1

CE-888
Watershed Management

Lecture: 7
Watershed Management and modelling
 Watershed Modeling
2

 A water resources manager need to solve various problems of

watershed with large number of complex variables.
 A tool is needed for providing the required information prior to
making the decision. Some examples include:
• To assess runoff in a watershed for given rainfall conditions.
• To estimate the groundwater available in a watershed.
• To assess groundwater pollution problem in a watershed area.
• Sometimes, models are used to provide Information required by
regulations.
 System Approach
3

System Approach: problems involves 3 steps;

i)Describe the system – involves modeling the watershed system;
ii)Describe the objective function normally stated in terms of economic
terms (eg.flooding;
iii)Optimize the system
Design problems classified: a) Long-run – design of multiple purpose
reservoir system – huge capital investment – benefits after & over a
long time
Intermediate run – irrigation & cultivation for a season
Short- run – how much water to be released for flood control
 Each require – hydrologic modeling
 Most situations – alternative models
 Models: criteria – accuracy, simplicity, consistency & sensitivity
 Watershed Model
4

Main tool in addressing a wide spectrum of environmental and water

resources problems. It can be used in:
water resources planning, development, design, operation, and
management; flooding; droughts; upland erosion; stream bank
erosion; coastal erosion; sedimentation; nonpoint source pollution;
water pollution from industrial, domestic, agricultural, and energy
industry sources; migration of microbes; deterioration of lakes;
desertification of land; degradation of land; decay of rivers; irrigation
of agricultural lands; conjunctive use of surface and groundwater;
reliable design of hydraulic structures and river training works etc.
 5
Watershed Management & Modeling
 As water resources manager two question arises:
1. What is a "Management problem"?
2. Why do we need models to solve a management problem?
In a management problem, decisions have to be made concerning
- Planning and operation of a considered (managed) system in order to
achieve certain goals, without violating technical and nontechnical
constraints that are imposed on the managed system.
• Example can be The case of construction of a check dam at the outlet
of a watershed.
Decisions to undertake are
- Where the check dam to be constructed?
- What is the size/ height?
- How much storage possible?
 6
Watershed Modeling
• A model may be defined as a selected simplified version of a real
system, which approximately simulates the latter's excitation-
response relations that are of interest .
• Predicting the behavior of a considered system in response to
excitations that stem from the implementation of management
decisions.
• Obtaining a better understanding of the considered system from
the geological, hydrological, and chemical points of view.
• Providing information required to comply with regulations
• Providing information for the design of observation networks, by
predicting the system's future behavior.
• Providing information for the design of field experiments.
 7
What are steps involved in Modeling procedure
Steps in general modeling process
Step 1: Identification of the information required for management
decisions.
For example for groundwater flow modeling, the information required
include:
• water levels at selected points.
• spring discharge.
• boundary discharge or recharge.
• concentration at specified points.
• quantity and quality of water pumped.
• geological parameters.
• hydrology parameters etc.
 8
What are steps involved in Modeling procedure
Step 2: Development of a conceptual model
To simplify the complex real system for our understanding of nature
and its behavior some simplifying assumptions are introduced are with
respect to:
•The geometry of investigation domain
• Effect of heterogeneity at different scales
•Nature of solid and fluid phases involved
•Flow regimes of fluids
•The various physical, chemical and biological processes
 9
What are steps involved in Modeling procedure

• The selection of the appropriate conceptual model for a

particular case depends on three main factors:
• The objectives of the investigations, i.e. what kind of information
is the model expected to provide for the purpose of making
management decisions.
• The available resources (including expertise, skilled personnel,
field data, and computers) required for constructing and solving
the model. This includes the ability to understand and describe
processes that take place and the data required for validating the
model and determining the numerical values of its coefficients.
• The legal and regulatory framework which pertains to the
considered case.
 10
What are steps involved in Modeling procedure
Step 3: Development of a mathematical model. In this step
conceptual model is expressed in the form of mathematical model
which consist of
•A definition of the geometry of the surfaces.
•Equations that express the balances.
•Flux equations that relate the fluxes.
•Constitutive equations.
•Sources and sinks.
•Initial conditions.
•Boundary conditions.
 11
What are steps involved in Modeling procedure
Step 4: Model Development
Step 5: Model validation/ verification
Step 6: Model calibration and parameter estimation
Various techniques exist for determining the 'best', or 'optimal', values of
various coefficients.
• Basic trial-and-error approach.
•Sophisticated optimization methods
•Priori estimates of values to be expected for the coefficients, as well as
information about lower & upper bound
•When conditions as described above do not exist, they can be created as
a field or laboratory experiment for determining the sought coefficients
 12
What are steps involved in Modeling procedure
Step 7: Model applications
Once we have calibrated the model for a considered problem, the
model is ready for use. Computer runs can be carried out employing
appropriate data to obtain the required results.
Step 8: Analysis of model uncertainty and stochastic modeling
•Uncertainty about many elements associated with the model.
•Is the selected conceptual model appropriate?
•Are the values of various coefficients used are correct?
• Are the selected boundaries & conditions appropriate?
These considerations pave the way for the development of stochastic
models.
 13
What are steps involved in Modeling procedure
Step 9: Summary, conclusions and reporting
The summary and conclusions should include:
• Information that the model was expected to provide
• Additional information concerning the accuracy of the information
•Uncertainty involved, suggested follow-up work, or other activities.
•The report on the modeling activities may be part of a larger report
on solving the considered problem, say as an appendix, or a report
that stands on its own.
 14
What are steps involved in Modeling procedure
Preprocessing and Post-processing
While using models for watershed modeling, the computer model
generally consists of:
•Preprocessing
•Processing
•Post-processing.
Model limitations:
•Limitations – conceptual or application related
•Conceptual limitations – relate to representation of actual process
or system with a model
•Application limitations – data limitations, model familiarity
 15
Classification of Models
Models can be classified at spatial and temporal scale:
Spatial Scale:
The watershed is represented in a model by one or more spatial computational units:
Lumped: Watershed is a single spatial unit
Semi-distributed: Watershed is divided into a number of spatial units, each of which
encompasses an area that is considered to be "homogeneous" according to some
specified criteria
Fully Distributed: Watershed is represented by a grid
Temporal Scale
The computational time step of a model must be matched to its spatial scale and
process descriptions. Time steps can range from minutes to a year. Typical
hydrologic simulation models have a time step of one day or less.
 16
Parameters Comparison
 17

Hydrologic Models - Types

There is a spectrum of models, depending on physical basis and
mathematical form:

Empirical: Based on highly simplified relationships, usually data-driven,

including statistical models
Conceptual: Based on mathematically convenient components that
loosely correspond to, but are abstractions of, physical processes
Physically-based: Based closely on mathematical representations of
the process physics
 18

Hydrologic Models - Types

Event vs. Continuous models
Event model : represents a single runoff event occurring over a period of
time ranging from about an hour to several days
Accuracy of the model output - Depend on the reliability of initial
conditions
Continuous watershed model: will determine flow rates and conditions
during both, runoff periods and periods of no surface runoff
Initial conditions must be known or assumed
Utilize runoff components: direct or surface runoff, shallow surface flow
(interflow) and groundwater flow
An event model may omit one or both of the subsurface components and
also evapotranspiration
 19

Hydrologic Models - Types

Complete vs. Partial Models

Complete or comprehensive watershed models

•Solves the water balance equation
•Represents more or less all hydrologic processes
•Increases the accuracy of the model

Partial Models
• Represents only a part of the overall runoff process
•Ex: Water yield model gives runoff volumes but no peak discharges
 20

Hydrologic Models - Types

Calibrated Parameter vs. Measured Parameter Models:
Calibrated parameter model:
•One or more parameters that can be evaluated only by fitting computed
hydrographs to the observed hydrographs
• Necessary - If the watershed component has any conceptual
component models
•Period of recorded flow is needed for estimating parameter values
Measured parameter model:
•Determination of parameters from known watershed characteristics
• Area and channel length – Maps and channel cross sections measured
in the field
•Usually applied to totally ungauged watersheds
 21

Hydrologic Models - Types

Lumped vs. Distributed Models
Lumped models
•Implicitly take into account the spatial variability of inputs, outputs, or
parameters
•Utilize average values of the watershed characteristics affecting runoff
•lead to significant error- due to nonlinearity and threshold values
Distributed models
•Include spatial variation in inputs, outputs, and parameters.
•Division of watershed area into a number of elements and calculation
of runoff volumes for each element
22
 23
Stochastic Model Vs Deterministic
• A system is a system. This is neither deterministic nor stochastic.
However, if we want describe the development of a (dynamic) system,
we use a model, and such a model (description) can be deterministic
or stochastic.
• In deterministic models, the output of the model is fully determined
by the parameter values and the initial conditions. A deterministic
model is a model that gives you the same exact results for a
particular set of inputs, no matter how many times you re-calculate it.
An example of a deterministic model is a calculation to determine
the return on a 5-year investment with an annual interest rate of
7%, compounded monthly
• Stochastic models possess some inherent randomness. The same
set of parameter values and initial conditions will lead to an ensemble
of different outputs eg stochastic dynamic programming (SDP) .
 24
Stochastic Model Vs Deterministic
• A stochastic model is a tool for estimating probability distributions of
potential outcomes by allowing for random variation in one or more
inputs over time. The random variation is usually based on
fluctuations observed in historical data for a selected period using
standard time-series techniques.
• Stochastic models possess some inherent randomness. The same set
of parameter values and initial conditions will lead to an ensemble of
different outputs.
• In general, stochastic is a synonym for random. For example,
a stochastic variable is a random variable. A stochastic process is
a random process. 
 25
Deterministic Hydrologic Model
• Three main categories: Lumped, Semi distributed & Distributed
• Lumped models: Parameters do not vary spatially within the
basin & response is evaluated only at the outlet, without
explicitly accounting for the response of individual subbasins.
• Parameters do not represent physical features of hydrologic
processes; model parameters – area weighted average
• Not applicable to event based processes
• Discharge prediction at outlet only
• Simple & minimal data requirements, easy use
• Eg. SCS-CN based models; IHACRES, WATBAL etc.
 26
Deterministic Hydrologic Model
Semi – distributed models: parameters are partially allowed to vary
in space by dividing the basin into a number of smaller subbasins
•Mainly two types: Kinematic wave theory models (eg. HEC-HMS
model) - simplified version of surface flow equations of physically
based model.
•Probability distributed models – spatial resolution is accounted for by
using probability distributions of input parameters across the basin.
•Advantage: structure is more physically based than lumped models.
•Less demanding on input data than distributed models.
•SWMM HEC-HMS TOPMODEL SWAT etc
 27
Deterministic Hydrologic Model
Distributed models: parameters are fully allowed to vary in space at a
resolution chosen by the user.
•Attempts to incorporate data concerning the spatial distribution of
parameters variation together with computational algorithms
•Requires large amount of data
•Governing physical processes are modeled in detail.
•Results at any location & time
•Highest accuracy in the rainfall-runoff modeling – if accurate data is
available
•High computational time, Cumbersome, experts required
•Eg. HYDROTEL; MIKE11/SHE, WATFLOOD etc.
 28
Physically based Watershed Model
Aim: Gain better understanding of hydrologic phenomena operating in a
watershed and how changes in watershed may affect these phenomena
•Complex processes
simplified by lumping process in space and time
•Laws of Physics
Conservation of mass, momentum and energy. Continuity equation,
equation of motion, equation of energy
One or more of these laws and several empirical relations are used in
physical model development
•Models may be fully distributed or semi distributed.
 29
Physically based Watershed Model
Scope of physical modeling
• Occurrence, movement, distribution and storage of water and their
variability in space and time
Technology of Physical Modeling
•Hydrodynamic Models
•theoretical, physical based or hydraulic models. e.g Dynamic wave
model for overland flow
•Overland / channel flow
• By continuity equation and momentum equation
• Modeling: May be in 1D, 2D or 3D
•Based on requirements & data availability
 30
Use of Numerical Simulation Models

• Hydrologic simulation models use mathematical equations to

calculate results like runoff volume or peak flow.
• Computer models allows parameter variation in space and time with
use of numerical methods
• Ease in simulation of complex rainfall patterns and heterogeneous
watersheds
• Evaluation of various design controls and schemes
• Effective use of land use and land cover parameters
• Improves quality of modeling using spatial characteristics
 31
Watershed Simulation Analysis
Steps for watershed simulation analysis
•Selection of model
• Input data collection: rainfall, infiltration, physiography, land use,

channel characteristics etc.

•Evaluate the study objectives under various watershed simulation
conditions.
•Selection of methods for obtaining basin hydrographs and channel
routing.
•Calibration and verification of model
•Model simulations for various conditions
•Sensitivity analysis
• Evaluate usefulness of model and comment on needed changes.
 32
Watershed Simulation Analysis
Model selection criteria
•Assumptions & conceptualization.
•Ability of model to predict variables required by the project.
•Hydrologic processes that need to be modeled to estimate the desired
outputs adequately (single-event or continuous processes)
•Availability of input data
•Expertise available & computational facility
•Price
 33
Surface Runoff estimation
Rational method (Empirical model)
Q=C i A or q=0.0028 C i A
where, q=design peak runoff rate m3/s;
C=Runoff coefficient;
i=rainfall intensity (mm/hr) for design return period and for a
duration equal to time of concentration of the watershed;
A=watershed area in ha; C= Runoff coefficient (rate of peak
runoff rate to rainfall intensity, (dimensionless));
C varies as per slope, land use etc.
 Surface Runoff estimation
34

In case of watershed comprising of different variable soil type, slope

and vegetative cover, the weighted average value of C is computed as
follows:
 35
Surface Runoff estimation
 36

Assumption and Limitation of Rational method

Watershed area < 200 acres
• The method is applicable if time of concentration (tc) for the drainage
area is less than the duration of peak rainfall intensity.
The time of concentration (tc) is the time required for water to travel from
the hydraulically most remote point of the basin to the point of interest.
The calculated runoff is directly proportional to the rainfall intensity.
Rainfall intensity is uniform throughout the duration of the storm.
The frequency of occurrence for the peak discharge is the same as the
frequency of the rainfall producing that event.
Rainfall is distributed uniformly over the drainage area.
The minimum duration to be used for computation of rainfall intensity is
10 minutes.
The rational method does not account for storage in the drainage area.
Available storage is assumed to be filled.
 37

Curve Number methods

Curve Number methods
38
Curve Number methods
39
 40
Assignment No: 2 Submission deadline 17th Nov
1. What are the important empirical equations used to calculate the
runoff?
2. Discuss why models are used for watershed planning and
management?
3. Calculate the peak runoff rate from a watershed of 100 ha for a
rainfall intensity of 1 hour duration of 60 mm per hour. The
watershed of sandy loam texture comprises of 30 ha cultivated land
on 2% slope, 40 ha pasture on 6% slope and the rest 30 ha wood
land of 8 % slope. The maximum length of flow path is 2500 m with
average slope of 4%.
NATIONAL UNIVERSITY 41

OF SCIENCES & TECHNOLOGY

THANK YOU