwater flow simulation, and it is anticipated that non-zero
concentration values can occur at the prescribed-head cells where incoming fluid flow is
possible.
In the HCN1 Package input, the user is required to specify concentrations of the
species entering the flow domain through specified-head cells. These concentrations are
multiplied by the influx of water to provide the contaminant mass influx as a third-type
boundary condition. However, if a specified-head cell behaves as an outflow boundary,
it is unnecessary to input concentration data for that cell. Note that concentration data
provided here will not be used in the model calculation if a boundary cell represents an
outflow boundary. This is because concentrations at outflow cells are actually unknown
and must be determined by the code. The data input for providing concentrations to
constant-head boundaries follows.
Note that it may not be known apriori which constant head nodes are inflow nodes,
and hence concentrations should be supplied at all suspected inflow nodes to provide
contaminant information to the system. Otherwise, zero contaminant flux is assumed at
the inflow nodes. If the node happens to be an outflow node during the simulation, the
concentration of exiting water is the correct boundary condition, which is automatically
determined by the code.
1. Data: IHCNCC
Format: I10
** Enter items 2 and 3 for each stress period.
2. Data: ITMP
Format: I10
**Enter item 3 only if ITMP > 0.
3. Data: Layer Row Column [Conc(I), I = 1,NSPECI]

Format: I10 I10 I10 5F10.0
(Input item 3 consists of one record for each specified-head boundary cell in the model
grid, at which inflow is anticipated. If ITMP is negative or zero, item 3 is not read.)
IHCNCC -- is a flag and a unit number.
If IHCNCC > 0, it is the unit number on which contaminant mass-
flux-terms of each species at all the specified-head

cells will be recorded whenever ICCCFL (see output
control) is set.
If IHCNCC = 0, mass-flux-terms of species will not be printed or
recorded.
If IHCNCC < 0, mass-flux-terms at all the specified-head cells will be
printed whenever ICCCFL is set.
If ITMP < 0, concentration data at the specified-head boundaries
from the last stress period will be reused.
If ITMP $ 0, ITMP is the number specified-head boundary cells in
the current stress period, which are possible inflow
cells and thus require contaminant concentration data.
Layer -- is the layer number of the specified-head boundary cell.
Row -- is the row number of the specified-head boundary cell.
Column -- is the column number of the specified-head boundary cell.
Conc(I) -- is the concentration of species I. Note that NSPECI is the total number
of species in the transport simulation.
3.4.5 Prescribed Concentration Boundary (PCN1) Package Input
Input for the new PCN Package is read from the unit specified in IUNIT(23).
However, this input is required only if prescribed concentration (first-type) boundary
conditions exist. Recall that the IUNIT array, which is read by the Basic Flow (BAS)
package, determines the options or packages used in a MODFLOW-
SURFACT/MODHMS simulation. (See the BAS Package input.) The PCN input data
is required to incorporate first-type boundary conditions in a transport simulation, (i.e.,
to specify constant concentration boundaries). For a given stress period, each record
specifies the information on the location and prescribed concentration value of a species.
PCN boundaries can be assigned to any cell in the domain including the boundary cells.
Note that the prescribed concentration boundary condition supersedes any other boundary
condition that may have inadvertently specified on the same node.
1. Data: MXPCONC IPCNCC

Format: I10 I10
2. Data: ITMP
Format: I10
3 Data: Layer Row Column Conc Species

Format: I10 I10 I10 F10.0 I10

MXPCONC -- is the maximum number of prescribed concentration entries in the
simulation.

IPCNCC -- is a flag and a unit number.
If IPCNCC > 0, it is the unit number on which mass-flux-terms of
each species at prescribed concentration cells will
be recorded whenever ICCCFL (see the output
control [OC] package input) is set.
If IPCNCC = 0, mass-flux-terms of species will not be printed or
recorded.
If IPCNCC < 0, mass-flux-terms at prescribed concentration cells
will be printed whenever ICCCFL is set.
If ITMP < 0, prescribed concentration data from the last stress
If ITMP $ 0, ITMP will be the number of records (used in
item 3) for the current stress period.
Layer -- is the layer number of the model cell that contains the prescribed
concentration boundary.
Row -- is the row number of the model cell that contains the prescribed
Column -- is the column number of the model cell that contains the prescribed
Conc -- is the value of prescribed concentration for the model cell.
Species -- is the identification number of the species for which concentration is
prescribed in this record. 1 # Species # NSPECI, where NSPECI is
the total number of species in the transport simulation.
3.4.6 Well (WEL) Package Input
Input for the Well (WEL) Package is read from the unit specified in IUNIT(2), as
requested by the flow simulation. Note that the extra input variables required for
transport are shown in shaded boxes.
FOR EACH SIMULATION
1. Data: MXWELL IWELCB IWELCC

Format: I10 I10 I10
2. Data: ITMP
Format: I10
3. Data: Layer Row Column Q Conc(I), I=1,NSPECI

Format: I10 I10 I10 F10.0 5F10.0
(Input item 3 consists of one record for each well. If ITMP is negative or zero,
item 3 is not read.)

MXWELL -- is the maximum number of wells used at any time.
IWELCB -- is a flag and a unit number for flow simulation.
If IWELCB > 0, it is the unit number on which cell-by-cell flow
If IWELCB = 0, cell-by-cell flow terms will not be printed or
recorded.
If IWELCB < 0, well recharge will be printed whenever
ICBCFL is set.
IWELCC -- is a flag and a unit number for transport simulation.

If IWELCC > 0, it is the unit number on which mass-flux-terms of
species at well cells will be recorded whenever
ICCCFL (see the Output Control [OC] package
input) is set.
If IWELCC = 0, mass-flux-terms of species will not be printed or
recorded.
If IWELCB < 0, mass-flux-terms at well cells will be printed
whenever ICCCFL is set.

If ITMP < 0, well data from the last stress period will be
reused.
If ITMP $ 0, ITMP will be the number of wells active
during the current stress period.
Layer -- is the layer number of the model cell that contains the well.
Q -- is the volumetric recharge rate. A positive value indicates recharge
and a negative value indicates discharge.
Conc(I) - is the concentration of species I. The input value is used only if the
cell represents an injection well (i.e., when Q is positive). Note that
NSPECI is the total number of species in the transport simulation.
3.4.7 Drain (DRN) Package Input

Input to the Drain (DRN) Package is read from the unit specified in IUNIT(3), as
requested by the flow simulation. Note that the extra input variable required for transport
is shown in a shaded box. Since the drains represent outflow boundaries, concentration
data at the drain cells is not required.
1. Data: MXDRN IDRNCB IDRNCC

Format: I10 I10 I10

2. Data: ITMP
Format: I10
3. Data: Layer Row Column Elevation Cond

Format: I10 I10 I10 F10.0 F10.0
(Input item 3 normally consists of one record for each drain. If ITMP is negative or zero,
item 3 will not be read.)
MXDRN -- is the maximum number of drain cells active at one time.
IDRNCB -- is a flag and a unit number for flow simulation.
If IDRNCB > 0, it is the unit number on which cell-by-cell flow
terms will be recorded whenever ICBCFL (see the
Output Control [OC] package input) is set.
If IDRNCB = 0, cell-by-cell flow terms will not be printed or
recorded.
If IDRNCB < 0, drain leakage for each cell will be printed
IDRNCC -- is a flag and a unit number for transport simulation.

If IDRNCC > 0, it is the unit number on which mass-flux-terms
of species at drain cells will be recorded
whenever ICCCFL (see Output Control (OC)
package input) is set.
If IDRNCC = 0, mass-flux-terms will not be printed or
recorded.
If IDRNCC < 0, mass-flux-terms of species at each drain cell

If ITMP < 0, drain data from the last stress period will be reused.
If ITMP $ 0, ITMP will be the number of drains active during the
Layer -- is the layer number of the cell containing the drain.
Row -- is the row number of the cell containing the drain.
Column -- is the column number of the cell containing the drain.
Elevation -- is elevation of the drain.
Cond -- is the hydraulic conductance of the interface between the aquifer and the
drain.

3.4.8 River (RIV) Package Input
Input to the River (RIV) Package is read from the unit specified in IUNIT(4), as
requested by the flow simulation. Note that the extra input variables required for transport
are shown in shaded boxes.
1. Data: MXRIVR IRIVCB IRIVCC

Format: I10 I10 I10
2. Data: ITMP
Format: I10
3. Data: LAYER ROW Column Stage Cond Rbot [Conc(I),I=1,NSPECI]

Format: I10 I10 I10 F10.0 F10.0 F10.0 5F10.0
(Input item 3 normally consists of one record for each river reach. If ITMP is negative
or zero, item 3 is not read.)
MXRIVR -- is the maximum number of river reaches active at one time.
IRIVCB -- is a flag and a unit number for flow simulation.
If IRIVCB > 0, it is the unit number on which cell-by-cell flow
If IRIVCB = 0, cell-by-cell flow terms will not be printed or
recorded.
If IRIVCB < 0, river leakage for each reach will be printed
IRIVCC -- is a flag and a unit number for transport simulation.

If IRIVCC > 0, it is the unit number on which mass-flux-terms of
species at river reaches will be recorded
whenever ICCCFL is set.
If IRIVCC = 0, mass-flux-terms will not be printed or recorded.
If IRIVCC < 0, mass-flux-terms of species at each river reach

If ITMP < 0, river data from the last stress period will be reused.
If ITMP $ 0, ITMP will be the number of reaches active during the
Layer -- is the layer number of the cell containing the river reach.
Row -- is the row number of the cell containing the river reach.

Column -- is the column number of the cell containing the river reach.
Stage -- is the head in the river.
Cond -- is the riverbed hydraulic conductance.
Rbot -- is the elevation of the bottom of the riverbed.
Conc(I) -- is the concentration of species I. Note that NSPECI is the total
3.4.9 General-Head Boundary (GHB) Package Input

Input for the GHB Package is read from the unit specified in IUNIT(7), as
requested by the flow simulation. Note that the extra input variables required for transport
are shown in shaded boxes.
1. Data: MXBND IGHBCB IGHBCC

Format: I10 I10 I10
2. Data: ITMP
Format: I10
1. Data: MXBND IGHBCB IGHBCC

Format: I10 I10 I10
2. Data: LAYER ROW Column Boundary Head Cond [Conc(I),I=I,NSPECI]

Format: I10 I10 I10 F10.0 F10.0 5F10.0
(Input item 3 normally consists of one record for each GHB. If ITMP is negative or zero,
item 3 is not read.)
MXBND -- is the maximum number of general-head-boundary cells at one time.
IGHBCB -- is a flag and a unit number for flow simulation.
If IGHBCB > 0, it is the unit number on which cell-by-cell flow
If IGHBCB = 0, cell-by-cell flow terms will not be printed or
recorded.
If IGHBCB < 0, boundary leakage for each cell will be printed
(If IFWLCB=-1, flux-contribution of each FWL

IGHBCC - is a flag and a unit number for transport simulation.
If IGHBCC > 0, it is the unit number on which mass-flux-terms
of species at general-head boundary cells will
be recorded whenever ICCCFL (see the Output
Control [OC] package) is set.
If IGHBCC = 0, mass-flux-terms will not be printed or
recorded.
If IGHBCC < 0, mass-flux-terms of species at each general-head
boundary cell will be printed whenever
ICCCFL is set.

If ITMP < 0, GHB data from the preceding stress period will be
reused.
If ITMP $ 0, ITMP is the number of general-head boundaries during
the current stress period.
Layer -- is the layer number of the cell affected by the head-dependent boundary.
Row -- is the row number of the cell affected by the head-dependent boundary.
Column -- is the column number of the cell affected by the head-dependent boundary.
Boundary head -- is the head on the boundary.
Cond -- is the hydraulic conductance of the interface between the aquifer cell and
the boundary.
Conc(I) -- is the concentration of species I. Note that NSPECI is the total

3.4.10 Recharge-Seepage Face (RSF4) Package Input

Input to the RSF4 Package is read from the unit specified in IUNIT(8), as
requested by the flow simulation. Note that the extra input variables required for the
variably saturated flow modules of MODFLOW-SURFACT/MODHMS are shown as
shaded,
whereas those required for transport are shown in enclosed shaded boxes. Seepage face
cells correspond to outflow boundary cells and, therefore, do not require concentration
data.
1. Data: NRCHOP IRCHCB IUNCNF MXSEEP MXZRC IPNDPOR IRCHCC

H
Format: I10 I10 I10 I10 I10 I10
I10
2. Data: INRECH INIRCH INPNDEL INSEEP INCONC

Format: I10 I10 I10 I10 I10
**Enter item 3a only if INRECH $ 0.

3a. Data: RECH (NCOL, NROW)
two-dimensional array (i.e., RECH for item 3a).
**Enter item 3b only if INRECH > 0 and MXZRCH > 0. The tabulated zonal recharge
values are added to the default recharge values read above in item 3a.
3b. Data: IZNRCH (NCOL, NROW)

Utility Module: U2DINT.
**Enter item 3c only if . Item 3c consists of one two-dimensional array for

each species I. These arrays should be entered for each species, from 1 through NSPECI,
in sequential order.
3c. Data:
**Enter item 4 only if the recharge option (NRCHOP) is equal to 2 and INIRCH $ 0.
4. Data: IRCH(NCOL, NROW)

**Enter item 5 only if the unconfined option is used (IUNCNF> 0) and INPNDEL $ 0.
5. Data: PNDEL(NCOL, NROW)

**Enter item 6 only if INSEEP > 0.
6. Data: Layer Row Column Elevation

Format: I10 I10 I10 F10.0
(Input item 6 normally consists of one record for each seepage face boundary cell. If
INSEEP is negative or zero, item 6 is not read.)

NRCHOP -- is the recharge option code (flag). Recharge rates are defined in a two-
dimensional array, RECH, with one value for each vertical column.
Accordingly, recharge is applied to one cell in each vertical column, and
the option code determines which cell in the column is selected for
recharge.
1 - Recharge is only to the top grid layer.
2 - Vertical distribution of recharge is specified in array IRCH.
3 - Recharge is applied to the highest active cell in each vertical
column. A constant-head node intercepts recharge and prevents
deeper infiltration.
IRCHCB -- is a flag and a unit number for flow simulation.
If IRCHCB > 0, it is the unit number on which cell-by-cell flow
If IRCHCB # 0, cell-by-cell flow terms will not be printed or
recorded.
IUNCNF -- is the unconfined option code (flag).

If IUNCNF > 0, recharge seepage face boundary condition is
employed and recharge (RECH) will be applied until
water table reaches the ponding elevations described
in PNDEL array.
If IUNCNF # 0, the specified recharge (RECH) is applied always like
in a confined case.
MXSEEP -- is the maximum number of seepage boundary cells active at one time.
MXZRCH -- is a zonal time-series flag for input of recharge

If MXZRCH = n, recharge varies as a time series, and n is the total
number of recharge zones used in the simulation
If MXZRCH = 0, MODFLOW’s input structure for recharge is used
with variations only at stress periods.
Note 1: MXZRCH > 0 indicates that recharge varies at a different (possibly

much smaller) time scale than the stress period of MODFLOW, and that
an independent time series for recharge will be provided (in file *.RTS)
for each zone as detailed in Section 4.5.2.
Note 2: A zonal index array identifies the grid blocks which lie within a
particular zone. An index value of zero allows for the default
application of MODFLOW’s recharge for that stress period (as read into
the RECH array).
Note 3: Zonal locations for the recharges in the time series may be varied
between stress periods. Thus, the zones themselves may be varied on
a stress period basis. This is controlled by the flag INRECH, which
also controls reading of recharge rates of each stress period.
Note 4: The recharge time-series file (RTS) associated with zonal recharge input
is read as per instructions provided in Section 4.5.2 of Volume I of this
document.
IPNDOR -- is a flag for reading the ponding porosity

If IPNDPOR = 0, ponding porosity is not read and the full cell area
is assumed to be ponded.
If IPNDPOR = 1, an array of ponding porosity is read. The
ponding porosity is the fraction of the total cell
area that is wetted by lakes/ponds, and is used to
determine the total ponded storage, if ponding
occurs.
IRCHCC -- is a flag and a unit number for transport simulation.

If IRCHCC > 0, it is the unit number on which mass-flux-terms of
each species at recharge seepage boundaries will
be recorded whenever ICCCFL (see the Output
Control [OC] package) is set.
If IRCHCC # 0, mass-flux-terms will not be printed or recorded.
INRECH -- is the RECH read flag.

If INRECH $ 0, an array of recharge rates, (RECH) is read.
If INRECH < 0, recharge rates from the preceding stress period are
used.
INIRCH -- is the IRCH read flag. When NRCHOP is two,
If INIRCH $ 0, an array of layer numbers (IRCH) is read.
If INIRCH < 0, the array (IRCH) used in the preceding stress
period is reused.
Note: When NRCHOP is one or three, INIRCH is ignored.
INPNDEL -- is the PNDEL read flag, when the unconfined option is used (IUNCNF
> 0).
If INPNDEL $ 0, an array of ponding elevations (PNDEL) is read.
If INPNDEL < 0, ponding elevations from the preceding stress
period are used.
INSEEP -- is a flag and a counter of seepage boundary cells.
If INSEEP < 0, seepage face boundary data from the preceding
If INSEEP $ 0, INSEEP is the number of seepage face
boundaries during the current stress period.

requested by the flow simulation. The required input data primarily consists of well
locations, top and bottom layers in which each well is screened, bottom elevation and
radius of the wells, withdrawal (–ve) or injection (+ve) rates, and well-bore specific
storage (L–1). In addition, concentrations of injected fluid are specified for a transport
simulation, for each species. Information on each well is specified in one record in a
given stress period. The number of records in a stress period depends on the number
wells in the domain. Note that the extra input variables required for transport are shown
in a shaded boxes.
1. Data: MXFWEL IFWLCB FCONST IFWLCC

Format: I10 I10 F10.0 I10
**Enter items 2 and 3 for each stress period
2. Data: ITMP
Format: I10
3. Data: LSTART LEND Row Column Elevation Radius Q Storage Conc(I), I=1, NSPECI
Format: I10 I10 I10 I10 F10.0 F10.0 F10.0 F10.0 5F10.0
(Input item 3 consists of one record for each well. If ITMP is negative or zero, item 3 is
not read.)
MXFWEL -- is the maximum number of wells in the simulation. These include wells
that may not be active (sink or source) but still act as a vertical
IFWLCB -- is a flag and a unit number for flow simulation.
If IFWLCB > 0, it is the unit number on which cell-by-cell flow
terms for the well will be recorded whenever
ICBCFL (see Output Control) is set.
If IFWLCB = 0, cell-by-cell flow terms for the well will not be
If IFWLCB < 0, well injection rate will be printed whenever
ICBCFL is set.
(If IFWLCB=-1, flux-contribution of each FWL
IFWLCC - is a flag and a unit number for transport simulation.

If IFWLCC > 0, it is the unit number on which mass-flux-
terms of species from wells (Sink/Source)
will be recorded whenever ICCCFL (see
If IFWLCC = 0, mass-flux-terms will not be printed or
recorded.
If IFWLCC < 0, mass-flux-terms of species from wells
(Sink/Source) will be printed whenever
ICCCFL is set.
FCONST -- is the multiplying factor used to compute hydraulic conductivity of the

well given by:
KR = FCONST * (Radius)2
ρg
where FCONST =
8µ


⎡M ⎤
: = dynamic viscosity of water
⎢ LT ⎥ .
⎣ ⎦
The value of FCONST depends on the length and time units used in preparing
input data sets. The following table presents FCONST values in different length and time
units:
⎡ 1 ⎤
FCONT Value in dimensions
⎢ LT ⎥
⎣ ⎦
VALUE OF FCONST FOR VARIOUS LENGTH AND TIME UNITS
Length Time Units

Meters 1.23E06 7.36E07 4.41E09 1.06E11 3.87E13

Feet 3.74E05 2.24E07 1.35E09 3.23E10 1.18E13
table, equivalent value of FCONST must be computed using the relation given above.
If ITMP < 0, Fracture well data from the last stress period will be
reused.
abandoned but unplugged) in the current stress period.
Note: Unless either a new well is drilled or an old well is plugged, ITMP
remains constant and is equal to MXFWEL. In other words, an existing
well that is not discharging/recharging must also be physically included
in a stress period with its Q value set to zero.
screen.
Radius -- is the radius of the well in the same units as any other length units used
in the simulation. Setting the radius of the well to zero, effectively
disconnects the well fracture elements, since the fracture well conductivity
(KR in equations (1) and (2)) becomes zero. (The storage of the well also
becomes zero because of the well volume being zero). The well action of
preventing unphysical overpumping, however, is still active.
Q -- is the volumetric flux (L3 T–1) into the well. A positive value indicates
injection of water in the well and a negative value indicates discharge
from the well.

Storage -- is the well-bore specific storage (Sfs ) given by the following formula: Sfs
= (1/Ls) (rc2/rw2), for a saturated simulation assumption for the aquifer,
where L s is the total screen length, rc is the radius of standing water pipe
(or casing), and rw is the screen radius. Note that the effect of well-bore
storage is considerable for large diameter wells, and particularly at earlier
times of a simulation. Well-bore storage effects may be neglected by
setting the Storage term to zero. For unconfined simulations, the well
bore storage resulting from desaturation of the well is automatically
incorporated for the well in equation (4) the FWL4 Package (Volume I)
and Storage represents any compressibility effects of the wellbore.
Conc(I) -- is the concentration of Species I. The input value is used only if it is

an injection well (i.e., when Q is positive). Note that NSPECI is the
total number of species in the transport simulation.

requested by the flow simulation (on file with default extension WL5). Input well
properties for all wells is first provided, consisting of well locations, top and bottom layers
in which each well is screened, bottom elevation and radius of the wells, and well-bore
specific storage (L–1). All wells are set to the inactive state at the start of the simulation.
Then, for each stress period, any number of wells may be activated or subsequently de-
activated. For each well that is activated at any stress period, additional information is
sought for the pumping rate of the well (negative for withdrawal and positive for
injection), a flag to denote inclusion/reading of skin-effect conductance, and a flag
indicating input of initial head conditions in a well. Additional respective input is
requested if these flags are active. Note that initial heads in the well need to be input
whenever an inactive state is made active to prevent unphysical values (HDRY) from
expressing the pressure state of the well. The FWL5 package performs these checks and
will abort a simulation with the appropriate message in the output listing file if it detects
a well being activated without associated initial head input. In addition, concentrations of
injected fluid are specified for a transport simulation, for each species. Note that the extra
input variables required for transport are shown in shaded boxes.
1. Data: MXFWEL5 IFWL5CB FCONST5 IFWL5HD IFWL5CC

Format: I10 I10 F10.0 I10 I10
2. Data: LSTART LEND Row Column Elevation Radius Storage

Format: I10 I10 I10 I10 F10.0 F10.0 F10.0
** Enter the remaining items (3, 4, 5 and 6) for each stress period.
3. Data: ITMP
Format: I10
** Enter items 4, 5 and 6 only if ITMP >0, for each of the ITMP = NCHANGE number
of wells that are activated or de-activated.
4. Data: LWELL IBOUNDF5 SKINI INHD FQ Conc(I), I=1, NSPECI

Format: I10 I10 I10 I10 F10.0 200F10

** Enter item 5 only if SKINI < 0 (also ITMP and IBOUNDF5 for this well have to be
greater than zero).
5. Data: SKNF (K), K=LSTART, LEND

Format: 8F10.3
Note that LSTART and LEND are starting and ending layers of well LWELL and are
defined for all wells via data item 2.
** Enter item 6 only if INHD = 1 (also ITMP and IBOUNDF5 for this well have to be
greater than zero).
6. Data: HDFWL5 (K), K=LSTART, LEND

Format: 8F10.3
** Enter item 7 only if INHD = 1 and also if ITRAN ≠ 0. (also ITMP and IBOUNDF5
for this well have to be greater than zero).
7. Data: CONFWLF (K, LSPEC), K=LSTART, LEND

Format: 8F10.3
Note: Item 7 is to be repeated (for each species) for the LSPEC = 1, NSPECI contaminant
species of simulation.
MXFWEL5 -- is the maximum number of wells in the simulation. These include wells
that may not be active (sink or source) but still act as a vertical
IFWL5CB -- is a flag and a unit number for flow simulation.
If IFWLCB > 0, it is the unit number on which cell-by-cell flow
terms for the well will be recorded whenever
ICBCFL (see Output Control) is set. Default
extension of file is CW5.
If IFWL5CB = 0, cell-by-cell flow terms for the well will not be
If IFWL5CB < 0, well injection rate will be printed whenever
ICBCFL is set.
(If IFWL5CB=-1, flux-contribution of each FWL
FCONST5 -- is the multiplying factor used to compute hydraulic conductivity of the
well given by:
KR = FCONST5 * (Radius)2
ρg
where FCONST =
8µ

⎡M ⎤
: = dynamic viscosity of water
⎢ LT ⎥ .
⎣ ⎦
The value of FCONST5 depends on the length and time units used in preparing
input data sets. The following table presents FCONST5 values in different length and time
units:
⎡ 1 ⎤
FCONT5 Value in dimensions
⎢ LT ⎥
⎣ ⎦
VALUE OF FCONST5 FOR VARIOUS LENGTH AND TIME UNITS
Length Time Units

Meters 1.23E06 7.36E07 4.41E09 1.06E11 3.87E13

Feet 3.74E05 2.24E07 1.35E09 3.23E10 1.18E13
table, equivalent value of FCONST5 must be computed using the relation given above.
IFWL5HD -- is a flag and unit number for saving FWL5 heads (and concentrations for
transport) to a separate file .
If IFWL5HD>0 it is the unit number on which a separate file with
default extension FH5 is opened for saving heads (and concentrations) at
all FWL5 nodes.
If IFWL5HD = 0 the FWL5 heads (and concentrations) are saved to the
HDS (and CON) files.
If IFWL5HD < 0 Heads (and concentrations) of FWL5 nodes are not
saved.
IFWLCC - is a flag and a unit number for transport simulation.

If IFWL5CC > 0, it is the unit number on which mass-flux-
terms of species from wells (Sink/Source)
will be recorded whenever ICCCFL (see
Output Control) is set. Default extension of
file is TW5.
If IFWL5CC = 0, mass-flux-terms will not be printed or
recorded.
If IFWL5CC < 0, mass-flux-terms of species from wells
(Sink/Source) will be printed whenever
ICCCFL is set.
screen.

Radius -- is the radius of the well in the same units as any other length units used
in the simulation. Setting the radius of the well to zero, effectively
disconnects the well fracture elements, since the fracture well conductivity
(KR in equations (1) and (2)) becomes zero. (The storage of the well also
becomes zero because of the well volume being zero). The well action of
preventing unphysical overpumping, however, is still active.
Storage -- is the well-bore specific storage (Sfs) given by the following formula: Sfs
= (1/Ls) (rc2/rw2), for a saturated simulation assumption for the aquifer,
where L s is the total screen length, rc is the radius of standing water pipe
(or casing), and rw is the screen radius. Note that the effect of well-bore
storage is considerable for large diameter wells, and particularly at earlier
times of a simulation. Well-bore storage effects may be neglected by
setting the Storage term to zero. For unconfined simulations, the well
bore storage resulting from desaturation of the well is automatically
incorporated for the well in equation (4) the FWL4 Package (Volume I)
and Storage represents any compressibility effects of the wellbore.
If ITMP < 0, Fracture well data from the last stress period will be
reused.
abandoned but unplugged) that require input for the
LWELL -- is the well number (entered sequentially in identifying the wells).
IBOUNDF5 -- is the ‘IBOUND’ value of the well indicating if the well is active or
inactive for this stress period.
If IBOUNDF5 > 0, the well is active.
If IBOUNDF5 = 0, the well is inactive.
SKINI -- is an index or value of skin factor for a well.
If SKINI = 0 then all skin factors of the well are set to zero.
If SKINI > 0 then the value of SKINI is used as the skin factor for all
layers of the well.
If SKINI < 0 then the skin factors will be read separately for each layer
of the well.
INHD -- is the index for reading initial heads (or concentrations for transport) of
the well.
If INHD = 0 then do not read initial heads. The well is probably already
active with a change of pumping rate (or concentrations for transport)or
is being deactivated.
If INHD = 1 then read initial head array for all layers of this well (useful
for restart simulations).
If INHD = 2 then the initial head array (or concentrations for transport)
for this well is set to the value of the adjacent porous matrix block, for all
well layers.
If INHD = 3 then the initial head array (or concentrations for transport)
for this well is set to the value of the bottom most porous matrix block,
for all well layers.
Note: INHD should not be zero when an inactive cell is made active.
FQ -- is the volumetric flux (L3 T–1) into the well. A positive value indicates
injection of water in the well and a negative value indicates discharge
from the well.
Conc(I) -- is the concentration of Species I. The input value is used only

if it is an injection well (i.e., when Q is positive) and denotes
the concentration of injected water. Note that NSPECI is the
total number of species in the transport simulation.
SKNF(K) -- is the value of the skin factor for all layers (from LSTART to
LEND) of the well.
Note: Setting a very high value for the skin factor (1020) effectively disconnects that layer’s
porous matrix block from the well thereby simulating casing effects within the layer.
HDFWL5(K) -- is the initial head value for all layers (from LSTART to LEND)
of the well
CONFWL5
(K,LSPEC) -- is the initial concentration value for all layers (from LSTART to
LEND) of species LSPEC, of the well.
3.4.13 Preconditioned Conjugate Gradient (PCG4) Package Input
Input required by the PCG4 package is similar to that of the PCG2 package
implemented in MODFLOW by Hill (1994). The PCG4 iterative solver is a simple
solution scheme requiring only closure tolerance limits and a maximum limit of solver
iterations as input. There are no iteration parameters to be estimated, nor other choices
to be made by the user. A solver-point-control-flag and maximum outer iterations (> 1
for unconfined systems) are also requested as input, to be consistent with the other solver
packages of MODFLOW. The PCG4 Package is required for a transport simulation, even
if it is not used for the flow simulation. Thus, two solver packages may be active for flow
and transport simulations. Note that some space conservation occurs when the PCG4
Package is used for both flow and transport.
Basic Package input file, as requested by the flow simulation. The input is as follows:

Format: I10 I10 I10 F10.0 F10.0 I10
2. Data: HCLOSE MUTPCG MNORTH RCLOSE

Format: F10.3 I10 I10 F10.3

routine. For a linear problem MXITER should be 1, unless more than
50 inner iterations are required, when MXITER could be as large as 10.
A larger number (generally less than 100) is required for a nonlinear
problem.
ITERI usually ranges from 6 to 600, depending on the problem, and
may be set at 600 with closure controlled by the tolerance limits.

unsaturated cases.

criterion is not met. BFACT must be greater than zero and less
than one.

algorithm forces a residual reduction at every Newton iteration.
For RESRED greater than unity, increases in residual are allowed
by the prescribed factor to allow the Newton algorithm to move
easily out of local minima. As RESRED increases, there are less
chances for backtracking. The user should determine which value
of this parameter gives optimal convergence for the particular
problem. See the NRBI package documentation in Chapter 7 Vol.
1, for suggested usage.

between iterations.
Note: This dampening may be effective in achieving convergence
for aquifer systems with highly irregular layering geometries that
system.
Commonly, HCLOSE is set equal 0.01.
iteration summary is suppressed, if MUTPCG = 2, the solver iteration
Orthomin solver to solve the transport matrix equation. A suggested
value of MNORTH ranges from 5 to 10, but lower values may be
used to conserve array space. Smaller MNORTH may, however,
weaken the solver’s performance. MNORTH larger than 10 may help
the solver to handle difficult problems but, in general, diminishing
returns are expected for MNORTH much larger than 10.
RCLOSE -- is the relative closure criterion for convergence. This parameter is

used for convergence of inner iterations, when the relative head
change is less than the prescribed value. Note that HCLOSE is also
used to check for inner iteration convergence, and convergence is
achieved when either RCLOSE or HCLOSE values have been met
by the respective changes at any inner iteration.
Note: If RCLOSE is set to zero, its value is taken as HCLOSE x 10-3.
3.4.14 Output Control (OC) Package Input

The OC package of the original MODFLOW (McDonald and Harbaugh, 1988) is
extended to include addition output control parameters required for a transport simulation.
Input to Output Control module is read from the unit specified in IUNIT(12), as requested
by the flow simulation. Note that the Adaptive Time-Stepping and Output Control (ATO4)
Package, which is read from the unit specified in IUNIT(22), serves as an alternative to
this Output Control package. (See ATO4 Package described in the next section for
details.) If both units [i.e., IUNIT(12) and IUNIT(22)] are activated, IUNIT(12) is
ignored and the new ATO4 Package is used to control the time-stepping scheme and output
print/save options. If both IUNIT(12) and IUNIT(22) are zero, no output control data are
read, and default output control is used. Under the default, head, concentration, and total
budgets for flow and transport are printed at the end of every stress period. Additionally,
if starting heads are saved (ISTRT is not 0), drawdown is printed at the end of every stress
period. The default printout format for head, drawdown, and concentration is 10G11.4.
All printer output goes to unit 6 as specified in the main program. If necessary, the unit
number for printer output can be changed to meet the requirements of a particular
computer. Note that the extra input parameters required for transport are shown in shaded
boxes. The transport output control data are ignored if the simulation is only for flow
(ITRAN=0 in BAS Package input).
Format: I10 I10 I10 I10 I10 I10
**Enter items 2 and 3 for each time step.

Format: I10 I10 I10 I10 I10 I10 I10
**Enter item 3 only if INCODE $0.

Format: I10 I10 I10 I10 I10 I10
(Record 3 is read 0,1, or NLAY times, depending on the value of INCODE.)

IHEDFM -- is a code for the format in which heads will be printed.
IDDNFM -- is a code for the format in which drawdowns will be printed.
Format codes have the same meaning for head, drawdown, and
concentration. A positive format code indicates that each row of
data is printed completely before starting the next row. This means
that when there are more columns in a row than will fit on one line,
additional lines are used as required to complete the row. This
format is called the wrap format. A negative format code indicates
that the printout is broken into strips where only that number of
columns that will fit across one line are printed in a strip. As many
strips are used as are required to print the entire model width. This
format is called the strip format. The absolute value of the format
code specifies the printout format as follows.
0 - (10G11.4) 5 - (15F7.3) 10 - (20F5.3)
1 - (11G10.3) 6 - (15F7.4) 11 - (20F5.4)
2 - (9G13.6) 7 - (20F5.0) 12 - (10G11.4)
3 - (15F7.1) 8 - (20F5.1)
4 - (15F7.2) 9 - (20F5.2)
Note that if IREALSL … 0, the drawdown is a meaningless parameter and the switch
IDDNFM is used to indicate format for printout of saturations of the active phase.
IHEDUN -- is the unit number to which heads will be written if they are saved
on disk.
IDDNUN -- is the unit number to which drawdowns will be written if they are
saved on disk.
Note that if IREALSL … 0, the drawdown is a meaningless parameter and the switch
IDDNUN is used to indicate unit number for printout of saturations of the active
phase.
ICONFM - is a code for the format in which concentrations are to be printed.

- The format codes are listed above (see the explanation of
IDDNFM).
ICONUN - is the unit number to which concentrations of each species are to

- be written if they are saved on disk.
INCODE -- is the head/drawdown/concentration output code. It determines the

number of records in input item 3.
If INCODE < 0, layer-by-layer specifications from the last
time steps are used. Input item 3 is not read.
If INCODE = 0, all layers are treated the same way. Input
item 3 will consist of one record.
If INCODE > 0, input item 3 will consist of one record for
each layer.

IHDDFL -- is a head and drawdown output flag.
If IHDDFL = 0, neither heads nor drawdowns (or saturations)
will be printed or saved on disk.
If IHDDFL … 0, heads and drawdowns (or saturations) will be

printed or saved according to the flags for
each layer specified in input item 3.
If IBUDFL … 0, overall volumetric budget will be printed.
(Note that the overall volumetric budget will always be printed at

the end of a stress period, even if the value of IBUDFL is zero.)
ICBCFL -- is a cell-by-cell flow-term flag for flow simulation.
If ICBCFL = 0, cell-by-cell flow terms are not saved or
printed.
If ICBCFL … 0, cell-by-cell flow terms are printed or recorded on

disk depending on flags set in the component of flow packages, i.e.,
IWELCB, IRCHCB, etc.
ICONFL -- is a concentration output flag.

If ICONFL = 0, concentrations are not printed or saved on
disk.
If ICONFL … 0, concentrations of each species are printed

or saved according to the flags for each
layer specified in input item 3.
ICONBD -- is a budget print flag for transport simulation.

If ICONBD … 0, transport budget is printed

ICCCFL -- is a cell-by-cell mass-flux-term flag for transport simulation.

printed.

etc.


Ddpr -- is the output flag for drawdown (or saturation) printout.

If Ddpr = 0, drawdown (or saturation) is not printed for the
corresponding layer.
If Ddpr … 0, drawdown (or saturation) is printed for the

Ddsv -- is the output flag for drawdown (or saturation) save.

If Ddsv = 0, drawdown (or saturation) is not saved for the
If Ddsv … 0, drawdown (or saturation) is saved for the

If Conpr = 0, concentration is not printed for the

If Conpr … 0, concentration is printed for the corresponding

layer.

If Consv = 0, concentration is not saved for the corresponding
layer.
If Consv … 0, concentration is saved for the corresponding

layer.
3.4.15 Adaptive Time-Stepping and Output Control (ATO4) Package

Input to the ATO4 Package is read from the unit specified in IUNIT (22), as
requested by the flow simulation. The original output control option, which is
specified in IUNIT (12), is ignored if the new ATO4 Package is activated. If both
IUNIT (12) and IUNIT (22) are zero, no output control data is read, and default
output control is used. Under the default, head, concentration, and total budgets for
flow and transport are printed at the end of every stress period. Additionally, if
starting heads are saved (ISTRT not 0), drawdown is printed at the end of every stress
period. Similar to the output control (OC) package, all printer output goes to unit 6
as specified in the main program. If necessary, the unit number for printer output can
be changed to meet the requirements of a particular computer. Note that the extra
input parameters required for transport are shown in shaded boxes.
The ATO package further includes flags and parameters to restart (or continue)
a simulation from a non-zero time value. This capability is active when the variable
TCNTNU is non-zero and represents the time value (in simulation units of time) from
which the simulation is restart. The first time-step size value of the restart simulation

is also then input in variable DTCNTNU (in simulation units of time). When this is
the case, the simulation will wind itself up to time TCNTNU (appropriately winding
stress period information on boundary conditions and time values of the recharge or
ET time-series files) and proceed from that point onwards with the first time-step size
being DTCNTNU. Initial heads (and concentrations for transport) as well as values
of initial storage in interception in the IPT package and initial values of the ON/OFF
switch of regulated structures in the CHF package (for MODHMS) are obtained from
the respective initial conditions as input for the simulation. Further flags are available
for direct input of starting heads from the binary HDS file (and/or starting
concentrations from the CON file) of a previous MODHMS simulation for the same
problem. When the flag (and unit number) ICNTNU is set to a positive number, the
file containing starting heads (all binary head outputs of a previous run) is opened on
the specified unit number (and with default extension of HDC—which may be obtained
by renaming the previous simulation’s HDS file to extension HDC) and the file is
wound up till the specified time (TCNTNU) is reached. The head values (in all
simulated domains) available in the file at this time value (or greater, if data is not
available for time TCNTNU) are used for starting the current restart simulation.
Similarly, when the flag and unit number, ICNTNUC is set to a positive value, and
when a transport simulation is performed, the file containing starting concentrations
is opened on the specified unit number (and containing default extension of
COC—which may be obtained by renaming the previous simulation’s CON file to
extension COC). This file is then wound up to the time value of TCNTNU (or
greater) to obtain initial concentrations for all simulation domains (which include
immobile and mobile domains in the subsurface for dual porosity cases, and, the OLF
surface and the CHF domain as appropriate) and all species of the restart simulation.
Format: I10 I10 I10 I10 I10 I10
TCNTNU DTCNTNU ICNTNU ICNTNUC IUNITMB

F10.3 F10.3 I10 I10 I10
Format: I10 I10 I10 I10 I10 I10 I10
3. Data: DELT TMIN TMAX TSMULT TSDIV NPRTS NPSTP
Format: F10.0 F10.0 F10.0 F10.0 F10.0 I10 I10
**Enter item 4 only if NPRTS > 0.

4. Data: TMVEC(IPR), IPR = 1,NPRTS
Format: 8F10.0 (number of records = NPRTS/8 + 0 or 1)
**Enter item 5 only if INCODE $ 0.
Format: I10 I10 I10 I10 I10 I10
(Record 5 is read 0, 1, or NLAY times, depending on the value of INCODE.)

IHEDFM -- is a code for the format in which heads will be printed.
IDDNFM -- is a code for the format in which drawdowns will be printed. Format
codes have the same meaning for head, drawdown, and concentration.
A positive format code indicates that each row of data is printed
completely before starting the next row. This means that when there
are more columns in a row than will fit on one line, additional lines
are used as required to complete the row. This format is called the
wrap format. A negative format code indicates that the printout is
broken into strips where only that number of columns that will fit
across one line are printed in a strip. As many strips are used as are
required to print the entire model width. This format is called the
strip format. The absolute value of the format code specifies the
printout format as follows:
0 - (10G11.4) 5 - (15F7.3) 10- (20F5.3)
1 - (11G10.3) 6 - (15F7.4) 11- (20F5.4)
2 - (9G13.6) 7- (20F5.0) 12- (10G11.4)
3 - (15F7.1) 8- (20F5.1)
4 - (15F7.2) 9- (20F5.2)
Note that if IREALSL … 0, the drawdown is a meaningless

parameter and the switch IDDNFM is used to indicate format for
printout of saturations of the active phase.
IHEDUN -- is the unit number to which heads will be written if they are saved
on disk.
IDDNUN -- is the unit number to which drawdowns will be written if they are
saved on disk.

parameter and the switch IDDNUN is used to indicate unit number
for printout of saturations of the active phase.
ICONFM -- is a code for the format in which concentrations will be printed.

The format codes are listed above (see the explanation of
IDDNFM).
ICONUN -- is the unit number to which concentrations of each species will

be written if they are saved on disk.
TCNTNU – Flag and time value for continuation of a previous simulation. If
TCNTNU = 0, the simulation is NOT a continuation of a previous
simulation. If TCNTNU > 0, the simulation continues from a
previous simulation, from the time value provided for this input.
DTCNTNU - time-step size value used for the continuation simulation from time
value of TCNTNU. If the simulation is not a continuation simulation,
this variable is not required/used.

ICNTNU – Flag and unit number for reading restart heads from an external file
If ICNTNU = 0, do not look for restart heads in a file created by a
previous simulation.
If ICNTNU > 0, it is the Fortram unit number from which to read
heads for the restart simulation (on file with extension HDC). The
file will be wound to the appropriate time value (TCNTNU) to obtain
the required starting heads.
ICNTNUC – Flag and unit number for reading restart concentrations from an
external file created by a previous simulation (when a transport
simulation is performed).
If ICNTNUC = 0, do not look for restart concentrations in a file
If ICNTNUC > 0, it is the Fortram unit number from which to read
concentrations for the restart simulation (on file with extension COC).
The file will be wound to the appropriate time value (TCNTNU) to
obtain the required starting concentrations.
IUNITMB -- Flag and unit number for saving nodal mass balance terms to a binary
file.
If IUNITMB # 0 Node-by-node flux balances are not saved.
If IUNITMB > 0 Node-by-node flux balances are saved to a binary
file with default extension NFB and the value of IUNITMB is the
Fortran Unit number on which the file is written.
INCODE -- is the head/drawdown/concentration output code. It determines the
number of records in input item 5.
If INCODE < 0, layer-by-layer specifications from the last time
steps are used. Input item 5 is not read.
If INCODE = 0, all layers are treated the same way. Input item 5
will consist of one record.
If INCODE > 0, input item 5 will consist of one record for each
layer.
IHDDFL -- is a head and drawdown output flag.
If IHDDFL = 0, neither heads nor drawdowns (or saturations) will
be printed or saved on disk.
If IHDDFL … 0, head and drawdowns (or saturations) will be

printed or saved according to the flags for each
layer specified in input item 5.
If IBUDFL < 0, overall volumetric budget will be printed every
“n” timesteps where n=abs (IBUDFL)

If IBUDFL > 0, overall volumetric budget will be printed as
specified by either NPRT or NPSTP in Item 5
input.
(Note that the overall volumetric budget will always be printed at the
end of a stress period, even if the value of IBUDFL is zero.)
ICBCFL -- is a cell-by-cell flow-term flag for flow simulation.
If ICBCFL = 0, cell-by-cell flow terms are not printed or saved.
If ICBCFL … 0, cell-by cell flow terms are printed or recorded on

disk depending on flags set in the component of
flow packages, i.e., IWELCB, IRCHCB, etc.
ICONFL -- is a concentration output flag simulation.

If ICONFL = 0, concentrations will NOT be printed or saved
on disk.
If ICONFL … 0, concentrations of each species will be

printed or saved according to the flags for
each layer specified in input item 5.
ICONBD -- is a budget print flag for transport simulation.

If ICONBD < 0, transport budget is printed every ‘n’
timesteps where n = abs (ICONBD)
If ICONBD > 0, transport budget is printed as specified by
either NPRTS or NPSTP in Item 5 input.
ICCCFL -- is a cell-by-cell mass-flux-term flag for transport simulation.

printed.

etc.
DELT -- is the first time step size of the stress period.
TMIN -- is the minimum value of time step size allowed in the stress period.
TMAX -- is the maximum value of time step size allowed in the stress period.
TSMULT -- is the time step multiplier. TSMULT should NOT be less than 1.0.
This value overwrites the TSMULT value read in item 9 of Basic
Package Input.
TSDIV -- is the reduction factor of time step size. TSDIV must be greater than
1.0.

NPRTS -- is the number of print times specified in item 4. If no print time is
specified (i.e., NPRTS=0), overall budgets for flow and transport
and heads/drawdowns/ concentrations will be printed or saved at the
end of the stress period according to the flags specified in input items
2 and 5. If NPRTS=0, item 4 must be omitted.
NPSTP -- is an output printout control parameter. NPSTP=n means that the
overall budgets for flow and transport and
heads/drawdowns/concentrations will be printed or saved at every nth
time step of the stress period according to the flags specified in input
items 2 and 5. Set NPSTP equal to zero if the step wise
printing/saving is NOT required.
TMVEC -- is a one-dimensional array containing NPRTS number of print time
values at which printout is desired in a particular stress period. Time
is assumed to be zero at the beginning of a stress period. The print
times must be specified in ascending order of magnitude and the
difference between two successive values must be at least TMIN.
Note: This item must be omitted if NPRTS=0.

Ddpr -- is the output flag for drawdown printout.
If Ddpr = 0, drawdown is not printed for the corresponding layer.
If Ddpr … 0, drawdown is printed for the corresponding layer.

parameter and the switch Ddpr is used as an output flag for
printout of saturations of the active phase.

Ddsv -- is the output flag for drawdown save.
If Ddsv = 0, drawdown is not saved for the corresponding layer.
If Ddsv … 0, drawdown is saved for the corresponding layer.

parameter and the switch Ddsv is used as a flag for saving of
saturations of the active phase.

If Conpr = 0, concentration is not printed for the corresponding
layer.
If Conpr … 0, concentration is printed for the corresponding layer.

If Consv = 0, concentration is not saved for the corresponding
layer.
If Consv … 0, concentration is saved for the corresponding layer.
3.4.16 Evapotranspiration (EVT) Package Input

Input to the Evapotranspiration (EVT) Package is read from the unit specified
in IUNIT (5). For contaminant transport simulations, the default (MODFLOW) data
sets allow for water to leave the system, but contaminant mass remains behind. An
option is provided with the flag IFRACET whereby the ET process also removes
component species mass. When IFRACET is set to one, additional input is requested
for each species (I) of simulation, to provide the root uptake factor, ETFRAC(I),
defined as the fraction of mass of component species I that is removed with the ET
water. Thus, for an ET rate of q, with component species concentration of Ci, the
mass removal rate is equal to qCi ETFRAC(I).
FOR EACH SIMULATION
1. Data: NEVTOP IEVTCB MXZETS2 IFRACET IEVTCC
Format: I10 I10 I10 I10 I10
**Enter item 2 only if IFRACET=1

2. Data: ETFRAC(I), I=1, NSPECI
Format: 200 F10.0

EVT1RP
3. Data: INSURF INEVTR INEXDP INIEVT
**Enter item 4 only if INSURF $ 0
4. Data: SURF
Module: U2DREL
**Enter item 5a only if INEVTR$0 and MXZETS2=0
5a. Data: EVTR
Module: U2DREL
**Enter item 5b only if INEVTR$0 and MXZETS2…0

5b. Data: IZNETS2
Module: U2DREL
Note that either item 5a or 5b is read depending on whether a time-series for maximum ET flux
(EVTR) is requested or not

**Enter item 6 only if INEXDP$0
6. Data: EXDP
Module: U2DREL
**Enter item 7 only if NEVTOP=2 and INIEVT$0
7. Data: IEVT
Module: U2DINT

NEVTOP -- is the evapotranspiration (ET) option code. ET parameters (ET
surface, maximum ET rate, and extinction depth) are specified
in two-dimensional arrays, SURF, EVTR, and EXDP, with one
value for each vertical column. Accordingly, ET is calculated
for one cell in each vertical column. The option codes
determine for which cell in the column ET will be calculated.
1 - ET is calculated only for cells in the top grid layer.
2 - The cell for each vertical column is specified by the
user in array IEVT.
3 - ET is calculated for highest active cell.
IEVTCB -- is a flag and a unit number.
If IEVTCB > 0, it is the unit number on which cell-by-cell
flow terms will be recorded whenever
ICBCFL (see Output Control) is set.
If IEVTCB # 0, cell-by-cell flow terms will not be printed
or recorded.
MXZETS2 -- Index for input of zonal time-series of the maximum ET flux
(EVTR) via a separate ETS file.
= 0 if no, maximum ET flux (EVTR) is read as per stress
periods.
= n if yes, a separate file is used to read time-series of
maximum ET flux (EVTR), and n is the total number of ET
zones used in the simulation.
Note: The ET time-series (ETS) associated with zonal input of
ETmax is read as per instructions provided in Appendix G of
Volume I of this document.
IFRACET -- is a flag (used for transport simulations) indicating if ‘uptake
factors’ for the chemical components are to be read.
If IFRACET = 0, uptake factors are not read, and ET
removes water without removal of
chemical mass.
If IFRACET = 1, uptake factors are read to allow chemical
component mass removal with ET.

IEVTCC -- is a flag and unit number for output of ET mass fluxes. If
IEVTCC > 0, it is the unit number on which the cell-by-cell
mass flux terms will be recorded whenever ICCCFL (see ATO)
is set.
If IEVTCC #0, cell-by-cell mass flux terms for ET will not

be recorded.
ETFRAC(I) -- is the root uptake factor for chemical species I. This is defined
as the fraction of chemical component I, that can leave the
system with ET flux. For example when ETFRAC(I) = 0.0,
ET removes water but leaves behind species I. If ETFRAC(I)
= 1.0, all of component I present, can leave with the calculated
ET flux.
INSURF -- is the ET surface (SURF) read flag.
If INSURF $ 0, an array containing the ET surface

elevation will be read.
If INSURF < 0, the ET surface from the preceding stress
INEVTR -- is the maximum ET rate (EVTR) read flag.
If INEVTR $ 0, an array containing the maximum ET rate

will be read.
IF INEVTR < 0, the maximum ET rate from the preceding
IZNETS2 -- is the integer array defining the zone value of each node, for
zonal input of the ET time series.
Note that the applied ET will be zero in cells with a zone value
of zero.
INEXDP -- is the extinction depth (EXDP) read flag.
If INEXDP $ 0, an array containing the extinction depth

(EXDP) will be read.
If INEXDP < 0, the extinction depth from the preceding
INIEVT -- is the layer indicator (IEVT) read flag. It is used only if the ET
option (NEVTOP) is equal to two.
If INIEVT $ 0, an array containing the layer indicators

(IEVT) will be read.
If INIEVT < 0, layer indicators used during the preceding
SURF -- is the elevation of the ET surface.
EVTR -- is the maximum ET rate (volume of water per unit area (Lt-1)).
EXDP -- is the ET extinction depth.

IEVT -- is the layer indicator array. For each horizontal location, it
indicates the layer from which ET is removed. It is needed
only if the ET option is equal to two.
3.4.17 Evapotranspiration Segments (ET2) Package
Input to the ET2 Package is read from the file that is type “ET2” in the name
file. All single-valued variables are free format if the option “FREE” is specified in
the Basic Package input file; otherwise, the variables have 10-character fields. Arrays
are read by an array-reading utility module, either U2DREL or U2DINT (Harbaugh
and others, 2000), as indicated. Optional variables are shown in brackets.
FOR EACH SIMULATION
0. #Text
Item 0 is optional - “#” must be in column 1. Item 0 can be repeated as many
times a desired.
1. NETSOP IETSCB NPETS NETSEG MXZETS1 IFRACETS IETSCC
2. INETSS INETSR INETSX [INIETS [INSGDF] ]
3. ETSS (NCOL, NROW) - U2DREL - If INETSS $0
4a. ESTR (NCOL, NROW) - U2DREL - If INETSR $0 and MXZETS1=0
4b. IZNETS1 (NCOL, NROW) - U2DREL - If INETSR $0 and MXZETS1=0.
Note that either item 4a or 4b is read depending on whether a time series for
maximum ET flux (ETSR) is requested or not.
5. ETSX (NCOL, NROW) - U2DREL - If INETSX $0
6. IETS (NCOL, NROW) - U2DINT - If NETSOP = 2 and if INIETS $0
7. PXDP (NCOL, NROW) - U2DREL If NETSEG > 1 and INSGDF $0

8. PETM (NCOL, NROW) - U2DREL If NETSEG > 1 and INSGDF $0
If NETSEG >1, (NETSEG - 1) repetitions of Items 7 and 8 are read. If NETSEG

>2, Items 7 and 8 are read for the uppermost segment intersection, followed by
repetitions of Items 7 and 8 for successfully lower intersections.
Explanation of Variables Read by the Evapotranspiration Segments Package
Text - is a character variable (79 characters) that starts in column 2. Any
characters can be included in Text. The “#” character must be in
column 1. Lines beginning with “#” are restricted to these first lines
of the input file. Text is written to the LIST output file when the
input file is read and provides an opportunity for the user to include
information about the model both in the input file and the associated
output file.
NETSOP - is the evapotranspiration (ET) option code. ET variables (ET surface,
maximum ET rate, and extinction depth) are specified in layer

variables, ETSS, ETSR, and ETSX, with one value for each vertical
column of cells in the model grid. Accordingly, ET is calculated for
one cell in each vertical column. The option codes determine the cell
within a column for which ET will be calculated. If NETSOP = 1,
ET is calculated only for cells in the top grid layer.
If NETSOP - 2, the cell for each vertical column is specified by the
user in variable IETS.
If NETSOP = 3, ET is calculated for the top most active grid block.
IETSCB - is a flag and unit number.
If IETSCB > 0, it is the unit number to which ETS1-Package cell-by-
cell flow terms will be written when “SAVE BUDGET” or a non-
zero value for ICBCFL is specified in Output Control (Harbaugh and
others, 2000). IETSCB must be a unit number associated with a file
listed with type “DATA (BINARY)” or “DATAGLO (BINARY)” in
the name file.
If IETSCB #0, ETS1-Package cell-by-cell flow terms will not be
written.
NPETS - is the number of evapotranspiration-segments parameters. No used.
Set equal to zero.
NETSEG - is the number of segments used to define the relation of
evapotranspiration rate to hydraulic head in the interval where the
evapotranspiration rate is variable.
MXZETS - Index for input of zonal time-series of the maximum ET flux

(ETSR) via a separate ETS file
= 0 if no, maximum ET flux (ETSR) is read as per stress-
periods.
= n if yes, a separate file is used to read time-series of maximum
ET flux (ETSR), and n is the total number of ET zones used in
the simulation.
Note: The ET time-series file (ETS) associated with zonal input of
ETmax is read as per instructions provided in Appendix G of
Volume I of the document.
IFRACETS - is a flag (used for transport simulations) indicating if ‘uptake

factors’ for the chemical components are to be read.
If IFRACETS = 0, uptake factors are not read and ET removes
water without removal of chemical mass.
If IFRACETS = 1, uptake factors are read to allow chemical
component mass removal with ET.
IETSCC - is a flag and unit number for output of ET mass fluxes.

If IETSCC > 0, it is the unit number on which the cell-by-
cell mass flux terms will be recorded
whenever ICCCFL (see ATO) is set.

If IETSCC # 0, cell-by-cell mass flux terms for ET will
not be recorded.
INETSS - is the ET surface (ETSS) read flag.
If INETSS $ 0, a layer variable containing the ET surface (ETSS)

If INETSS < 0, the ET surface from the preceding stress period will
be reused.
INETSR - is the ESTR read flag. Its function depends on whether or not
parameters are being used.
If INETSR $0, a layer variable containing the maximum ET rate

(ETSR) will be read from Item 6 in the ETS1 input file.
If INETSR <0, the maximum ET rate from the preceding stress
IZNETS1 - Integer array defining the zone value of each node, for zonal input of
the ET time series.
Note that the applied ET will be zero in cells with a zone value of
zero.
INETSX - is the extinction depth (ETSX) read flag.
If INETSX $0, a layer variable containing the extinction depth

(ETSX) will be read from Item 8 in the ETS1 input file.
If INETSX <0, the extinction depth from the preceding stress period
will be reused.
INIETS - is the layer indicator (IETS) read flag. It is read if the option
(NETSOP) is equal to two or if NETSEG >1. If NETSEG >1 and
NETSOP is not equal to two, INIETS is ignored and IETS is not
read.
If INIETS $0, a layer variable containing the layer indicator (IETS)

If INIETS < 0, layer indicators used during the preceding stress
INSGDF - is the segment definition read flag. It is read only if NETSGEG >1.
If INSGDF $0, two layer variables to define PXDP and PETM for
each fo (NETSEG - 1) segment intersections are read from Items 10
and 11, respectively, of the ETS1 input file.
If INSGDF <0, PXDP and PETM from the preceding stress period
will be reused.
ETSS - is the elevation of the ET surface (L).
ETSR - is the maximum ET flux (volumetric flow rate per unit area, LT1)
Pname - is the name of a parameter that will be used to define the ETSR the
ETSR variable in the current stress period.

Iname - is an instance name that is ready only if Pname is a time-varying
parameter. Multiple instances of the same time-varying parameter are
not allowed in a stress period.
IETSPF - is an optional format code for printing the ETSR variable after it has
been defined by parameters. The format codes are the same as those
used in the U2DREL array reading utility module (Harbaugh and
others, 2000).
ETSX - is the ET extinction depth (L). This variable is read only if INETSX
$0.
IETS - is the layer indicator variable. For each horizontal location, IETS
indicates the layer from which ET is removed. It is read only if the
ET option (NETSOP) is equal to two and if INIETS $0.
PXDP - is a proportion of the extinction depth (dimensionless), measured
downward from the ET surface, which, with PETM, defines the
shape of the relation between the evapotranspiration rate and head.
The value of PXDP must be between 0.0 and 1.0, inclusive.
Repetitions of PXDP and PETM are read in sequence such that the
first occurrence represents the bottom of the first segment, and
subsequent repetitions represent the bottom of successively lower
segments. Accordingly, PXDP values for later repetitions
(representing lower segments) should be greater than PXDP values
for earlier repetitions.
PETM - is a proportion of the maximum evapotranspiration rate
(dimensionless) which, with PXDP, defines the shape of the relation
between the evapotranspiration rate and head. The value of PETM
should be between 0.0 and 1.0, inclusive. Repetitions of PXDP and
first segment, and subsequent repetitions represent the bottoms of
successively lower segments. Accordingly, PETM values for later
repetitions (representing lower segments) generally would be less than
PETM values for earlier repetitions.
3.4.18 Input Instructions for FHB Package
Input instructions for the FHB package are an extension to input instructions
for the FHB package specified by Leake and Lilly (1997). Input for FHB is read
from IUNIT(30), specified in the Basic Package input (McDonald and Harbaugh,
1988, chap. 4, p. 9–11). All input is free format, which requires each of the numbered
data groups to start on a new input record. More than one record can be used for any
data group and numbers within data groups must be separated by at least one space or
a comma. Integer data types cannot include a decimal point. Blank spaces are not
treated as zeros. The FHB package has been extended to include a step function
capability in addition to interpolation of the prescribed time-series of head or flux at
any node. For a transport simulation, the FHB package further includes possible
temporal variations of concentrations of species at the flow and head boundary nodes.
For flow boundary nodes, the parameter IAUX of each node determines if the
concentrations of species are input or whether a zero value is applied by default. For
head boundary nodes, the parameter IAUX further indicates if concentrations from the
HCN package are to be used.

FOR EACH SIMULATION
1. Data: NBDTIM NGWFLW NGWHED IFHBSS IFHBCB NFHBX1 NFHBX2
Type: Integer Integer Integer Integer Integer Integer Integer
NOLFLW NOLHED NCHFLW NCHHED ITHTFHB IFHBCC

Integer Integer Integer Integer Integer Integer
Note that the shaded items, NOLFLW, NOLHED, NCHFLW and NCHHED are used
only with MODHMS and may be set to zero for a SURFACT simulation.
Omit data item 2 if NFHBX1=0. Input item 2 consists of one record for each of
NFHBX1 auxiliary variables.
Omit data item 3 if NFHBX2=0. Input item 3 consists of one record for each of
NFHBX2 auxiliary variables.
Data items 4a and 4b are required for all simulations. Include NBDTIM times in data
item 4b.
4b. Data: BDTIM(NBDTIM)
Type: Real
Omit data item 5a if all NGWFLW, NOLFLW and NCHFLW=0.
Omit data item 5b if NGWFLW=0. Input item 5b consists of one set of numbers for
each of NGWFLW cells. Each set of numbers includes layer, row and column indices,
an integer auxiliary variable, and NBDTIM values of specified flow.
5b. Data: Layer Row Column IAUX FLWRAT(NBDTIM)
Omit data item 5bii if ITRAN = 0 or IAUX = 0
5bii. Data: FBCONC (NSPECI, NBDTIM)
Type: Real
Note that data item 5bii consists of one entry for each of the NBDTIM interpolation
time values for a species. Data item 5bii is repeated for each species associated with
transport. Also note that if IAUX (of data item 5b) = 0, the concentration of all
Omit data item 5c if NOLFLW=0. Input item 5c consists of one set of numbers for each
of NOLFLW cells. Each set of numbers includes overland row and column indices, an
5c. Data: Row Column IAUX FLWRAT(NBDTIM)


Omit data item 5cii if ITRAN = 0 or IAUX = 0
5cii. Data: FBCONC (NSPECI, NBDTIM)
Type: Real
Note that data item 5cii consists of one entry for each of the NBDTIM
interpolation time values for a species. Data item 5cii is repeated for each
species associated with transport. Also note that if IAUX (of data item 5c) = 0,
the concentration of all species is taken as zero for this node.
Omit data item 5d if NCHFLW=0. Input item 5d consists of one set of numbers
for each of NCHFLW cells. Each set of numbers includes channel segment
number, an integer auxiliary variable, and NBDTIM values of specified flow.
5d. Data: Segment IAUX FLWRAT(NBDTIM)

Omit data item 5dii if ITRAN = 0 or IAUX = 0

5dii. Data: FBCONC (NSPECI, NBDTIM)
Type: Real
Note that data item 5dii consists of one entry for each of the NBDTIM
interpolation time values for a species. Data item 5dii is repeated for each
species associated with transport. Also note that if IAUX (of data item 5d)=0,
the concentration of all species is taken as zero for this node.
Omit data items 6a and 6b if NFHBX1=0 or if all NGWFLW, NOLFLOW and

NCHFLOW=0. Include one set of data items 6a and 6b for each of NFHBX1
auxiliary variables. Input item 6b consists of one set of numbers for each of
NGWFLW, NOLFLW and NCHFLW cells. Each set includes NBDTIM values of the
variable.
Type: Real
Omit data item 7a if all NGWHED, NOLHED and NCHHED=0.
Omit data item 7b if NGWHED=0. Input item 7b consists of one set of numbers for
each of NGWHED cells. Each set of numbers includes layer, row, and column
indices, an integer auxiliary variable, and NBDTIM values of specified head.
7b. Data: Layer Row Column IAUX SBHED(NBDTIM)
Omit data item 7bii if ITRAN = 0 or IAUX # 0
7bii. Data: HBCONC (NSPECI, NBDTIM)
Type: Real

Note that data item 7bii consists of one entry for each of the NBDTIM interpolation
time values for a species. Data item 7bii is repeated for each species associated with
transport. Also note that if IAUX (of data item 7b) = 0, the concentration of all
species is taken as zero for this node. Further, if IAUX (of data item 7b)<0, the
Omit data item 7c if NOLHED=0. Input item 7c consists of one set of numbers
for each of NOLHED cells. Each set of numbers includes overland, row, and
column indices, an integer auxiliary variable, and NBDTIM values of specified
head.
7c. Data: Row Column IAUX SBHED(NBDTIM)

Omit data item 7cii if ITRAN = 0 or IAUX # 0

7cii. Data: HBCONC (NSPECI, NBDTIM)
Type: Real
Note that data item 7cii consists of one entry for each of the NBDTIM
interpolation time values for a species. Data item 7cii is repeated for each
species associated with transport. Also note that if IAUX (of data item 7c) = 0,
the concentration of all species is taken as zero for this node. Further, if IAUX
(of data item 7c)<0, the concentrations input in the HCN package are used for
the simulation.
Omit data item 7d if NCHHED=0. Input item 7d consists of one set of numbers
for each of NCHHED cells. Each set of numbers includes channel segment
number, an integer auxiliary variable, and NBDTIM values of specified head.
7d. Data: Segment IAUX SBHED(NBDTIM)

Omit data item 7dii if ITRAN = 0 or IAUX # 0

7dii. Data: HBCONC (NSPECI, NBDTIM)
Type: Real
Note that data item 7dii consists of one entry for each of the NBDTIM
interpolation time values for a species. Data item 7dii is repeated for each
species associated with transport. Also note that if IAUX (of data item 7d)=0,
the concentration of all species is taken as zero for this node. Further, if IAUX
(of data item 7d)<0, the concentrations input in the HCN package are used for
the simulation.
Omit data items 8a and 8b if NFHBX2=0 or if NGWHED, NOLHED and NCHHED

all =0. Include one set of data items 8a and 8b for each of NFHBX2 auxiliary
variables. Input item 8b consists of one set of numbers for each of NGWHED,
NOLHED, NCHEED cells. Each set includes NBDTIM values of the variable.

Type: Real
NBDTIM -- is the number of times at which flow and head will be
specified for all selected cells.
If NBDTIM = 1, specified flow and head values will
remain constant for the entire
simulation.
If NBDTIM > 1, specified flow and head values will
be computed for each time step using
linear interpolation.
NGWFLW -- is the number of groundwater cells at which flows will be
specified.
NGWHED -- is the number of groundwater cells at which head will be
specified.
IFHBSS -- is the FHB steady-state option flag. If the simulation is
transient, the flag is read but not used. For steady-state
simulations, the flag controls how specified-flow,
specified-head, and auxiliary-variable values will be
computed for each steady-state solution.
If IFHBSS = 0, values of flow, head, and auxiliary
variables will be taken at the starting
time of the simulation. This results
in use of the first value in arrays
FLWRAT, SBHED, and AuxVar for
each respective boundary cell.
If IFHBCB … 0, values of flow, head, and auxiliary

variables will be interpolated in the
same way that values are computed
for transient simulations.
IFHBCB -- is a flag and unit number.
If IFHBCB > 0, it is the unit number on which
cell-by-cell flow terms will be
recorded whenever ICBCFL is set
(see McDonald and Harbaugh, 1988,
chap. 4, p. 14–15).
If IFHBCB # 0, cell-by-cell flow terms will not be

recorded.
computed for each time step for each specified-flow cell.
computed for each time step for each specified-head cell.

NOLFLW -- is the number of overland cells at which flows will be
specified.
NOLHED -- is the number of overland cells at which heads will be
specified.
NCHFLW -- is the number of channel cells at which flows will be

specified.
NCHHED -- is the number of channel cells at which heads will be
specified.
ITHTFHB -- is an index for controlling time-stepping such that tabular
entries for the time inputs are exactly met by the adaptive
time-stepping routines.
IFITHTFHB = 0, the adaptive time-stepping is independent
of tabular time inputs for FHB.
If ITHTFHB=1, the time steps are adapted to honor the
tabular time value entries of the FHB input.
If ITHTFHB=2, the head or flux values are provided as step
functions instead of interpolating between tabular time values
entries. Thus, the head or flux value associated with a time
entry is maintained till the next time entry at which stage the
head or flux values associated with the new time entry are
implemented.
Note that time steps are adapted to honor the tabular time
values when ITHTFHB=2 to accurately capture each step
change in head or flux.
IFHBCC -- is a flag and unit number for printing mass flux output of a
transport simulation.
If IFHBCC>0, it is the unit number on which cell-by-cell
mass flux terms will be recorded wherever ICCCFL is set in
the ATO or OC packages.
If IFHBCC#0, cell-by-cell mass flux terms will not be
recorded.
VarName -- is the name of an auxiliary variable. Name can include up to
16 characters with no embedded blank characters.
Weight -- is the time-weighting factor for an auxiliary variable
specifying the fraction of each time step at which the value of
the variable will be interpolated. Value must be in the range
from 0.0 to 1.0.
IFHBUN -- is the unit number on which data lists will be read. The same
or different unit numbers can be used to read lists in data
items 4b, 5b, 5c and 5d, 6b, 7b, 7c and 7d, and 8b.
CNSTM -- is a constant multiplier for data list BDTIM (data item 4b),
FLWRAT (part of data item 5b, 5c and 5d), SBHED (part of
data item 7b, 7c and 7d), and auxiliary variables in data items
6b and 8b.
IFHBPT -- is a flag for printing values of data lists in items 4b, 5b, 5c
and 5d, 6b, 7b, 7c and 7d, and 8b.
If IFHBPT > 0 data list read at the beginning of the
simulation will be printed.
If IFHBCB < 0 data list read at the beginning of the
simulation will not be printed.
BDTIM -- is simulation time at which values of specified flow and (or)
values of specified head will be read. NBDTIM values are
required.
Layer -- is the layer index of specified-flow cell (data item 5b, 5c and
5d) or specified-head cell (data item 7b, 7c, and 7d).
Row -- is the row index of specified-flow cell (data item 5b, 5c and
5d) or specified-head cell (data item 7b, 7c and 7d).
Column -- is the column index of specified-flow cell (data item 5b and
5c) or specified-head cell (data item 7b and 7c).
Segment -- is the channel segment number of specified-flow cell (data
item 5d and 7d).
IAUX -- is a dummy parameter.
FLWRAT -- is volumetric rate of flow at specified-flow cells. A list of
NBDTIM values must be specified for each of NFLW
specified-flow cells.
FBCONC -- is the concentration of species for fluid entering the domain
via the specified-flow cells. If a value is provided for a node
during outflow conditions, the value is ignored and the
outflow concentration is determined by the concentration of
water leaving the domain at the specified-flux cell.
AuxVar -- is value of real auxiliary variable at specified-flow and
specified-head cells. A list of NBDTIM values must be
specified for each of NFLW specified-flow cells and for each
of NHED specified-head cells.
SBHED -- is an array containing NBDTIM values of the head for each
specified-head cell.
HBCONC -- is the concentration of species for fluid entering the domain
via the specified-head cells. If a value is provided for a node
during outflow conditions, the value is ignored and the
outflow concentration is determined by the concentration of
water leaving the domain at the specified-head cell.

4.1 GENERAL
This chapter presents three levels of testing used to demonstrate verification and
application of MODFLOW-SURFACT/MODHMS to groundwater contamination
problems. The first level of testing verifies the coding of single-species and multi-
component transport formulations and assesses the accuracy of various numerical
schemes by solving relatively simple problems for which analytical solutions are
available. The second level proceeds to deal with practical problems with complexities
(e.g., heterogeneity, transient flow, and water-table conditions) which preclude
analytical solution. This allows testing of the code’s utility and flexibility and
identification of pertinent model features that are necessary for realistic field
simulations. The multiphase transport modules of Version 2.2 require this second
level of verification to demonstrate use and assess accuracy. The third level of testing
focuses on actual field application of the code. Three input data sets of the test
simulations, selected one from each level, are presented in Appendices A through C.
4.2 LEVEL I: PROBLEMS WITH ANALYTICAL SOLUTIONS

The first level of testing consists of five test problems. These problems are
discussed in Sections 4.2.1 through 4.2.5. They include:
• One-dimensional single-species transport in a uniform flow field
• One-dimensional three-member chain decay transport with retardation
• Transport due to an injection/withdrawal well
• Two-dimensional transport from a point source in a uniform flow field
• Transport due to an injection/withdrawal well pair.
Numerical solutions produced by MODFLOW-SURFACT/MODHMS have been
compared with analytical solutions, as well as results from other numerical codes.
The comparative study demonstrates the superior performance of the TVD schemes
used in the transport simulator of MODFLOW-SURFACT/MODHMS.
4.2.1 One-Dimensional, Single-Species Transport
This problem concerns one-dimensional advection-dispersion of a conservative
solute species through a semi-infinite porous medium (Figure 4.1). As illustrated, the
contaminant is released from a channel fully penetrating a shallow confined aquifer.
An analytical solution of the problem was developed by Ogata and Banks (1961). This
solution also can be found in Bear (1979). For the cases documented herein, the inlet
was a first-type boundary with prescribed relative concentration of 1.0 mg/m3. The
values of Darcy velocity and effective porosity were specified as 1 m/d and 0.25,
respectively. The modeled length of the aquifer domain was specified as 400 m. A
uniform rectangular mesh was used consisting of 40 grid blocks, each with )x = 10
m. Three cases were studied with different values of longitudinal dispersivity ("L).
The first case had a dispersivity value of 5 m, the second case of 2 m, and the third
case of zero. For the first two cases, the simulation was performed for 20 time steps
with )t kept constant and equal to 2.5 days. A constant time step size of 1.25 days
was used for the third case. Simulated concentration distributions using the adaptive
implicit TVD scheme with the optimal time stepping are plotted in Figures 4.2, 4.3,
and 4.4 at t = 50 d for the three cases
Chapter 4: Verification and Application 4-1

Figure 4.1 One-Dimensional Transport of a Conservative Species.
4-2 Chapter 4: Verification and Application

One-Dimensional Transport "L=5 m (Pe=2).


for which the grid Peclet numbers are Pe=2, 5, and 4, respectively. The grid Peclet
number is defined as Pe=)x/"L.
As can be seen, the numerical and analytical solutions are in excellent agreement
for all cases. Upstream and central difference results obtained using FTWORK code
(Faust et al., 1993) are also plotted on each of the three figures for comparison. For
a low Peclet number of 2, the central difference scheme performed comparable to the
TVD scheme of MODFLOW-SURFACT/MODHMS, but gave oscillatory results for
higher Peclet numbers of 5 and infinity. The upstream scheme produced smeared
concentration fronts for all three cases. Note that for the case of pure advection
(Figure 4.4) the TVD scheme of MODFLOW-SURFACT/MODHMS was able to
resolve the sharp front within 3 grid-blocks.
4.2.2 One-Dimensional Transport of a Three-Member Radioactive Decay
Chain
This problem is depicted schematically in Figure 4.5. The problem was selected
to test implementation of the chain decay routines in MODFLOW-
SURFACT/MODHMS. It concerns transport of a chain of three radionuclides
released from a decaying source located at x = 0 in a steady-state flow field within a
confined porous medium reservoir. Analytical solutions can be found in Coats and
Smith (1964) and Lester et al. (1975). The flow region and the reservoir properties
used in the present simulation also are depicted in Figure 4.5. The properties of the
three radionuclide components and the initial and third-type decaying source boundary
conditions employed are given in Table 4.1. For comparison purposes, these data were
selected to correspond to the data used in a previous study on a three-member chain
transport simulation reported by Reeves and Cranwell (1981). The adaptive implicit
TVD scheme was selected for the simulation.
The flow region was discretized into 42 grid blocks of equal length, thus giving
)x = 16.4 ft (5 m) and a mesh Peclet number Pe = 1.93. (The mesh Peclet number
is defined as Pe =)x/"L.). The problem was solved for 15 time steps, with a constant
time step size value of 18.26 yr. Figure 4.6 shows the concentration distributions of
all three components after 273.9 years of simulation. Agreement with the analytical
solution is good. Excellent mass balance results were obtained with the normalized
mass balance error being less than a fraction of a percent for all three species.
4.2.3 Transport Due to an Injection/Withdrawal Well
This example problem considers an injection/pumping cycle for a fully
penetrating well in a confined aquifer with zero ambient flow field (Figure 4.7).
Water of a constant concentration (C@) is injected into the well in the center of the
domain for the first stress period. For the second stress period, the flow is reversed,
and the contaminated groundwater is pumped out. The flow field is assumed to reach
a steady state instantaneously for each stress period. An approximate analytical
solution for this problem is given by Gelhar and Collins (1971).
A numerical model consisting of 31 columns, 31 rows, and 1 layer was used to
simulate the concentration change at the injection/extraction well for comparison with
the solution of Gelhar and Collins (1971). An axi-symmetric simulation with a refined
grid at the vicinity of the well was also performed to test this option of the code.
Model parameters used in the simulation are provided in Table 4.2. A constant head
of 50 ft was supplied along the outer boundary of the domain to produce a steady-state
flow field. The first stress period was divided into 10 equal time steps, and the second
stress period was divided into 28 equal time steps. The input data used for the
MODFLOW-SURFACT/MODHMS simulation are presented in Appendix A. The
adaptive implicit TVD scheme was used to solve the problem.

Figure 4.5 One-Dimensional Transport of a Three-Member Radioactive Decay
Chain.

Table 4.1 Parameter Values and Initial and Boundary Conditions for
Simulation of One-Dimensional Transport of a Three-Member
Radioactive Decay Chain.
Radionuclide properties*
Component R t½,R yr 8R, yr-1 RR coR T, yr
1 433 0.0016 9351 1 273.9

2 15 0.0462 9351 0 273.9
3 6540 0.0001 9351 0 273.9
Type of Condition Mathematical Expressions
Initial
cR(x,t = 0) = 0R = 1,2,3
Inlet boundary
∂cl
Vcl − D = V cl l = 1, 2, 3
∂x
c1 = c1o e − λ1t
λ1c1o
c2 = c e o
2
− λ2 t
+
λ1 − λ 2
(e −λ t e −λ t )
2 1
λ2 c2o −λ t −λ t
c3 = c e o −λ3t
3 +
λ2 − λ3
(e − e ) 3 2
⎡ e − λ1t e − λ2 t e − λ3t ⎤
+ λ1λ 2 c1o ⎢ − + ⎥
⎣ (λ1 − λ 2 )(λ1 − λ3 ) (λ1 − λ 2 )(λ 2 − λ3 ) (λ1 − λ3 )(λ 2 − λ3 )⎦
*
Note coR = concentration of inflowing water at t = 0+ and
T = time period of release.

Figure 4.6 Species Concentration Distributions Showing a Comparison of
Analytical and MODFLOW-SURFACT/MODHMS Solutions for
One-Dimensional Transport of a Three-Member Radioactive Decay
Chain.
Figure 4.7 Transport Due to an Injection/Withdrawal Well.

Table 4.2 Model Parameters for Simulations of Transport From an
Injection/Extraction Well.
Parameter Value
Cell width along rows, )x 900 ft

Cell width along columns, )y 900 ft
Layer thickness, )z 20 ft
Hydraulic conductivity of the aquifer 0.005 ft/s
Porosity, N 0.3
Longitudinal dispersivity, "L 100 ft
Transverse dispersivity, "T 100 ft
Volumetric injection (+)/extraction (–) 1 ft3/s
rate for first/second stress periods
1.0 at well node during the
Concentration
first stress period
Length of the injection period 2.5 years
Length of the extraction period 7.5 years

Simulation results are compared with the analytical solution in Figure 4.8, as
a breakthrough curve of contaminant concentration at the well. The match between
our numerical solution and the approximate analytical solution is good. Cumulative
mass balance errors for the simulation are negligible. Also depicted in Figure 4.8 are
results from the MOC and MT3D codes. The MOC solution is based on the method
of characteristics (MOC) and presented by El Kadi (1988). The MT3D solution is
based on the modified method of characteristics (MMOC) presented by Zheng (1990).
Note that both the MOC and MT3D solutions show considerable numerical
oscillations. The oscillatory behavior of well concentration breakthrough curve is
attributable to the particle tracking schemes used in MOC and MT3D, which are prone
to numerical difficulties in handling point sinks and/or sources of variable strength.
4.2.4 Two-Dimensional Transport from a Point Source in a Steady
Uniform Flow Field
This problem concerns two-dimensional dispersion of solute in a uniform and
steady groundwater flow field. In practice, the situation may correspond to that
involving the areal migration of contaminants continually released into an extensive
aquifer from an injection well treated as a point source in the areal plane (Figure
4.9a). Assuming that the injected rate is small so that the natural groundwater flow
is virtually undisturbed, the analytical solution of the problem can be found in Wilson
and Miller (1978). Two cases involving conservative and non-conservative solute
species were simulated using MODFLOW-SURFACT/MODHMS. The model
parameters, shown in Table 4.3, are taken from Huyakorn et al. (1984). The selected
parameter values for Case 1 were based on data from the field study of hexavalent
chromium contamination reported by Perlmutter and Lieber (1970). For Case 2, the
values of retardation and decay constant were chosen arbitrarily to test the
performance of MODFLOW-SURFACT/MODHMS for a non-conservative species.
Numerical simulations of both cases were performed using a rectangular grid that
represents the domain shown in Figure 4.9b. The grid consists of 741 cells, each
having ∆x = 30 m, and ∆y = 30 m. For both cases, the simulations were performed
for 14 time steps with )t kept constant and equal to 100 days. Concentration profiles
along the x-axis at t = 1400 d are plotted in Figure 4.10 for both Cases 1 and 2. The
MODFLOW-SURFACT/MODHMS solutions were obtained using the adaptive
implicit TVD scheme. Numerical results from MODFLOW-SURFACT/MODHMS
are compared with the analytical solution and the finite-difference solution of
FTWORK (Faust et al., 1993). Both numerical solutions are in fairly good agreement
with the analytical solution. Note that for both cases MODFLOW-
SURFACT/MODHMS produces slightly more accurate solutions than FTWORK.
4.2.5 Transport Due to an Injection-Withdrawal Well Pair
This problem concerns solute transport between a pair of recharging and
discharging wells operating at a constant flow rate. Both wells fully penetrate a
constant-thickness confined aquifer that is assumed to be of infinite areal extent,
homogeneous, and isotropic. The flow field is assumed to be in a steady state with
typical flow lines and solute fronts depicted in Figure 4.11.
An analytical solution has been developed for transport in this situation
(Hoopes and Harleman 1967). A numerical model consisting of 31 columns, 33 rows,
and 1 layer was used to predict the concentration breakthrough curve at the well for
comparison with the Hoopes and Harleman (1967) solution. Model parameters used
for the simulation are shown in Table 4.4. These parameters are taken from Huyakorn
et al. (1984). The model domain encompasses –145 # x # 145 cm, and

Figure 4.8 Breakthrough Curve from Simulation of Transport Due to an
Injection/Withdrawal Well.
Table 4.3 Model Parameters For Simulation of Two-Dimensional
Transport From a Point Source in a Steady, Uniform Flow
Field.
Parameter Value
Darcy velocity, V 0.161 m/d

Porosity, N 0.35
Longitudinal dispersivity, "L 21.3 m
Transverse dispersivity, "T 4.3 m
Aquifer saturated thickness, b 33.5 m
Contaminant mass flux, QcO 704 g/(m.d)
(per unit thickness of aquifer)
Case 1:
Linear adsorption coefficient, Kd 0.0
ρΒ Kd 1.0
Retardation Coefficient, R = 1+
ϕ
0.0
Decay constant, 8
Case 2:
Linear adsorption coefficient, Kd 0.14 m3/Kg
Bulk density of soil, DB 2.5 Kg/m3
ρΒ Kd
Retardation coefficient, R = 1+ 2.0
ϕ
Decay constant, 8 0.00019/d
Selected grid
)x = 30 m, )y = 30 m, 19 x 39 grid
Source location = Row 10, Column 7
Time steps
)t = 100 d (used 14 time steps)

Figure 4.9 Two-Dimensional Transport from a Point Source in a Uniform Flow
Field: (a) Problem Description, and (b) Model Discretization.

Figure 4.10 Concentration Profiles Along the Center Line of the Plume
Obtained from Simulation of Two-Dimensional Transport from a
Point Source in a Steady Uniform Flow Field.
Figure 4.11 Problem Description of Transport Due to an Injection-Withdrawal
Well Pair.
–145 # y # 145 cm. The origin of the coordinates axes is at the mid-point of the line
joining the injection and pumping wells. Only half the domain (17 rows, 31 columns)
was used in the simulation due to symmetry along the x- axis passing through the well
doublet. Initial concentrations in the domain were set to zero, and a unit concentration
value was prescribed at the injection well location. Fifty time steps were used for a
total simulation time of 8 x 10 4s (0.926 days) with time steps increasing in size by a
factor of 1.2 throughout the simulation. Mass balance errors were below a fraction
of a percent for the entire MODFLOW-SURFACT/MODHMS simulation.
Concentration breakthrough at the pumping well is compared between analytical and
numerical solutions in Figure 4.12 for an adaptive implicit TVD scheme. The TVD
results from MODFLOW-SURFACT/MODHMS compare very favorably with the
analytical solution. Note that the TVD numerical solution exhibits slight smearing
near the toe of the breakthrough curve. The TVD solution starts to depart from the
analytical solution at t >20,000 seconds (0.231 days). This departure is due to the
model boundary effects on the predicted concentration. The analytical solution
assumes an areally infinite domain whereas the numerical model is based on a finite
grid encompassing an area of only 290 x 290 cm2. Results using Galerkin finite
element schemes of the SEFTRAN code (Huyakorn, et al., 1984) are also shown in
Figure 4.12. The Galerkin solution is noted to be unstable for this problem.
4.3 LEVEL II: PRACTICAL PROBLEMS WITH COMPLEXITIES THAT

PRECLUDE ANALYTICAL SOLUTION
The second level of testing consists of seven test problems. These example
problems are discussed in Sections 4.3.1 through 4.3.7. The first three problems
concern transport in confined or unconfined aquifers. The next four examples
demonstrate transport capabilities in a multiphase situation. They include:
• Two-dimensional (areal) transport of a contaminant plume in a heterogeneous
confined aquifer with a pair of injection and withdrawal wells and strong
ambient groundwater flow.
• One-dimensional transport in a contaminated unconfined aquifer being
remediated using a system of gallery and shallow trench.
• Two-dimensional transport of a contaminant plume in a heterogeneous
unconfined aquifer.
• Two-dimensional transport of contaminant in the water phase of an
unsaturated rectangular soil slab.
• Two-dimensional transport of contaminant in an active water phase with
partitioning and diffusion in the air phase.
• Two-dimensional transport of contaminants that dissolve from an immobile
NAPL phase into an active water phase with volatization and diffusion in the
air phase.
• Three-dimensional study of air sparging at a NAPL contaminated site.
The MT3D code documented by Zheng (1990) is selected as a comparative
code for Level II Testing of MODFLOW-SURFACT/MODHMS for confined or
unconfined problems. With identical finite-difference grids, both MODFLOW-
SURFACT/MODHMS and MT3D were used to perform the transport simulations.
Comparative tests show that the TVD scheme of MODFLOW-SURFACT/MODHMS
produces strictly mass-conserved and stable numerical solutions whereas the MMOC
scheme of MT3D produces severely oscillatory concentration breakthrough curves

Table 4.4 Model Parameters for Simulation of Transport with an
Injection-Withdrawal Well Pair.
Parameter Value
Well flow rate, Q 2.339 cm3/sec

Well spacing, 2xO 61.0 cm
Thickness of aquifer model, b 8.9 cm
Porosity, N 0.374
Retardation factor, R 1
Decay constant, 8 0
Longitudinal dispersivity, "L 0.294 cm
Lateral dispersivity, "T 0

Figure 4.12 Breakthrough Concentrations at Withdrawal Well Obtained from
Simulation of Transport with an Injection-Withdrawal Well Pair.
and substantial mass balance errors. The MAGNAS and STAFF3D codes are selected
as comparative tools for problems involving multiphase conditions. MAGNAS and
STAFF 3D use a mesh centered finite element discretization and the example problems
were analyzed by MAGNAS, STAFF3D, and MODFLOW-SURFACT/MODHMS
using similar grid spacings. The comparative tests demonstrate the applicability and
accuracy of MODFLOW-SURFACT/MODHMS’s multiphase transport simulations
options.
4.3.1 Two-Dimensional Transport in a Heterogeneous Confined Aquifer
with a Steady-State Flow Field
This problem, taken from Zheng (1990), is depicted in Figure 4.13. It
concerns contaminant transport from an injection to a pumping well under a strong
ambient, steady-state groundwater flow field. The flow domain is bounded by no-flow
boundaries on the east and west sides. The north side is a constant-head boundary and
the south side is a specified-head boundary with a linear decline of hydraulic head
toward the west. The contaminant is injected into the aquifer through an injection well
for a duration of 1 year. The total contaminant mass introduced into the aquifer
system is Mi = Qiciti , where Qi is the injection flow rate, ci is the concentration of the
injected fluid, and ti is the injection duration. Using the injection well data given in
Figure 4.13, the calculated value of Mi is 1825 kg.
Contaminant injection is zero for the remaining 5 years of simulation again
under steady-state flow conditions. A pumping well located downstream operates at
a constant rate of Qp = 0.0189 m3/s through the entire simulation period of 6 years.
A low conductivity zone lies between the two wells. The aquifer hydraulic
conductivities and well flow rates also are depicted in Figure 4.13. For both zones,
the effective porosity, and longitudinal and transverse dispersivities are set equal to
0.3, 20 ft, and 4 ft, respectively.
A uniform grid consisting of 14 columns, 18 rows, and 1 layer was used to
discretize the study domain. The grid spacings were
MODFLOW-SURFACT/MODHMS was run using the TVD scheme and 128 time
steps to cover the simulation time of 6 years. The input data used for the simulation
are presented in Appendix B. The concentration breakthrough curve at the pumping
well is plotted in Figure 4.14. Also included is the corresponding result obtained from
the MT3D code using the Hybrid Method of Characteristics (HMOC) scheme,
recommended by Zheng (1990). Note that the MODFLOW-SURFACT/MODHMS
solution is stable and smooth with slight front smearing, whereas the MT3D solution
exhibits serious numerical oscillations between adjacent time steps. There is,
however, a reasonable overall agreement between the two solutions (if the oscillatory
behavior of MT3D is filtered).
The spatial distribution of contaminant at one year of simulation is shown in
Figure 4.15. In general, the patterns of the plumes produced by MODFLOW-
SURFACT/MODHMS and MT3D are similar. However, the concentration
distributions inside the plumes are significantly different. Owing to numerical
oscillations, the MT3D plume exhibits a multi-peak behavior despite the fact that
there is only one contaminant source and the source (injection well) is still operating
at t = 1 year. Note that the local downstream peaks are simply artifacts of the
numerical oscillations. On the other hand, MODFLOW-SURFACT/MODHMS
produces a realistic plume, without oscillation and having only one peak at the source
location. Some numerical dispersion is encountered with the TVD scheme of
MODFLOW-SURFACT/MODHMS (because of the coarseness of the grid) which
results in a wider lateral extent of the plume boundary.

Heterogeneous Confined Aquifer.
Figure 4.14 Pumping Well Concentrations Obtained from Simulation of Two-
Dimensional Transport in a Heterogeneous Aquifer.
C (ppm)
MODFLOW-SURFACT Plume MT3D Plume
Correct Single-Peak Behavior Incorrect Multi-Peak Behavior
4.00
at the Source Location due to Numerical Oscillations
3.53
3.06
2.59
2.13
1.66
1.19
0.72
0.25
Figure 4.15 Concentration Distribution at One Year of Simulation of Two-

Dimensional Transport in a Heterogeneous Confined Aquifer, (a)
MODFLOW-SURFACT/MODHMS, and (b) MT3D.
As the problem of interest concerns aquifer remediation, cumulative mass
budget computations were also performed. Shown in Figure 4.16 are three
components of the mass budgets produced by MODFLOW-SURFACT/MODHMS
and MT3D. These mass budget components are defined in the conventional manner
as presented in Section 2.5. To recap, I is the total mass introduced into the
groundwater flow system as a result of contributions from external sources, O is the
mass out of the groundwater flow system through sinks, and )S is the net gain in
storage of contaminant in the groundwater flow system. As can be seen from Figure
4.16, the mass budget of MODFLOW-SURFACT/MODHMS is physically consistent,
satisfying the mass balance requirement (I-O-)S=Zero) at all time values. On the
other hand, the mass budget of MT3D displays a physically unrealistic behavior,
having the cumulative outflow (O) in excess of the total mass introduced into the
groundwater system (I) by almost 10 percent at the end of the simulation time.
The mass budgets are also presented (Figure 4.17) using the definition of IN
and OUT components given by Zheng (1990). According to this definition, IN is the
total mass into the groundwater flow system from external sources + mass release
from storage as a result of the decrease in solute and sorbed concentrations, and OUT
is the total mass out of the groundwater flow system though sinks + mass
accumulation in storage as a result of the increase in solute and sorbed concentrations.
Note that IN = I + )Sin , and OUT = O + )Sout , where )Sin is the mass release
from storage as a result of concentration decreases, and )Sout is the mass accumulation
in storage as a result of concentration increases. It follows that IN – OUT = I – O
– )S, where )S = )Sin = )Sout.
As can be seen from Figure 4.17, the mass budget components of
MODFLOW-SURFACT/MODHMS closely satisfy the following mass conservation
requirements: (1) the difference between IN and OUT is zero, and (2) neither IN nor
OUT exceed 3650 kg (twice the total mass injected into the groundwater flow system).
It is evident that the second mass conservation requirement is seriously violated by the
mass budget of MT3D. The IN and OUT terms of MT3D exceed the required ceiling
by as much as 120 percent at the end of the simulation period.
Although the IN and OUT components of the MT3D mass budget are grossly
in error, their errors happen to be compensating. As a result, the mass balance error
information provided by MT3D is artificially subdued. The dimensional and
dimensionless (normalized) mass balance errors are plotted in Figure 4.18 for both
MT3D and MODFLOW-SURFACT/MODHMS. At all time values, MODFLOW-
SURFACT/MODHMS produces negligible mass balance errors. Despite the offsetting
effect of gross errors in its IN and OUT terms, MT3D still produces almost 8 percent
error at early time values and close to 2 percent at later time values.
4.3.2 One-Dimensional Transport in Transient Unconfined Flow
This problem is depicted in Figure 4.19a. The scenario of interest concerns
remediation of a shallow unconfined aquifer contaminated with a soluble chemical
from a leaky pipe. A remedial system consisting of a fresh-water injection gallery and
a trench is used to flush the aquifer of contaminant. Initially, the water table is 4
meters below the land surface and the concentration in the saturated zone (covering
the length of 70 m upstream of the trench) is 1 kg/m3. The initial contaminant mass
per unit width may be determined as Mo = (1 kg/m3) x (70 m) x (8 m) x 0.3 = 168
kg/m (conc x Volume x Porosity). A constant water level of 1.5 m above the aquifer
base is maintained for the trench. The gallery fresh-water injection scheme is depicted
in Figure 4.19b.

Figure 4.16 Mass Budget from Simulation of Two-Dimensional Transport in a
Heterogeneous Confined Aquifer.
Figure 4.17 IN-OUT Mass Budget from Simulation of Two-Dimensional
Transport in a Heterogeneous Confined Aquifer.
Figure 4.18 Mass Balance Errors from Simulation of Two-Dimensional
Transport in a Heterogeneous Confined Aquifer, (a) Dimensional
Error, and (b) Normalized Error.
Figure 4.19 One-Dimensional Transport in Transient Unconfined Flow,
(a) Problem Description, and (b) Gallery Operation.
The aquifer domain between the gallery and trench was discretized into 50
cells (50 columns, 1 row, and 1 layer). Transient flow followed by transient transport
simulations were performed using MODFLOW-SURFACT/MODHMS and MT3D in
conjunction with MODFLOW. Shown in Figure 4.20 are simulated water-table
profiles at typical time values. The highly transient nature of the unconfined flow may
be noted. Using the same transient flow field, transport simulations were conducted
for the duration of 850 days. The TVD and HMOC schemes were used in running the
MODFLOW-SURFACT/MODHMS and MT3D, respectively. Concentration
distributions and mass budgets are shown in Figures 4.21 and 4.22, respectively. As
can be seen, there are some significant differences between the results from
MODFLOW-SURFACT/MODHMS and MT3D. At early time values (during the
first 10 days of simulation), concentration profiles predicted by both codes are in good
agreement (Figure 4.21). The two numerical solutions deviate from each other at later
time values. MT3D underpredicts concentration values during the subsequent stage
of remediation (t $ 270 days).
An examination of the mass budget plot (Figure 4.22) indicates that once again
the TVD scheme of MODFLOW-SURFACT/MODHMS rigorously conserves mass,
whereas the HMOC scheme of MT3D seriously violates mass conservation
requirements. At all times, the IN and OUT components of the MODFLOW-
SURFACT/MODHMS mass budget are virtually identical in absolute values. At the
end of simulation (t = 380 days), both components approach the value of 168 kg,
which is the initial mass of contaminant in the aquifer system. On the other hand, the
OUT component of the MT3D mass budget is slightly overestimated while the IN
component is grossly underestimated. This results in a large mass balance error as can
be seen in Figure 4.23.
4.3.3 Two-Dimensional Transport in a Heterogeneous Unconfined Aquifer
with a Transient Flow Field
This problem, shown in Figure 4.24, is adapted from the previous problem
presented in Section 4.3.1. We now switch from a confined to an unconfined aquifer.
Additionally, transient unconfined flow is created via a cyclic operation of the
pumping well. The unconfined aquifer parameters, boundary conditions, and well
data are depicted in Figure 4.24. Note that the total contaminant mass injection into
the aquifer is = 0.15 (m3/s) x 100 (ppb) x 1 (year) = 473.05 kg. The
specific yield of the aquifer is taken to be the same as its effective porosity. For both
zones, the porosity, and longitudinal and transverse dispersivities are set equal to 0.25,
20 ft, and 4 ft, respectively.
Transient flow followed by transient transport simulations were performed
using a uniform grid identical to that described in Section 4.3.1. Shown in Figure
4.25 is the simulated hydraulic head at the pumping well. The well head response is
consistent with its pumping schedule (Figure 4.24b). The concentration at the
pumping well versus time is plotted in Figure 4.26. In general, the breakthrough
curves obtained from MODFLOW-SURFACT/MODHMS and MT3D have similar
trends. The multi-peak behavior of the MODFLOW-SURFACT/MODHMS
breakthrough curve is a result of the transient pumping schedule of the recovery well.
In terms of concentration values, there are, however, significant differences
between the numerical solutions obtained from MODFLOW-SURFACT/MODHMS
and MT3D. The MODFLOW-SURFACT/MODHMS solution has the desirable
features of being smooth and strictly mass-conservative (Figures 4.26 through 4.28).
The MT3D solution exhibits severe numerical oscillations over several time intervals.
Furthermore, the mass conservation requirements are seriously violated. As can be

Figure 4.20 Water-Table Profiles Obtained from One-Dimensional Unconfined
Flow Simulation.
Figure 4.21 Concentration Profiles from Simulation of One-Dimensional
Transport in Transient Unconfined Flow.
Figure 4.22 Mass Budget from Simulation of One-Dimensional Transport in
Transient Unconfined Flow.
Figure 4.23 Normalized Mass Balance Error from Simulation of One-
Dimensional Transport in Transient Unconfined Flow.
Heterogeneous Unconfined Aquifer, (a) Model Domain, and (b)
Pumping Schedule.
Figure 4.25 Temporal Distribution of Hydraulic Head at the Pumping Well
from Simulation of Transport in a Heterogeneous Unconfined
Aquifer.
Figure 4.26 Pumping Well Breakthrough Curve from Simulation of Transport
in a Heterogeneous Unconfined Aquifer.
Figure 4.27 Mass Budget from Simulation of Transport in a Heterogeneous
Unconfined Aquifer.

Figure 4.28 Normalized Mass Balance Error from Simulation of Transport in
a Heterogeneous Unconfined Aquifer.
seen, both the IN and OUT components of the MT3D mass budget are grossly
overestimated.
For the simulated case, the ceiling of the IN and OUT absolute values is equal
to twice the total amount of contaminant mass injected into the pristine groundwater
system (i.e., 2 x 473.05 = 946.1 kg). However, the absolute values of IN and OUT
given by MT3D at the end of the simulation (t = 7 years) are 1283.1 kg and 1214.4
kg, respectively. Although the aquifer systems are not completely cleaned up at t =
7 years, the IN and OUT terms of the MT3D model have already exceeded the
maximum allowable value by 35.6 and 28.3 percent, respectively. The corresponding
IN and OUT values of MODFLOW-SURFACT/MODHMS are 915.1 kg, slightly
below the ceiling of 946.1 kg.
The mass balance behavior of both transport models are shown in Figure 4.28.
MODFLOW-SURFACT/MODHMS produces virtually zero mass balance error at all
times. On the other hand, MT3D exhibits substantial mass balance errors up to 43
percent during the early part of the simulation. At later time values (t > 1 yr), the
MT3D mass balance errors are subdued (and eventually declining to approximately 5
percent) because of the offsetting effect of the gross errors in the IN and OUT terms.
4.3.4 Two-Dimensional Transport in the Water Phase of an Unsaturated
Rectangular Soil Slab
This problem corresponds to the two-dimensional unsaturated flow problem
discussed in Section 2.3B.2 of Volume I. It concerns transport of a nonconservative
solute in a transient flow field in a two-dimensional, cross-sectional, rectangular soil
slab. As illustrated in Figure 4.29, the soil is initially dry and water with dissolved
solute is allowed to enter the system at the upper portion of the left hand boundary
(x=0, 6#y#10 cm). Inflow head is 6 cm with a prescribed solute concentration of
1 ppm. Outflow occurs along the entire right hand side boundary under initial
pressure conditions (head of -90 cm). The remaining boundaries of the domain are
under no-flow, zero concentration gradient conditions. Physical parameter values of
the system are shown in Table 4.5. Grid and time discretization are discussed in
Section 2.3B.2 of Volume I. Water saturations and Darcy velocities at different times
were obtained by performing the flow computations for the corresponding time step
(as discussed in Section 2.3B.2 of Volume I) prior to the respective transport
computations.
Figure 4.30 shows contaminant concentration contours at 0.508 d as simulated
by MODFLOW-SURFACT/MODHMS and the finite element model STAFF3D
(HydroGeoLogic, 1992) which has the required physical simulation capabilities
necessary for solving this problem. The results compare well indicating that the
unsaturated transport equation for contaminant in the active water phase of
MODFLOW-SURFACT/MODHMS is implemented correctly.
4.3.5 Two-Dimensional Transport in an Active Water Phase with Mass
Transfer and Diffusion in the Inactive Air Phase
The problem scenario is depicted in Figure 4.31. A two-dimensional,
rectangular, cross-section of a water-table aquifer is analyzed for contamination via
a surface-applied, volatile pesticide, along a 200 m long strip. The water-table aquifer
bottom is an assumed no-flow condition at an elevation of 10 m above sea level (ASL).
Land surface is at 28 m ASL. The water-table enters the 675 m long domain at an
elevation of 20.16 m ASL and exits at 13.3 m ASL. Average precipitation recharge
along land surface is 6.2 cm/yr (1.966x10-9 m/s). The leading edge of the
contaminated strip of land is 175 m downstream of the left edge of the domain. A
steady-state flow field is assumed under these conditions to observe

Figure 4.29 Geometry and Boundary Conditions for Simulation of Solute
Transport in an Unsaturated Rectangular Soil Slab.
Table 4.5 Physical Parameter Values for Simulation of Transport in
an Unsaturated Rectangular Soil Slab
Parameter Value
Porosity, N 0.45
Initial Concentration, Co 0
Longitudinal Dispersivity, "L 1 cm
Transverse Dispersivity, "T 0
Molecular Diffusion, Do 0.01 cm2/d
Decay Coefficient, 8 0.001d-1
Bulk density, DB 1.46 g/cm3
Distribution Coefficient, Kd 0.308 cm3/g
Note that flow parameters for this problem are provided in Section 2.3B.2, Volume
I of the MODFLOW-SURFACT/MODHMS document.

Figure 4.30 Simulated Contaminant Concentrations in an Unsaturated
Rectangular Soil Slab at 0.508d.
Figure 4.31 Problem Description for Leaching of Volatile Pesticide from the Soil
Surface with Air Phase Diffusion.
Table 4.6 Material Parameter Values for Simulation of Transport in
an Active Water Phase with Volatization and Diffusion in
the Air Phase
Parameter Value
Saturated Hydraulic Conductivity, K 3.9x10-4 m/s

Porosity, N 0.45
van Genuchten unsaturated soil parameter, " 0.5 m
van Genuchten unsaturated soil parameter, $ 1.454
van Genuchten relative permeability function is used --
Residual Water Saturation 0.1756
Longitudinal Dispersivity, "L 0.1071
Transverse Dispersivity, "T 0
Bulk density of soil, DB 1650 kg/m3
Bulk diffusion coefficient of pesticide in air, Do 0.5001x10-2 m2/s
Distribution Coefficient, Kd 0.434x10-4 m3/kg
Concentration Based Air-Water Partition Coefficient, Kaw 4.394

leaching and transport of pesticide in the subsurface. Table 4.6 depicts the soil
parameters and transport parameters of the system.
A uniform grid containing 18 layers and 27 columns was used to discretize the
domain with )x=25 m and )y=1 m. After solution to the steady-state flow field, the
transport solution was obtained with a starting time-step value of 100 s which was
increased by a factor of 2 up to a maximum time-step size of 2.59x106s. The transport
simulation was performed for 100 years.
Figures 4.32a through 4.32c show contaminant profiles in the subsurface at
1 day, 60 days, and 20 years respectively as obtained from the MODFLOW-
SURFACT/MODHMS simulation as well as from a MAGNAS (HydroGeoLogic,
1994) simulation of the same problem. The results compare well, indicating that the
multiphase transport modules of MODFLOW-SURFACT/MODHMS with mass
transfer via equilibrium partitioning, and diffusion in the inactive phase are operating
appropriately. A further understanding of this situation may be obtained by observing
simulated conditions under assumptions of no mass partitioning to the inactive phase.
Figure 4.33 shows pesticide concentration profiles obtained using the no-partitioning-
to-air-assumption in both MODFLOW-SURFACT/MODHMS and MAGNAS. The
two codes again compare well in their results. Partitioning to air and air phase
diffusion are noted to have a significant effect in spreading contaminant in the
unsaturated zone. Partitioning also retards contaminant movement vertically through
the unsaturated zone, and the contaminant diffused in the air phase is noted to dissolve
in the recharge water and leach to the water table at locations away from the pesticide
source. It may thus be noted that the assumption of no-mass transfer to the air phase
may have a significant impact on contaminant distributions and movement in the
subsurface, for sufficiently volatile contaminants. A further understanding of this
situation can be obtained by a MAGNAS simulation with both air and water phases
active, and air at constant atmospheric conditions at the land surface. Contaminant
distributions for this case are similar to Figures 4.32 indicating that the inactive-air-
phase assumption of solving the Richard’s equation is sufficient for examining volatile
contaminant transport (i.e., only diffusion in the air-phase) with water flowing in the
unsaturated zone.
4.3.6 Dissolution and Volatization of an Immobile Multi-Component NAPL
with Advective and Dispersive Transport
This problem consists of two parts, both of which use the same problem
geometry. The first problem examines the dissolution of a NAPL mass, consisting of
PCE, TCE, DCE, and VC, located below the water table with water as the active
phase. The second problem examines vaporization of the same NAPL mass, were it
located above the water table, with air as the active phase. Details concerning the
implementation and results of the two problems are described below.
4.3.6.1 NAPL Dissolution—Active Water Phase
This problem considers a domain which is 20 m × 1 m × 5 m in the x-, y-,
and z-dimensions and is discretized using 20×1×10 cells of dimension 1 m × 1 m
× 0.5 m. Water is the active phase and groundwater flow is established by specifying
a hydraulic head of 3 m at the inflow boundary and 2.5 m at the outflow boundary.
Both constant head boundary conditions are specified over an elevation interval of 0
m to 2 m with the remainder of the boundaries being impermeable to groundwater
flow. The steady-state hydraulic head and water saturation distribution are shown in
Figure 4.34. Physical properties of the porous medium are listed in Table 4.7. The
perimeter of the NAPL source zone is indicated by the black box in Figure 4.34 and
is a 2.0 m × 1.0 m × 1.0 m area containing NAPL at a saturation of 0.2. The
NAPL consisted of PCE, TCE, DCE, and VC each at a mass fraction of 0.25 resulting
in an initial inventory of 40.75 kg of PCE, 36.0 kg of TCE, 30.5 kg of DCE, and
22.7 kg of VC. The initial NAPL conditions are implemented with

Figure 4.32 Simulated Pesticide Concentrations for Problem 4.3.5 (a) 1 d, (b)
60 days, and (c) 20 yrs.
Figure 4.33 Simulated Pesticide Concentrations for Problem 4.3.5 with No Mass
Transfer to the Air Phase (a) 1 d, (b) 60 d, and (c) 20 yrs.
a)
head (m)
5
elevation (m)
3
4 2.9
3 2.8
2 2.7
1 2.6
0 2.5
0 5 10 15 20
x-direction (m)
b) water
saturation
elevation (m)
5 1
4 0.8
3 0.6
2 0.4
1 0.2
0 0
0 5 10 15 20
x-direction (m)
Figure 4.34 Hydraulic Head and Water Saturation Profiles for NAPL
Dissolution Example 4.3.6.1.
102 102
1 1
10 10
0 0
10 10
Mass of PCE (kg)
Mass of TCE (kg)

-1 -1
10 10
-2 -2
10 10
total mass
-3 -3
10 total mass 10 mass in water
mass in water mass on soil
-4 mass on soil -4 mass in air
10 mass in air 10 mass in napl
mass in napl mass left boundary
-5 mass left boundary -5
10 10
-6 -6
10 10
102 102
1 1
10 10
0 0
10 10
Mass of DCE (kg)
Mass of VC (kg)
-1 -1
10 10
-2 -2
10 10
-3 total mass -3
10 mass in water 10 total mass
mass on soil mass in water
-4 mass in air -4 mass on soil
10 mass in napl
10 mass in air
mass left boundary mass in napl
-5 -5 mass left boundary
10 10
-6 -6
10 -2 -1 0 1 2 3 4 10 -2 -1 0 1 2 3 4
10 10 10 10 10 10 10 10 10 10 10 10 10 10
Time (days) Time (days)

dissolution time.
Table 4.7 Physical Properties of Soil
Property Value
Hydraulic conductivity
Kx = K y = K z 1.0 m/d
Porosity
N 0.25
Dispersivity
"L 0.5 m
"TH 0.05 m
"TV 0.005 m
Mass Density of Soil

Db 1,600 kg/m3
van Genuchten Parameters

" 5.0 m-1
$ 1.9
Swr 0.2

IINIT=2 and contaminant mass is redistributed to all phases present, with a flash
calculation at the beginning of the simulation. Chemical properties of the four
contaminants are listed in Table 4.8. The dissolution of these four contaminants from
the NAPL source zone is governed by their solubility and respective mass fractions
in the NAPL phase according to Raoults Law. The resulting contaminant mass
distribution for all four contaminants between the water, soil, air, and NAPL phases,
as a function of time, is shown in Figure 4.35. VC was depleted most rapidly from
the NAPL phase because it is the most soluble of the four components. The NAPL
phase was completely dissolved by approximately 8,000 days at which point it
consisted entirely of PCE which is the least soluble component. Figure 4.36 shows
the water phase concentration distribution of all four contaminants at a time of 1,000
days. These results indicate that the majority of transport occurs due to advection and
dispersion in the active water phase while some of the contaminant has diffused
upwards into the inactive gas phase. The largest mass balance error occurred for PCE
at a value of 1.12 ×10-4% indicating that contaminant mass was conserved for all four
components during the dissolution and transport processes.
4.3.6.2 NAPL Volatization—Active Air Phase
This problem considers the same problem geometry as used in the previous
NAPL dissolution problem, 4.3.6.1. In this case, air is the active phase and gas phase
flow is established by specifying a constant gas phase pressure of 101.3 kCa
(equivalent hydraulic head h=0 m) at the inflow boundary and a constant gas phase
pressure of 91.3 kCa (equivalent hydraulic head h= -1 m) at the outflow boundary.
Both boundary conditions extended from an elevation of 3.5 m to 4.5 m with the
remainder of the boundaries being impermeable. Atmospheric gas phase pressure is
101.3 kCa. The vacuum applied at the outflow boundary has raised the water table,
as shown in Figure 4.37, approximately 1 m relative to the initial water table condition
shown in Figure 4.34. The head contours, depicted in Figure 4.37, represent gas
phase pressure converted to equivalent hydraulic head values. Physical properties of
the porous media and of the gas phase specific to this active air phase problem, are
listed in Table 4.9 and supplement those provided in Table 4.7. The perimeter of the
initial NAPL emplacement zone for this case is indicated by the black box in Figure
4.37 and contains the same chemical mass and saturation distribution used in the
previous dissolution problem, 4.3.6.1. The flag IINIT=2 is used to redistribute the
initial NAPL inventory into all phases, with a flash calculation at the beginning of the
simulation. The distribution of contaminant mass for all four components between the
water, soil, air, and NAPL phases, as a function of time, is shown in Figure 4.38.
The time required to completely vaporize the NAPL phase is approximately 7 days
which is significantly faster than during the dissolution problem. This is due to the
preference for the contaminants to partition into the gas phase relative to the water
phase and the increased gas phase pore velocity resulting in greater mass transfer of
contaminants away from the NAPL source. Figure 4.39 shows the concentration
distribution for all four contaminants at a time of 0.1 days. These results indicate that
the majority of transport occurs in the unsaturated zone with advection of the gas
phase while some diffusion has occurred down below the water table. The largest
mass balance error occurred for VC at a value of 4.15 × 10-5% indicating that
contaminant mass was conserved for all four components during the vaporization and
transport processes.
4.3.7 Three-Dimensional Study of Air Sparging at a NAPL Contaminated
Site
This problem corresponds to the verification example for air phase flow
discussed in Section 2.3C, Volume I of the MODFLOW-SURFACT/MODHMS
document. The site has been contaminated by gasoline from several leaking
underground storage tanks which have since been removed. Residual levels of NAPL
exist in the subsurface at an approximate NAPL saturation of 0.1, and gasoline
components exist under equilibrium conditions, in all phases in the subsurface. The
flag IINIT=1 is used to indicate that initial NAPL components are in equilibrium with

Table 4.8 Chemical Properties of Contaminants.
Property Value
Diffusion coefficient in water phase

DwPCE = DwTCE = DwDCE = DwVC 1×10-9 m2/d
Diffusion coefficient in air phase
DaPCE = DaTCE = DaDCE = DaVC 1×10-5 m2/d
Diffusion coefficient in NAPL phase
DnPCE = DnTCE = DnDCE = DnVC 1×10-9 m2/d
Gas-Water Partitioning Coefficient
KgwPCE 0.09043
KgwTCE 0.08348
KgwDCE 0.1053
KgwVC 0.1222
NAPL-Water Partitioning Coefficient
KnwPCE 6792.0
KnwTCE 1028.0
KnwDCE 404.0
KnwVC 329.0
Soil-Water Partitioning Coefficient
KdPCE 4.51 × 10-5 m3/kg
KdTCE 1.56 ×10-5 m3/kg
KdDCE 8.06 ×10-6 m3/kg
KdVC 3.70 ×10-6 m3/kg
Standard Mass Density of Contaminant
DPCE 1630.0 kg/m3
DTCE 1440.0 kg/m3
DDCE 1220.0 kg/m3
DVC 908.0 kg/m3

a)
5 PCE conc.
elevation (m)
(kg/m3)
4 1
3 0.1
2
0.01
1
0.001
0
0.0001
0 5 10 15 20
b)
5 TCE conc.
elevation (m)
(kg/m3)
4
3 1
2 0.1
1 0.01
0 0.001
0.0001
0 5 10 15 20
c)
5 DCE conc.
elevation (m)
(kg/m3)
4
1
3
0.1
2
0.01
1
0 0.001
0.0001
0 5 10 15 20
d)
5 VC conc.
elevation (m)
(kg/m3)
4
3 1
2 0.1
1 0.01
0 0.001
0.0001
0 5 10 15 20
x-direction (m)
(d) VC at 1,000 days.
a)
head (m)
5
elevation (m)
0
4 -0.2
3 -0.4
2 -0.6
1 -0.8
0 -1
0 5 10 15 20
x-direction (m)
b) water
saturation
elevation (m)
5 1
4 0.8
3 0.6
2 0.4
1 0.2
0 0
0 5 10 15 20
x-direction (m)
Figure 4.37 (a) Equivalent hydraulic head contours and (b) water saturations of
the active gas phase flow problem. The black box indicates the
perimeter of the NAPL source zone.
Table 4.9 Physical Properties Controlling Active Gas Phase Flow.
Property Value
Mass Density of Water

Dw 1000.0 kg/m3
Mass Density of Air
Dw 1.777 kg/m3
Viscosity of Water
µW 5.0 × 10-4 Pa@s
Viscosity of Air
µA 1.983 × 10-8 Pa@s
Compressibility of Water
6W 1 × 10-8 Pa-1
Compressibility of Air
6A 1.777 × 10-5 Pa
Atmospheric Gas Phase Pressure
PgATM 101.3 kCa
gravitational Acceleration
g 9.8066 m/s2

102 102
1 1
10 10
0 0
10 10
Mass of PCE (kg)
Mass of TCE (kg)

-1 -1
10 10
-2 -2
10 10
-3 -3
10 10
total mass total mass
-4 -4
10 mass in water 10 mass in water
mass on soil mass on soil
mass in air mass in air
-5 -5
10 mass in napl 10 mass in napl
mass left boundary mass left boundary
-6 -6
10 10
102 102
1 1
10 10
0 0
10 10
Mass of DCE (kg)
Mass of VC (kg)
-1 -1
10 10
-2 -2
10 10
-3 -3
10 10
total mass total mass
-4 mass in water -4 mass in water
10 10
mass on soil mass on soil
mass in air mass in air
-5 -5
10 mass in napl 10 mass in napl
mass left boundary mass left boundary
-6 -6
10 -2 -1 0 1 10 -2 -1 0 1
10 10 10 10 10 10 10 10
Time (days) Time (days)

vaporization time.
a)
5 PCE conc.
elevation (m)
(kg/m3)
4 1
3 0.1
2
0.01
1
0.001
0
0.0001
0 5 10 15 20
b)
5 TCE conc.
elevation (m)
(kg/m3)
4
3 1
2 0.1
1 0.01
0 0.001
0.0001
0 5 10 15 20
c)
5 DCE conc.
elevation (m)
(kg/m3)
4
1
3
0.1
2
0.01
1
0 0.001
0.0001
0 5 10 15 20
d)
5 VC conc.
elevation (m)
(kg/m3)
4
3 1
2 0.1
1 0.01
0 0.001
0.0001
0 5 10 15 20
x-direction (m)
Figure 4.39 Concentration contours of (a) PCE, (b) TCE, (c) DCE, and (d) VC
at 0.1 days.
all phases present. A rectangular domain of 30 m x 20 m x 3 m around the
contaminated soils was set up for investigation of the effectiveness of NAPL removal
by air sparging. A steady-state air flow field is set up by injecting air in the
subsurface, which flows to the land surface which is under constant atmospheric
conditions. Flow parameters are discussed in Section 2.3C, Volume I of the
MODFLOW-SURFACT/MODHMS document, which also discusses the flow field
results. The radius of influence of the sparge is about 10 m at steady-state conditions,
at the well elevation. Table 4.10 shows the transport material parameter values for
the simulation. Phase properties of benzene are assumed to represent the gasoline
mixture which is simulated as a single component. The total transport simulation time
is 10 years. Figure 4.40 shows the gasoline saturations along a horizontal plane
through the well at 10 years of simulation. The gasoline has been completely removed
within a radius of about 5 m from the well. Figure 4.41 shows the mass loss of
gasoline in various phases in the subsurface, over the 10 year simulation period. Mass
in the NAPL phase decreases, as the domain is stripped of gasoline by the sparge
process. Diminishing returns are noted after 1.2 years of simulation, when NAPL is
stripped in the immediate vicinity of the sparge well. At this time, alternative
methods, or additional sparge wells would be more cost effective in cleaning up the
site.
4.4 LEVEL III: ACTUAL FIELD SIMULATION PROBLEMS
The third level of testing consists of two field simulation problems. These
problems are discussed in sections 4.4.1 and 4.4.2. The simulations include:
• Cross-sectional transport of chloride at the Borden site
• Three-dimensional transport of radionuclide from a waste disposal site.
Both simulation examples demonstrate the capability of MODFLOW-
SURFACT/MODHMS to handle site-specific conditions and produce simulation
results that are consistent with field observations.
4.4.1 Cross-Sectional Simulation of Chloride Transport at the Borden Site
The selected example concerns the unconfined aquifer system underneath the
Canadian Forces Base Borden Site in Ontario, Southern Canada. The site has been
extensively studied in a controlled subsurface environment. Frind and Hokkanen
(1987) provide details of the site characteristics, plume delineation, and
conceptualization of the system at the site. The case investigated here considers one
of the finite-element simulations reported by Frind and Hokkanen (1987) for
comparison of performance and accuracy with MODFLOW-SURFACT/MODHMS.
The study area is located near a landfill operated as a refuse dump from 1940 to 1976
(Figure 4.42). The landfill occupies an area of about 4 ha and is situated over an
unconfined sand aquifer that thins northward, in the direction of groundwater flow.
A chloride plume has been delineated to exist up to about 750 m north of the landfill
(Figure 4.42). Cross section (AAN) of the aquifer along the center line of the plume
is shown in Figure 4.43. Groundwater flow in the aquifer has been conceptualized by
Frind and Hokkanen (1987) using a homogeneous hydraulic conductivity, constant
lateral boundary water table levels, no-flow across the bottom boundary and a steady
but non-uniform recharge condition along the top of the domain. The finite-difference
grid and recharge distribution used in our MODFLOW-SURFACT/MODHMS
simulation are shown in Figures 4.44 and 4.45, respectively. Although this grid
seems complicated, it can be easily generated using any WINDOWS-based graphical
user interface customized for the MODFLOW-SURFACT/MODHMS modeling code.
Input data generated for the MODFLOW-SURFACT/MODHMS simulation are
presented in Appendix C.
A steady flow followed by transient transport simulation was performed with
MODFLOW-SURFACT/MODHMS. The flow and transport parameters of the model

Table 4.10 Material Parameter Values for Air Sparging Simulation of
Example 4.3.7.
Parameters Values
Longitudinal Dispersivity, "L 2m
Transverse Dispersivity, "T 0.2 m
Porosity, N 0.35
Molecular Diffusion Coefficient in water, Dw 1 x 10-9 m2
Molecular Diffusion Coefficient in air, Da 1 x 10-5 m2
Molecular Diffusion Coefficient in NAPL, Dn 1 x 10-9 m2
Equilibrium partitioning coefficient between gas 0.182
and water, Kgw
Equilibrium partitioning coefficient between 130
NAPL and water, Knw
Mass density of NAPL, Dn 800 kg/m3
Adsorption coefficient, kd 0 m3/kg
Decay coefficient, 8 0 d-1

Figure 4.40 Gasoline Saturations Along a Horizontal Plane Through the Well
at 10 Years.
Figure 4.41 Mass Loss of Gasoline from the Various Phases in the Subsurface
Over the 10 year Simulation Period.
Figure 4.42 Location of Borden Landfill Showing Monitoring Network. Cross-
section A-AN is along Longitudinal Plume Axis (from Frind and
Hokkanen, 1987).
Figure 4.43 Cross-section A-AN Showing Aquifer Depth and Location of
Monitoring Devices (from Frind et al., 1985).

Figure 4.44 Finite-Difference Grid Used in Borden Flow and Transport
Simulation.
Figure 4.45 Recharge Distribution Used for the Flow Simulation.
Table 4.11 Parameters for the Borden Site Groundwater Flow and
Chloride Transport Simulation.
Parameters Values
Upstream hydraulic head, hU 22.38 m

Downstream head, hD 18.25 m
Horizontal conductivity, Kxx 10.02 m/d
Vertical conductivity, Kzz 0.50 m/d
Effective porosity, N 0.38
Source concentration, Cs 250 mg/R
Longitudinal dispersivity, "L 10.0 m
Vertical dispersivity, "V 0.01 m
Molecular diffusion coefficient, Do 5 x 10-5 m2/d
Retardation Coefficient, R 1.0
Degradation Constant, 8 0.0

are provided in Table 4.11. A third-type contaminant flux condition was applied over
the landfill source. Hydraulic head distributions simulated by MODFLOW-
SURFACT/MODHMS and the finite-element model of Frind and Hokkanen (1987)
are compared in Figure 4.46. An excellent agreement between the two numerical
solutions may be observed. Shown in Figure 4.47 is a comparison of the simulated
and observed chloride plumes. Note that the observed concentration distribution
exhibits two local peaks downstream from the landfill. This behavior could not be
produced by the model because of its assumptions of single source and homogeneous
aquifer characteristics.
4.4.2 Three-dimensional Transport of Radionuclide at a Waste Disposal
Site
This example concerns the simulation of three-dimensional (3-D) groundwater
flow and radionuclide transport from a waste disposal site. Figure 4.48 shows a 3-D
visualization of the modeled domain representing a heterogeneous system with an
irregular topography. The depicted site area was used to bury low-level waste
containers, that have leached and contaminated the groundwater. Figure 4.49 shows
a schematic cross-section of the aquifer system underneath the site. The unconfined
aquifer beneath the facility is separated from the confined aquifer under it by a patchy
aquitard unit comprised of extensive clay lenses. Another aquitard unit separates this
semi-confined aquifer from the bottommost water producing unit which is underlain
by bedrock. Field characterization of the site has been performed using geophysical
and hydrogeological investigation techniques. Additionally, data fusion modeling
techniques (Jones and Gibbs, 1995) have been employed recently to quantify spatial
variability in hydraulic conductivity of the upper aquifers. Contaminant inventories,
leaching rates, and migration characteristics have also been measured or estimated at
the site.
To demonstrate 3-D flow and transport modeling capability of MODFLOW-
SURFACT/MODHMS, simulations were conducted using available data. The domain
was discretized into 27 columns, 27 rows, and 21 layers of nodes. Shown in Figure
4.48 is a 3-D view of the grid, vertically distorted to mimic the stratigraphy of the
aquifer system. The bottommost aquifer was assumed to be homogeneous with a
horizontal conductivity value of 40 ft/d and a horizontal to vertical anisotropy of
Kxx/Kzz=3. The upper units were treated as heterogeneous and anisotropic with
hydraulic conductivity values ranging from 10–5 ft/d to 40 ft/d and vertical to horizontal
anisotropies as large as 10 5. The groundwater flow system was assumed to be at steady-
state with lateral flow boundaries being constant head conditions supplied from field
observations and fusion modeling results. Surface boundary features modeled include
precipitation recharge with resultant seepage face conditions when water table elevations
reach the surface, and a stream running across the southern portion of the domain.
Owing to severe nonlinearities of the steady flow problem, a transient simulation
approach was used. Time marching was performed until a steady-state was reached.
Figure 4.50 shows simulated water-table profiles and surface water features. Water-
level elevations are above the top of the domain in the stream, and perched water exists
on top of the patchy low conductivity zones in the north-east of the domain. Figure
4.51 shows 3-D pathlines of particles released near the surface in the north of the
domain. The particle movement is typically down to the water table and south towards
the stream. These flow conditions were incorporated into the related transient
simulation of radionuclide (Tritium) transport from the disposal site for 78 years. Table
4.12 lists the transport parameters used in the simulation. To reflect the leaching
conditions in the field, the contaminant flux at the source was determined by taking the
product of constant volumetric fluid flux and time-dependent leaching concentrations.
Simulated concentration distributions are shown in Figures 4.52 and 4.53 at 27 and 78
years of simulation, respectively. The flow and transport results from MODFLOW-
SURFACT/MODHMS simulations compare favorably with the previous modeling study
using the VAM3DF finite-element model (Jones and Gibbs, 1995).

Figure 4.46 Comparison of Simulated Hydraulic Heads Produced by (a) Finite-
Element Model of Frind and Hokkanen (1987), and (b)
MODFLOW- Model.
Figure 4.47 Comparison Between (a) Observed Chloride Plume in 1979, and (b)
Simulated Plume at 40 Years.
Figure 4.48 Domain Discretization, Geometry and Horizontal Conductivity
Distribution for Simulation of Three-Dimensional Transport of
Radionuclide at a Waste Disposal Site.
Figure 4.49 Schematic Cross-Section of the Modeled Aquifer System for
Simulation of Three-dimensional Transport of Radionuclide at a
Waste Disposal Site.
Figure 4.50 Steady-State Water-Table Profiles Obtained from Simulation of 3-D
Groundwater Flow in the Aquifer System Underlying Waste
Disposal Site.
Figure 4.51 Particle Pathlines Obtained from Simulation of 3-D Groundwater
Flow in the Aquifer System Underlying Waste Disposal Site.
Table 4.12 Transport Parameters for Simulation of Three-Dimensional
Transport of Radionuclides at a Waste Disposal Site.
Parameters Values
Longitudinal dispersivity, "L 20 ft
Transverse dispersivity, "T 5 ft
Vertical dispersivity, "V 2 ft
Soil porosity, M 0.25
Diffusion coefficient, DE 1 × 10-6 ft2 d-1
Decay coefficient, 8 1.544 × 10-4 d-1
Adsorption coefficient, Rd 0
Numerical Options
Time discretization factor, THRTRD 0.5 (i.e., Crank-Nicolson time discretization)
Space discretization factor, IACLVL -2 (TVD scheme)
Non-linear iterations, MXITERC 2 (FCT scheme)

Figure 4.52 Concentration Distribution at 27 Years of Simulation of Transport
of Radionuclide at a Waste Disposal Site.
Figure 4.53 Concentration Distribution after 78 Years of Simulation of
Transport of Radionuclide at a Waste Disposal Site.
Anderson, D.V., 1983. ILUCG3: Subprograms for the solution of a linear
asymmetric matrix equation arising from a 7, 15, 19, or 27 point 3D
discretization, Computer Physics Communications, 30:43-49.
Anderson, D.V., 1983. ICG3: Subprograms for the solution of a linear symmetric
matrix equation arising from a 7, 15, 19, or 27 point 3D discretization,
Computer Physics Communications, 30:51-57.
Arminjon, P., and Dervieux, A., 1994. Construction of TVD-like artificial
viscosities on two-dimensional arbitrary FEM grid, J. Comp. Phys., 106:176-
198.
Bear, J., 1979. Hydraulics of Groundwater, McGraw-Hill, New York, 569 pp.
Bear, J., Corapcioglu, M.Y., and Balakrishna, J., 1984. Modeling of centrifugal
filtration in unsaturated deformable porous media, Advances in Water
Resources, 7:150-167.
Bear, J., and Pinder, G.F., 1978. Porous media deformation in multiphase flow,
Journal of the Engineering Mechanics Division, ASCE, 104:891-894.
Behie, G.A., and Vinsome, P.K.W., 1982. Block iterative methods for fully implicit
reservoir simulation, Soc. Petrol. Eng. J., Oct.:658-668.
Blunt, M., and Rubin, B., 1992. Implicit flux limiting schemes for petroleum
reservoir simulation, Journal of Computational Physics, 102:194-210.
Boris, J.P., and Book, D.L., 1973. Flux corrected transport, I, SHASTA, a fluid
transport algorithm that works, J. Comp. Phys., 11:38-39.
Brutsaert, W., 1964. The propagation of elastic waves in unconsolidated unsaturated
granular mediums, Jour. Geophys. Res., 69:243-257.
Burnett, R.D., and Frind, E.O., 1987. Simulation of contaminant transport in three
dimensions, 2. Dimensionality effects, Water Resources Research, 23(4):695-
705.
Coats, K.H., and Smith B.H., 1964. Dead-end pore volume and dispersion in porous
media, Soc. Petro. Engr. J., 4(4):73-93.
Cox, R.A., and Nishikawa, T., 1991. A new total variation diminishing scheme for
the solution of advective-dominant solute transport, Water Resour. Res.,
27(10):2645-2654.
Domenico, P.A., and Schwartz, F.W., 1990. Physical and Chemical Hydrogeology,
John Wiley and Sons, New York.
El-Kadi, A., 1988. Applying the USGS mass-transport model (MOC) to remedial
actions by recovery wells, Ground Water, 26(3):281-288.
Forsyth, P.A., 1994. Three-dimensional modeling of steam flush for
DNAPL site remediation, Int. J. Num. Meth. in Fluids, 19:1055-
1081.
Frind, E.O., Matanga, G.B., and Cherry J.A., 1995. The dual formulation for flow
for contaminant transport modeling, 2. The Borden aquifer, Water Resources
Res., 21:170-182.
Frind, E.O., and Hokkanen, G.E., 1987. Simulation of the borden plume using
alternating direction galerkin technique, Water Resourc. Res., 23(5):918-930.

Gelhar, L.W., and Collins M.A., 1971. General analysis of longitudinal dispersion
in nonuniform flow, Water Resources Research, 7(6):1511-1521.
Harten, A., 1983. High resolution schemes for hyperbolic conservation laws, Journal
of Computational Physics, 49:357-393.
Hoopes, J.A., and Harleman, D.R., 1967. Wastewater recharge and dispersion in
porous media, Journal of the Hydraulics Division, ASCE, 93( HY5):51-71.
Huyakorn, P.S., Kretschek, A.G., Broome, R.W., Mercer, J.W., and Lester, B.H.,
1984. Testing and Validation of Models for Simulating Solute Transport in
Groundwater: Development, Evaluation, and Comparison of Benchmark
Techniques. International Ground Water Modeling Center. HRI Report No.
35, Nov. 1984.
Jones, W.F., and Gibbs, B.P., 1995. Flow and Transport Model of the SRS Old
Burial Grounds, Using Data Fusion Modeling, Technical Report Submitted to
U.S. Department of Energy, Office of Technology Development, Coleman
Research Corporation, CMB/95-328, November, 1995.
Leake, S.A., Prudic, D.E., 1988. Documentation of a Computer Program to
Simulate Aquifer-System Compaction using the Modular Finite-Difference
Ground-water Flow MODEL, Techniques of Water-Resources Investigations
of the United States Geological Survey, Chapter A2, Book 6.
Lester, D.H., Jansen G., and Burkholder, H.C., 1975. Migration of radionuclide
chains through an adsorbing medium, AICHE Symposium Series, 71(152):202-
213.
Millington, R.J., and Quirk, J.M., 1961. Permeability of Porous Solids, Trans.
Faraday Society, 57:1200-1207.
Ogata, A., and Banks, R.B., 1961. A Solution of the Differential Equation of
Longitudinal Dispersion in Porous Media, U.S. Geological Survey
Professional Paper No. 411-A.
Panday, S., Huyakorn, P.S., Therrien, R., and Nichols, R.L., 1993. Improved three-
dimensional finite-element techniques for field simulation of variably saturated
flow and transport, Journal of Contaminant Hydrology, 12:3-33.
Panday, S., and Corapcioglu, M.Y., 1991. Solute rejection in freezing soils, Water
Resources Research, 27(1):99-108.
Perlmutter, N.M., and Lieber, M., 1970. Dispersal of Plating Wastes and Sewage
Contaminants in Groundwater and Surface Water, South Famingdale-
Massapequa Area, Nassau County, New York, U.S. Geological Survey Water
Supply Paper 1879-G.
Reeves, M., and Cranwell, R.M., 1981. User’s Manual for the Sandia-Waste
Isolation Flow and Transport Model, (SWIFT) 4-81, NUREG/CR-2324, U.S.
Nuclear Regulatory Commission, 145 pp.
Roe, P.L., 1981. Approximate Riemann solvers, parameter vectors, and difference
schemes, Journal of Computational Physics, 43:357-372.
Scheidegger, A.E., 1961. General theory of dispersion in porous media, Jour. of
Geophysical Research, 66:3273-3278.
Sweby, P.K., 1984. High resolution schemes using flux limiters for hyperbolic
conservation laws, Siam J. Number. Anal., 21(5):995-1011.
5-2 Chapter 5: References

Unger, A.J.A., Forsyth, P.A., and Sudicky, E.A., 1996. Variable spatial and
temporal weighting schemes for use in multi-phase compositional problems,
Advances in Water Resources, 19(1):1-27.
van Leer, B., 1977. Towards the ultimate conservative difference scheme. IV. A
new approach to numerical convection, Journal of Computational Physics,
23:276-299.
van Leer, B., 1979. Towards the ultimate conservative difference scheme. V. A
second-order sequel to Godunov’s method, Journal of Computational Physics,
32:101-136.
Wilson, J.L., and Miller, P.J., 1978. Two-dimensional plume in uniform ground-
water flow, Journal of the Hydraulics Division, ASCE, 104(HY4):503-514.
Yee, H.C., 1987. Construction of explicit and implicit symmetric TVD schemes and
their applications, Journal of Computational Physics, 68:151-179.
Zheng, C., 1990. A Modular Three-Dimensional Transport Model for Simulation of
Advection, Dispersion and Chemical Reactions of Contaminants in
Groundwater Systems, prepared for USEPA Kerr Environmental Research
Laboratory, Ada, OK, 74820.

CHAPTER 6: ADDENDUM FOR DUAL DOMAIN SIMULATIONS
6.0 Addendum For Dual Domain Simulations . . . . . . . . . . . . . . . . . 6-1
6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-1
6.2 Formulation of The Dual Domain Equations . . . . . . . . . . 6-3
6.3 Numerical Implementation . . . . . . . . . . . . . . . . . . . . . . 6-5
6.4 Input/output Considerations . . . . . . . . . . . . . . . . . . . . . 6-7
6.5 Verification Example . . . . . . . . . . . . . . . . . . . . . . . . . 6-8
6.6 Input Instructions For Dual Domain Simulations . . . . . . . 6-10
6.6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 6-10
6.6.2 Basic Flow (BAS Package Input) . . . . . . . . . . . . 6-10
6.6.3 BCF4 Input . . . . . . . . . . . . . . . . . . . . . . . . . . 6-17
6.6.4 Basic Transport (BTN1) Package Input . . . . . . . . 6-28
6.7 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-39
6.0 ADDENDUM FOR DUAL DOMAIN SIMULATIONS
6.1 INTRODUCTION
Contaminant transport behavior as expressed by equation (1) of Chapter 2,
Volume II (within a steady-state or transient flow field of equation (1) Volume I), misses
an important significant mechanism—that of a dual domain—which produces the highly
nonsigmoidal and asymmetrical distributions of tailing of contaminant noticed in several
experimental studies as well as in observations of contaminant breakthrough in the field.
The dual domain mechanism is significant, for several conditions listed by van Genuchten
and Wierenga (1976). Van Genuchten (1981) showed that the same general equation
applied irrespective of how the processes or mechanisms are conceptualized. Thus, the
same set of dimensionless variables represent a) linear equilibrium adsorption, physical
kinetic model; b) physical kinetic model in presence of anion exclusion; c) 2-site kinetic
adsorption model; and d) 1-site kinetic adsorption model. The term ‘dual domain’
therefore represents retarding mechanisms involving both equilibrium and kinetic
adsorption processes, without explicitly defining the mechanism.
Figure 6.1 shows some physical situations where the dual domain concept may
contribute significantly to transport behavior. Figure 6.1a shows the situation of an
aggregated soil wherein flow and advective-dispersive solute transport occur in the mobile
domain, but no flow occurs in or out of the aggregated domain (soil clumps). Diffusion
is the kinetic process which transfers contaminant mass in and out of the aggregates.
Equilibrium adsorption onto soil grains occurs on soil within the mobile domain, as well
as on soil within the aggregates. Figure 6.1b shows that water in dead-end regions, and
immediately surrounding soil grains in a porous medium is immobile. The kinetic process
of diffusion transfers contaminant mass from the mobile region to the immobile region
(and vice versa), from which it can adsorb (assuming equilibrium conditions) onto soil
grains. Figure 6.1c shows a fractured medium system, for which these exists extensive
literature involving the dual domain conceptualization, along with its own terminology.
Flow and advective-dispersive solute transport occur through the fracture domain, with
diffusion of contaminant into the matrix. HydroGeoLogic (1992) correlates the kinetic
mass transfer rate between domains, to the effective diffusion coefficient of the
contaminant within the matrix system and the surface area to volume ratio of the matrix
(or aggregates of Figure 6.1a). They further simplify these equations for cubic matrix
block or spherical matrix block assumptions with a cubic packing. Equilibrium adsorption
occurs on the soil within the matrix as well as along the fracture walls in the mobile
domain. Figure 6.1d shows the situation where flow and advective-dispersive transport
occurs through an extensive network of macropores, earthworm channels or solution
channels, with diffusive mass transfer into the adjacent medium. Equilibrium adsorption
occurs along channel walls, and on soil in the immobile region within the medium. Figure
6.1e shows the conceptual representation of all these dual region systems. Both mobile
and immobile domains consist of soil, and voids (filled with water). The mobile fraction,
f (equivalent to ‘fracture porosity’ of a dual porosity system), is the fraction of the total
domain that lies in the mobile region (i.e., f=Vm/VT where Vm is the mobile domain
volume and VT is the total domain volume). The immobile fraction, 1-f, is therefore the
ratio of the volume of the immobile domain (Vim) to total volume. The porosity within the
mobile domain, Nm, is the volume of voids in the mobile domain, Vvm, divided by the
volume of the mobile domain, Vm. If, further, the water saturation in these
Chapter 6: Addendum For Dual Domain Simulations 6-1

M:\Projects\MODFLOW Manuals\SURFACT and HMS Ver.3\Vol II\WordPerfect Files\Chapter_6 new.wpd
Figure 6.1 Examples and schematic of dual domain system.
6-2 Chapter 6: Addendum For Dual Domain Simulations

voids is unity, Nm=Vwm/Vm, where Vwm is the volume of water in the mobile domain.
Analogously, and assuming that the immobile domain is fully saturated with water, we
have the porosity of the immobile domain as Nim=Vwim/Vim, where Vwim is the volume of
water in the immobile domain.
6.2 FORMULATION OF THE DUAL DOMAIN EQUATIONS

In a dual domain system, equation (1) of Volume I describes the variably-saturated
transient flow of water through the mobile domain (of volume Vm) with no flow occurring
into or through the immobile domain (of volume Vim). The equation can be cast in terms
of the total system volume (from the mobile domain volume) by multiplying equation (1)
by the mobile fraction, f, to give
∂ ⎛ ∂h ⎞ ∂ ⎛ ∂h ⎞ ∂ ⎛ ∂h ⎞
⎜ Κ xx k rw ⎟ + ⎜⎜ Κ yy k rw ⎟⎟ + ⎜ Κ zz k rw ⎟ − Wf
∂x ⎝ ∂x ⎠ ∂y ⎝ ∂y ⎠ ∂z ⎝ ∂z ⎠
(1)
∂S ∂h
= fφ m w + fS w S s
∂t ∂t
where
Κ ii = fΚ ii (ii = xx, yy, or zz ) (2)
is the hydraulic conductivity of the total system, as may be measured by pump tests in the
field, and Kii is the conductivity of the mobile domain (e.g., conductivity of a fracture if
theoretically calculated for fracture/matrix systems from fracture aperture and some flow
theory like Hagen-Poiseuille laminar flow, or conductivity of a soil sample from the
mobile region outside of the aggregates if measured in the laboratory). The other terms
are defined after equation (1) of Volume I. It should be noted that the saturation Sw above
them denotes the saturation within the mobile region (Vw/Vvm), and Ss is the specific storage
of the (soil within the) mobile region. No water flows into or through the immobile
region, thus a flow equation for the immobile region is not required. The Darcy flux
components, qi, per unit total volume may be computed from Darcy’s law as
∂h
qi = Κ ij k rw (3)
∂x j
MODFLOW-SURFACT/MODHMS has the option of allowing air to be the active
modeled phase instead of water. In that case, equation (1) may be expressed for air
instead of water, following the derivations in section 2.2c for solving the air flow
equations within the MODFLOW framework.
For the dual domain system, the transport equation for contaminants in the active
phase " within the mobile domain may be written per unit total volume as
∂ ⎛ k ∂cmk ⎞ ∂ ∂ ∂
∂xi
⎜ Dij
⎜
⎟−
∂x j ⎟⎠ ∂xi
qi cmk =( ∂t
)
fφ m Sαm cmk +(∂t
fρ Βm csm
k
) ( )
⎝
+ fλαk φ m Sαm cmk + fλκs ρ Βm csm k − qck* + fΓ k (4)
NPAR NPAR
− ∑ ξ kj fλαj φm Sαm cmj −
j =1
∑ξ
j =1
kj fλsj ρ Βm csmj + Rmk →im − Rimk →m

where the terminology has been defined after equation (1) of Volume II, f and qi are
defined above, and all terms with subscript m refer to the respective values and properties
in the mobile domain. The decay rate is assumed to be the same for both immobile and
mobile domains, therefore, this term is not individually subscripted. Rmk →im is the rate
of mass transfer of contaminant k from the mobile domain to the immobile domain per unit
total volume, and Rimk →m is the mass transfer rate per unit total volume of contaminant
k from the immobile to the mobile domain. The mass balance equation for contaminant
k within the immobile domain may be written per unit total volume as
∂
{1 − f }φ im cimk + ∂ {1 − f }ρ Βim c sim
( ) k
( + {1 − f }λαk φ im cim
k
)
∂t ∂t
NPAR
+ {1 − f }λ ρ k
s
k
Βim sim c − ∑ ξ {1 − f }λα φ
kj
j
c j
im im (5)
j =1
NPAR
− ∑ ξ {1 − f }λ
j =1
kj
j
s ρ Βim c sim
j
− Rmk →im + Rimk →m = 0
where it is assumed that the active phase totally occupies the void spaces within the
immobile domain (S"im=1). Effective properties for Nim and DBim may be used if S"im…1.
Equilibrium adsorption in the immobile domain is still considered to be mediated through
the water phase, when the active phase is air. Further, in consistency with the assumption
of no flow into or out of the immobile domain of the active phase, the immobile domain
is incompressible, and Nim and f remain constant. The rate of mass transfer from the
mobile domain depends on the mobile domain concentration and is expressed by a first
order kinetic term
Rmk →im = α mk →im c mk (6)
where α m→
k -1
im (t ) is the first order kinetic rate constant. Similarly, the rate of mass
transfer from the immobile domain depends on its concentration and is expressed by a first
order kinetic term as
Rimk →m = α imk →m cim

k
(7)
where α im→
k
m (t-1) is the first order kinetic rate constant. In (6) and (7) it is assumed that
the mass transfer for each contaminant k between the two domains is not affected by the
presence of the other contaminants, thus α im→
k
m α m→
and
k
im are constants. At
equilibrium, we have
k
cim = c mk and Rmk →im = Rimk →m , thus α imk →m = α mk →im = α k
which is the first order kinetic constant for mass transfer of contaminant k between
domains. Equation (6) minus Equation (7) then gives
Rmk →im − Rimk →m = α k c mk − cim

k
( ) (8)
Equation (8) may be used to replace the last two terms in equations (4) and (5) to
give two equations in the two unknown concentrations cmk and ckim for any contaminant
component k. The mass transfer rates coefficient may be related to the effective diffusivity
of contaminant within the immobile domain as

∂cim
α (cmk − cimk ) = Dim* (1 − f )σ (9)
∂χ
*
where Dim is the effective diffusion within the immobile domain, P the length dimension
within the immobile domain, and F is the specific surface between the two domains
defined as the area of contact divided by the volume of the immobile domain. Thus, for
a rectangular slab representation of the immobile domain, we have (HydroGeoLogic,
1992).
f = b / (b + b′) (10)
and
σ = 2 / b′ (11)
where b and bN are the dimensions of the mobile and immobile domains respectively (see
Figure 6.2). Using (10) and (11) in (9) and rearranging gives " for a rectangular slab
block representation of the immobile domain as
α = 4 Dim* / b ′ 2 (12)
Similarly, for spherical blocks with a cubic packing we have
f = b / (2a + b ) (13)
and
3
σ = (14)
a
where Figure 6.2 defines the dimensions. Using (12) and (13) in (9) yields
α = 3Dim* / a 2 (15)
assuming that the dimension over which the diffusion is measured is from the block
surface to its center. These relations may be used to estimate " when field data is lacking.
6.3 NUMERICAL IMPLEMENTATION

The flow equation for the dual domain system is identical to that of the single
continuum representation, with the right hand side of equation (1) containing the mobile
fraction, f, which may be multiplied into the primary and secondary storage terms after
they are read. The transport equations (4) and (5), with equation (8) providing the mass
transfer are discretized into two equations in two unknowns (for each component),
k k
cmk and cim
k
(Note that c sm and c sim may be obtained from the equilibrium Freundlich
isotherms for the two domains). In general, the discretized equations may be represented
as
A1k c k + B1k cim

k
= R1k (16)

Figure 6.2 Typical matrix block fracture units: (a) slab matrix block
representation, and (b) spherical matrix block representation.

A2k c k + B2k cim
k
= R2k (17)
where B1k , A2k , and B2k are diagonal matrices since there is no flow in the immobile
domain, and C ( = C m )
k k
represents the concentration in the mobile domain.
Manipulating equation (17) gives
R2k A2k k
c k
im = k − kc (18)
B2 B2
which may be inserted into (16) to give
⎡ k A2k ⎤ k ⎡ k R2k k ⎤
⎢ A1 − k ⎥ c = ⎢ R1 − k B1 ⎥ (19)
⎣ B2 ⎦ ⎣ B2 ⎦
The matrix and right hand side vector may be manipulated according to equation
k
(19) before each solve for c k (at any time step or iteration). cim may then be obtained
from back substitution of the calculated ck into equation (18). Note that the second term
of equation (5) is expanded using the modified-Picard method due to the non-linear
retardation term, as is done for the second term on the right hand side of equation (4) for
the mobile domain.
6.4 INPUT/OUTPUT CONSIDERATIONS

The dual domain system requires additional input to define the two domains. A
switch IDUAL is read in the basic input file (.BAS) to activate dual domain simulations.
This switch is further used to define and adsorption options within the immobile domain.
Thus, when IDUAL=0, a single continuum model is used as provided in Chapters 1 of
Volumes I and II. For IDUAL > 0, a dual continuum model is used. For these cases,
the fraction of the total domain occupied by the mobile domain, f, is also required as input.
This is provided in the block centered flow (.BCF) input file. Now for transport
considerations, if IDUAL=1, no equilibrium adsorption is assumed for the immobile
domain. Hence, the mass transfer rate term ", and the porosity within the immobile
domain, Nim are required as input and are provided in the basic transport (.BTN) input
package. The porosity read therein in the array PHI is now considered to be the initial
porosity of the mobile domain, Nm. Note that equation (25) of Chapter 2, Volume II,
which defines the compressibility, is applicable only to the mobile domain (consistent with
the flow assumptions) therefore the porosity Nm may change in a transient flow situation
due to compressibility effects. If IDUAL=2, then equilibrium adsorption occurs in the
immobile domain with the same isotherm constant (K) [and exponent values for non-linear
cases] as for the mobile domain. No additional input is therefore required to define
adsorption in the immobile domain, than for the case of IDUAL=1. If IDUAL=3 then
equilibrium adsorption in the immobile domain is different from the mobile domain, and
K imk (and η im for non-linear cases) are required as additional inputs from the basic
k
thus
transport (.BTN) input package for dual domain simulations. The retardation index,
LINR, governs the type of retardation occurring in both domains. Thus if LINR=1 then
retardation is linear, and if LINR=2 then retardation is non-linear in both domains.
Finally, the initial concentrations of each component k in the immobile domain are also
requested as input from the .BTN file for dual domain simulations (IDUAL > 0).

Additional output for dual domain simulations includes concentrations of the
contaminants in the active phase within the immobile region, as well as the mass balance
components of storage, decay, and generation of contaminants in the immobile domain.
Mass transfer rates (and cumulative masses) are also reported in the mass balance output.

The dual domain simulation capability is verified by comparing numerical results
with the analytical solution of van Genuchten and Wierenga (1976). The problem
considered involves calculating the effluent curves for a 30 cm long soil column (L = 30
cm) leached with 3 pore volumes of a chemical (T1=qt/{L [fNm + (1-f)Nim]}) at a constant
flux q of 10 cm/day. Table 6.1 provides the material properties used in the simulations.
The soil column is discretized into 1 layer, 1 row, and 30 columns of dimensions 1 cm x
1 cm x 1 cm each. The fully saturated, steady-state flow field is setup by providing
constant head conditions at each end of the column, of 10 cm and 9.7 cm, with an effective
conductivity of the medium of 1,000 cm/d to provide the q=10 cm/d flux through the
column.
Van Genuchten and Wierenga (1976) express their dual domain equations with
different parameter definitions from those provided here. These definitions are presented
here to show the parameter conversions between the systems of equations. They define
2m and 2im as the fractions of the soil filled with mobile and stagnant water, respectively.
Thus, referring to Figure 6.1e, we have
θ m = Vwm / VT (20)
and
θ im = Vwim / VT (21)
Thus, 2, the fraction of soil filled with (mobile plus immobile) water is given as
θ = θ m + θ im (22)
They further define the fraction of the total water that is mobile,
Φ = θ m /θ , (23)
as a dimensionless parameter for which numerical experiments are performed. Van

Genuchten and Wierenga (1976) then define the soils conditions in terms of 2, f, and M.
Using (23), we therefore get
θ m = Φθ (24)
and using (22) and (24) we have
θm
θ im = − θ m = θ − Φθ (25)
Φ
Further, by definition we have [See Figure 6.1e]
fφ m S m = θ m (26)
and
(1 − f )φ im = θ im (27)

Table 6.1 Material properties for dual domain column simulations
Parameter Value
Constant flux through soil column, q 10 cm3/cm2/d
Fraction of total water that is mobile, M 0.65
Fraction of domain filled with water, 2 0.4
Mobile fraction of domain, f 0.4
Porosity within mobile domain, Nm (=M2/f) 0.65
Porosity within immobile domain, Nim (={2-M2}/{1-f}) 0.23333
Index IDUAL indicating that equilibrium adsorption 2
characteristics of both mobile and immobile domain are
the same
Index LINR indicating linear equilibrium adsorption 1
Bulk density of soil, DB (=DBim) 1.3 g/cm3
Linear equilibrium adsorption coefficient, kd (=kdim) 0.5 cm3/g
Diffusion coefficient with tortuosity, D 30 cm2/d
Effective Diffusion coefficient for dual domain, Deff 30 cm2/d*0@65*0@4=7.8 cm2/d
(=2mD=M2D)
Bulk diffusion coefficient, Do (=D/{Sw10/3 Nm})
12 cm2/d
Dimensionless loading time
3 pore volumes
qt
T1
L[φ m f + φ im (1 − f )]
Time of source release, t 3.6 days
First order, kinetic mass transfer coefficient, ":
case 1, no mass transfer =0
case 2, slight mass transfer = 0.15
= 100
case 3, very large mass transfer

Substituting (24) into (26) and rearranging gives (for Sm=1 i.e., a saturated system)
Φθ
φm = (28)
f
Substituting (25) into (27) and rearranging gives
(θ − Φθ )
φim = (29)
(1 − f )
Equations (28) and (29) may be used to provide the parameters used in
MODFLOW-SURFACT/MODHMS (Nm, Nm f) from those of van Genuchten and
Wierenga (2, f, M).
Results of the simulations are provided in Figure 6.3, for breakthrough of
contaminant at the downstream end (L=30 cm) of the soil column. The numerical results
compare well with curves from the analytical solution of van Genuchten and Wierenga
(1976). The influence of the mass transfer coefficient " is examined by studying
breakthrough for three different values of ". When "=0, there is no diffusion of
contaminant into the immobile domain and the effluent curve acquires a more or less
symmetrical shape. For small values of ", there is a slow exchange of contaminant
between mobile and immobile domains causing significant decrease in peak concentration,
and considerable tailing. With large values of the mass transfer coefficient, the rate of
exchange between the mobile and immobile domains increases eventually leading to
equilibrium conditions between the two domains, where Cm=Cim at any time. With the
lack of kinetic effects, the effluent curve reverts to a more or less symmetrical shape with
peak concentrations approaching unity. Breakthrough is delayed for this case, as
compared to the equilibrium case of "=0, due to the larger total pore volume encountered
by the solutes.
6.6 INPUT INSTRUCTIONS FOR DUAL DOMAIN SIMULATIONS

6.6.1 Introduction
Additional input required for dual domain simulations are supplied via three
MODFLOW-SURFACT/MODHMS packages: BAS, BCF, and BTN. The required input
is logically incorporated into the simulation via these three files. The basic flow package
input (BAS) provides the simulation option switch, IDUAL, which indicates whether a
dual domain simulation is to be performed. If this index is on, then the fraction of the
domain which lies within the mobile region, f, is provided via the block centered flow
package input (BCF). The subsequent transport simulation also requires Nim, the porosity
within the immobile domain along with Nm already supplied by the basic transport input
package. Depending on the equilibrium adsorption options, the Freundlich adsorption
coefficient and exponent for the immobile domain may also be required as input. Decay
in the mobile and immobile domains is considered to occur at the same rate, hence, decay
coefficients in soil and water read in for the mobile domain, apply to the immobile domain
as well. Initial concentrations within the immobile domain are also supplied via the BTN
input package for dual domain simulations. Since all other input for dual domain
simulations is the same as provide in Volumes I and II of the MODFLOW-
SURFACT/MODHMS document, only the input instructions for the BAS, BCF, and BTN
packages are provided below, describing the additional input requirements.
6.6.2 Basic Flow (BAS Package Input)
Input for the BAS package is read from Unit 1 as specified in the main program
of MODFLOW. The BAS package has been modified to require additional flags
indicating that a dual domain transport simulation is requested and to open three additional

Chapter 6: Addendum For Dual Domain Simulations
1.0
Analytical Solution
(van Genuchten and
Wierenga, 1976)
.8 Numerical Solution
α = 0.0
α = 0.15
Relative Concentration, cm
α =∞
.6
.4
α =∞
.2 α = 0.15
α = 0.0
0
0 1 2 3 4 5 6 7
Pore Volume -T
6-11
Figure 6.3 Effluent curves for a sorbing medium showing a comparison of numerical and analytical solutions for mass
transfer coefficient (") values to 0.0, 0.15, and 4.
files that are needed. These flags correspond to the unit numbers on IUNIT(17),
IUNIT(23), and IUNIT(24). The flag ITRAN indicates the transport simulation
requirements. If the flag is zero, no transport simulation is required and only a flow
simulation is performed. If the flag is unity, a dual flow-and-transport simulation is
required. The flag IDUAL indicates that dual domain transport simulations will be
performed (if IDUAL > 0). Note that the user may turn off the flag ITRAN at any stage
of the modeling study to perform the flow simulation only. No other data need be
disturbed. The capability for performing steady-state flow analysis for some stress periods
and transient analysis for others, has been included from a similar concept used in
MODFLOW 2000. The entire input data structure is presented here for the BAS package
for completeness. The additional variables required for transport analysis are shown in
shaded boxes.
1. Data: HEADNG(32)
Format: 20A4
2. Data: HEADNG (continued)
Format: 12A4
3. Data: NLAY NROW NCOL NPER ITMUNI ITRAN IDUAL MPHFLAG

Format: I10 I10 I10 I10 I10 I10 I10 I10
4. Data: IUNIT(60)
Format: 60I3
(BCF WEL DRN RIV EVT TLK GHB RSF SIP DE4 SOR OC PCG
STR PG2 HFB BTN LAK2 IBS CHD FWL4 ATO PCN HCN OLF CHF PRL
OBS GFD FHB IPT RCT SLC LMG PCG5 RES FWL5 SFR1 LAK3 GAGE
ETS1)
5. Data: IAPART ISTRT

Format: I10 I10
6. Data: IBOUND(NCOL, NROW)

Utility Module: U2DINT. Note that U2DINT is the utility module that reads
a two-dimensional integer array (i.e., IBOUND for item 6).
7. Data: HNOFLO DATUM

Format: F10.3 F10.3
8. Data: SHEAD(NCOL, NROW)

**Enter item 9 for each stress period.

9. PERLEN NSTP TSMULT SS/TR
F10.0 I10 F10.0 8X, A2
HEADNG -- is the simulation title that is printed on the printout. It may be up to 128
characters long; 80 in the first record and 48 in the second. Both
records must be included even if they are blank.
NLAY -- is the number of model layers.
NROW -- is the number of model rows.
NCOL -- is the number of model columns.
NPER -- is the number of stress periods in the simulation.
ITMUNI -- indicates the time unit of model data. (It is used only for printout of
elapsed simulation time. It does not affect model calculations.)
0 - undefined 3 - hours
1 - seconds 4 - days
2 - minutes 5 - years
The unit of time must be consistent for all data values that involve time.
For example, if years is the chosen time unit, stress-period length, time-
step length, transmissivity, etc., must all be expressed using years for
their time units. Likewise, the length unit must also be consistent.
ITRAN -- is a flag that determines if the simulation is either flow only or flow
and transport.
If ITRAN = 0, the simulation is for flow only.
If ITRAN = 1, the simulation is for both flow and transport.
IDUAL -- is the switch indicating if a dual domain transport simulation

is required.
If IDUAL=0, the simulation is for a single domain continuum as
expressed in Volume II of the MODFLOW-SURFACT/MODHMS
document.
If IDUAL=1, the simulation is for a dual domain representation
with no equilibrium adsorption occurring in the immobile domain.
If IDUAL=2, the simulation is for a dual domain representation
with equilibrium adsorption characteristics of the immobile domain
being equal to that of the mobile domain at any node in the system.
IF IDUAL=3, the simulation is for a dual domain representation
with equilibrium adsorption characteristics of the immobile domain
being distinct from those of the mobile domain, requiring additional
input of these properties.
MPHFLAG -- When MPHFLAG > 0 and ISS = 0, the ATO4 package is used,
MODFLOW-SURFACT/MODHMS writes out a binary file x.mph,
which is used by MODPATH to determine time step information.

IUNIT -- is a 60-element table of input units for use by all major options. The
additional packages required for flow and transport simulation are shown
in shaded boxes. Note that the Preconditioned Conjugate Gradient
(PCG4) solver specified in IUNIT(13) or the PCG5 solver specified in
IUNIT (35) must be used for the transport simulation (i.e.,
ITRAN=1). For the flow simulation, any solver including (SIP,
SSOR, or PCG2), may be used.
IUNIT LOCATION MAJOR OPTION

1 Block-Centered Flow (BCF4) Package
2 Well (WEL) Package
3 Drain (DRN) Package
4 River (RIV) Package
5 Evapotranspiration (EVT) Package
6 Transient Leakage Package (TLK)
7 General-Head-Boundary (GHB) Package
8 Recharge-Seepage Face Boundary (RSF4)
Package
9 Strongly Implicit Procedure (SIP) Package
10 Direct Solver (DE4) Package
11 Slice Successive Over-Relaxation (SSOR)
Package
12 Output Control (OC) Package
13 Preconditioned Conjugate Gradient

(PCG4) Package
14 Streamflow-Routing (STR) Package
15 Preconditioned Conjugate Gradient
(PCG2) Package
16 Horizontal-Flow Barrier (HFB) Package
17 Basic Transport (BTN1) Package

18 Lake (LAK2) Package
19 Interbed-Storage (IBS) Package
20 Time-Variant Specified-Head (CHD)
Package
21 Fractured-Well (FWL4) Package
22 Adaptive Time-Stepping and Output
Control (ATO4) Package

23 Prescribed-Concentration Boundary
(PCN1) Package
24 Specified-head Concentration Boundary

(HCN1) Package
25 Overland Flow Package (OLF)1
26 Channel Flow Package (CHF)1
27 Analytical Infiltration Package (PRL)1
28 Observation Node Package (OBS)
29 General Finite Differences (GFD)
30 Flow and Head (time-varying) Boundary
Package (FHB)
31 Interception of Precipitation and
Evapotranspiration Package (IPT)1
32 Reactive Transport Package (RCT)
33 Saltchem Module (SLC)
34 Link - AMG Package (LMG)2
35 Preconditioned Conjugate Gradient 5
Package (PCG5)
36 Reservoir Package (RES)
37 Multi-Node Fracture Well Package
(FWL5)
38 Steam-Flow Routing Package (SFR1)
39 Lake 3 Package (LAK3)
40 Gage Package (GAGE)
41 Evapotranspiration Segments Package
(ETS1)
1. Note: OLF, CHF, PRL and IPT packages are not available with MODFLOW-
SURFACT/MODHMS.
2. Note: LMG Package links are provided, however the AMG solver is not available.
If IUNIT(n) # 0, the corresponding major option is not being used.
If IUNIT(n) > 0, the corresponding major option is being used and data
for that option will be read from the unit number
contained in IUNIT(n). The unit numbers in IUNIT
should be integers from 1 to 98. It is recommended that
a different number be used for each major option.
Printer output is assigned to unit 6 (unless it is changed
to meet computer requirements). That unit number
should not be used for any other input or output. The

user is also permitted to assign unit numbers for output.
Those numbers should be different from those assigned
to input. The Basic Package reads input from unit 1
(unless it is changed to meet computer requirements).
It is permissible but unwise to use that unit for other
major options. Note that if both IUNIT(12) and
IUNIT(22) are greater than zero, IUNIT(12) is set equal
to zero. (Refer to OC Package input or ATO4 Package
input for details.)
IAPART -- indicates whether array BUFF is separate from array RHS.
If IAPART = 0, the arrays BUFF and RHS occupy the same space. This
option conserves space. This option should be used
unless some other package explicitly says otherwise.
If IAPART … 0, the arrays BUFF and RHS occupy different space. This
option must be used if input LAYCON value of any model
layer is either 31, 33, 40, or 43. Note that LAYCON
values for layers are input in item 2 of the BCF4
Package.
ISTRT -- indicates whether starting heads are to be saved. If they are saved, they
will be stored in array STRT. They must be saved if drawdown is
calculated.
If ISTRT = 0, starting heads are not saved.
If ISTRT … 0, starting heads are saved.
IBOUND -- is the boundary array.
If IBOUND(J,I,K) < 0, cell J,I,K has a constant head where J,I,K are the
column, row, layer numbers of the cell,
respectively.
If IBOUND(J,I,K) = 0, cell J,I,K is inactive.
If IBOUND(J,I,K) > 0, cell J,I,K is active.
HNOFLO -- is the value of head to be assigned to all inactive cells (IBOUND = 0)
throughout the simulation. Since heads at inactive cells are unused, this
does not affect model results but serves to identify inactive cells when
head is printed. This value is also used as drawdown at inactive cells if
the drawdown option is used. Even if the user does not anticipate having
inactive cells, a value for HNOFLO must be submitted.
DATUM -- is a shift value applied to the datum of all head and elevation
calculations. This amount is subtracted from all head and elevation
values internally, to minimize computational round-off errors. Note
that heads and elevations of a simulation are reported with respect
to the original datum and this entry is only used internally for
scaling purposes.
SHEAD -- is head at start of the simulation. Regardless of whether starting head is
saved, these values must be input to initialize the solution.
PERLEN -- is the length of a stress period. It is specified for each stress period.

NSTP -- is the number of time steps in a stress period. Note that if ATO4
Package is used (i.e., IUNIT(22) > 2), the number of time steps in a
stress period are determined by the code.
TSMULT -- is the multiplier for the length of successive time steps. The length of the
first time step DELT(1) is related to PERLEN, NSTP, and TSMULT by
the relation
DELT(1) = PERLEN(1-TSMULT) / (1-TSMULT**NSTP).
Note that if ATO4 Package is used (i.e., IUNIT(22) > 0), TSMULT value
read from BAS input will be overwritten by that read from ATO4 Package.
SS/TR -- is a character variable that indicates if the stress period is transient or
stead state. The only allowed options are ‘SS’ and ‘TR’. If left blank,
the entire simulation proceeds in steady or transient mode according to
flag ISS in the BCF file. Note that if this option is used, the flag ISS
should be set to zero in the BCF file.
6.6.3 BCF4 Input
Input for the BCF4 Package is read from the unit specified in IUNIT (1). A brief
explanation of the input variables is provided in the next subsection. The data are read
once at the beginning of the simulation. IUNIT array, which is read by the Basic Package,
determines the options or packages used in a MODFLOW simulation. (See Basic Package
of the original MODFLOW document for details.) Note that the extra input variable
required by BCF4 is shown in a shaded box.
1. Data: ISS IBCFCB HDRY IWDFLG WETFCT IWETIT IHDWET IAXIS IREALSL ICNTRL IVHYC IANIXY ICURVL IOLDHDR
Format: I10 I10 F10.0 I10 F10.0 I10 I10 I10 I10 I10 I10 I10 I10 I10
2. Data: LAYCON (NLAY) (Maximum of 80 layers)

Format: 40I2
(If there are 40 or fewer layers, use one record; otherwise, use two records.)
3. Data: RHOWP RHOAP VISW VISG COMPWAT COMPAIR ATMGP GRAV INITW
Format: E10.3 E10.3 E10.3 E10.3 E10.3 E10.3 E10.3 E10.3 I10
**Enter item 4a only if IANIXY=0
4a. Data: TRPY (NLAY)

the real one-dimensional (1-D) array (i.e., TRPY
for item 3).
**Enter item 4b only if IANIXY=1

4b. Data: TRPYNU (NCOL, NROW)
Utility Module: U2DREL. Note that item 4b is entered once for each layer of
the model grid.
**Enter item 5a only if ICURVL =0

5a. Data: DELR (NCOL)
**Enter item 5b only if ICURVL = 1

5b. Data: DELRCV (NCOL, NROW)
**Enter item 6a only if ICURVL = 0
6a. Data: DELC (NROW)

**Enter item 6b only if ICURVL = 1

6b. Data: DELCCV (NCOL, NROW)
**Enter item 7a only if INITW=0
7a. Data: ZWT (NROW*NCOL)
**Enter item 7b only if INITW=1

7b. Data: PWC (NCOL, NROW)
Utility Module: U2DREL. Note that item 7b consists of one array for each
layer in the model grid. Arrays for layers 1
A subset of the following two-dimensional arrays is used to describe the properties
of each layer of the model grid. The arrays needed for each layer depend on the input
LAYCON value (see item 2) of each layer, whether the simulation is transient (ISS = 0)
or steady-state (ISS not 0), and if the wetting capability is active (IWDFLG not 0) for the
wetting schemes of BCF2 and BCF3. Note that IWDFLG should be set to zero for the
new variably saturated flow scheme introduced in the BCF4 Package (i.e., input
LAYCON=43). If an array is not needed, it must be omitted. There are no situations in
which all arrays will be required. The required arrays (items 8-15) for layer 1 are read
first; then the arrays for layer 2, etc., until the property arrays for all layers are input.
**Enter item 8 only if the simulation is transient (ISS = 0).


**Enter item 9 only if the input LAYCON value for this layer is one of the following:
0, 2, 10, 12, 20, 22
9. Data: TRAN (NCOL, NROW)

Note that for input LAYCON=40 or 43, the hydraulic conductivity, bottom and top
elevations are input in items 10, 11, and 14, respectively.
**Enter items 10 and 11 only if the input LAYCON value for this layer is one of the
following:
1, 3, 11, 13, 21, 23, 31, 33, 40, 43
10. Data: HY (NCOL, NROW)

**Enter item 11 only if vertical hydraulic conductivities are to be input. (i.e., if

IVHYC>0)
11. Data: VHY (NCOL, NROW)
12. Data: BOT (NCOL, NROW)

**Enter item 13 only if this layer is not the bottom most layer of the and if IVHYC=0.
grid
13. Data: VCONT (NCOL, NROW)

**Enter item 14 only if the simulation is transient (ISS = 0) and the input LAYCON value
for this layer is one of the following: 2, 3, 12, 13, 22, 23, 33, 43

**Enter item 15 only if the input LAYCON value for this layer is one of the following: 2,
3, 12, 13, 22, 23, 33, 40, 43
15. Data: TOP (NCOL, NROW)

**Enter item 16 only if the wetting capability is active (IWDFLG is non-zero) and input
LAYCON value for this layer is one of the following: 1, 3, 11, 13, 21, 23, 31, 33
16. Data: WETDRY (NCOL, NROW)


Note: The WETDRY array is irrelevant for input LAYCON = 40 or 43, and, therefore,
item 13 is not needed for the new variably saturated flow scheme.
17. Data: VANAL (NCOL, NROW)

Note: Enter item 17 only if the input LAYCON value is 43 and IREALSL…0.
18. Data: VANBT (NCOL, NROW)

Note: Enter item 18 only if the input LAYCON value is 43, and IREALSL…0.
19. Data: VANSR (NCOL, NROW)

Note: Enter item 19 only if the input LAYCON value is 43, and IREALSL…0.
20. Data: BROOK (NCOL, NROW)

Note: Enter item 20 only if the input LAYCON value is 43, and IREALSL=2 or 4.
**Enter item 21 only if IDUAL (read in from the Basic (BAS) Package) is greater than
0:
21. Data: PHIF (NCOL, NROW)
ISS -- is the steady-state flag
If ISS = 0, the simulation is transient.
If ISS is not 0, the simulation is steady-state.
IBCFCB -- is a flag and a unit number.
If IBCFCB > 0, it is the unit number in which cell-by-cell flow terms
will be recorded whenever ICBCFL is set. (See
Output Control in McDonald and Harbaugh [1988].)
If IBCFCB = 0, cell-by-cell flow terms will not be printed or
recorded.
If IBCFCB < 0, flow for each constant-head cell will be printed
HDRY -- is the head assigned to cells that are converted to dry during a simulation.
Although this value plays no role in the model calculations, it is useful as
an indicator when examining the simulated heads that are output from the
model. HDRY is thus similar to HNOFLO in the Basic (BAS) Package,

which is the value assigned to cells that are no-flow cells at the start of a
model simulation.
Note: When input LAYCON value for the layer is 40 or 43, the HDRY
value is not assigned to cells that dry during the simulation like in the
other LAYCON options 0, 1, 2, 3, 10, 11, 12, 13, 20, 21, 22, 23, 31,
and 33.
IWDFLG -- is a flag that determines if the wetting capability is active.
If IWDFLG = 0, the wetting capability is inactive.
If IWDFLG is not 0, the wetting capability is active.
Note: IWDFLG is not used if ANY input LAYCON value in item 2 is 40
or 43).
WETFCT -- is a factor that is included in the calculation of the head that is initially
established at a cell when it is converted from dry to wet. (See
IHDWET.) Leave blank when wetting capability is inactive (i.e.,
IWDFLG = 0).
Note: This flag is irrelevant to input LAYCON = 40 or 43 conditions,
and may be left blank for these options.
IWETIT -- is the iteration for attempting to wet cells. Cell wetting is attempted
during each of the IWETIT iterations. If IWETIT is 0, it is changed to
1. Leave blank when wetting capability is inactive (i.e., IWDFLG = 0).
IHDWET -- is a flag that determines which equation is used to calculate the initial head
at cells that become wet.
If IHDWET = 0, equation (3a) of McDonald et al. (1991) is used:
h = BOT + WETFCT (hn - BOT).
If IHDWET is not 0, equation (3b) of McDonald et al. (1991) is used:
h = BOT + WETFCT (THRESH).
IAXIS -- is a flag that determines if the simulation is axi-symmetric. If IAXIS

>0, the axi-symmetric approach is employed.
Note: For the axi-symmetric simulation approach (IAXIS >0) NROW
must be set equal to 1 in the Basic Package Input.
IREALSL -- is a flag indicating that real soil moisture functions are used to define
flow in the unsaturated zone above the water-table.
If IREALSL=0, the pseudo-soil relations are used to define the

water-table.

If IREALSL is greater than 0, then flow in the unsaturated zone is
treated rigorously.
If IREALSL=2, van Genuchten function is used to characterize

moisture retention, and Brooks-Corey function
is used for relative permeability characteristics
of the unsaturated layers and water flow is
simulated.

If IREALSL=4, van Genuchten functions are used to
characterize moisture retention and Brooks-
Corey function is used for relative permeability
characteristics of the unsaturated layers and air
flow is simulated.
Note: This flag is required only for input LAYCON=43.
ICNTRL -- is a flag determining selection of mid-point or upstream weighting of

the relative permeability term.
If ICNTRL = 0, upstream weighting is used (suggested option).
If ICNTRL = 1, midpoint weighting is used.
Note: This flag is required only for input LAYCON=43.
VHYC -- is the vertical hydraulic conductivity input flag.

If IVHYC = 0, vertical hydraulic conductivities are not read.
Leakances are read for multi-layer simulations.
If IVHYC = 1 , vertical hydraulic conductivities are read.
Leakances are computed from vertical
conductivity of each layer as per a volume
weighted harmonic mean.
IANIXY -- is the index for variable anisotropy on a cell-by-cell basis.
If IANIXY = 0, horizontal anisotropy is uniform within each
layer as in the original MODFLOW formulation.
If IANIXY=1, horizontal anisotropy is non-uniform within a layer
and is input on a cell-by-cell basis for all layers.

ICURVL -- is the index for using a curvilinear grid in the areal plane.
If ICURVL =1, a curvilinear grid is used with input of DELR and
DELC required for every node within a layer.
Note that layers are stacked vertically below each
other, hence, DELR and DELC for a node in
lower layers are the same as for the node vertically
above it.
If ICURVL =0, a rectangular grid is used, as in the original
MODFLOW.
IOLDHDR -- is a flag indicating if old header formats (MODFLOW-style) are to be

used for saving cell-by-cell flow or budget terms in the binary file..
If IOLDHDR =1, the old MODFLOW-style headers are used for
cell-by-cell flow term output in the binary file.
If IOLDHDR =0, a new header format is used that includes the total
time value (TOTIM) and the stress-period time
value (PERTIM) in the header for cell-by-cell
budget term outputs.
LAYCON -- is the layer-type index array. Each element of LAYCON holds the
identification for the respective layer. Read one value for each layer.
There is a limit of 80 layers. If there are 40 or fewer layers, use one
record; otherwise, use two records. Leave unused elements in a record
blank. LAYCON is the parameter through which the new options of the
BCF4 Package may be implemented, as discussed below.
The BFC4 Package supports 16 different input LAYCON values. These values
are input in I2 format. From this value, the first digit (tens), if any, determines the
application of the type of interblock transmissivity function depending on the value of the
digit. The second digit (ones) of the input LAYCON value determines the layer
characteristics as in the original model. The meaning of the second digit is as follows
(McDonald and Harbaugh, 1988):
0 - Strictly confined - Transmissivity and storage coefficient of the layer
are kept constant for the entire simulation.
1 - Strictly unconfined - Transmissivity of the layer varies. It is calculated
from the saturated thickness and hydraulic
conductivity. The storage coefficient is constant in
time. This code is valid only for Layer 1.
2 - Confined/Unconfined - Transmissivity of the layer is kept constant. The
storage coefficient may alternate between confined
and unconfined values.
3 - Confined/Unconfined - Transmissivity of the layer is allowed to vary. It is
calculated from the saturated thickness and hydraulic
conductivity. The storage coefficient may alternate
between confined and unconfined values.
The above description can be summarized as follows (Goode and Appel, 1992):

The second digit of
the input LAYCON value Layer Characteristics
0 Strictly confined; T and S constant in time
1 Strictly unconfined; T depends on head, S constant in
time
2 Confined/Unconfined; T constant in time, S depends on
head because of storage conversion
3 Confined/Unconfined; T and S depend on head
Note that the new variably saturated flow option of the BCF4 Package are
implemented with input LAYCON = 40 or 43 for all the layers in the model grid. This
new option requires a non-zero value of IAPART. The flag IAPART, which is read from
item 5 of the BAS Package, indicates whether array BUFF is separate from array RHS.
The factor of ten represented by the first digit of the input LAYCON is stored in
the LAYAVG array. The second digit (ones) is stored in a LAYCON array. The stored
values of LAYCON and LAYAVG for an input LAYCON value are given in the following
table:
Input LAYCON Value Stored LAYCON Value Stored LAYAVG Value

0 0 0
1 1 0
2 2 0
3 3 0
10 0 10
11 1 10
12 2 10
13 3 10
20 0 20
21 1 20
22 2 20
23 3 20
31 1 30
33 3 30
40 0 40
43 3 40
**Note that input LAYCON values 30, 32, 41, 42, are not allowed by MODFLOW-
SURFACT/MODHMS.
The meaning of the LAYAVG (tens digit with a factor of ten) is as follows:

Stored
LAYAVG Value Meaning
0 Harmonic mean interblock transmissivity (McDonald and

Harbaugh, 1988)
10 Arithmetic mean interblock transmissivity (Goode and
Appel, 1992)
20 Logarithmic mean interblock transmissivity (Goode and
Appel, 1992)
30 Arithmetic mean saturated thickness times logarithmic mean
hydraulic conductivity (Goode and Appel, 1992)
40 Harmonic mean interblock hydraulic conductivity
introduced in BCF4 Package. Note that this LAYAVG
value is to be used when the rigorous unconfined storage
schemes are implemented.
Thus, the tens part of the input LAYCON value is stored in array LAYAVG, and
the ones part of the input LAYCON is stored in LAYCON array. The stored values of
LAYAVG determines the method of computing interblock transmissivity (when LAYAVG
= 10, 20, 30, or 40).

VISW -- Viscosity of water.
VISG -- Viscosity of air.
ATMGP -- Standard atmospheric pressure.
Values for RHOWP, RHOAP, VISW, VISG, COMPWAT, COMPAIR,
ATMGP, and GRAV depend on the units used in preparing input data sets. The
following table presents values for these parameters, at standard conditions, in different
sets of units commonly used.
Parameter SI Units cgs Units fps Units

RHOWP 1000 kg/m3 1 g/cm3 1.939 slugs/ft3
RHOAP 1.777 kg/m3 1.777×10-3 g/cm3 .003445 slugs/ft3
VISW 1×10-3 N.s/m2 1.0 cP 6.72×10-4 Rb/ft.s (Rbf.s/ft2)
VISG 1.983×10-5 N.s/m2 1.983×10-2 cP 1.3327×10-5 Rb/ft.s (Rbf.s/ft2)
COMPWAT 1-8 Pa-1 10-9 cm2/dyn 4.788×10-7 ft2/Rbf
COMPAIR 1.777×10-5 Pa-1 1.777×10-6 cm2/dyn 8.5083×10-4 ft2/Rbf
ATMGP 1.01×105 Pa 1.01×106 dyn/cm2 2.1094×103 Rbf/ft2
GRAV 9.8066 m/s2 980.66 cm/s2 32.2 ft/s2

TRPY -- is a 1-D array containing an anisotropy factor for each layer. It is the
ratio of transmissivity or hydraulic conductivity (whichever is being used)
along a column to transmissivity or hydraulic conductivity along a row.
Set to 1.0 for isotropic conditions. This is a single array with one value
per layer. Do not read an array for each layer; include only one array
control record for the entire array. If the value is the same for all layers,
the entire array can be specified by setting LOCAT to 0 and setting
CNSTNT to the value that applies to all layers. See below for explanation
of LOCAT and CNSTNT.
Note: TRPY array is not needed for an axi-symmetric simulation
(IAXIS>0).
TRPYNU -- is a 2-D array containing the cell-by-cell anisotropy factors within each
layer. It is the ratio of transmissivity or hydraulic conductivity
(whichever is being used) along a column to transmissivity or hydraulic
conductivity along a row. Set to 1.0 for isotropic conditions. One
array is read for each layer.
Note: TRPYNU array is not needed for an axi-symmetric simulation
(IAXIS>0).
LOCAT -- is an index for the location of values to fill-in the array of concern.
<0, read an unformatted record containing the array values
= 0, set all the array values equal to constant (CNSTNT)
>0, read the formatted records containing the array values.
CNSTNT -- is a constant to which all array values are set if LOCAT is equal to zero
or by which all array values are multiplied if LOCAT is not equal to zero.
DELR -- is the cell width along rows. Read one value for each of the NCOL
columns. This is a single array with one value for each column. For an
axi-symmetric simulation (IAXIS>0), DELR represents radial width ()r)
of a cylindrical cell.
DELRCV -- is the cell width along rows for a curvilinear grid. This is a 2-D
array containing one value for each NCOL column, for each of the
NROW rows of a layer.
DELC -- is the cell width along columns. Read one value for each of the NROW
rows. This is a single array with one value for each row.
Note: DELC array is NOT read for an axi-symmetric simulation
(IAXIS>0).
DELCCV -- is the cell width along columns for a curvilinear grid. This is a 2-D
array containing one value for each of the NCOL column, for each
of the NROW rows of a layer.
ZWT -- is the elevation of water table in the domain.
SF1 -- is the primary storage coefficient, the storability. Read only for a
transient simulation (steady-state flag, ISS, is 0). For input LAYCON
equal to 1, 11, 21, or 31, SF1 will always be specific yield, while for

input LAYCON equal to 2, 12, 22, 3, 13, 23, 33, 40, or 43, SF1 will
always be confined storage coefficient. For input LAYCON equal to 0,
10, or 20, SF1 would normally be confined storage coefficient. The
primary storage coefficient divided by the block thickness is equal to the
specific storage (Ss) of equations (1) and (2).
Note (1): An input LAYCON value of 0, 10, or 20 can also be
used to simulate water-table conditions where
drawdowns are expected to remain everywhere a small
fraction of the saturated thickness and where there is no
layer above, or flow from above is negligible. In this
case, specific yield values would be entered for SF1.
TRAN -- is the transmissivity along grid rows. TRAN is multiplied by TRPY to
obtain transmissivity along columns. TRAN is read only for a layer
represented by one of the following input LAYCON values: 0, 2, 10,
12, 20, 22.
HY -- is the hydraulic conductivity along grid rows. HY is multiplied by
TRPY to obtain hydraulic conductivity along columns. HY is read only
for a layer represented by one of the following input LAYCON values:
1, 3, 11, 13, 21, 23, 31, 33, 40, 43.
VHY -- is the vertical hydraulic conductivity of each cell. VHY is read only
for a layer represented by one of the following input LAYCON
values: 1, 3, 11, 13, 21, 23, 31, 33, 40, 43.
BOT -- is the elevation of the aquifer bottom. Read only for a layer represented
by one of the following input LAYCON values: 1, 3, 11, 13, 21, 23, 31,
33, 40, 43.
VCONT -- is the vertical hydraulic conductivity divided by the thickness, from one
layer to the layer below. The value for a cell is the hydraulic
conductivity divided by thickness for the material between the node in
that cell and the node in the cell below. Because there is not a layer
beneath the bottom layer, VCONT is not specified for the bottom layer.
SF2 -- is the secondary storage coefficient, the specific yield. SF2 is read only
if the simulation is transient (steady-state flag, ISS, is 0) and input
LAYCON value for the layer is one of the following: 2, 3, 12, 13, 22,
23 ,33, 43 .
TOP -- is the elevation of the aquifer top. TOP is read only for a layer
represented by one of the following input LAYCON values: 2, 3, 12, 13,
22, 23 ,33, 40, 43.
WETDRY -- is a combination of the wetting threshold and a flag to indicate which
neighboring cells can cause a cell to become wet. If WETDRY < 0,
only the cell below a dry cell can cause the cell to become wet. If
WETDRY > 0, the cell below a dry cell and the four horizontally
adjacent cells can cause a cell to become wet. If WETDRY is 0, the cell
cannot be wetted. The absolute value of WETDRY is the wetting
threshold. When the sum of BOT and the absolute value of WETDRY
at a dry cell is equaled or exceeded by the head at an adjacent cell, the
cell is wetted. WETDRY is read only if IWDFLG is not 0 and input
LAYCON value is one of the following: 1, 3, 11, 13, 21, 23, 31, 33.

VANAL -- is the van Genuchten parameter " for unsaturated soils. VANAL is
read only when the input LAYCON value for the layer is 43, and
the real soil moisture functions are requested with IREALSL…0.
VANBT -- is the van Genuchten parameter $ for unsaturated soils. VANBT is

read only when the input LAYCON value for the layer is 43, and
the real soil moisture functions are requested with IREALSL…0.
VANSR -- is the residual saturation level for unsaturated soil. VANSR is read
only when the input LAYCON value for the layer is 43, and the
BROOK -- is the Brooks-Corey exponent for computing relative permeability

of unsaturated soil. BROOK is read only when the input LAYCON
value for the layer is 43, and the real soil moisture functions, with
Brooks-Corey permeability relations are requested with
IREALSL=2 or 4.
PHIF -- is the fraction of mobile region in the model domain, f, defined as

the volume of the mobile domain divided by the entire domain
volume. PHIF is read in only when IDUAL [from the Basic (BAS)
package] is greater than 0.
6.6.4 Basic Transport (BTN1) Package Input

input parameters.

Format: F10.0 I10 I10 I10 F10.0
active phase).

Format: F10.0 F10.0 F10.0 F10.0 F10.0 F10.0
**Enter item 3 only if IACLVL = –1.

3. Data: ALFX ALFY ALFZ
Format: F10.0 F10.0 F10.0
** Each of the following items (5 through 21) consists of NLAY two-dimensional arrays,
must be omitted.
requested below.

Utility Module: U2DREL. Note that U2DREAL is the utility module that
reads a two-dimensional array.







**Enter item 12 only if IDUAL > 0.
12. Data: PHIIM (NCOL, NROW)


**Enter item 14 only if IDUAL=3.
14. Data: RHOBIM (NCOL, NROW)


Format: 5F10.0

Format: 5F10.0

Format: 5F10.0
*Enter item 16a only if IEQPART >0 or if IREALSL >2.

Format: 5F10.0

Format: 5F10.0

Format: 5F10.0


**Enter item 20 only if IDUAL=3
20. Data: KDIM (NCOL, NROW)



***Enter item 22 only if IDUAL=3 and LINR=2
22. Data: ETAIM (NCOL, NROW)




* Enter item 24c only if ILAMWS = 0.

Note that 24c consists of one array for each layer in the model grid. Arrays for layers 1

**Enter items 26 and 27 only if IDUAL > 0.
26. Data: DUALRATE (NCOL, NROW)


27. Data: SIMCONC (NCOL, NROW)

**Enter items 28a, 28b, and 28c only if ICHAIN=2
28a. Data: NPAREN(I)

Format: I10
28b. Data: KPAREN(J), J=1, NPAREN

Format: 5I10
28c. Data: SFACTR(J), J=1, NPAREN

Format: 5F10.3

radioactive decay).
If IDCYTP = 0, the degradation is both species and soil dependent
(e.g., biodegradation).
including soil.
limiter.
upstream factors.

Note that the TVD scheme is highly accurate but more
computationally demanding than the other upstream weighting
schemes. If steep concentration fronts are expected, the TVD
scheme should be selected. On the other hand, the fully upstream
weighting scheme is the most efficient but only first-order accurate.
If the grid is sufficiently refined and numerical dispersion is not
expected to be significant, the fully upstream weighting scheme
should be selected. The user who is familiar with upstream
weighting may prefer partial to full-upstream weighting to control
numerical oscillations while avoiding too much numerical
dispersion. This is achieved with IACLVL=–1 or 1. With
IACVL=–1, the scheme becomes a central difference scheme if
zero upstream factors are specified.
simulation.
ICHAIN -- is the index for chain-decay conditions.
If ICHAIN = 2, Chain decay or transformation with daughter
products is accounted for with a branched chain
where one parent can have several daughter
products, or one daughter product may have
several parent components.
products is taken into account , with one unit of
parent decaying to one unit of daughter in a
straight chain.
omitted.
active phase.
transport.

NDENS -- Flag for use with density-dependent module.
by saturation as:
approach as:
Cs = Kd * Cw

present.
active phase.

Note: Input instructions for the BCF4 package supply standard
values for RHOWP, RHOAP, COMPAIR, ATMGP, and GRAV in
several commonly used sets of units.
CNOFLO -- is the value of concentration to be assigned to all inactive cells (i.e.,
THETRD -- is a control parameter for time weighting factor.
If THETRD = 0.0, C r a n k -Nicolson factor is select e d
automatically.
If IBCFCC > 0, it is the unit number on which cell-by-cell storage
and decay terms will be recorded whenever
ICCCFL (see Output Control) is set.

NOMATRIX -- is the index for not writing coefficient matrix.
If NOMATRIX = 1, coefficient matrix is not written.
If NOMATRIX = 0, coefficient matrix is written.
PHIM -- Porosity of the immobile domain Nim. It is the void space in the
immobile domain per unit volume of immobile domain. PHIM is
read only if IDUAL > 0.
RHOB -- is bulk density of soil ( ) . Read only if LINR > 0. For dual
M
L3
RHOBIM -- is the bulk density of the immobile domain. Read only if
IDUAL=3. (i.e., if equilibrium adsorption in the two domains are
different).
DIFF(I) -- is the molecular diffusion coefficient (L2/T) of species I in the active

IEQPART…0.
when IEQPART=2.
EQPARTAW(I) -- is the gas-water partitioning coefficient for species I, defined as
KD -- is soil partition coefficient ( ).

L3
M Read only if LINR > 0. For
dual domain simulations, KD is the partition coefficient in the
mobile domain, if immobile domain has distinct adsorption
then adsorption is assumed to occur only in the immobile domain,
and this parameter defines the mobile domain adsorption
characteristics.
KDIM -- is the soil partition coefficient of the immobile domain. Ready only
if IDUAL=3 (L3/M).
ETA -- is the Freundlich isotherm exponent (0) in equation (2) of Chapter
2. Read only if LINR = 2. For dual domain simulations, ETA is
the Freundlich exponent of the mobile domain, if immobile domain
has distinct adsorption characteristics from the mobile domain
(IDUAL=3). If IDUAL=2 then both domains have similar
adsorption characteristics and this parameter defines the adsorption
for both domains. If IDUAL=1, then adsorption is assumed to
occur only in the immobile domain, and this parameter defines the
mobile domain adsorption characteristics.
ETAIM -- is the Freundlich isotherm exponent (0) for the immobile domain.
Ready only if IDUAL=3 and LINR=2.

= 1 in item 1).
SCONC -- is concentration at start of the simulation (M/L3) (within the mobile
domain, for dual domain simulations.
DUALRATE -- is a first-order coefficient describing the rate of mass transfer
between the mobile and immobile regions (1/T). Read only if
IDUAL > 0.
SIMCONC -- is the concentration in the immobile domain at the start of the
simulation (M/L3). Read only if IDUAL > 0
NPAREN(I) -- is the number of parent species of component I
KPAREN(J) -- Component index of the J th parent that decays to component I (for
each J=1, NPAREN(I)).
SFACTR(J) -- Fraction of parent component J that decays to component I. (for
each J=1, NPAREN(I)). Decay of one unit of J produces SFACTR
units of component I, therefore SFACTR may also be used to scale
units in a decay process.
6.7 REFERENCES
HydroGeoLogic, Inc., 1992. STAFF3D - A Three-Dimensional Finite-Element Code for
Simulating Fluid Flow and Transport of Radionuclides in Fractured Porous Media,
version 2.0, 1992.
van Genuchten, M.Th., 1981. Non-equilibrium transport parameters from miscible
displacement experiments. Res. Rep. 119, U.S. Salinity Laboratory, Riverside,
Calif., 88 pp.
van Genuchten, M.Th. and Wierenga, P.J., 1976. Mass transfer studies in sorbing porous
media, I. Analytical solutions. Soil Sci. Soc. Am. J., 40: 473-480.

CHAPTER 7: ADDENDUM FOR DENSITY-DEPENDENT FLOW/TRANSPORT
SIMULATIONS
7.0 Addendum For Density-Dependent Flow/Transport Simulations . . 7-1
7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-1
7.2 Formulation of the Density-Dependent Flow/Transport
Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-1
7.3 Numerical Implementation and Solution Options . . . . . . . . 7-2
7.4 Input/Output Considerations . . . . . . . . . . . . . . . . . . . . . 7-5
7.5 Verification Example . . . . . . . . . . . . . . . . . . . . . . . . . 7-6
7.6 Input Instructions for Density-Dependent Simulations . . . . 7-6
7.6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . 7-6
7.6.2 Basic Transport Package Input (.BTN) . . . . . . . . 7-10
Input . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-22
7.7 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-23
7.0 ADDENDUM FOR DENSITY-DEPENDENT FLOW/TRANSPORT
SIMULATIONS
7.1 INTRODUCTION
Fluid flow and contaminant transport equations discussed earlier (in Volumes I and
II of this document respectively) assume that there is no coupling between the flow and
solute transport processes. This coupling can become significant in several instances,
specifically when the density of solutes significantly impacts the flow field as solutes move
within a domain. Variable density fluid flow and solute transport processes are most
significant in analysis of coastal aquifer systems where the heavier seawater intrudes
landward in deeper portions of the aquifer. Bear (1979) provides theoretical details and
solutions to saltwater intrusion and upconing problems in various unconfined and confined
aquifer systems.
This section details the flow and transport coupling in MODFLOW-
SURFACT/MODHMS resulting from the density effects of solutes on water. The density
effects of more than one component species may be included in a simulation. Further, the
density-dependent flow and transport simulation option may be used along with all the
other features of flow and transport available in MODFLOW-SURFACT/MODHMS,
including unsaturated flow, reactive transport and dual porosity media. An active air-
phase simulation may also include the density term to incorporate the density effects of
component vapors on the flow field of air. Note that density effects are incorporated only
into the active flow phase for multipahse transport investigations with MODFLOW-
SURFACT/MODHMS, and only diffusion occurs in the inactive flow phases of a
simulation.
7.2 FORMULATION OF THE DENSITY-DEPENDENT FLOW/TRANSPORT

EQUATIONS
The governing equation for flow of a mixture fluid of variable density in an
aquifer system can be written in terms of an equivalent freshwater head as
(HydroGeoLogic, 1994.)
NDENS
η c =
T T
∑ηc
i =1
i i
(1)
where h is now defined as an equivalent freshwater head, z is the length in the vertical
direction, 0TcT is the total density factor, the subscript w represents the active fluid phase
(which is water for most typical simulations of coastal aquifer systems but could also be
air in dense vapor simulation cases), and the other terms in the equation are defined in
Volume I where the governing fluid flow equation is presented. The terms 0TcT in
equation (1) are the additional buoyancy terms required by the flow equation for density-
dependent simulations. The total density factor is the sum of the density factors of each
of the component species whose density effect is significant. Hence
∂h ∂S w ∂η T c T ∂ ⎡ ⎧⎪ ∂h ∂z ⎫⎪⎤
SwSs +φ + S wφ = ⎢ K ij ⎨ + η T cT ⎬⎥ − W (2)
∂t ∂t ∂t ∂xi ⎢⎣ ⎩ j ⎪ ∂x ∂x j ⎪
⎭⎥⎦
Chapter 7: Addendum For Density Dependent Flow/Transport Simulations 7-1

M:\Projects\MODFLOW Manuals\SURFACT and HMS Ver.3\Vol II\WordPerfect Files\Chapter_7..wpd
where NDENS is the total number of component species whose density effects are
included in a simulation, 0i is the buoyancy term for species i, and ci is the concentration
of species i. The buoyancy term for each species is defined as
η i
=
(ρ i
s − ρo ) (3)
c si ρ o
where Do is the reference (freshwater) density, and Dis is the density of a solution
containing concentration of solutes equal to cis. Note that equation (3) assumes a linear
variation of density with concentration for each species of concern, and equation (2)
assumes that the buoyancy effects for each species are additive. Boundary conditions for
the flow equation can include all of MODFLOWs available boundaries. Note however,
that prescribed head conditions should be provided in terms of the equivalent freshwater
head which is the variable of solution for the density-dependent flow formulation. Also,
the general head type boundaries of MODFLOW-SURFACT/MODHMS do not include
the density term (between the bed of the general head condition and the aquifer within the
domain) which should be accommodated in the input if required. After solution to the
flow equation, the fluxes may be computed as
⎧⎪ ∂h ∂z ⎫⎪
qi = K ij ⎨ + η T cT ⎬ (4)
⎪⎩ ∂x j ∂x j ⎪⎭
Transport of all solutes occurs under the flow field computed from equation (4)
above. The transport equations are unaltered from those discussed in Volume II of the
MODFLOW-SURFACT/MODHMS document, for the density-dependent flow/transport
case. However, the flow field (equations 1 and 4) is dependent on the concentration of
solutes within the domain and the concentrations are conversely dependent on the flow
field, causing a coupling of the two equation sets.
Additional coupling of the two equations is optionally provided via the hydraulic
conductivity of the medium, which could also depend on density and viscosity of the fluid
which in turn are related to the fluid concentrations. The hydraulic conductivity of a
medium (Ks) with fluid density Ds and viscosity µs, is related to the reference hydraulic
conductivity (K) for reference (freshwater) density Do and viscosity µo as:
ρs µo
Ks = K (5)
ρo µ s
where the relationship between concentration of each species and density is assumed linear
as discussed earlier. Assuming a similar linear relationship between fluid viscosity and
concentration with ideal mixing law for viscosities gives
⎡⎛ µ si − u o ⎞ i ⎤
µ s = ∑ ⎢⎜⎜ i
⎟c + µ o ⎥ / i
⎟ (6)
i ⎢
⎣⎝ c s ⎠ ⎥⎦
where µ s , is the viscosity of fluid with concentration of species i at c si
i
, for all species
i that include density and viscosity coupling.
7.3 NUMERICAL IMPLEMENTATION AND SOLUTION OPTIONS

Two considerations are involved with numerical implementation of the density-
dependent flow and transport solution into MODFLOW-SURFACT/MODHMS. The first
7-2 Chapter 7: Addendum For Density Dependent Flow/Transport Simulations

requires implementation of the additional buoyancy terms into the flow equation, and the
second involves iterative coupling of the flow and transport equations.
The flow equation (1) contains additional terms for the buoyancy effects of
freshwater, as compared to the standard flow equation discussed earlier in Volume I of the
MODFLOW-SURFACT/MODHMS document. Since an iterative coupling is employed
between the flow and transport equations, the additional buoyancy terms expressed as
functions of solute concentrations, are added to the right-hand-side of the assembled flow
equations. The additional mass stored due to component density effects is expressed in
finite difference form as
∑ [(η c ) )]
∂η T c T φS w
(
NDENS
i t + ∆t
φS w − η ici
t
= i
(7)
∂t ∆t i =1
where t is the known time level and t + )t is the new time level at which solution is
sought. In the Picard iterative procedure, (ci)t+)t may be estimated from a previous iterate
value. Note that this term is therefore neglected if a time-lagged solution scheme is
employed. The buoyancy flux term may be expressed in finite difference form, for flow
between any two nodes 1 and 2 as
∂z K NDENS
K ijη c T
= 12
T
∂x j ∆l 12
∑ η c (z
i =1
i i
av 1 − z2 ) (8)
where K12 is the conveyance between nodes 1 and 2, )R12 is their separation distance,
is the volume weighted average concentration of species ‘i’ of the adjacent cells, and z1 and
z2 are the mean node elevations. The additional terms of equations (7) and (8) are placed
on the right-hand-side of the assembled flow equations for porous medium as well as
fracture well connections, and are subsequently also incorporated into the mass balance
calculations of the flow equation, once solution is obtained for h and c. After solution of
h at each iteration, the fluxes within the system are obtained using equation (4), for use
by the transport equation. Equation (4) consists of the traditional flow component of
MODFLOW-SURFACT/MODHMS and the buoyancy component which is expressed in
the form of equation (8). Hence, equation (8) needs to be added to MODFLOW-
SURFACT/MODHMS for calculating the flux that includes buoyancy components.
It should be noted that the buoyancy terms are implemented in the fracture well
package of MODFLOW-SURFACT/MODHMS as well, to provide full flexibility in
simulation capabilities. Further, fully dry cells do not include the buoyancy term for
unconfined flow conditions and height adjustments are made for cells that contain the
water table, to include only the saturated thickness portion of a cell in the
i
c av computation.
The hydraulic conductivity of the system may optionally be dependent on fluid

density and viscosity. The flag KEFFECT determines this coupling. If KEFFECT is
zero, the hydraulic conductivity is not dependent on density or viscosity. If KEFFECT
= 1, the density effect on K is included, and if KEFFECT = 2, the viscosity effect is also
i
included. The volume weighted average concentration c av is used to compute the density
or viscosity of the fluid, with equation (5) providing the new hydraulic conductivity.
Coupling of the flow and transport equations is performed using the Picard
iterative scheme. For every time-step, the flow equation is first solved and fluxes within
the domain and across boundaries are computed to include the density terms. Next, the
transport equations are solved for the density components (NDENS). This sequence of

solving flow and transport equations is repeated till the solution converges for heads and
concentrations of all NDENS components. This flow field is then used for solving
transport of the remaining components (if any) whose density effects are not significant
to the simulation, before proceeding to the next time-step. A one-step steady-state solution
may also be attempted for the coupled flow and transport situation. Eliminating the
storage terms for the NDENS components in the transport equation effectively renders it
steady-state. The flow-field achieved by the one-step steady-state solution may then be
used to solve for transient transport of the remaining component species. The Cooley
(1983) under-relaxation formula is applied to heads and concentrations of the NDENS
species of simulation at every density iteration to dampen oscillations. Additional under-
relaxation is performed on the NDENS transport equations, to dampen the maximum
change of concentration at any iteration. Note that the density iteration is embedded
within each non-linear flow iteration, for non-linear flow situations (which may use a
Newton-Raphson formulation to resolve the flow non-linearities).
Certain options have been provided in MODFLOW-SURFACT/MODHMS to
facilitate solution of the density-dependant problem. Firstly, it is not known apriori, if a
prescribed head node is an inflow or an outflow node. During solution of the transport
equation at every iteration, the physically correct boundary condition requires that if a
node is an inflow boundary, the concentration of inflow water be provided as a first type
or third type condition. If, however, a node is an outflow node, the prescribed
concentration should not be provided. The density-dependent solution of flow and
transport in MODFLOW-SURFACT/MODHMS checks for this condition at every
iteration and implements the boundary condition accordingly. However, this scheme may
cause a flip-flop condition at the boundary which may discourage convergence, hence
several options are provided to deal with this situation. A flag ISWAB has been provided,
that controls implementation of the transport boundary condition, depending on flow
conditions (inflow or outflow) at a prescribed head node. If ISWAB=1, a transport
boundary condition coincident with a prescribed head node, checks for inflow or outflow
at every iteration. If ISWAB = 2, this check is time-lagged for a transient simulation.
Thus, the inflow/outflow determined from a previous time-step (or from initial conditions
for the first time-step) is used to determine whether to apply prescribed concentrations at
prescribed head nodes. If ISWAB=0, this check is not performed, and the transport
boundary conditions are always applied, as determined by the user. If ISWAB=-1, the
first iteration’s inflow/outflow conditions are used for determining the transport boundary
conditions at prescribed head nodes for that time-step. It should be noted that ISWAB=1
provides the rigorous physical condition; however, ISWAB=2 or -1 provide much better
convergence characteristics, and converged solutions that are acceptable in practice (only
a few ‘outflow’ nodes in a domain might have prescribed concentrations from the list
provided by the user.) The option of ISWAB=0 requires the user to ensure that the
supplied list of prescribed concentration conditions, does not contain outflow nodes in the
domain.
Another option to assist density-dependent simulations is provided by the flag
IMOVEON. When IMOVEON=0, if density-dependent iterations fail to converge, the
time-step size is cut and calculations are attempted using the reduced time-step. However,
if IMOVEON=1, the non-converged solution is accepted and the simulation proceeds to
the next time-step. This option can be useful when time-marching to steady-state, where
the non-convergence of intermediate solutions is of little consequence in achieving a
converged steady-state solution. A user should however ensure that such a solution is
truly converged, if IMOVEON=1 option is used. Note that the IMOVEON option works
only with adaptive time-stepping. If the traditional MODFLOWs time-stepping and output

control routines are used instead, failed iterations at any time-step will abort the
simulation.
A third option that is also useful for density-dependent simulations, is provided by
the flag ILAG. If ILAG=0, the density-dependent flow and transport iterations are
performed as discussed above. However, if ILAG=1, the buoyancy terms are updated
in the flow equation in a time-lagged manner. Thus for non-linear flow cases, the flow
iterations are performed till convergence, using buoyancies calculated from concentrations
of the previous time-step. Transport calculations on all components (those with or without
density effects) are performed only after a converged flow solution. Use of this option
speeds up computations (even more so for saturated or confined flow conditions);
however, care should be taken to select a sufficiently small maximum time-step size, such
that the time-lagging does not introduce significant errors in the solution.
The density-dependent flow and transport solution lays the structure for a Picard
coupling of the flow and transport equations. Other coupling processes (for instance solute
viscosity effects on the flow field) can be easily incorporated into the structure above.
7.4 INPUT/OUTPUT CONSIDERATIONS

All additional input required for density-dependent flow and transport simulations
may be read from the Basic Transport Input File (having default extension .BTN). An
additional parameter, NDENS, is read in the first group of data in this file. If
NDENS=0, density-dependent considerations are skipped and MODFLOW-
SURFACT/MODHMS ignores the routines discussed in this chapter. A value of NDENS
> 0, indicates that density-dependent flow and transport considerations are to be included
in the simulation, and the integer value of NDENS represents the number of component
species whose density effects are considered. Note that NDENS should be less than or
equal to the total number of component species of simulation, NSPECI.
If NDENS > 0, additional input is required for the buoyancy parameters and the
density-dependent simulation options. The buoyancy parameters include the density of
freshwater [RDENS], and the reference density of a solution of component i
[RHODENS(I)], at reference concentration [CDENS(I)], for all NDENS components.
Other input parameters for density-dependent simulations include the simulation option
flags ISWAB, IMOVEON, and ILAG discussed in Section 7.3, and a concentration
convergence criterion for the density iterations, DCLOSE, applied to all NDENS
components.
Output of density-dependent flow and transport simulations is the same as for
regular flow and transport simulations. Heads and concentrations of all species of
simulation are provided in the ASCII output file along with mass balance information for
the flow and transport equations. Binary files containing heads, concentrations of all
species, and cell-by-cell flow terms (for flow and contaminant flux) are generated in an
identical manner to regular flow/transport simulations. Note that mass-balance
information at any time-step provided in the ASCII output file for the NDENS species that
include density effects, may be placed in a different location than for non-density
flow/transport situations, due to iteration considerations.
Density-dependent flow/transport computations use the concept of equivalent
freshwater heads for determining the flow-field. An environmental head is a more
measurable quantity, therefore, an option has been provided to compute and print
environmental heads for density-dependent flow/transport simulation cases. The
environmental head is computed from the equivalent freshwater head at any location, by
subtracting the effects of density from that point up to the surface. Therefore, this is the

head that would be measured in a well that has an open hole to the surface. A flag,
NOBOY is included in the .BTN file which indicates whether computation and printout
of environmental heads is required. If NOBOY=1, the environmental heads are calculated
and printed (or saved) in accordance with the other flags for recording computed
concentrations.

The density-dependent flow and transport simulation capability is verified using
an analytical solution known as the Henry’s seawater intrusion problem. This problem
was also used to verify the various simulation options, and comparisons were made with
another finite-element numerical code, DSTRAM, for both steady-state and transient
cases. Figure 7.1 provides a schematic description of the problem and lists the material
parameters of the simulations. The material parameters were selected so that the cases
analyzed correspond to those solved by other workers (HydroGeoLogic, 1994; Pinder &
Cooper, 1970; Lee and Cheng, 1974; Segol et al., 1975, and Frind, 1982). As depicted
in Figure 7.1, freshwater enters the aquifer on the left face, with the coastal side
corresponding to the right face. The coastal boundary condition allows convective mass
transport out of the system over the top portion (80#y#100m) of the right face, where the
normal concentration gradient is zero. This is consistent with the condition used by Segol
et al. (1975) and Frind (1982), but is a modification of those used by Lee and Cheng
(1974) and Henry (1964). It should be noted that the vertical extent of the outlet portion
of the seaward boundary is not known apriori, but can be determined by trial and error,
or by effective use of the ISWAB options. The remaining portion of the seaward boundary
(0#y#800 m) has a prescribed relative concentration of unity, from where seawater enters
the simulation domain. The simulated aquifer is fully confined from top and bottom,
where no flow, zero-concentration gradient conditions are applied. The domain of
simulation is discretized with 10 layers, 15 columns, and 1 row of grid blocks. A steady-
state analysis is first performed, and Figure 7.2 shows a comparison of simulation results,
with results from DSTRAM (HydroGeoLogic, 1994). These results also match closely
with the results of the other analyses (Segol et al., 1975; Frind, 1982). A transient
simulation was also performed for the same problem, and results for this case are
compared with DSTRAM in Figure 7.3. The simulation results of the two codes compare
well for all times investigated. The ILAG, IMOVEON, and ISWAB options were also
tested with this problem. Results for all cases were practically identical to the
corresponding cases in Figures 7.2 and 7.3.
7.6 INPUT INSTRUCTIONS FOR DENSITY-DEPENDENT SIMULATIONS

7.6.1 Introduction
Additional input required for density-dependent flow and transport simulations are
all supplied via the basic transport input package (having default extension .BTN). The
required additional input for density effects of the components is logically incorporated
into the data file, and is controlled by the flag NDENS. If density effects are to be
included, a value of NDENS greater than zero is set. The numeric value also indicates the
number of component species whose density effects are to be included. Additional input
required for the NDENS density components, then follows. Further input is also required
in the prescribed concentration file (default extension of .PCN) to identify nodes that are
prescribed head conditions (for flow) which might be outflow nodes at any iteration or
time-step. Depending on the value of the simulation flag ISWAB, these nodes are not
provided with the prescribed concentration value, if it is an outflow node (at any

Vn = Mc/Mn = 0
h = ,(b-y)
20m
K = 1 m/d Mc/Mn = 0
U = 0.066 m/d
, = 0.0245
Vn = U 80m
c=0 n = 0.35
h = ,(b-y)
Do = 0.066 m2/d c=1
Vn = Mc/Mn = 0
200m
Figure 7.1 Problem description of seawater intrusion in a coastal confined aquifer.

120
100
DSTRAM
MODFLOW-SURFACT
80
ELEVATION, m
60
0.1
0.5
40
0.9
20
0
0 50 100 150 200
DISTANCE, m
Figure 7.2 Chloride concentrations for steady-state simulation of saltwater

intrusion.

100
Time = 682 days

80
DSTRAM
MODFLOW-SURFACT
ELEVATION, m
60
40
0.1
0.5
20
0.9
0
0 50 100 150 200
DISTANCE, m
100
Time = 2948 days

80 DSTRAM
MODFLOW-SURFACT
ELEVATION, m
60
0.1
40
0.5
0.9
20
0
0 50 100 150 200
DISTANCE, m
100
Time = 32000 days

80
DSTRAM
MODFLOW-SURFACT
ELEVATION, m
60
0.1
40 0.5
0.9
20
0
0 50 100 150 200
DISTANCE, m
Figure 7.3 Chloride concentrations for transient analyses of saltwater intrusion.

iteration or time-step). Input data for all other MODFLOW-SURFACT/MODHMS files
remain unchanged. Hence, the next section provides only the modified input file
instructions, i.e., for the .BTN and .PCN packages. Input instructions for the remaining
modules of MODFLOW-SURFACT/MODHMS may be referenced from the respective
chapters documented earlier in Volumes I and II.
7.6.2 Basic Transport Package Input (.BTN)
input parameters.

Format: F10.0 I10 I10 I10 F10.0
active phase).

Format: F10.0 F10.0 F10.0 F10.0 F10.0 F10.0
**Enter item 3a only if IACLVL = –1.
3a. Data: ALFX ALFY ALFZ

Format: F10.0 F10.0 F10.0
**Enter item 3b only if NDENS > 0 (for density-dependent flow/transport)
3b. Data: RFRESH IMOVEON ISWAB DCLOSE ILAG NOBOY KEFFECT EMUO
Format: F10.3 I10 I10 F10.3 I10 I10 I10 I10
* Each of the following items (5 through 21) consists of NLAY two-dimensional arrays,
must be omitted.

requested below.

Utility Module: U2DREL. Note that U2DREAL is the utility module that
reads a two-dimensional array.






**Enter item 12 only if IDUAL > 0.

12. Data: PHIIM (NCOL, NROW)

**Enter item 14 only if IDUAL=3.

14. Data: RHOBIM (NCOL, NROW)

Format: 5F10.0

Format: 5F10.0

Format: 5F10.0
**Enter item 16a only if IEQPART >0 or if IREALSL >2.

Format: 5F10.0

Format: 5F10.0

Format: 5F10.0

**Enter item 19 only if NDENS > 0. Note that item 19 consists of one entry for each
of the NDENS species whose density effects are to be incorporated in flow, in sequential
order of the components.
19. Data: CDENS RHODENS VISCDENS

Format: F10.3 F10.3 F10.3
Note that VISCDENS is used only if KEFFECT = 2 (i.e., when fluid viscosity affects the
hydraulic conductivity).

**Enter item 21 only if IDUAL=3

21. Data: KDIM (NCOL, NROW)

***Enter item 23 only if IDUAL=3 and LINR=2

23. Data: ETAIM (NCOL, NROW)



** Enter item 25c only if ILAMWS = 0.

Note that 25c consists of one array for each layer in the model grid. Arrays for layers 1

**Enter items 27 and 28 only if IDUAL > 0.

27. Data: DUALRATE (NCOL, NROW)
28. Data: SIMCONC (NCOL, NROW)

**Enter items 29a, 29b, and 29c only if ICHAIN=2

29a. Data: NPAREN(I)

Format: I10
29b. Data: KPAREN(J), J=1, NPAREN

Format: 5I10
29c. Data: SFACTR(J), J=1, NPAREN

Format: 5F10.3

radioactive decay).
If IDCYTP = 0, the degradation is both species and soil dependent
(e.g., biodegradation).
including soil.
limiter.
upstream factors.
Note that the TVD scheme is highly accurate but more computationally
demanding than the other upstream weighting schemes. If steep
concentration fronts are expected, the TVD scheme should be selected.
On the other hand, the fully upstream weighting scheme is the most
efficient but only first-order accurate. If the grid is sufficiently refined
and numerical dispersion is not expected to be significant, the fully
upstream weighting scheme should be selected. The user who is
familiar with upstream weighting may prefer partial to full-upstream
weighting to control numerical oscillations while avoiding too much
numerical dispersion. This is achieved with IACLVL=–1 or 1. With
IACVL=–1, the scheme becomes a central difference scheme if zero
upstream factors are specified.

simulation.
ICHAIN – is the index for chain-decay conditions.
products is taken into account.

omitted.
active phase.
transport.
NDENS -- is the number of components whose density effects are to be
included for flow. Note that in the sequential numbering of
component species, the components whose density effects are to be
included for flow should be ordered first, followed by components
whose density effects are not considered.
by saturation as:
approach as:
Cs = Kd * Cw
present.

active phase.
Note: Input instructions for the BCF4 package supply standard values
for RHOWP, RHOAP, COMPAIR, ATMGP, and GRAV in several
commonly used sets of units.
RFRESH -- Reference density of freshwater (used only for density-dependent
flow and transport).
IMOVEON -- is an index determining acceptance of non-converged iterations for
density-dependent flow and transport.
If IMOVEON = 1, Non-converged solution is accepted and
solution proceeds to next time-step. This
option is useful for seeking long term
solutions or time-marching to steady-state,
where intermediate results are of little
consequence.
If IMOVEON = 0, The time-step size is cut for non-
converged flow/transport iterations, to
attempt solution for the smaller time-step.
Note that the IMOVEON flag is active only if the adaptive time-
stepping option of the .ATO package is used. For MODFLOW’s
time-stepping with output control using the .OC package, the
simulation will be aborted at any time-step that convergence does
not occur within the prescribed maximum number of iterations.
ISWAB -- is an index determining handling of prescribed concentration
boundaries along prescribed head nodes of a density-dependent
flow/transport simulation (identified as such in the .PCN package).
This switch helps in obtaining converged solutions, since the
inflow/outflow state of a prescribed head node is not known apriori,
and may change from iteration to iteration.
If ISWAB = 1, the transport boundary condition on a
prescribed head node checks for inflow/outflow
at every iteration. The prescribed

concentration is then applied only at inflow
nodes.
If ISWAB = 0, the check is not applied, and transport
boundary conditions of prescribed
concentrations are applied everywhere as
determined by the user in the .PCN file.
If ISWAB = 2, the check for inflow/outflow is time-lagged for
a transient simulation.
If ISWAB = -1, the first iterations inflow/outflow conditions
are used for applying prescribed concentrations
for a particular time-step.
DCLOSE -- is the convergence criterion for closure on concentrations, for the
density-dependent flow and transport simulation case.
ILAG -- is an index determining if buoyancy terms in the flow equation may
be updated in a time-lagged manner.
If ILAG = 1, the buoyancy term is time-lagged. The user
should ensure that the time-lagged solution is
accurate.
If ILAG = 0, the buoyancy terms are updated in an iterative
fashion for rigorous solution.
NOBOY -- is an index determining if environmental heads are computed and
printed/saved as part of the simulation output.
If NOBOY = 0, environmental heads are not printed or saved.
If NOBOY = 1, environmental heads are printed and/or saved in
accordance with the concentration output flags set
in the ATO or OC packages. Further, if
environmental heads output in binary format is
requested, it is saved in the .CON (default
extension) file after saving all concentration
output.
KEFFECT – is an index determining if hydraulic conductivity values are affected
by density or viscosity.
KEFFECT = 0 is no, there is no effect on K.
= 1 is yes, K is affected by density of fluid.
= 2 is yes, K is affected by density as well as viscosity
of fluid.
EMUO – Reference viscosity of freshwater (used only if KEFFECT = 2) in
density dependent flow and transport.
CNOFLO – is the value of concentration to be assigned to all inactive cells (i.e.,

THETRD – is a control parameter for time weighting factor.
If THETRD = 0.0, Crank-Nicolson factor is selected
automatically.
If IBCFCC > 0, it is the unit number on which cell-by-cell storage
and decay terms will be recorded whenever
ICCCFL (see Output Control) is set.
NOMATRIX – is the index for not writing coefficient matrix.

If NOMATRIX = 1, coefficient matrix is not written, and is
computed whenever it is required.
If NOMATRIX = 0, coefficient matrix is written to be used
subsequently, if possible.
Note that one may be faster than the other depending on CPU speed
and disk writing speed of the computational environment.
PHIM -- Porosity of the immobile domain Nim. It is the void space in the
immobile domain per unit volume of immobile domain. PHIM is
read only if IDUAL > 0.

RHOB -- is bulk density of soil . Read only if LINR > 0. For dual

RHOBIM – is the bulk density of the immobile domain. Read only if
IDUAL=3. (i.e., if equilibrium adsorption in the two domains are
different).
DIFF(I) – is the molecular diffusion coefficient (L2/T) of species I in the active
IEQPART…0.
when IEQPART=2.
EQPARTAW(I) – is the gas-water partitioning coefficient for species I, defined as
CDENS -- maximum solution concentration of a component i.
RHODENS – reference fluid density of a solution of component i at maximum
concentration [CDENS].
VISCDENS – reference fluid viscosity of solution of component i at maximum
concentration [CDENS].
L3
KD -- is soil partition coefficient M . Read only if LINR > 0. For
dual domain simulations, KD is the partition coefficient in the
mobile domain, if immobile domain has distinct adsorption
then adsorption is assumed to occur only in the immobile domain,
and this parameter defines the mobile domain adsorption
characteristics.

KDIM -- is the soil partition coefficient of the immobile domain. Ready only
if IDUAL=3 (L3/M).
ETA -- is the Freundlich isotherm exponent (0) in equation (2) of Chapter
2. Read only if LINR = 2. For dual domain simulations, ETA is
the Freundlich exponent of the mobile domain, if immobile domain
has distinct adsorption characteristics from the mobile domain
(IDUAL=3). If IDUAL=2 then both domains have similar
adsorption characteristics and this parameter defines the adsorption
for both domains. If IDUAL=1, then adsorption is assumed to
occur only in the immobile domain, and this parameter defines the
mobile domain adsorption characteristics.
ETAIM – is the Freundlich isotherm exponent (0) for the immobile domain.
Ready only if IDUAL=3 and LINR=2.
= 1 in item 1).
SCONC – is concentration at start of the simulation (M/L3) (within the mobile
domain, for dual domain simulations.
DUALRATE – is a first-order coefficient describing the rate of mass transfer
between the mobile and immobile regions (1/T). Read only if
IDUAL > 0.
SIMCONC – is the concentration in the immobile domain at the start of the
simulation (M/L3). Read only if IDUAL > 0
NPAREN(I) – is the number of parent species of component I
KPAREN(J) – Component index of the J th parent that decays to component I (for
each J=1, NPAREN(I)).
SFACTR(J) – Fraction of parent component J that decays to component I. (for
each J=1, NPAREN(I)). Decay of one unit of J produces SFACTR
units of component I, therefore SFACTR may also be used to scale
units in a decay process.

7.6.3 Prescribed Concentration Boundary (PCN1) Package Input
Input for the new PCN Package is read from the unit specified in IUNIT(23).
However, this input is required only if prescribed concentration (first-type) boundary
conditions exist. Recall that the IUNIT array, which is read by the Basic Flow (BAS)
package, determines the options or packages used in a MODFLOW-
SURFACT/MODHMS simulation. (See the BAS Package input.) The PCN input data
is required to incorporate first-type boundary conditions in a transport simulation, (i.e.,
to specify constant concentration boundaries). For a given stress period, each record
specifies the information on the location and prescribed concentration value of a species.
PCN boundaries can be assigned to any cell in the domain including the boundary cells.
Note that the prescribed concentration boundary condition supersedes any other boundary
condition that may have inadvertently specified on the same node.
1. Data: MXPCONC IPCNCC

Format: I10 I10
2. Data: ITMP
Format: I10

3 Data: Layer Row Column Conc Species ICNSHD
Format: I10 I10 I10 F10.0 I10 I10

MXPCONC -- is the maximum number of prescribed concentration entries in the
simulation.
IPCNCC -- is a flag and a unit number.
If IPCNCC > 0, it is the unit number on which mass-flux-terms of
each species at prescribed concentration cells will
be recorded whenever ICCCFL (see the output
control [OC] package input) is set.
If IPCNCC = 0, mass-flux-terms of species will not be printed or
recorded.
If IPCNCC < 0, mass-flux-terms at prescribed concentration cells
If ITMP < 0, prescribed concentration data from the last stress
If ITMP $ 0, ITMP will be the number of records (used in
item 3) for the current stress period.
Layer -- is the layer number of the model cell that contains the prescribed
Row -- is the row number of the model cell that contains the prescribed
Column -- is the column number of the model cell that contains the prescribed
Conc -- is the value of prescribed concentration for the model cell.
Species -- is the identification number of the species for which concentration is
prescribed in this record. 1 # Species # NSPECI, where NSPECI is
the total number of species in the transport simulation.

ICNSHD -- is an index.
If ICNSHD =1, this is a prescribed head B.C. for density-
dependent transport and a check should be
internally performed on this node to see if it
could be an outflow node. (If so, prescribed
concentration is not applied at this outflow node).
If ICNSHD = 0, no checks are performed, this node is ALWAYS
a prescribed concentration boundary.
7.7 REFERENCES
Bear, J., 1979. Hydraulics of Groundwater, McGraw-Hill, New York, 567 p.
HydroGeoLogic, Inc., 1994. DSTRAM: Density-dependent Solute Transport Analysis
Model-Documentation and User’s Guide, Version 4.1.
Frind, E. O., 1982. Simulation of long-term transient density-dependent transport in
groundwater, Advances in Water Resources, vol. 5, pp. 73-88.
Henry, H. R., 1964. Effects of dispersion on salt incroachment in coastal aquifers,
Seawater in Coastal Aquifers (H.H. Cooper et al.) U.S. Geological Survey, Water
Supply Paper 1613-C.
Lee, C. and R. T. Cheng, 1974. On seawater encroachment in coastal aquifers, Water
Resources Research, vol. 10, no. 5, pp. 1039-1043.
Pinder, G. F. and H. H. Cooper, Jr., 1970. A numerical technique for calculating the
transient position of the saltwater front, Water Resources Research, vol. 6, no. 3,
pp. 875-880.
Segol, G., G. F. Pinder, and W. G. Gray, 1975. A Galerkin-finite element technique for
calculating the transient position of the saltwater front, Water Resources Research,
vol. 11, no. 2, pp. 343-347.

CHAPTER 8: REACTION (RCT) PACKAGE
Chapter 8: Reaction (RCT) Package . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
8.0 Reaction (RCT) Package . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-1
8.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-1
8.2 Formulation of the Reaction Module . . . . . . . . . . . . . . . . . . . . . . . . . . 8-1
8.2.1 Module 1: Instantaneous Aerobic Decay of BTEX . . . . . . . . . . . 8-2
8.2.2 Module 2:Instantaneous Degradation of BTEX using Multiple Electron
Acceptors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-4
8.2.3 Module 3: Kinetic-limited degradation of BTEX using multiple electron
acceptors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-6
8.2.4 Module 4: Double Monod Model . . . . . . . . . . . . . . . . . . . . . . 8-10
8.2.5 Module 5: Sequential Decay Reactions . . . . . . . . . . . . . . . . . . . 8-12
8.2.6 Module Aerobic/Anaerobic Model for PCE/TCE Degradation . . . . 8-13
8.2.7 User Defined Module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-16
8.3 Reaction Package Input . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-20
8.4 Model Verification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-22
8.4.1. One-Dimensional Transport of a Three-Member Radioactive Decay Chain
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-22
8.4.2. Chain Decay of NAPL in a Single Cell . . . . . . . . . . . . . . . . . . . 8-23
8.5 Model Simulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-25
8.5.1. Module 1: Instantaneous Aerobic Decay of BTEX . . . . . . . . . . . 8-25
8.5.2. Module 2: Instantaneous Degradation of BTEX using Multiple Electron
Acceptors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-26
8.5.3 Module 3: Kinetic-limited Degradation of BTEX Using Multiple Electron
Acceptors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-27
8.5.4 Module 4: Double Monod Model . . . . . . . . . . . . . . . . . . . . . . 8-31
8.5.5 Module 5: Sequential Decay Reactions . . . . . . . . . . . . . . . . . . . 8-33
8.6 Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-37
8.7 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-38
8.0 REACTION (RCT) PACKAGE
8.1 GENERAL
The contaminant transport equations discussed earlier only allow for linear
reaction and equilibrium adsorption processes. The reaction package (RCT) expands this
capability by allowing for a more general set of non-linear reactions or kinetic adsorption
processes to be analyzed. The RCT package has been implemented as a multiphase
extension to the reaction modules provided in the popular RT3D code (Clement, 1997)
with further application to dual domain (dual porosity) systems. The reaction modules
included are:
1. Instantaneous Aerobic Decay of BTEX - which may be generalized to model
instantaneous reactions between any two components.
2. Instantaneous Degradation of BTEX using Multiple Electron Acceptors - which
may be generalized to model instantaneous reactions between any component and
a sequence of electron acceptors.
3. Kinetic Degradation of BTEX using Multiple Electron Acceptors
4. Double Monod Model
5. Sequential Decay Reactions - This capability is included with the general reaction
package, however, sequential decay reactions are solved in a numerically more
efficient manner using the linear reaction capability of the basic transport module.
6. Aerobic/Anaerobic PCE/TCE Degradation Model
7. User Defined Module - Any generalized set of reaction equations may be provided
by the user for immobile and mobile components that participate in transport.
This Chapter details the mathematical model and numerical solution methods for
transport of multiple reactive components. Details may also be found in Clement et al.,
1996a, 1996b, 1997, 1998. The reaction modules currently included in the RCT package
are detailed, followed by data input instructions and example problems to verify and
benchmark the new modules and to demonstrate applicability.
8.2 FORMULATION OF THE REACTION MODULE

The partial differential equation governing 3-D transport of contaminants in a
variably saturated porous medium with an immobile NAPL phase from which contaminant
components may dissolve into water and volatilize into air C wk may be expressed as:
∂
( ⎛
)
(Rk )eff C wk = ∂ ⎜⎜ Dijk ( ) ∂C wk ⎞⎟ ∂ vi C wk
−
( )
+ q ss C ssk + rk (1)
∂t ∂xi ⎝ eff
∂x j ⎟⎠ ∂x j
Where t is the time, vi is the Darcy velocity, is the water phase concentration of
contaminant component k, rk is the reaction rate (negative for depletion or decay), qss is
the volumetric flow rate via sources or sinks and c ssk the solute concentration of the
sources or sinks per unit volume of the porous medium. Effective retardation factor (Rk)eff
and effective dispersion tensor (D ) k
ij eff , are defined as:
Chapter 8: Reaction (RCT) Package 8-1

(Rk )eff = φ (S w + K awk S a + K nwk S n ) + ρ b k dk (C wk )η −1 (2)
(D ) k
ij eff ( )
= Dijk w
+ K aw
k
Dijk( ) a
+ K nw
k
( )
Dijk n
(3)
Where N is the porosity, Sw, Sa, and Sn are saturations of water, air, and NAPL; (D ) k
ij w ,
(D ) , and (D )
k
ij a
k
ij n are the apparent hydrodynamic tensors in water, air, and NAPL;
k k
K aw , and K nw are partitioning coefficients between water and gas and between water
and NAPL, respectively; k dk is the partitioning coefficient between water and soil; Db is
the bulk density of the porous medium; and 0 is the Freundlich isotherm exponent. For
the simplicity of the equations, the effective term eff will be dropped hereinafter. Note that
actually c wk denotes concentration of k in the active phase and similar equations to (2) and
(3) can be written with partitioning expressed from the air phase for an air-active system.
An operator-split (OS) numerical strategy is used to solve multi-species

contaminant transport in a multi-phase porous medium. Employing the OS strategy, the
solution algorithm initially solves the advection, dispersion, and source/sink terms for all
mobile components for a transport time step. Then, the coupled reaction terms are solved
implicitly using a differential equation solver. The solver automatically computes the
time-step sizes required to precisely integrate the reaction equations. Iterations may also
be performed to achieve convergence between the transport and reaction routines (to
handle larger time-steps than the OS method). The parameter MXOUTITR determines
whether iterations are performed or not. If MXOUTITR = 1 then iterations are not
performed between transport and reaction calculations in the classic OS method. If
MXOUTITR = 2 then a second iteration is performed to provide correction as in Flux
Corrected Transport (FCT) schemes. If MXOUTITR is greater than two, then the solution
iterates between transport and the reaction module till convergence is achieved. The
following sections describe the details of all the reaction modules currently available,
followed by the detailed input data structures.
8.2.1 Module 1: Instantaneous Aerobic Decay of BTEX
The purpose of this module is to simulate aerobic degradation of BTEX using an

instantaneous reaction model in both saturated and unsaturated porous media in which the
immobile NAPL phase may be present. The model is similar to those simulated by
BIOPLUME-II (Rafai et al., 1987) with the inclusion of multiphase considerations.
The overall aerobic reaction stoichiometry for a fuel hydrocarbon compound (e.g.,
benzene) can be written as (Lu et al, 1999):
C 6 H 6 + 7.5O2 → 6CO2 + 3H 2 O (4)
8-2 Chapter 8: Reaction (RCT) Package

The transport equations of hydrocarbon and oxygen in the active phase for this model are:
∂ (R H H w ) ∂ ⎛⎜ H ∂H w ⎞⎟ ∂ (vi H w )
= Dij − + q ss H ss + rH (5)
∂t ∂xi ⎜⎝ ∂x j ⎟⎠ ∂xi
∂ (RO Ow ) ∂ ⎛⎜ O ∂Ow ⎞⎟ ∂ (vi Ow )

= Dij − + q ss Oss + rO (6)
∂t ∂xi ⎜⎝ ∂x j ⎟⎠ ∂xi
Where H and O represent hydrocarbon and oxygen, respectively, and other terms are
defined following the definitions explained in the previous section.
At each time step, an instantaneous reaction algorithm is used to model the

removal rates. According to this algorithm, either hydrocarbon or oxygen mass
(whichever is limiting) will be reduced to zero within a grid cell, after a reaction time step.
Total masses of hydrocarbon and oxygen in a grid cell are:
M H = RH H w (7)
M O = RO Ow (8)
Where MH and MO are the masses of hydrocarbon and oxygen, respectively. The reaction
algorithm based on the total mass is written as:
M H (t + 1) = M H (t ) − M O (t ) / F and M O (t + 1) = 0, when M H (t ) > M O (t ) / F (9)
and
M O (t + 1) = M H (t ) − M H (t )F and M H (t + 1) = 0, when M O (t ) > M H (t )F (10)
Where t refers to a particular time step and F is the stoichiometric ratio.
Concentrations of hydrocarbon and oxygen after the reaction can be obtained as:
H w = M H (t + 1) / RH (11)
and
Ow = M O (t + 1) / RO (12)
For non-linear sorption, an iteration method is used to obtain accurate Hw and Ow.
Details of the Reaction Module:
Total number of components (ncomp) = 2

The component names are: BTEX and Oxygen
Number of input parameters = 1

Parameter 1: Stoichiometric ratio F (Oxygen to BTEX) = 3.14.
By changing the value of F instantaneous reaction between any other chemicals can be
simulated using this module.
8.2.2 Module 2:Instantaneous Degradation of BTEX using Multiple Electron

Acceptors
This reaction model simulates instantaneous biodegradation of BTEX via five
different degradation pathways: aerobic respiration, denitrification, iron reduction, sulfate
reduction, and methanogenesis. Thus, hydrocarbon uses, in sequential order, oxygen,
nitrate, ferric iron (iron III), sulphate, and carbon-dioxide as electron acceptors for
degradation. The reaction algorithm used in this model is similar to the one in BIOPLUME
III (Rafai et al., 1998). The biochemical reactions (with benzene considered here as an
example) are (Lu et al, 1999):
7.50 2 + C6 H 6 → 6CO2 + H 2O (13)
6 NO3− + 6 H + + C 6 H 6 → 6CO2 + 6 H 2 O + 3 N 2 (14)
30 Fe(OH )3 + 60 H + + C 6 H 6 → 6CO2 + 78 H 2 O + 30 Fe 2+ (15)
3.75SO42− + 7.5 H + + C 6 H 6 → 6CO2 + 3H 2 O + 3.75 H 2 S (16)
4.5 H 2 O + C 6 H 6 → 2.25CO2 + 3.75CH 4 (17)
The reactions are assumed to occur in a sequential order, as listed above. Since ferric iron
is immobile, it is simulated as an initial condition available within the domain which is
consumed to produce ferrous iron (iron II) that is mobile (the rate of ferrous iron
production is the same as the rate of ferric iron reduction). Therefore only transport of
ferrous iron is simulated while the solid phase ferric iron available for reaction is updated
internally. Similarly, for methanogenesis, methane is produced and its transport in the
aquifer is tracked. Reduction by methanogenesis is assumed to occur only up to a
maximum methanogenesis potential, Epot, which is the total capacity for methanogenesis
of the domain. Beyond this level, the hydrocarbon is not reduced further. The coupled
transport equations are therefore written for the mobile components as:
∂ (R H H w ) ∂ ⎛⎜ H ∂H w ⎞⎟ ∂ (vi H w )
= Dij − + q ss H ss + rH (18)
∂t ∂xi ⎜⎝ ∂x j ⎟⎠ ∂xi
and
∂ (RO Ow ) ∂ ⎛⎜ O ∂Ow ⎞⎟ ∂ (vi Ow )

= Dij − + q ss Oss + rO (19)
∂t ∂xi ⎜⎝ ∂x j ⎟⎠ ∂xi
and

∂ (R N N w ) ∂ ⎛⎜ N ∂N w ⎞⎟ ∂ (vi N w )
= Dij − + q ss N ss + rN (20)
∂t ∂xi ⎜⎝ ∂x j ⎟⎠ ∂xi
and
∂ (R F Fw ) ∂ ⎛⎜ F ∂Fw ⎞⎟ ∂ (vi Fw )
= Dij − + q ss Fss + rF (21)
∂t ∂xi ⎜⎝ ∂x j ⎟⎠ ∂xi
and
∂ (R P Pw ) ∂ ⎛⎜ P ∂Pw ⎞⎟ ∂ (vi Pw )
= Dij − + q ss Pss + rP (22)
∂t ∂xi ⎜⎝ ∂x j ⎟⎠ ∂xi
and
∂ (R E E w ) ∂ ⎛⎜ E ∂E w ⎞⎟ ∂ (vi E w )
= Dij − + Qss E ss + rE (23)
∂t ∂xi ⎜⎝ ∂x j ⎟⎠ ∂xi
Where H, O, N, F, P, and E represent hydrocarbon, oxygen, nitrate, ferrous iron, sulfate,

and methane, respectively.
The biodegradation of hydrocarbon using the aerobic and anaerobic electron

acceptors is simulated using the instantaneous reaction and the principle of superposition
as in BIOPLUME III (Rifai et al., 1998). The following general instantaneous reaction
algorithm is applied to utilize different electron acceptors sequentially:
M D (t + 1) = M D (t ) − M A (t ) / FA / D and M A (t + 1) = 0, when M D (t ) > M A (t ) / FA / D (24)

and
M A (t + 1) = M A (t ) − M D (t ) FA / D and M D (t + 1) = 0, when M A (t ) > M D (t ) F A / D (25)
where, t refers to a particular time step, MD and MA are the masses of the electron donor
and electron acceptor, and FA/D is the stoichiometric ratio between the electron acceptor
and the electron donor.
Note that the rate of ferrous iron production is equal to the rate of ferric iron reduction and
hence
rF = −rFe (26)

The component names are: BTEX, Oxygen, Nitrate, Fe2+, Sulfate, Methane

Parameter 1: Fe3+ available for biodegration (initial condition, as concentration in
water phase)
Parameter 2: Methanogenesis potential (initial condition, as concentration in water
phase)
The other five constants are fixed values (the values assume concentrations in
mg/L):
Parameter 3: stoichiometric ratio of oxygen (consumed) to BTEX = 3.14
Parameter 4: stoichiometric ratio of nitrate (consumed) to BTEX = 4.9
Parameter 5: stoichiometric ratio of Fe3+ (consumed) to BTEX = 21.8
Parameter 6: stoichiometric ratio of sulfate (consumed) to BTEX = 4.7
Parameter 7: stoichiometric ratio of methane (produced) to BTEX = 0.78
8.2.3 Module 3: Kinetic-limited degradation of BTEX using multiple electron

acceptors
The reactive transport model considered here describes the transport and rate-
limited degradation of hydrocarbon through five different degradation pathways: aerobic
respiration, denitrification, iron reduction, sulfate reduction, and methanogenesis
(Clement, 1997). The stoichiometry of the degradation reactions is described by equations
(13) through (17), and the relevant transport equations are provided by equations (18)
through (23). Note again, as in module 2, that the immobile ferric iron is simulated as an
initial condition which is consumed to produce ferrous iron that is transported through the
domain. The reactions for this case, however, are kinetic-limited. The rate of hydrocarbon
decay is assumed to be directly proportional to hydrocarbon concentration (i.e., the decay
is a first-order reaction). A Monod-type term is used to account for the presence (or
absence) of various electron acceptors, and an inhibition model is used to simulate the
sequential electron acceptor utilization process. The electron acceptors are, in sequential
order, oxygen, nitrate, ferric iron, sulphate and carbon dioxide. The kinetic expressions
used for modeling hydrocarbon decay are (Lu et al, 1999):
Ow Os
rH ,O = − k wO H w φS − k O
H ρb
K wO + Ow K sO + Os
w s s
(27)
O O
− k H a O a φS a − k nO H n O n φS n
O
K a + Oa K n + On
a
Nw K iO, w Ns K iO, s
rH , N = −k H w N
N
φS w − k s H s N
N
ρb
K w + N w K iO, w + Ow K s + N s K iO, s + Os
w
(28)
Na K iO,a Nn K iO,n
− ka H a N
N
φS a − k n H n N
N
φS n
K a + N a K iO,a + Oa K n + N n K iO,n + On

Fw K iO,w K iN,w
rH ,F = −k H w F
F
φS w
K w + Fw K iO,w + Ow K iN,w + N w
w
Fs K iO,s K iN,s
− k Hs F
F
ρb
K s + Fs K iO,s + Os K iN,s + N s
s
(29)
Fa K iO,a K iN,a
k aF H a φS a
K + Fs K iO,a + Oa K iN,a + N a
F
a
Fn K iO,n K iN,n
− k Hn F
F
φS n
K n + Fn K iO,n + On K iN,n + N n
n
Pw K iO, w K iN, w K iF, w

rH , P = −k H w P
P
φS w
K w + Pw K iO, w + Ow K iN, w + N w K iF, w + Fw
w
Ps K iO, s K iN, s K iF, s

− k Hs P
P
ρb
K s + Ps K iO, s + Os K iN, s + N s K iF, s + Fs
s
(30)
O N F
Pa K K K
− k aP H a
i ,a
φS a
i ,a i ,a
K aP + Pa K iO,a + Oa K iN,a + N a K iF,a + Fa

Pn K iO,n K iN,n K iF,n
− k Hn P
P
φS P
K n + Pn K iO,n + On K iN,n + N n K iF,n + Fn
n
Cw K iO, w K iN, w K iF, w K iP, w

rH ,C = −k H w C
C
φS w
K w + C w K iO, w + Ow K iN, w + N w K iF, w + Fw K iP, w + Pw
w
Cs K iO, s K iN, s K iF, s K iP, s

− k Hs C
C
ρb
K s + C s K iO, s + Os K iN, s + N s K iF, s + Fs K iP, s + Ps
s
(31)
Ca K iO,a K iN,a K iF,a K iP,a
− Ka Ha C
C
φS a
K a + C a K iO,a + Oa K iN,a + N a K iF,a + Fa K iP,a + Pa
Cn K iO,n K iN,n K iF,n K iP,n
− k Hn C
C
φS n
K n + C n K iO,n + On K iN,n + N n K iF,n + Fn K iP,n + Pn
n
where rH,O is the hydrocarbon destruction rate utilizing oxygen; rH,N is the destruction rate
utilizing nitrate, rH,F is the destruction rate utilizing Fe3+; rH,P is the destruction rate
utilizing sulfate; rH,C is the destruction rate via methanogenesis; kwo, kso, kao, kno are the
first-order degradation rate constants for hydrocarbon utilizing oxygen as the electron

acceptor in water, on solid, in gas, and in NAPL, respectively; Kwo, Kso, Kao, and Kno are
the Monod half-saturation constants in water, on solid, in gas, and in NAPL, respectively;
Ki,wo, Ki,so, Ki,ao, and Ki,no, are the oxygen inhibition constant in water, on solid, in gas, and
in NAPL, respectively; and similar nomenclature is used for all subsequent reactions with
superscripts N, F, P, and C denoting nitrate, ferric iron, sulphate, and carbon dioxide,
respectively. Note that by setting the half-saturation constant to an arbitrarily small value,
we simulate zero-order dependency with respect to the electron donor and thus, a first-
order dependency with respect to hydrocarbon. The inhibition constants are typically set
to some arbitrarily small values to simulate the sequential electron acceptor utilization
process.
The concentration of carbon dioxide, C in equation (32), is related to the methane

concentration (E) via the methanogenesis potential Epot, as:
C = E pot − E (32)
Thus, methanogenesis is maximum when methane concentrations are zero and drop to zero
when methane concentration reaches the methanogenesis potential. The total rate of
hydrocarbon destruction rH via all decay processes is written as:
rH = rH ,O + rH , N + rH , F + rH , P + rH ,C (33)
Rates of electron acceptor utilization (or product formation) are given by the rates of
hydrocarbon destruction multiplied by an appropriate yield coefficient (Y) as:
rO = YO / H rH ,O (34)
R N = YO / H rH ,O (35)
rF = YO / H rH ,O (36)
rP = YO / H rH ,O (37)
rC = YC / H rH ,C (38)

The component names are: BTEX, Oxygen, Nitrate, Iron (Fe2+), Sulfate,
Methane
Parameter 1: Fe3+ available for biodegradation (initial condition, as
concentration in water phase)
Parameter 2: Methanogenesis potential (initial condition, as concentration in water
phase)
Parameter 3: Hydrocarbon decay rate via aerobic process in water
Parameter 4: Hydrocarbon decay rate via denitrification in water
Parameter 5: Hydrocarbon decay rate via iron reduction in water

Parameter 6: Hydrocarbon decay rate via sulfate reduction in water
Parameter 7: Hydrocarbon decay rate via methanogenesis in water
Parameter 8: Hydrocarbon decay rate via aerobic process on solid
Parameter 9: Hydrocarbon decay rate via denitrification on solid
Parameter 10: Hydrocarbon decay rate via iron reduction on solid
Parameter 11: Hydrocarbon decay rate via sulfate reduction on solid
Parameter 12: Hydrocarbon decay rate via methanogenesis on solid
Parameter 13: Hydrocarbon decay rate via aerobic process in gas
Parameter 14: Hydrocarbon decay rate via denitrification in gas
Parameter 15: Hydrocarbon decay rate via iron reduction in gas
Parameter 16: Hydrocarbon decay rate via sulfate reduction in gas
Parameter 17: Hydrocarbon decay rate via methanogenesis in gas
Parameter 18: Hydrocarbon decay rate via aerobic process in NAPL
Parameter 19: Hydrocarbon decay rate via denitrification in NAPL
Parameter 20: Hydrocarbon decay rate via iron reduction in NAPL
Parameter 21: Hydrocarbon decay rate via sulfate reduction in NAPL
Parameter 22: Hydrocarbon decay rate via methanogenesis in NAPL
Parameter 23: Half saturation constant for oxygen (0.5 mg/L) in water
Parameter 24: Half saturation constant for nitrate (0.5 mg/L) in water
Parameter 25: Half saturation constant for Fe3+ (0.5 mg/L) in water
Parameter 26: Half saturation constant for sulfate (0.5 mg/L) in water
Parameter 27: Half saturation constant for methane (0.5 mg/L) in water
Parameter 28: Half saturation constant for oxygen (0.5 mg/L) on solid
Parameter 29: Half saturation constant for nitrate (0.5 mg/L) on solid
Parameter 30: Half saturation constant for Fe3+ (0.5 mg/L) on solid
Parameter 31: Half saturation constant for sulfate (0.5 mg/L) on solid
Parameter 32: Half saturation constant for methane (0.5 mg/L) on solid
Parameter 33: Half saturation constant for oxygen (0.5 mg/L) in gas
Parameter 34: Half saturation constant for nitrate (0.5 mg/L) in gas
Parameter 35: Half saturation constant for Fe3+ (0.5 mg/L) in gas
Parameter 36: Half saturation constant for sulfate (0.5 mg/L) in gas
Parameter 37: Half saturation constant for methane (0.5 mg/L) in gas
Parameter 38: Half saturation constant for oxygen (0.5 mg/L) in NAPL
Parameter 39: Half saturation constant for nitrate (0.5 mg/L) in NAPL
Parameter 40: Half saturation constant for Fe3+ (0.5 mg/L) in NAPL
Parameter 41: Half saturation constant for sulfate (0.5 mg/L) in NAPL
Parameter 42: Half saturation constant for methane (0.5 mg/L) in NAPL
Parameter 43: Inhibition coefficient for oxygen reaction (0.001 mg/L) in water
Parameter 44: Inhibition coefficient for nitrate reaction (0.001 mg/L) in water
Parameter 45: Inhibition coefficient for Fe3+ reaction (0.001 mg/L) in water
Parameter 46: Inhibition coefficient for sulfate reaction (0.001 mg/L) in water
Parameter 47: Inhibition coefficient for oxygen reaction (0.001 mg/L) on solid
Parameter 48: Inhibition coefficient for nitrate reaction (0.001 mg/L) on solid
Parameter 49: Inhibition coefficient for Fe3+ reaction (0.001 mg/L) on solid
Parameter 50: Inhibition coefficient for sulfate reaction (0.001 mg/L) on solid
Parameter 51: Inhibition coefficient for oxygen reaction (0.001 mg/L) in gas
Parameter 52: Inhibition coefficient for nitrate reaction (0.001 mg/L) in gas
Parameter 53: Inhibition coefficient for Fe3+ reaction (0.001 mg/L) in gas
Parameter 54: Inhibition coefficient for sulfate reaction (0.001 mg/L) in gas
Parameter 55: Inhibition coefficient for oxygen reaction (0.001 mg/L) in NAPL
Parameter 56: Inhibition coefficient for nitrate reaction (0.001 mg/L) in NAPL
Parameter 57: Inhibition coefficient for Fe3+ reaction (0.001 mg/L) in NAPL

Parameter 58: Inhibition coefficient for sulfate reaction (0.001 mg/L) in NAPL
Parameter 59: Oxygen yield ( Y
O 2 / HC
= 3.14 )
Parameter 60: Nitrate yield ( YNO 3 / HC = 4.9 )
Parameter 61: Iron (Fe2+) production yield ( YFe 2 + / HC = 21.8 )
Parameter 62: Sulfate yield ( YSO 4 / HC = 4.7 )
Parameter 63: Methane production yield ( YCH 4 / HC = 0.78 )
8.2.4 Module 4: Double Monod Model

This model simulates the reaction between an electron donor and an electron
acceptor mediated by actively growing bacteria cells living in aqueous, soil, air, and
NAPL phases. Applications of different types of double-Monod models are discussed by:
Molz et al. (1986); Rafai and Bedient (1990); Taylor and Jaffe (1990); Clement et al.
(1996a); Clement et al. (1996b); Clement et al. (1997); and Clement et al. (1998).
Assuming an equilibrium model for sorption and a Monod kinetic model for
biological reactions (Rafai and Bedient, 1990; Clement et al., 1996a), the fate and
transport of an electron donor (hydrocarbon, for example) in a multi-dimensional saturated
porous medium can be written as:
∂R D Dw ∂ ⎛⎜ ∂Dw ⎞⎟ ∂ (vi Dw )
= Dij − + q ss Dss + rD (39)
∂t ∂xi ⎜⎝ ∂x j ⎟⎠ ∂xi
∂R A Aw ∂ ⎛⎜ A ∂Aw ⎞⎟ ∂
= Dij − (vi Aw ) + q ss Ass + rA (40)
∂t ∂xi ⎜⎝ ∂w j ⎟⎠ ∂xi
∂R X X w ∂ ⎛⎜ ∂X w ⎞⎟ ∂ (vi X w )
= X ij − + q ss X ss + rX (41)
∂t ∂xi ⎜⎝ ∂x j ⎟⎠ ∂xi
where D, A, and X represent the electron donor, electron acceptor, and bacterial cells,
respectively; and subscripts w and s represent water and solid phases, respectively. The
reaction terms are defined as:
∂ρ b X s
= rX s (42)
∂t
Dw K wA Ds K sA
rD = − µ mw X w φS − µ s
X ρb
K wD + Dw K wA + Aw K sD + Ds K sA + As
w m s
(43)
D K aA Dn K nA
− µ ma X a D a φS − µ n
X φS n
K a + Da K aA + Aa K nD + Dn K nA + An
a m n

rA = Y A / D rD (44)
Dw K wA Da K aA
rX = −Y X / D µ X w Dw
φS w − Y X / D µ m X a D
a
φS a
K w + Dw K wA + Aw K a + Da K aA + Aa
m
Dn K nA (45)
− Y X / D µ mn X n φS n
K nD + Dn K nA + An
( )
− K att + K we X wφS w + K det X s ρ b − K ae X aφS a − K ne X nφS n
rX s = Y X / D µ sm X s
Ds As
( )
ρ b − K det + K se X s ρ b + K att X wφS w (46)
K + Ds K s + As
D
s
A
Where KwD, KsD, KaD, and KnD are the half saturation coefficients in water, on solid, in gas,
in NAPL for the electron donor; KwA, KsA, KaA, and KnA are the half saturation coefficients
in water, on solid, in gas, in NAPL for the electron acceptor; :mw, :ms, :ma, and :mn are
the contaminant utilization rates in water, on solid, in gas, and in NAPL, Katt is the
bacterial attachment coefficient; Kdet is the bacterial detachment coefficient; and Kwe, Kse,
Kae, and Kne are the endogenous cell death or decay coefficients in water, on solid, in gas,
and in NAPL.
The coupled governing equations are solved employing the OS technique.

Mobile components (mcomp) = 3
The component names are: ED (Electron Donor), EA (Electron Acceptor), AB
(Aqueous phase mobile Bacteria), SB (Soil-phase immobile
Bacteria).
Parameter 1 = Specific utilization rate in water
Parameter 2 = Specific utilization rate on solid
Parameter 3 = Specific utilization rate in gas
Parameter 4 = Specific utilization rate in NAPL
Parameter 5 = Monod half saturation constant for electron donor in water
Parameter 6 = Monod half saturation constant for electron donor on solid
Parameter 7 = Monod half saturation constant for electron donor in gas
Parameter 8 = Monod half saturation constant for electron donor in NAPL
Parameter 9 = Monod half saturation constant for electron acceptor in water
Parameter 10 = Monod half saturation constant for electron acceptor on solid
Parameter 11 = Monod half saturation constant for electron acceptor in gas
Parameter 12 = Monod half saturation constant for electron acceptor in NAPL
Parameter 13 = Biomass produced per unit amount of electron donor utilized,
YX/D
Parameter 14 = Electron acceptor used per unit amount of electron donor
utilized, YA/D
Parameter 15 = Endogenous cell death or decay coefficients in water
Parameter 16 = Endogenous cell death or decay coefficients on solid
Parameter 17 = Endogenous cell death or decay coefficients in air

Parameter 18 = Endogenous cell death or decay coefficients in NAPL
Parameter 19 = First-order, bacterial attachment rate
Parameter 20 = First-order, bacterial detachment rate
Parameter 21 = Soil bulk density, rb
8.2.5 Module 5: Sequential Decay Reactions
This model simulates reactive transport coupled by a series of sequential

degradation reactions (up to four components). The example considered here focuses on
modeling dechlorination of PCE and its daughter products under anaerobic conditions.
Assuming first-order decay kinetics, the transport and transformation of a

sequential decay chain can be simulated by solving the following set of partial differential
equations:
∂R A Aw ∂ ⎛⎜ ∂A ⎞ ∂ (vi Aw )
= Dij w ⎟ − + qss Ass + rA (47)
∂t ∂xi ⎜⎝ ∂x j ⎟⎠ ∂xi
∂RB Bw ∂ ⎛⎜ B ∂Bw ⎞⎟ ∂(vi Bw )

= Dij − + q ss Bss + rB (48)
∂t ∂xi ⎜⎝ ∂x j ⎟⎠ ∂xi
∂RC C w ∂ ⎛⎜ C ∂C w ⎞⎟ ∂ (vi C w )
= Dij − + q ss C ss + rC (49)
∂t ∂xi ⎜⎝ ∂x j ⎟⎠ ∂xi
∂RD Dw ∂ ⎛⎜ ∂Dw ⎞⎟ ∂ (vi Dw )

= Dij − + q ss Dss + rD (50)
∂t ∂xi ⎜⎝ ∂x j ⎟⎠ ∂xi
where A, B, C, and D represent contaminant species. The reaction terms are defined as:
rA = − K wA AwφS w − K sA As ρ b − K aA AaφS a − K nA AnφS n (51)
(
rB = YB / A K wA AwφS w + K sA As ρ b + K aA AaφS a + K nA AnφS n )
( )
(52)
− K wB BwφS w + K sB Bs ρ b + K aB BaφS a + K nB BnφS n
(
rC = YC / B K wB BwφS w + K sB Bs ρb + K aB BaφS a + K nB BnφS n )
( )
(53)
− K wC CwφS w + K sC Cs ρb + K aC CaφS a + K nC CnφS n
(
rD = YD / C K wC C wφS w + K sC C s ρ b + K aC C aφS a + K nC C nφS n )
( )
(54)
− K DwφS w + K Ds ρ b + K DaφS a + K DnφS n
D
w
D
s
D
a
D
n

Where KwA, KsA, KaA, and KnA are the first order decay rate coefficients in water, on solid,
in gas, in NAPL for species A; Y represents different stoichiometric yield values. Similar
nomenclature is used for other species.

Mobile components (mcomp) = 4
The component names are: PCE, TCE, DCE, and VC
Number of input parametera = 19
Parameter 1 = PCE first-order degradation rate in water
Parameter 2 = TCE first-order degradation rate in water
Parameter 3 = DCE first-order degradation rate in water
Parameter 4 = VC first-order degradation rate in water
Parameter 5 = PCE first-order degradation rate on solid
Parameter 6 = TCE first-order degradation rate on solid
Parameter 7 = DCE first-order degradation rate on solid
Parameter 8 = VC first-order degradation rate on solid
Parameter 9 = PCE first-order degradation rate in gas
Parameter 10 = TCE first-order degradation rate in gas
Parameter 11 = DCE first-order degradation rate in gas
Parameter 12 = VC first-order degradation rate in gas
Parameter 13 = PCE first-order degradation rate in NAPL
Parameter 14 = TCE first-order degradation rate in NAPL
Parameter 15 = DCE first-order degradation rate in NAPL
Parameter 16 = VC first-order degradation rate in NAPL
Parameter 17 = Yield, YB/A (Ytce_pce=0.792)
Parameter 18 = Yield, YC/B (Ydce_tce=0.738)
Parameter 19 = Yield, YD/C (Yvc_dce=0.644)
The yield values are on a mg/L basis; to be consistent, the user must use mg/L units for
all concentrations.
8.2.6 Module Aerobic/Anaerobic Model for PCE/TCE Degradation
This model simulates degradation of PCE/TCE and their degradation products via
both aerobic and anaerobic pathways. The conceptual model used to model all chlorinated
solvent degradation reactions, mediated by aerobic and anaerobic dechlorination processes,
is described in the figure below:
Anaerobic Anaerobic Anaerobic Anaerobic

PCE TCE DCE VC ETH
Ae
Ae
Ae
Ae
Ae
ro b
ro b
ro b
ro b
ro b
ic
ic
ic
ic
ic
4Cl- 3Cl- 2Cl- Cl- Products
Figure 8.1 Conceptual Model for Chlorinated Solvent Biodegradation (after

Clement, 1997).

Assuming first-order biodegradation kinetics, transport and transformation of
PCE, TCE, DCE, VC, ETH, and CI can be simulated by solving the following set of
partial differential equations:
∂RP Pw ∂ ⎛⎜ P ∂Pw ⎞⎟ ∂ (vi Pw )

= Dij − + q ss Pss + rP1 + rP 2 (55)
∂t ∂xi ⎜⎝ ∂x j ⎟⎠ ∂xi
∂RT Tw ∂ ⎛⎜ T ∂Tw ⎞⎟ ∂ (viTw )

= Dij − + q ssTss − YT / P rP1 + rT 1 + rT 2 (56)
∂t ∂xi ⎜⎝ ∂x j ⎟⎠ ∂xi
∂RD Dw ∂ ⎛⎜ D ∂Dw ⎞⎟ ∂ (vi Dw )

= Dij − + q ss Dss − YD / T rT 1 + rD1 + rD 2 (57)
∂t ∂xi ⎜⎝ ∂x j ⎟⎠ ∂xi
Where P, T, D, V, E, and C represent PCE, TCE, DCE, VC, Ethene, and chloride,
respectively; subscripts 1 and 2 represent anaerobic and aerobic pathways; and the reaction
terms are defined as:
∂RV Vw ∂ ⎛⎜ V ∂Vw ⎞⎟ ∂ (viVw )

= Dij − + q ssVss − YV / D rD1 + rV 1 + rV 2 (58)
∂t ∂xi ⎜⎝ ∂x j ⎟⎠ ∂xi
∂RE E w ∂ ⎛⎜ E ∂E w ⎞⎟ ∂ (vi E w )
= Dij − + q ss E ss − YE / V rV 1 + rE1 + rE 2 (59)
∂t ∂xi ⎜⎝ ∂x j ⎟⎠ ∂xi
∂RC C w ∂ ⎛⎜ C ∂C w ⎞⎟ ∂ (vi C w )
= Dij − + q ss C ss + YC / P (rP ,1 + 4rP , 2 )
∂t ∂xi ⎜⎝ ∂x j ⎟⎠ ∂xi (60)
+ YC / T (rT ,1 + 3rT , 2 ) + YC / D (rD ,1 + 2rD , 2 ) + YC / V (rV ,1 + rV , 2 )
rP ,1 = − K wP,1 PwφS w − K sP,1 Ps ρ b − K aP,1 PaφS a − K nP,1 PnφS n (61)
rP , 2 = − K wP, 2 PwφS w − K sP, 2 Ps ρ b − K aP, 2 PaφS a − K nP, 2 PnφS n (62)
rT ,1 = − K wT ,1TwφS w − K sT,1Ts ρ b − K aT,1TaφS a − K nT,1TnφS n (63)
rT , 2 = − K wT , 2TwφS w − K sT, 2Ts ρ b − K aT, 2TaφS a − K nT, 2TnφS n (64)

rD ,1 = − K wD,1 DwφS w − K sD,1 Ds ρ b − K aD,1 DaφS a − K nD,1 DnφS n (65)
rD , 2 = − K wD, 2 DwφS w − K sD, 2 Ds ρ b − K aD, 2 DaφS a − K nD, 2 DnφS n (66)
rV ,1 = − K wV,1VwφS w − K sV,1Vs ρ b − K aV,1VaφS a − K nV,1VnφS n (67)
rV , 2 = − K wV, 2VwφS w − K sV, 2Vs ρ b − K aV, 2VaφS s − K nV, 2VnφS n (68)
rE ,1 = − K wE,1 E wφS w − K sE,1 E s ρ b − K aE,1 E aφS a − K nE,1 E nφS n (69)
rE , 2 = − K wE, 2 E wφS w − K sE, 2 E s ρ b − K aE, 2 E aφS a − K nE, 2 E nφS n (70)
Where Kw,1P, Ks,1P, Ka,1P, and Kn,1P are the PCE anaerobic first order decay rate coefficients
in water, on solid, in gas, in NAPL; Kw,2P, Ks,2P, Ka,2P, and Kn,2P are the PCE aerobic first
order decay rate coefficients in water, on solid, in gas, in NAPL; Y represents different
stoichiometric yield values. Similar nomenclature is used for other species.

The component names are: PCE, TCE, DCE, VC, ETH, and C1
Parameter 1 = Anaerobic decay rate for PCE in water
Parameter 2 = Anaerobic decay rate for TCE in water
Parameter 3 = Anaerobic decay rate for DCE in water
Parameter 4 = Anaerobic decay rate for VC in water
Parameter 5 = Anaerobic decay rate for ETH in water
Parameter 6 = Anaerobic decay rate for PCE on solid
Parameter 7 = Anaerobic decay rate for TCE on solid
Parameter 8 = Anaerobic decay rate for DCE on solid
Parameter 9 = Anaerobic decay rate for VC on solid
Parameter 10 = Anaerobic decay rate for ETH on solid
Parameter 11 = Anaerobic decay rate for PCE in gas
Parameter 12 = Anaerobic decay rate for TCE in gas
Parameter 13 = Anaerobic decay rate for DCE in gas
Parameter 14 = Anaerobic decay rate for VC in gas
Parameter 15 = Anaerobic decay rate for ETH in gas
Parameter 16 = Anaerobic decay rate for PCE in NPAL
Parameter 17 = Anaerobic decay rate for TCE in NAPL
Parameter 18 = Anaerobic decay rate for DCE in NAPL
Parameter 19 = Anaerobic decay rate for VC in NAPL
Parameter 20 = Anaerobic decay rate for ETH in NAPL
Parameter 21 = Aerobic decay rate for PCE in water
Parameter 22 = Aerobic decay rate for TCE in water
Parameter 23 = Aerobic decay rate for DCE in water
Parameter 24 = Aerobic decay rate for VC in water
Parameter 25 = Aerobic decay rate for ETH in water
Parameter 26 = Aerobic decay rate for PCE on solid
Parameter 27 = Aerobic decay rate for TCE on solid

Parameter 28 = Aerobic decay rate for DCE on solid
Parameter 29 = Aerobic decay rate for VC on solid
Parameter 30 = Aerobic decay rate for ETH on solid
Parameter 31 = Aerobic decay rate for PCE in gas
Parameter 32 = Aerobic decay rate for TCE in gas
Parameter 33 = Aerobic decay rate for DCE in gas
Parameter 34 = Aerobic decay rate for VC in gas
Parameter 35 = Aerobic decay rate for ETH in gas
Parameter 36 = Aerobic decay rate for PCE in NPAL
Parameter 37 = Aerobic decay rate for TCE in NAPL
Parameter 38 = Aerobic decay rate for DCE in NAPL
Parameter 39 = Aerobic decay rate for VC in NAPL
Parameter 40 = Aerobic decay rate for ETH in NAPL
Parameter 41 = Yield, YT/P (YTCE/PCE=0.792)
Parameter 42 = Yield, YD/T (YDCE/TCE=0.738)
Parameter 43 = Yield, YV/D (YVC/DCE=0.645)
Parameter 44 = Yield, YE/V (YETH_VC=0.449)
Parameter 45 = Yield, YC/P (YCl/PCE=0.214)
Parameter 46 = Yield, YC/T (YCl/TCE=0.270)
Parameter 47 = Yield, YC/D (YCl/DCE=0.366)
Parameter 48 = Yield, YC/V (YCl/VC=0.567)
8.2.7 User Defined Module
The user may write a subroutine containing all reactions pertinent to the modeling
effort, which may be linked into the executable. Note that with the operator splitting
technique used here, the equation for concentration change due to reaction at a time-step
is written as:
∂C wk rk
= (71)
∂t ( Rk ) eff
Which is the equation required by the reaction module. The effective retardation term,
(Rk)eff is given by equation (2) and can be calculated by a subroutine MASSC. Other
subroutines / functions available to the user include subroutine GETCON that calculates
the concentration of a component given it’s total mass and effective retardation coefficient,
and a real*4 function CS which calculates the last term of equation (2) for adsorption
considerations. The subroutine statements are as follows:
• SUBROUTINE MASSC(CONC,LSPEC,AMASS) where CONC is the

concentration, LSPEC is the associated species, and AMASS is the
effective retardation term for the associated species.
• SUBROUTINE GETCON(TMASS,AMASS,CONC,LSPEC) where

TMASS is the total mass of component, AMASS is the effective
retardation term, CONC is the concentration, and LSPEC is the associated
species index.
• FUNCTION CS(RHOBK,F_CNEW,F_AETA,F_EQPAR,LINR) where

RHOBK is the adsorption coefficient times bulk density, F_CNEW is the
concentration of the species for which this function is called, F_AETA is
the Freundlich exponent, F_EQPAR is unity when water is the active

phase, and 1.0 / EQCOEF_GW(LSPEC) if air is the active phase, and
EQCOEF_GW(LSPEC) is the gas-water partitioning coefficient of species
LSPEC for which this term is being determined.
As an example , the following subroutine may be written to simulate sequential decay

reactions as discussed for Module 5. Note also, the common blocks available to the user
to transfer information from the transport module to the reaction module.
C-----Reaction Model#5-----------------------------------------------------------------------------C
c Four species (chlorinated solvents PCE -> TCE -> DEC -> VC)
c sequential, anaerobic, first-order, rate-limited, reductive dechlorination reactions.
C*********************************************************************C
SUBROUTINE RXNEQN7(NCOMP,T,Y,DYDT)
C*********************************************************************C
c
c Note that this subroutine is the same as RXNEQN5, for sequential chain decay.
C------------------------------------------------------------------------------------------------------------
c****Definition of values in the rc array and definition of y arrays**
C*** DECAY RATES
C .....in WATER PHASE
c rc(1) = PCE decay rate
c rc(2) = TCE decay rate
c rc(3) = DCE decay rate
c rc(4) = VC decay rate
C .....on SOLID PHASE
C .....in AIR PHASE
C .....in NAPL PHASE
c
C*** MASS RATIOS
c rc(17) = yield tce/pce (131.36/165.8) -- fixed value
c rc(18) = yield dce/tce (96.9/131.36) -- fixed value
c rc(19) = yield vc/dce (62.45/96.9) -- fixed value
C------------------------------------------------------------------------------------------------------------
C*** CONCENTRATIONS
c y(1) = PCE
c y(2) = TCE
c y(3) = DCE
c y(4) = VC
C------------------------------------------------------------------------------------------------------------
C*****FUNCTION CS evaluates retardation term for solid phase.

C
C dydt = Computed RHS of user defined differential equation [array variable
dydt(NCOMP)]
c
c NCOMP = total number of components for this reaction.
C------------------------------------------------------------------------------------------------------------
C*****Dimension required arrays and define variables*************************
REAL*8 T,Y(NCOMP),DYDT(NCOMP)
REAL*8 AMASS,AMASSP,AMASST,AMASSD,AMASSV
C
INTEGER*4 NCOMP,LSPEC,IREALSL,IEQPART
REAL*4 EQCOEF_NW(6),EQCOEF_GW(6),EQCOEF_NG(6)
REAL*4 RHOBK(6),AETA(6),EQPAR(6)
REAL*8
SATNEW,SATOLD,SATNEW2,SATOLD2,SATNEW3,SATOLD3,VOL,PHI_L
REAL*8 RC(100)
C------------------------------------------------------------------------------------------------------------
C******Common blocks used to pass parameters from transport routines to reaction
routines.
COMMON /PART/
EQCOEF_NW,EQCOEF_GW,EQCOEF_NG,IREALSL,IEQPART
COMMON /SORP/ RHOBK,AETA,LINR,CTOL,EQPAR
COMMON /SOIL/
SATNEW,SATOLD,SATNEW2,SATOLD2,SATNEW3,SATOLD3,
+ VOL,PHI_L
COMMON /RXNRC/ RC
C------------------------------------------------------------------------------------------------------------
C
C.....First, calculate phase contents for water, air and NAPL
CW = PHI_L*SATNEW
CA = PHI_L*SATNEW2
CN = PHI_L*SATNEW3
C------------------------------------------------------------------------------------------------------------
C
C.....Calculate Effective Retardation coefficient for all species
C
C... AMASSP = Retardation coefficient for PCE
CALL MASSC(Y(1),1,AMASS)
AMASSP = AMASS/VOL
C ... AMASST = Retardation coefficient for TCE
AMASST = AMASS/VOL
C ... AMASSD = Retardation coefficient for DCE
AMASSD = AMASS/VOL
C ... AMASSV = Retardation coefficient for VC
AMASSV = AMASS/VOL
C------------------------------------------------------------------------------------------------------------
C
C......Calculate DYDT
c

C.....for PCE
DYDT(1)= -(
+ RC(1)*Y(1) * CW
+ + RC(5)*CS(RHOBK(1),Y(1),AETA(1),EQPAR(1),LINR)*Y(1)
+ + RC(9)*EQCOEF_GW(1)*Y(1) * CA
+ + RC(13)*EQCOEF_NW(1)*Y(1) * CN
+ ) / AMASSP
C.....for TCE
DYDT(2)= (
+ -(
+ RC(2)*Y(2) * CW
+ + RC(10)*EQCOEF_GW(2)*Y(2) * CA
+ + RC(14)*EQCOEF_NW(2)*Y(2) * CN
+ )
+ +(
+ RC(1)*Y(1) * CW
+ + RC(9)*EQCOEF_GW(1)*Y(1) * CA
+ + RC(13)*EQCOEF_NW(1)*Y(1) * CN
+ ) * RC(17)
+ ) / AMASST
C.....for DCE
DYDT(3)= (
+ -(
+ RC(3)*Y(3) * CW
+ + RC(11)*EQCOEF_GW(3)*Y(3) * CA
+ + RC(15)*EQCOEF_NW(3)*Y(3) * CN
+ )
+ +(
+ RC(2)*Y(2) * CW
+ + RC(10)*EQCOEF_GW(2)*Y(2) * CA
+ + RC(14)*EQCOEF_NW(2)*Y(2) * CN
+ ) * RC(18)
+ ) / AMASSD
C.....for VC
DYDT(4)= (
+ -(
+ RC(4)*Y(4) * CW
+ + RC(12)*EQCOEF_GW(4)*Y(4) * CA
+ + RC(16)*EQCOEF_NW(4)*Y(4) * CN
+ )
+ +(
+ RC(3)*Y(3) * CW
+ + RC(11)*EQCOEF_GW(3)*Y(3) * CA
+ + RC(15)*EQCOEF_NW(3)*Y(3) * CN
+ ) * RC(19)
+ ) / AMASSV

C------------------------------------------------------------------------------------------------------------
C
RETURN
END
C------------------------------------------------------------------------------------------------------------
8.3 REACTION PACKAGE INPUT

Input for the Reaction (RXT) package is read from the unit specified in IUNIT(10)
for MODFLOW-SURFACT/MODHMS, and in IUNIT(32) for MODHMS.
1. Data: IDOMN
Format: Free
Note that IDOMN should be set to zero if dual domain simulations are not performed
2. Data: KINETICS NRXNC NRXNV MXOUTITR NNOTCVR NIMSPCS

Format: I5 I5 I5 I5 I5 I5
**Enter item 3 only if IDOMN = 1 or 2.
3. Data: KINETICS_MAT NRXNC_MAT NRXNV_MAT

MXOUTITR_MAT NNOTCVR_MAT NIMSPCS_MAT
Format: 6I5
**Enter Items 4 through 8 only if IDOMN = 0 or 2
4. Data: ISWITCH (I), I = 1, (NRXNC+NRXNV)

Format: I1
**Item 5 should be repeated for (NRXNC+NRXNV) times. For each parameter, If

ISWITCH =0, a constant value is required; if ISWITCH > 0, an array of (NODES)
values are required for each node.
5. Data: PARAMETER
Format: Free
**Enter item 6 only if KINETICS > 2. Item 6 should be repeated for all NCOMP
components of the simulation where NCOMP = NSPECI + NIMSPCS (mobile and
immobile components).
6. Data: ATOL RTOL

Format: F10.0 F10.0
**Enter item 7 only if MXOUTITR (or MXOUTIT) >2
7. Data: CATOL(I), I=1,NCOMP

Format: Free
Note that if MXOUTITR (or MXOUTIT) =1, the operator splitting technique is used,
whereby the reactions are solved after solving for transport for each time step. If
MXOUTITR (or MXOUTIT)=2, then a correction is done on the second iteration,
with solution proceeding as in FCT schemes. For MXOUTITR (or MXOUTIT)>2,
iterations are performed over transport and reaction, till convergence is achieved.

**Enter item 8 NIMSPCS times, for each immobile species involved in the reaction. Note
that item 8 consists of one array for each of the immobile species of reaction. The array
is of size NODES, with each element of the array representing the initial conditions at
each node.
8. Data: BIOS
Format: Free
**Enter Items 9 through 13 only if IDOMN = 1 or 2
9. Data: ISWITCH_MAT (I), I = 1, (NRXNC_MAT+NRXNV_MAT)

Format: I1
**Item 10 should be repeated for (NRXNC_MAT+NRXNV_MAT) times. For each

parameter, If ISWITCH_MAT = 0, a constant value is required; if ISWITCH_MAT >
0, an array of (NGW) values are required for each subsurface node (NGW = number of
subsurface nodes).
10. Data: PARAMETER_MAT

Format: Free
**Enter item 11 only if KINETICS_MAT > 2. Item 11 should be repeated for all
NCOMP components of the simulation where NCOMP = NSPECI + NIMSPCS (mobile
and immobile components).
11. Data: ATOL_MAT RTOL_MAT

Format: F10.0 F10.0
**Enter item 12 only if MXOUTITR_MAT (or MXOUTIT) >2
12. Data: CATOL_MAT(I), I=1,NCOMP

Format: Free
Note that if MXOUTITR_MAT (or MXOUTIT) =1, the operator splitting technique
is used, whereby the reactions are solved after solving for transport for each time
step. If MXOUTITR_MAT (or MXOUTIT)=2, then a correction is done on the
second iteration, with solution proceeding as in FCT schemes. For MXOUTITR_MAT
(or MXOUTIT)>2, iterations are performed over transport and reaction, till
convergence is achieved.
**Enter item 13, NIMSPCS times, for each immobile species involved in the reaction.
Note that item 13 consists of one array for each of the immobile species of reaction. The
array is of size NGW, with each element of the array representing the initial conditions
at each subsurface node.
13. Data: BIOS_MAT

Format: Free

IDOMN -- is the domain for which reactions are simulated

0=for active domain only (also for IDUAL = 0)
1=for inactive domain only
2=for both active and inactive domain of a dual domain
simulation
KINETICS -- is the reaction module number

1= instantaneous aerobic degradation of BTEX
2= instantaneous biodegradation of BTEX via five pathways
3= kinetic-limited biodegradation of BTEX via five pathways
4= Double Monod Model
5= 4 species sequential decay
6= aerobic/anaerobic model for PCE degradation
7= user defined model
NRXNC -- is the number of spatially uniform parameters
NRXNV -- is the number of spatially varying parameters
MXOUTITR -- is the number of outer iteration
NNOTCVR -- is the maximum number of time-step reductions allowed in the

iterative solution of transport and reactions, if iterative failure
occurs. The simulation is aborted after NNOTCVR reductions.
NIMSPCS -- is the number of immobile species participating in the reactions
ISWITCH -- indicates wether parameters are spatially uniform or varying.

If ISWITCH = 0, spatially uniform
If ISWITCH = 1, spatially varying
PARAMETER -- indicates required input parameters
ATOL -- indicates absolute tolerance for the differential equations solver
RTOL -- indicates relative tolerance for the differential equations solver
CTOL -- indicates tolerance concentration for the outer iteration
BIOS -- Initial concentration of immobile species involved in reaction - Note

that these immobile species (NIMSPCS of them) are not included
with the main transport routines.
All parameters are repeated for the immobile (matrix) domain of a dual domain simulation.
Parameters with the extension _MAT have same descriptions as above, but for the
immobile domain.

8.4 MODEL VERIFICATION
This section presents example problems used to verify the MODFLOW-
SURFACT/MODHMS Reaction Modules.
8.4.1. One-Dimensional Transport of a Three-Member Radioactive Decay

Chain
This problem is selected to test implementation of the reaction module since it can
be solved by both the reaction (RCT) module, as well as the basic transport (BTN)
module. The problem considers transport of a chain of three radionuclides released from
a decaying source located at x = 0 in a steady-state flow field within a confined porous
medium reservoir. Simulation conditions are described in Section 4.4.2, where this
problem has been solved directly within the BTN package.
Figure 8.2 shows the concentration distributions of all three components after
273.9 years of simulation. Agreement between the two modules is excellent.
1.0E+00
C1-RCT
1.0E-01
C2-RCT
Concentration
1.0E-02 C3-RCT
C1-BTN
1.0E-03
C2-BTN
1.0E-04 C3-BTN
1.0E-05
1.0E-06
1.0E-07
0 100 200 300 400 500
X (ft)
Figure 8.2. Comparison of results for BTN and RCT modules for three member
chain decay with decay in water and soil phases.
Another simulation is performed for the same setup, with decay occurring only in the
water phase (for the case of the problem discussed in Section 4.4.2, decay occurred in
water and soil phases at the same prescribed rate). Simulation results for this case are
shown in Figure 8.3. Again, the results show very close agreement indicating that the RCT
package is working correctly.

1.0E+00
C1-RCT
1.0E-01
C2-RCT
Concentration
1.0E-02 C3-RCT
C1-BTN
1.0E-03
C2-BTN
1.0E-04 C3-BTN
1.0E-05
1.0E-06
1.0E-07
0 100 200 300 400 500
X (ft)
Figure 8.3. Comparison of results for BTN and RCT modules for three member
chain decay with decay in water phase only.
8.4.2. Chain Decay of NAPL in a Single Cell
This problem considers chain decay of PCE dissolved from an immobile NAPL
initially consisting of PCE in a single cell. Initial PCE NAPL saturation was 0.2. First
order decay rates of PCE, TCE, DCE, and VE were 0.05, 0.03, 0.02, and 0.01 1/day,
respectively. Yield coefficients between species were assumed 1.0. The results are
compared with those obtained from the BTN Transport module for two cases of decay in
the water and solid phases, and of decay in water phase only. Figure 8.4 shows the
concentration distributions of components in the water phase vs. time. Agreement between
the two modules is excellent.
NAPL
NAPL Chain
Chain Decay
Decay (water
(water andonly)
solid)
3.0E-01
1.2E+00 PCE-MS
(mg/L)
PCE-MS
Concentration(mg/L)
1.0E+00 TCE-MS
TCE-MS
2.0E-01 DCE-MS
8.0E-01 DCE-MS
Concentration
VC-MS
VC-MS
6.0E-01 PCE-RM
1.0E-01 PCE-RM
4.0E-01 TCE-RM
TCE-RM
DCE-RM
2.0E-01 DCE-RM
VC-RM
0.0E+00 VC-RM
0.0E+00
0 2000 4000 6000 8000 1000
0 2000 4000 6000 8000 1000
0
0
Time (days)
Time (days)
Figure 8.4.Comparison of straight chain decay of PCE with presence of

NAPL.

8.5 MODEL SIMULATIONS
This section presents example problems used to demonstrate application of the
MODFLOW-SURFACT/MODHMS Reaction Modules. Six simulations corresponding to
the six reaction modules coded in the reaction model of MSVMS. All the simulations use
the same problem geometry.
These problems consider a domain which is 1005.9 m x 518.2 x 3 m in the x- y-

and z-dimensions and is discretized using 30x20x1 cells of dimension 33.53 m x 25.91 m
x 3 m respectively. Groundwater flow is established by specifying a hydraulic head of 43
m at the inflow boundary and 3 m at the outflow boundary. Physical properties of the
porous medium are: Hydraulic conductivity = 6.25 m/day; Longitudinal dispersivity =
8.23 m; Transverse dispersivity = 0.823 m; Vertical transverse dispersivity = 0.0823 m;
Porosity = 0.25; and Bulk density = 1.8 g/ cm3.
8.5.1. Module 1: Instantaneous Aerobic Decay of BTEX
This simulation considers instantaneous aerobic decay of BTEX and was conducted
using MSVMS reaction module 1. The BTEX was introduce by specifying a constant
concentration 8 mg/L at some source nodes. Initial and inflow Oxygen concentrations were
assumed 6 mg/L. Figure 8.5 shows the concentrations of BTEX and Oxygen at 1000 days.
Note that oxygen is immediately depleted where BTEX is present.
Concentration (mg/L)
0
1
2
3
4
5
6
7
8
(a) BTEX
0
1
2
3
4
5
6
7
8
(b) Oxygen
Figure 8.5. BTEX and oxygen concentrations at 1000 days obtained using
module 1.

8.5.2. Module 2: Instantaneous Degradation of BTEX using Multiple Electron
Acceptors
This simulation considers instantaneous decay of BTEX using multiple electron

acceptors. MSVMS reaction module 2 was used for this simulation. The BTEX was
introduce by specifying a constant concentration 8 mg/L at some source nodes. Initial and
inflow concentrations for BTEX, Oxygen, nitrate, ferrous iron, sulfate and methane were
assumed 0.0 6.0, 17.0, 0.001, 100.0, and 0.0001 mg/L, respectively. Maximum ferric
iron and carbon dioxide available for reactions were assumed 50.5 and 2.05 mg/L,
respectively. Figure 8.6 shows the concentrations of BTEX , Oxygen nitrate, ferrous iron,
sulfate and methane at 1000 days.
0
1
2
3
4
5
6
7
8
(a) BTEX
0
1
2
3
4
5
6
7
8
(b) Oxygen
0.001
3
6
9
12
15
18
(c) Nitrate

0.001
3
6
9
12
15
18
(d) Ferrous iron
0
10
20
30
40
50
60
70
80
90
100
(e) Sulphate
0.0001
0.1
0.2
0.3
0.4
0.5
0.6
(f) Methane
Figure 8.6. BTEX, oxygen, nitrate, ferrous iron, sulfate and methane
concentrations at 1000 days obtained using module 2.
8.5.3 Module 3: Kinetic-limited Degradation of BTEX Using Multiple Electron

Acceptors
This simulation considers kinetic-limited decay of BTEX using multiple electron

acceptors. MSVMS reaction module 3 was used for this simulation. The BTEX was
introduced by specifying a constant concentration 8 mg/L at some source nodes. Initial and
inflow concentrations for BTEX, Oxygen, nitrate, ferrous iron, sulfate and methane were
assumed 0.0 6.0, 17.0, 0.001, 100.0, and 0.0001 mg/L, respectively. Maximum ferric

iron and carbon dioxide available for reactions were assumed 50.5 and 2.05 mg/L,
respectively. The reaction variables are summarized in Table 8.1. Figure 8.7 shows the
concentrations of BTEX, oxygen, nitrate, ferrous iron, sulfate and methane at 1000 days.
Table 8.1. Reaction parameters used in 8.5.3
Parameter 1: Fe3+ available for biodegradation = 50.5 mg/L

Parameter 2: CO2 available for biodegradation = 2.05 mg/L
Parameter 3: Hydrocarbon decay rate via aerobic process in water = 0.051 1/day
Parameter 4: Hydrocarbon decay rate via denitrification in water = 0.031 1/day
Parameter 5: Hydrocarbon decay rate via iron reduction in water = 0.005 1/day
Parameter 6: Hydrocarbon decay rate via sulfate reduction in water = 0.004 1/day
Parameter 7: Hydrocarbon decay rate via methanogenesis in water = 0.002 1/day
Parameter 8: Hydrocarbon decay rate via aerobic process on solid = 0.0 1/day
Parameter 9: Hydrocarbon decay rate via denitrification on solid = 0.0 1/day
Parameter 10: Hydrocarbon decay rate via iron reduction on solid = 0.0 1/day
Parameter 11: Hydrocarbon decay rate via sulfate reduction on solid = 0.0 1/day
Parameter 12: Hydrocarbon decay rate via methanogenesis on solid = 0.0 1/day
Parameter 13: Hydrocarbon decay rate via aerobic process in gas = 0.0 1/day
Parameter 14: Hydrocarbon decay rate via denitrification in gas = 0.0 1/day
Parameter 15: Hydrocarbon decay rate via iron reduction in gas = 0.0 1/day
Parameter 16: Hydrocarbon decay rate via sulfate reduction in gas = 0.0 1/day
Parameter 17: Hydrocarbon decay rate via methanogenesis in gas = 0.0 1/day
Parameter 18: Hydrocarbon decay rate via aerobic process in NAPL = 0.0 1/day
Parameter 19: Hydrocarbon decay rate via denitrification in NAPL = 0.0 1/day
Parameter 20: Hydrocarbon decay rate via iron reduction in NAPL = 0.0 1/day
Parameter 21: Hydrocarbon decay rate via sulfate reduction in NAPL = 0.0 1/day
Parameter 22: Hydrocarbon decay rate via methanogenesis in NAPL = 0.0 1/day
Parameter 23: Half saturation constant for oxygen in water = 0.01 mg/L
Parameter 24: Half saturation constant for nitrate in water = 0.01 mg/L
Parameter 25: Half saturation constant for Fe3+ in water = 0.01 mg/L
Parameter 26: Half saturation constant for sulfate in water = 0.01 mg/L
Parameter 27: Half saturation constant for methane in water = 0.01 mg/L
Parameter 28: Half saturation constant for oxygen on solid = 0.01 mg/L
Parameter 29: Half saturation constant for nitrate on solid = 0.01 mg/L
Parameter 30: Half saturation constant for Fe3+ on solid = 0.01 mg/L
Parameter 31: Half saturation constant for sulfate on solid = 0.01 mg/L
Parameter 32: Half saturation constant for methane on solid = 0.01 mg/L
Parameter 33: Half saturation constant for oxygen in gas = 0.01 mg/L
Parameter 34: Half saturation constant for nitrate in gas = 0.01 mg/L
Parameter 35: Half saturation constant for Fe3+ in gas = 0.01 mg/L
Parameter 36: Half saturation constant for sulfate in gas = 0.01 mg/L
Parameter 37: Half saturation constant for methane in gas = 0.01 mg/L
Parameter 38: Half saturation constant for oxygen in NAPL = 0.01 mg/L
Parameter 39: Half saturation constant for nitrate in NAPL = 0.01 mg/L
Parameter 40: Half saturation constant for Fe3+ in NAPL = 0.01 mg/L
Parameter 41: Half saturation constant for sulfate in NAPL = 0.01 mg/L
Parameter 42: Half saturation constant for methane in NAPL = 0.01 mg/L
Parameter 43: Inhibition coefficient for oxygen reaction in water = 1.0mg/L
Parameter 44: Inhibition coefficient for nitrate reaction in water = 1.0mg/L
Parameter 45: Inhibition coefficient for Fe3+ reaction in water = 10.0mg/L
Parameter 46: Inhibition coefficient for sulfate reaction in water = 1.0mg/L

Parameter 47: Inhibition coefficient for oxygen reaction on solid = 1.0mg/L
Parameter 48: Inhibition coefficient for nitrate reaction on solid = 1.0mg/L
Parameter 49: Inhibition coefficient for Fe3+ reaction on solid = 10.0mg/L
Parameter 50: Inhibition coefficient for sulfate reaction on solid = 1.0mg/L
Parameter 51: Inhibition coefficient for oxygen reaction in gas = 1.0mg/L
Parameter 52: Inhibition coefficient for nitrate reaction in gas = 10.0mg/L
Parameter 53: Inhibition coefficient for Fe3+ reaction in gas = 1.0mg/L
Parameter 54: Inhibition coefficient for sulfate reaction in gas = 1.0mg/L
Parameter 55: Inhibition coefficient for oxygen reaction in NAPL = 1.0mg/L
Parameter 56: Inhibition coefficient for nitrate reaction in NAPL = 10.0mg/L
Parameter 57: Inhibition coefficient for Fe3+ reaction in NAPL = 1.0mg/L
Parameter 58: Inhibition coefficient for sulfate reaction in NAPL = 1.0mg/L
Parameter 59: Oxygen yield = 3.14
Parameter 60: Nitrate yield = 4.9
Parameter 61: Iron (Fe2+) production yield = 21.8
Parameter 62: Sulfate yield YSO4/HC = 4.7
Parameter 63: Methane production yield = 0.78
0
1
2
3
4
5
6
7
8
(a) BTEX
0
1
2
3
4
5
6
7
8
(b) Oxygen

0
3
6
9
12
15
18
(c) Nitrate
0
3
6
9
12
15
18
(d) Ferrous iron

0
10
20
30
40
50
60
70
80
90
100
(e) Sulphate
0.0001
0.1
0.2
0.3
0.4
0.5
0.6
(f) Methane
Figure 8.7. BTEX, oxygen, Nitrate, Ferrous iron, Sulfate, Methane

concentrations at 1000 days obtained using module 3.

8.5.4 Module 4: Double Monod Model
This simulation considers the reaction between an electron donor and an electron
acceptor mediated by actively growing bacteria cells living in aqueous, soil, air, and
NAPL phases. The reaction module number 4 was used for this simulation. The BTEX
was introduce by specifying a constant concentration 8 mg/L at some source nodes. Initial
and inflow concentrations for BTEX, Oxygen aqueous phase bacteria and adsorbed
bacteria were assumed 0.0, 6.0, 0.0001 and 0.001 mg/L, respectively. The reaction
variables are summarized in Table 8.2. Figure 8.8 shows the concentrations of BTEX,
Oxygen, aqueous phase bacteria and adsorbed bacteria at 1000 days.
Parameter 1 = Specific utilization rate in water = 514.4 1/day

Parameter 2 = Specific utilization rate on solid = 514.4 1/day
Parameter 3 = Specific utilization rate in gas = 514.4 1/day
Parameter 4 = Specific utilization rate in NAPL = 514.4 1/day
Parameter 5 = Monod half saturation constant for electron donor in water = 6.0 mg/L
Parameter 6 = Monod half saturation constant for electron donor on solid = 6.0 mg/L
Parameter 7 = Monod half saturation constant for electron donor in gas = 6.0 mg/L
Parameter 8 = Monod half saturation constant for electron donor in NAPL = 6.0 mg/L
Parameter 9 = Monod half saturation constant for electron acceptor in water = 1.0 mg/L
Parameter 10 = Monod half saturation constant for electron acceptor on solid = 1.0 mg/L
Parameter 11 = Monod half saturation constant for electron acceptor in gas = 1.0 mg/L
Parameter 12 = Monod half saturation constant for electron acceptor in NAPL = 1.0
mg/L
Parameter 13 = Biomass produced per unit amount of electron donor utilized, YX/D = 0.5
mg/L
Parameter 14 = Electron acceptor used per unit amount of electron donor utilized,
YA/D=0.32 mg/L
Parameter 15 = Endogenous cell death or decay coefficients in water = 0.15034 1/day
Parameter 16 = Endogenous cell death or decay coefficients on solid = 0.15034 1/day
Parameter 17 = Endogenous cell death or decay coefficients in air = 0.15034 1/day
Parameter 18 = Endogenous cell death or decay coefficients in NAPL = 0.15034 1/day
Parameter 19 = First-order, bacterial attachment rate = 6.9984 1/day
Parameter 20 = First-order, bacterial detachment rate = 0.59962 1/day
Parameter 21 = Soil bulk density = 1.8 mg/L
Initial solid phase bacterial concentration = 0.001 mg/L

0
1
2
3
4
5
6
7
8
(a) BTEX
0.0001
0.0010
0.0020
0.0030
0.0038
0.0040
0.0050
(b) Oxygen
0.0001
0.0010
0.0020
0.0030
0.0038
0.0040
0.0050
(c) Aqueous Bacteria

0
1
2
3
4
5
6
7
8
(c) Adsorbed
Bacteria
Figure 8.8.Electron donor, electron acceptor, and aqueous phase bacteria

concentration at 1000 days obtained using module 4.

This problem considers chain decay of dissolved PCE which was introduce by
specifying a constant concentration 0.24 mg/L at some source nodes. The reaction
variables are summarized in Table 8.3. Figure 8.9 shows the concentrations of PCE,
TCE, DCE, and VC at 1000 days.
Parameter 1 = PCE first-order degradation rate in water = 0.05 1/day

Parameter 2 = TCE first-order degradation rate in water = 0.03 1/day
Parameter 3 = DCE first-order degradation rate in water = 0.02 1/day
Parameter 4 = VC first-order degradation rate in water = 0.01 1/day
Parameter 5 = PCE first-order degradation rate on solid = 0.0 1/day
Parameter 6 = TCE first-order degradation rate on solid = 0.0 1/day
Parameter 7 = DCE first-order degradation rate on solid = 0.0 1/day
Parameter 8 = VC first-order degradation rate on solid = 0.0 1/day
Parameter 9 = PCE first-order degradation rate in gas = 0.0 1/day
Parameter 10 = TCE first-order degradation rate in gas = 0.0 1/day
Parameter 11 = DCE first-order degradation rate in gas = 0.0 1/day
Parameter 12 = VC first-order degradation rate in gas = 0.0 1/day
Parameter 13 = PCE first-order degradation rate in NAPL = 0.0 1/day
Parameter 14 = TCE first-order degradation rate in NAPL = 0.0 1/day
Parameter 15 = DCE first-order degradation rate in NAPL = 0.0 1/day
Parameter 16 = VC first-order degradation rate in NAPL = 0.0 1/day
Parameter 17 = Yield, YB/A =0.792
Parameter 18 = Yield, YC/B= 0.738
Parameter 19 = Yield, YD/C= 0.645
Note that this setup can also be simulated by the BTN package branched chain-decay
formulation with zero branching, and appropriate yield factors.

0.00001
0.0001
0.001
0.01
0.05
0.1
0.2
0.3
(a) PCE
0.00001
0.0001
0.001
0.01
0.05
0.1
0.2
0.3
(b) TCE
0.00001
0.0001
0.001
0.01
0.05
0.1
0.2
0.3
(c) DCE
0.00001
0.0001
0.001
0.01
0.05
0.1
0.2
0.3
(c) VC
Figure 8.9. PCE, TCE, DCE, and VC concentrations at 1000 days for module 5.

This problem considers degradation of PCE and their degradation products via
both aerobic and anaerobic pathways. PCE was introduce by specifying a constant
concentration 0.24 mg/L at some source nodes. The reaction variables are summarized in
Table 8.4. Figure 8.10 shows the concentrations of PCE, TCE, DCE, VC, ETH, and Cl
at 1000 days.
Parameter 1 = Anaerobic decay rate for PCE in water = 0.05 1/day

Parameter 2 = Anaerobic decay rate for TCE in water = 0.03 1/day
Parameter 3 = Anaerobic decay rate for DCE in water = 0.02 1/day
Parameter 4 = Anaerobic decay rate for VC in water = 0.01 1/day
Parameter 5 = Anaerobic decay rate for ETH in water = 0.02 1/day
Parameter 6 = Anaerobic decay rate for PCE on solid = 0.0 1/day
Parameter 7 = Anaerobic decay rate for TCE on solid = 0.0 1/day
Parameter 8 = Anaerobic decay rate for DCE on solid = 0.0 1/day
Parameter 9 = Anaerobic decay rate for VC on solid = 0.0 1/day
Parameter 10 = Anaerobic decay rate for ETH on solid = 0.0 1/day
Parameter 11 = Anaerobic decay rate for PCE in gas = 0.0 1/day
Parameter 12 = Anaerobic decay rate for TCE in gas = 0.0 1/day
Parameter 13 = Anaerobic decay rate for DCE in gas = 0.0 1/day
Parameter 14 = Anaerobic decay rate for VC in gas = 0.0 1/day
Parameter 15 = Anaerobic decay rate for ETH in gas = 0.0 1/day
Parameter 16 = Anaerobic decay rate for PCE in NPAL = 0.0 1/day
Parameter 17 = Anaerobic decay rate for TCE in NAPL = 0.0 1/day
Parameter 18 = Anaerobic decay rate for DCE in NAPL = 0.0 1/day
Parameter 19 = Anaerobic decay rate for VC in NAPL = 0.0 1/day
Parameter 20 = Anaerobic decay rate for ETH in NAPL = 0.0 1/day
Parameter 21 = Aerobic decay rate for PCE in water = 0.005 1/day
Parameter 22 = Aerobic decay rate for TCE in water = 0.003 1/day
Parameter 23 = Aerobic decay rate for DCE in water = 0.002 1/day
Parameter 24 = Aerobic decay rate for VC in water = 0.005 1/day
Parameter 25 = Aerobic decay rate for ETH in water = 0.002 1/day
Parameter 26 = Aerobic decay rate for PCE on solid = 0.0 1/day
Parameter 27 = Aerobic decay rate for TCE on solid = 0.0 1/day
Parameter 28 = Aerobic decay rate for DCE on solid = 0.0 1/day
Parameter 29 = Aerobic decay rate for VC on solid = 0.0 1/day
Parameter 30 = Aerobic decay rate for ETH on solid = 0.0 1/day
Parameter 31 = Aerobic decay rate for PCE in gas = 0.0 1/day
Parameter 32 = Aerobic decay rate for TCE in gas = 0.0 1/day
Parameter 33 = Aerobic decay rate for DCE in gas = 0.0 1/day
Parameter 34 = Aerobic decay rate for VC in gas = 0.0 1/day
Parameter 35 = Aerobic decay rate for ETH in gas = 0.0 1/day
Parameter 36 = Aerobic decay rate for PCE in NPAL = 0.0 1/day
Parameter 37 = Aerobic decay rate for TCE in NAPL = 0.0 1/day
Parameter 38 = Aerobic decay rate for DCE in NAPL = 0.0 1/day
Parameter 39 = Aerobic decay rate for VC in NAPL = 0.0 1/day
Parameter 40 = Aerobic decay rate for ETH in NAPL = 0.0 1/day
Parameter 41 = Yield, YT/P= 0.792
Parameter 42 = Yield, YD/T= 0.738

Parameter 43 = Yield, YV/D= 0.645
Parameter 44 = Yield, YE/V= 0.449
Parameter 45 = Yield, YC/P= 0.214
Parameter 46 = Yield, YC/T = 0.270
Parameter 47 = Yield, YC/D = 0.366
Parameter 48 = Yield, YC/V= 0.567
0.00001
0.0001
0.001
0.01
0.05
0.1
0.2
0.3
(a) PCE
0.00001
0.0001
0.001
0.01
0.05
0.1
0.2
0.3
(b) TCE
0.00001
0.0001
0.001
0.01
0.05
0.1
0.2
0.3
(c) DCE

0.00001
0.0001
0.001
0.01
0.03
0.05
(d) VC
0.00001
0.0001
0.001
0.01
0.03
0.05
(e) ETH
0.001
0.01
0.1
1
2
3
(f) Cl
Figure 8.10. PCE, TCE, DCE, VC, ETH, and Cl concentrations at 1000 days
obtained using module 6.
8.6 Acknowledgments
We would like to acknowledge the previous works of the RT3D (Clement 1997) team
which has lead to the development of this RCT package. We are also extremely grateful
to Dr. Prabhakar Clement for discussions and support on issues related to this package.

8.7 References
Clement, T. P., 1977. RT3D: A Modular Computer Code for Simulating Reactive Multi-
species Transport in 3-Dimensional Groundwater Aquifers, Pacific Northwest
National Laboratory, Richland, WS, USA. PNNL-11720. Found online at
http://bioprocess.pnl.gov/rt3d.htm.
Clement, T. P., Y. Sun, B. S. Hooker, and J. N. Petersen, 1998. Modeling Multi-species

Reactive Transport in Groundwater Aquifers, Groundwater Monitoring &
Remediation, 18(2): 79-92.
Clement, T. P., B. M. Peyton, R. S. Skeen, B. S. Hooker, J. M. Petersen, and D.

Jennings, 1997. Microbial Growth and Transport in Porous Media Under
Denitrification Conditions: Experiment and Simulation Results, Journal of
Contaminant Hydrology, 24: 269-285.
Clement, T. P., B. S. Hooker, and R. S. Skeen, 1996a. Numerical Modeling of

Biologically Reactive Transport From a Nutrient Injection Well, ASCE Journal
of Environmental Engineering, 122(9): 833-839.
Clement, T. P., B. S. Hooker, and R. S. Skeen, 1996b. Macroscopic Models for

Predicting Changes in Saturated Porous Media Properties Caused by Microbial
Growth, Groundwater, 34(5): 934-942.
Lu, G., T. P., Clement, C. Zheng, and T. H. Wiedemeier, 1999. Natural Attenuation of
BTEX Coumpouonds: Model Development and Field-Scale Application,
Groundwater, 37(5): 707-717.
Moltz, F. J., M. A. Widdowson, L. D. Benefield, 1986. Simulation of microbial growth

dynamics coupled to nutrient oxygen transport in porous media, Water Resources
Research, 22(8): 1207-1216.
Rafai, S. H., P. B. Bedient, R. C. Borden, and J. F. Haasbeek, 1987. BIOPLUME II -

Computer model of two-dimensional contaminant transport under the unfluence
of oxygen limited biodegradation in groundwater, National Center for Ground
Water Research, Rice University.
Rafai, S. H., C. J. Newell, J. R. Gonzales, S. Dendrou, B. Dendrou, L. Kennedy, J. T.

Wilson, 1998. BIOPLUME III - Natural Attenuation Decision Support System,
User’s Manual, Version 1.0, National Risk Management Research Laboratory,
U.S. EPA, Cincinnati, OH, USA, IAG#RW57936164.
Rafai, S. H., and P. B. Bedient, 1990. Comparison of biodegradation kinetics with an

instantaneous reaction model for groundwater, Water Resources Research, 26(4):
637-645.
Taylor, S. W., and P. R. Jaffe, 1990. Substrate and Biomass Transport in a Porous
Medium, Water Resources Research, 26: 2181-2194.


Modflowsurfact V3 PDF

Uploaded by

Copyright:

Available Formats

Modflowsurfact V3 PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Modflowsurfact V3 PDF

Uploaded by

Copyright:

Available Formats

HYDROGEOLOGIC, INC.

HGL SOFTWARE LICENSE

COPYRIGHT © 1996, BY HYDROGEOLOGIC, INC.

Figure 1 Sub-directory structure for example flow problems

Table 1 MODFLOW-SURFACT Input/Output Default File Extensions . . 1-13

This document provides a brief overview of procedures for installation,

which is corrected to be:

1.2 SYSTEM REQUIREMENTS

1.4 THE SURF PACKAGES

1.5 THE ACT PACKAGES

1.7 LIST OF FILES

Fig 1. Sub-directory structure for example problems provided by

1.9 VIRTUAL MEMORY USE

VOLUME I: FLOW MODULES

COPYRIGHT © 1996, BY HYDROGEOLOGIC, INC.

CHAPTER 2: THE BLOCK-CENTERED FLOW (BCF4) PACKAGE

CHAPTER 3: THE FRACTURE-WELL (FWL4) PACKAGE

CHAPTER 5: THE ADAPTIVE TIME-STEPPING AND OUTPUT CONTROL (ATO4)

CHAPTER 6: THE PRECONDITIONED CONJUGATE GRADIENT (PCG4) PACKAGE

CHAPTER 7: THE NEWTON-RAPHSON LINEARIZATION WITH BACKTRACKING

CHAPTER 8: OBSERVATION NODE (OBS1) PACKAGE

Chapter 1: Introduction 1-1

1-2 Chapter 1: Introduction

Chapter 1: Introduction 1-3

1-4 Chapter 1: Introduction

Chapter 2: The Block-Centered Flow (BCF4) Package 2-1

2.2A FORMULATION OF THE UNCONFINED FLOW EQUATION USING A

krw is the relative permeability, which is a function of water saturation;

N is the drainable porosity taken to be equal to the specific yield, ;

Sw is the degree of saturation of water, which is a function of the pressure head;

Ss is the specific storage of the porous material ( ); and

2-2 Chapter 2: The Block-Centered Flow (BCF4) Package

Chapter 2: The Block-Centered Flow (BCF4) Package 2-3

and (van Genuchten, 1977):

2-4 Chapter 2: The Block-Centered Flow (BCF4) Package

2.2C FORMULATION OF VARIABLY SATURATED AIR FLOW EQUATIONS

Chapter 2: The Block-Centered Flow (BCF4) Package 2-5

where Ssa is a volumetric specific storage to air defined analogously as:

where $w is the compressibility of water. Equations (12) and (13) yield

2-6 Chapter 2: The Block-Centered Flow (BCF4) Package

Chapter 2: The Block-Centered Flow (BCF4) Package 2-7

2.3A VERIFICATION EXAMPLE FOR UNCONFINED FLOW

Modeling Approaches and Results

2-8 Chapter 2: The Block-Centered Flow (BCF4) Package

Chapter 2: The Block-Centered Flow (BCF4) Package 2-9

Modeling Approach and Results

2-10 Chapter 2: The Block-Centered Flow (BCF4) Package

Chapter 2: The Block-Centered Flow (BCF4) Package 2-11

2.3C VERIFICATION EXAMPLE FOR AIR-PHASE FLOW SIMULATION

2-12 Chapter 2: The Block-Centered Flow (BCF4) Package

Chapter 2: The Block-Centered Flow (BCF4) Package 2-13

2-14 Chapter 2: The Block-Centered Flow (BCF4) Package

Modeling Approach and Results

Chapter 2: The Block-Centered Flow (BCF4) Package 2-15

Horizontal hydraulic conductivity, Kh (Kxx and Kyy) = ft/s

Vertical hydraulic conductivity, Kv (Kzz) = ft/s

2-16 Chapter 2: The Block-Centered Flow (BCF4) Package

Chapter 2: The Block-Centered Flow (BCF4) Package 2-17

2-18 Chapter 2: The Block-Centered Flow (BCF4) Package