Arab J Geosci

DOI 10.1007/s12517-014-1340-z

ORIGINAL PAPER

A fully distributed spreadsheet modeling as a tool for analyzing

groundwater level rise problem in Jeddah city
Amro M. M. Elfeki & Jarbou Bahrawi

Received: 21 October 2013 / Accepted: 18 February 2014

# Saudi Society for Geosciences 2014

Abstract In this paper, a spreadsheet modeling is introduced was 0.96 for the hydraulic head data. The results have been
as a tool for studying fully distributed real-world engineering displayed graphically in one, two, and three dimensions.
problems. The merits of using the spreadsheet are the follow- Various scenarios have been investigated to show solutions
ing: its simplicity, the spreadsheet is very well known to many of the problem. The results have shown that the spreadsheet
researchers and engineers; it performs fast calculations; it has modeling is a powerful, flexible, and a user-friendly tool in
a user-friendly graphic interface; and it can also provide modeling fully distributed real-world engineering problems.
animated graphs. These merits provide engineers and scien- Results can be transferred from spreadsheet to GIS software to
tists with a flexible tool to tackle environmental problems and present results over satellite images or base maps. Analysis of
quickly performing solution scenarios to their problems. A various solution scenarios can be done efficiently, and one can
real-world case study has been presented to illustrate the get quick answers to the decision makers at a preliminary
application of such spreadsheet models. The groundwater stage of a project before going to a sophisticated modeling
level rise in Jeddah city is one of the problems that Jeddah approach with specialized software.
city are facing. The reasons of groundwater level rise in
Jeddah are the following: the leakage from septic tanks, leak- Keywords Spreadsheet . Groundwater level rise .
age from freshwater pipe networks, landscape irrigation, an- Microcomputer modeling . Environmental models . Jeddah
nual rainfall, flash floods from eastern wadis, etc. These city
sources have been complied with a regional groundwater flow
model (the general form of the partial differential equation for
steady groundwater flow in a heterogeneous, unconfined Introduction
aquifer with a fully distributed recharge, evaporation and
pumping rates, and with a variable aquifer bottom elevation), Comprehensive numerical models have become very popular
discritized with a finite difference method, then formulated on tools to solve environmental issues. Nowadays, there are
a spreadsheet, and solved with the built-in macro-solver on the many software available to study environmental problems
Excel software. The spreadsheet model has been tested and such as MODFLOW (Harbaugh 2005) and MT3DMS
compared with analytical solution. The model has also been (Zhang and Wang 1999) among others. Solving problems
calibrated with observations, and the correlation coefficient with such software needs time and effort such as following
training courses to get acquainted with the software since this
software became sophisticated.
As experiences are gained by using spreadsheet modeling,
A. M. M. Elfeki (*) : J. Bahrawi
it becomes apparently a very powerful computational tool for
Department of Hydrology and Water Resources Management,
Faculty of Meteorology, Environment and Arid Land Agriculture, scientists and engineers (Liengme 2003; DeCoursey 2003).
King Abdulaziz University, P.O. Box 80208, Jeddah 21589, Saudi Thus, nowadays, a variety of analyses can be performed using
Arabia microcomputers that traditionally in the past required main-
e-mail: [email protected]
frame computers, supercomputers, or sophisticated software.
A. M. M. Elfeki In the literature, there are many research works in that specific
e-mail: [email protected] area. Olsthoorn (1985 and 1998) shows how to use
 Arab J Geosci

spreadsheets for groundwater modeling. Hancock and Heaney and/or geometry throughout the solution domain. However,
(1987) have shown applications of electronic spreadsheets for the variation of the martial properties in their technique con-
hydrological modeling. They analyzed the impact of well- siders only variation of the materials over zones such as dam
field development and surface drainage on the hydrology of core and dam toe. However, in the current study, the authors
the Cypress Creek Basin north of Tampa, FL, USA. Ribando considered a fully distributed martial property, climatic forc-
and Galbis-Reig (1998) and Ribando (1998) have shown ing, and pumping over the domain cells in the model design.
applications of the advanced spreadsheet features for heat Karahan (2007) has developed a numerical solution of the
and mass transfer problems with Visual Basic Applications ADE with a third-order upwind scheme by using spreadsheet
(VBA) on Excel; however, these applications are limited to simulation (ADE-TUSS). The solution is a user-friendly and a
simple cases and mainly for educational purposes. flexible solution algorithm for the numerical solution of the
Karahan and Ayvaz (2005a, b) have developed a time- one-dimensional (1-D) ADE. One of the most important ad-
dependent groundwater model using spreadsheet (TGMSS). vantages of ADE-TUSS algorithm is that it does not require
The model can be considered as a practical method that uses the matrix algebra at each time step of the transient solutions.
spreadsheets instead of the conventional solution methods. The ADE-TUSS model has been tested with analytical solu-
The model is designed for confined heterogeneous aquifer tions. It has been concluded that the use of the high-order
with horizontal bottom, and no climatic forcing is introduced schemes in the spreadsheet simulation is very applicable for
in the model. They compared their results with MODFLOW the numerical solution of ADE. However, the cases presented
and show good agreement in terms of the hydraulic heads. The in Karahan study are mainly hypothetical that assume constant
displayed examples are just hypothetical cases. However, in advection velocity and dispersion coefficient; it would be
the current study, the developed model is for unconfined good for educational purposes, while cannot be applied to
heterogeneous aquifer with variable aquifer thickness; vari- real-world problems.
able bottom topography, climatic forcing (such as infiltration Aravind et al. (2010) have applied the spreadsheet ap-
rates and evaporation rates), and pumping are included in a proach to clustering. Their work shows a simple approach of
fully distributed type of approach and applied on a real-world the clustering process using the simplest of clustering algo-
case study. rithms—the K-means. The novelty of their work comes from
Karahan (2006) showed applications of the spreadsheets the fact that it shows a way to perform clustering in Microsoft
for the implicit solution of the advection–diffusion equation Excel 2007 without using macros, through the innovative use
based on finite differences ( which stands for advection– of what-if analysis.
diffusion equation implicit spreadsheet simulation Bhattacharjya (2011) has developed a simple spreadsheet
(ADEISS)). The model is flexible in the sense that by chang- cell-oriented methodology for solving a groundwater flow
ing only the values of temporal and spatial weighted param- inverse problem. The groundwater flow problem is solved
eters with ADEISS implementation, solutions are implicitly using an embedded optimization approach and spreadsheet
obtained for backward in time centered in space (BTCS), solver. The simulation is compared with the result of a GMS
upwind, and Crank–Nicolson schemes. The ADEISS uses MODFLOW simulation. The spreadsheet model produces
iterative spreadsheet solution technique. Thus, it does not similar results with the GMS MODFLOW simulation.
require a solution of simultaneous equations for each time Bhattacharjya concluded that the methodology may be used
step using matrix algebra. It has been concluded that the by the practicing engineer for solving small-scale groundwater
ADEISS implementation is a computationally convenient problems.
procedure for the three well-known methods in the literature: Karahan (2012) has showed the use of spreadsheet solver
the BTCS, upwind, and Crank–Nicolson. in estimating the parameters for the Muskingum routing tech-
Ayvaz and Karahan (2007) have developed an application nique for educational and engineering purposes. Two solution
of three-dimensional free surface flow problems. The model is techniques are proposed, namely the trial–error with spread-
based on the finite difference equation that has been derived in sheet procedure and the nonlinear optimization method with
the general case of anisotropic and nonuniform hydraulic the Excel solver. Both of the proposed methods are tested on
conductivities with variable grid spacing. One of the main three samples given in literature.
advantages of their proposed method is that it eliminates the The objective of this paper is to show the powerfulness of
matrix algebra to a vector form of the governing equations and spreadsheet models in analyzing real-world engineering prob-
seepage problems may be iteratively solved with spreadsheets. lems, proposing solutions directly and quickly in an efficient
This technique has been adopted in the current study. Another manner, and displaying the solution scenarios graphically in
advantage of their proposed solution technique among previ- two-dimensional (2-D) and three-dimensional (3-D) formats
ous ones is that it considers the variation of material properties (using the built-in graphical interface) without programming
Arab J Geosci

with VBA language and by using only the built-in Excel The groundwater level rise in Jeddah was first inves-
iterative solution macros. tigated by Abu-Rizaizah (1996, 1997, 1999); latter, the
A real-world engineering problem, namely the groundwa- problem is investigated by Alquhtani and Shehata
ter level rise in Jeddah city, has been analyzed through the (2003). They predicted the rate of water rise of 36 to
development of a 2-D steady state groundwater flow in a 204 cm/year depending on the location. A survey by
heterogeneous unconfined aquifer with variable bottom to- Alquhtani and Shehata (2003) has confirmed these find-
pography and taking into account a fully distributed recharge, ings. Studies by Basamed (2001) showed that the change
evaporation rates, and pumping. The problem is formulated on in water table elevation is changing ±0.12 m/year.
multiple spreadsheets to solve a fully distributed groundwater Sources of groundwater level rise as stated by AL-
model. The spreadsheet model has been verified and calibrat- Sefry and Sen (2006) are the following: (1) leakage from
ed form field data. The application shows also how one can water supply systems which has been estimated to be
display solution scenarios to decision makers and how to 191,000 m3/day, (2) infiltration from cesspools is about
transfer results from Excel to GIS software to display results 89,250 m3/day, (3) groundwater recharge from rainfall is
on satellite images and base maps. about 0.37 m/year, (4) excess landscape irrigation is
about 26,300 m3/day, (5) leakage from underground
storage tanks cannot be estimated, and (6) subsurface
Environmental problem statement inflow from the eastern wadis is estimated to be
33,000 m3/day. By summing up all these quantities
The environmental problem that has been tackled in this paper and dividing by the Jeddah area under study that is
is the groundwater level rise in Jeddah city. Jeddah city is the about 1,765 km2, it ends up with an average infiltration
second major city in Saudi Arabia, and it is located on the Red rate of about 1.2 mm/day (44 cm/year). If one assumes
Sea coastal plains as shown in Fig. 1a. that this rate is equal to the rate of the water rise,

-4

-8

-12

-16

-20

-24
0 2 4 6 8 10 12 14 16 18 20 22 24

A B
Fig. 1 Satellite image of the study area with a model grid of 1 km×1 km (a) and the groundwater elevation contour map in Jeddah city within the study
area (adapted from AL-Sefry and Sen 2006; b)
 Arab J Geosci

therefore under this assumption, this value falls in the the underlying aquifer. The general steady state 2-D
range (36–204-cm/year rate of water rise) suggested by groundwater flow model with sources and sinks is given
Abu Rizaizah (1996, 1997, 1999). by Batelaan et al. (1998),
In this study, the effect of this recharge on the water



level variation in Jeddah will be studied through the ∂ ∂h ∂ ∂h
spreadsheet modeling, and various scenarios shall be K ðh−Z Þ þ K ðh−Z Þ þ R−E−W ¼ 0 ð1Þ
∂x ∂x ∂y ∂y
evaluated form the model results in the following sec-
tions. There are very few papers published about the
topic, which were hardly collected and analyzed to where K is the hydraulic conductivity (L/T), h is the hydraulic
obtain the necessary data for the modeling task. head (L), Z is the aquifer bottom elevation (L), R is the
Figure 2 is adapted from Alquhtani and Shehata groundwater recharge rate (L/T), E is the evaporation rate
(2003) who mapped the different zones in Jeddah city from the groundwater (L/T), and W is the pumping rate (L3/
based on aquifer hydraulic conductivity, aquifer thick- T/L2).
ness, ground surface topography, and sewage discharge Equation 1 can be discritized using block-centered finite
per litter per day per person. These data have been difference scheme as follows:
digitized and incorporated in Excel sheet for the model
applications. For the groundwater data, the work by AL-
  h −h  ∂
  h −h 
∂ i; j i; j
Sefry and Sen (2006) was adopted (see Fig. 1b) and K iþ 12 ; j hiþ1; j −Z iþ1; j iþ1; j þ K i; jþ 12 hi; jþ1 −Z i; jþ1 i; jþ1
∂x 2 2 Δx ∂y 2 2 Δy
digitized to get the necessary information of the water
þ Ri; j −E i; j −W i; j ¼ 0
table elevation, which has been used later in the coming
section for model calibration. ð2Þ

where i is the index for x-direction, j is the index for y-

Mathematical formulation direction, Δx is the interval in x-direction, and Δy is the
interval in the y-direction.
To study the groundwater rise problem in Jeddah city, a Further elaboration of the outer derivatives would lead to
model has been developed for the groundwater flow in the following expression:

 
h −h   
h −h 
i; j i−1; j
K iþ 12 ; j hiþ1; j −Z iþ1; j iþ1; j
−K i− 12 ; j hi−1; j −Z i−1; j i; j

2 2 Δx 2 2 Δx
 
h −h  Δx  
h −h  ð3Þ
i; j i; j
K i; jþ 12 hi; jþ1 −Z i; jþ1 i; jþ1
−K i; j− 12 hi; j−1 −Z i; j−1 i; j−1

Δy Δy
þ þ Ri; j −E i; j −W i; j ¼ 0
2 2 2 2

Δy

Simplifying the above equation leads to    

K iþ 12 ; j hiþ 12 ; j −Z iþ 12 ; j K i− 12 ; j hi− 12 ; j −Z i− 12 ; j
hiþ1; j þ hi−1; j
Δx2  Δy 
2

 h   h
K iþ 12 ; j hiþ1; j −Z iþ1; j
iþ1; j
− K h −Z
i; j K i; jþ 12 hi; jþ 12 −Z i; jþ 12 K i; j− 12 hi; j− 12 −Z i; j− 12
1
;
 hΔx 2 iþ 2 j

1
iþ ; j 1
iþ ; j
hΔx 2 þ hi; jþ1 þ hi; j−1
Δy2 Δy2
2 2 2 2
i; j i−1; j
−K i− 12 ; j hi−1; j −Z i−1; j þ K i− 12 ; j hi−1; j −Z i−1; j W i: j
Δx Δx þRi; j −E i; j − ¼
2 2
 2 2
 2 2

þK iþ 12 ; j hiþ 1; j −Z iþ1; j
hiþ1; j
−K iþ 12 ; j hiþ1; j −Z iþ1; j
hi; j ð4Þ 2  ΔxΔy   
Δy2 Δy2 K iþ 12 ; j hiþ 12 ; j −Z iþ 12 ; j K i− 12 ; j hi− 12 ; j −Z i− 12 ; j
 h  h
2 2 2 2

i; j i−1; j 4 þ
−K i− 12 ; j hi−1; j −Z i−1; j þ K i− 12 ; j hi−1; j −Z i−1; j Δx2 Δy2
2 2 Δy2 2 2 Δy2    3
þ Ri; j − E i; j − W i; j ¼ 0 K i; jþ 12 hi; jþ 12 −Z i; jþ 12 K i; j− 12 hi; j− 12 −Z i; j− 12
þ þ 5 hi; j
Δy2 Δy2
Factorization of the above equation together with
collecting common terms leads to ð5Þ
Arab J Geosci

cells) between the two neighboring cells (i + 1, j) and

(i,j) calculated as

−1
1 1
K iþ 12 ; j ¼ 2 þ ð7Þ
K iþ1; j K i; j

hiþ12; j is the arithmetic average hydraulic head between two

neighboring cells (i + 1,j) and (i, j) calculated as

1 
hiþ 12 ; j ¼ hiþ1; j þ hi; j ð8Þ
2

Z iþ12; j is the arithmetic average bottom elevation between

two neighboring cells (i + 1,j) and (i, j) calculated as
1 
Z iþ 12 ; j ¼ Z iþ1; j þ Z i; j ð9Þ
2

Equation 6 can be formulated in spreadsheet and solved

numerically by iterative solver in Excel® that will be ex-
plained in the next chapter.

Spreadsheet modeling
Fig. 2 Data on the characteristics of the groundwater aquifer under
Jeddah city (adapted from Alquhtani and Shehata 2003)
Spreadsheets are popular for many users, e.g., engineers and
scientists. The reason for its popularity is that it combines four
programs: the electronic spreadsheet, graphical interface, data
management, and functions’ library. In the current applica-
After the algebraic manipulation, we get the final equation tions, multiple sheets are used. Each sheet carries its own data
2     3 for the problem. The groundwater level data presented in
K iþ 12 ; j hiþ 12 ; j −Z iþ 12 ; j K i− 12 ; j hi− 12 ; j −Z i− 12 ; j Fig. 1b have been schematized and digitized into cells of
6 hiþ1; j þ hi−1; j 7
6 7 1 km2 in a spreadsheet called “GWMAP” as presented in
Δx Δy
2 2
6   7
6 7 Fig. 3 (top). The use of the graphical interface in the Excel
hi; j ¼ 6
6þ
K i; jþ 12 hi; jþ 12 −Z i; jþ 12 K i; j− 12 hi; j− 12 −Z i; j− 12 7
6 hi; jþ1 þ hi; j−1 7
7 software provides us the ability to present the sheet graphical-
6 Δy 2 Δy 2 7
4 W i; j 5 ly as a 2-D contour map displayed in Fig. 3 (bottom).
þRi; j − E i; j −
2  ΔxΔy    In the same manner, other sheets have also been generated
K iþ 12 ; j hiþ 12 ; j −Z iþ 12 ; j K i− 12 ; j hi− 12 ; j −Z i− 12 ; j such as the aquifer thickness and the sewage discharge as
4 þ shown in Fig. 4 (top) and (bottom), respectively.
Δx2 Δy2
   3 The spreadsheet model presented is a fully distributed
K i; jþ 12 hi; jþ 12 −Z i; jþ 12 K i; j− 12 hi; j− 12 −Z i; j− 12 groundwater model that uses multiple sheets (such as layers
þ þ 5
Δy2 Δy2 in terms of GIS technology) namely the following: (1) a sheet
for fully distributed hydraulic conductivity, (2) a sheet for
ð6Þ fully distributed natural recharge or infiltration, (3) a sheet
where K iþ1; j is the harmonic average hydraulic conductivity for fully distributed evaporation or exfiltration, (4) a sheet for
2
(to keep the mass balance for the flow between the fully distributed pumping, (5) a sheet for fully distributed
 Arab J Geosci

Fig. 3 The schematized color map that has been represented in the spreadsheet, with groundwater table elevations in the cells (top) and the 2-D graphical
representation in Excel (bottom)

bedrock elevation topography, (6) a sheet for fully distributed assumed given head boundaries where the groundwater
water surface elevation, (7) a sheet for defining the initial level is specified at these boundaries. However, the top
condition, and (8) a sheet for the definition of the boundary and bottom boundaries are assumed to be no flow
conditions. where there is no flow crossing these boundaries. This
Figure 5 shows the formulation of the equation in the feature has been presented in the Excel sheet by
Excel sheet and its surrounding cells. The black cells in assigning the values at the black cells on top and
the sheet represent boundary conditions. In the present- bottom to be equal to the head at the neighboring cells
ed sheet in the figure, the left and right boundaries are up or down the cell under consideration, respectively.
Arab J Geosci

Fig. 4 The schematic map that has been represented in the spreadsheet: aquifer thickness in meters (top) and the sewage discharge liter per day per
person (bottom) according to the data presented in Fig. 2

Fig. 5 The central cell and its four neighbors used for the calculations of given head boundary equals to 1 m, right side is a given boundary
the analytical solution case. The first condition in the model equation is a condition equals to 10 m, and the top and bottom boundaries are assumed
switch (IF statement) to start running the model, if the value at cell $E$11 no-flow boundaries. The model is initialized by zero values on the entire
is equal to “0,” the model is standby, while if the same cell is not “0,” the domain
model runs. The boundary conditions are assumed as follows: left side is a
 Arab J Geosci

Equation 6 is programmed in the Excel sheet as follows:

¼ IFðData!$E$6 ¼ 0; 0;

IFðQ15 < 0; 0; ðððð2=ð1=Permeability!Q15 þ 1=Permeability!R15ÞÞ*ð0:5*ðQ15 þ R15Þ‐0:5*ðBed!Q15 þ Bed!R15ÞÞ=Data!E112 *

R15 þ ðð2=ð1=Permeability!Q15 þ 1=Permeability!P15ÞÞ*ð0:5*ðQ15 þ P15Þ‐0:5*ðBed!Q15 þ Bed!P15ÞÞ=Data!E112 *

P15 þ ðð2=ð1=Permeability!Q15 þ 1=Permeability!Q16ÞÞ*ð0:5*ðQ15 þ Q16Þ‐0:5*ðBed!Q15 þ Bed!Q16ÞÞ=Data!$E$122 *

Q16 þ ðð2=ð1=Permeability!Q15 þ 1=Permeability!Q14ÞÞ*ð0:5*ðQ15 þ Q14Þ‐0:5*ðBed!Q15 þ Bed!Q14ÞÞ=Data!E122 *
ð10Þ
Q14 þ Recharge!Q15‐Evapo!Q15‐ðPumping!Q15=ðData!E11*Data!E12ÞÞÞ=ððð2=ð1=Permeability!Q15 þ 1=Permeability!R15ÞÞ*

ð0:5*ðQ15 þ R15Þ‐0:5*ðBed!Q15 þ Bed!R15ÞÞ=ðData!E11Þ2 þ ðð2=ð1=Permeability!Q15 þ 1=Permeability!P15ÞÞ*

0:5*ðQ15 þ P15Þ‐0:5*ðBed!Q15 þ Bed!P15ÞÞ=ðData!E11Þ2 þ ðð2=ð1=Permeability!Q15 þ 1=Permeability!Q16ÞÞ*

ðð2=
ð0:5*ðQ15 þ Q16Þ‐0:5*ðBed!Q15 þ Bed!Q16ÞÞ=ðData!E12 Þ2ð1=Permeability!Q15
þ þ 1=Permeability!Q14ÞÞ*

ð0:5*ðQ15 þ Q14Þ‐0:5*ðBed!Q15 þ Bed!Q14ÞÞ=ðData!E12Þ 2

Each term in the equation is illustrated by its corresponding the cells inside the solution domain. It controls the values in
formula and presented in Fig. 6. The first IF statement at cell cells of the solution domain from becoming negative to avoid
$E$6 is a switch to run the program. If the value in the cell is the divergence of the solution.
equal to “0,” the program is in standing by mode, while when The calculations of the model are performed using the
the value at the cell is not “0,” the program runs, and pushing Excel sheet presented in Fig. 7 that shows the number of
F9 several times is normally used to continue calculating until iterations assigned by the modeler (in that paper 1,000 itera-
the solution reaches convergence (i.e., the values in the do- tions are sufficient) and the corresponding accuracy that de-
main cells do not change anymore). The second IF statement, scribes the relative difference between two successive itera-
“Q15 < 0,” is just a control statement of the hydraulic head in tions (accuracy of 0.001 is sufficient).

Fig. 6 Formulation of the equation in the Excel sheet noting that the first term #1 is a switch (IF statement) to start running the model, if the value at cell
$E$6 is equal to “0,” the model is standby, while if the same cell is not “0,” the model runs
Arab J Geosci

Fig. 7 The Excel panel for setting up the maximum number of iterations calculation, Maximum Iterations: “1,000,” and Maximum change:
and the allowable criteria for the model conversion used in the spread- “0.001.” The rest are the default values
sheet calculations. Workbook calculation: “Automatic,” Enable iterative

Spreadsheet model testing

The model has been tested for the special case where analyt-
ical solution exists, i.e., in case of a homogenous and a
rectangular domain with no pumping.
Since there is no 2-D analytical solution exists for the
developed 2-D model, therefore the well-known analytical
solution case for the 1-D groundwater flow is used (Wang
and Anderson 1982).
The analytical solution gives the groundwater head distri-
bution for a homogeneous unconfined aquifer with a recharge
from the top and a horizontal aquifer bottom. This case is
expressed analytically as Wang and Anderson (1982),

sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
 Fig. 8 Comparison between analytical solution and numerical model
h21 −h22 R results for homogeneous aquifer with no recharge or pumping in a
hð x Þ ¼ h1 −
2
x þ ðL−xÞx ð11Þ rectangular domain and horizontal aquifer bottom under the boundary
L K conditions given in Fig. 5
 Arab J Geosci

Fig. 9 a Comparison of the hydraulic head distribution over the study head distribution over the satellite image of the study area between the
domain between the observations (A) and the spreadsheet model results observations (A) and the spreadsheet model results (B); the color scale
(B) using the Excel graphical interface. b Comparison of the hydraulic bar shows the hydraulic head values

where h2 hydraulic head at the downstream boundary (m)

R effective recharge rate (m/day)
h(x) hydraulic head at x location along the aquifer (m) K hydraulic conductivity (m/day)
h1 hydraulic head at the upstream boundary (m) L aquifer length (m)
Arab J Geosci

In this test case, it is assumed that the aquifer has a hydrau-

lic conductivity of 11.5 m/day, a length of 3,000 m, no
recharge is considered, and the aquifer bottom is assumed
horizontal. The boundary conditions for test case are assumed
as follows: Left side is a given head boundary equals to 1 m,
right side is a given boundary condition equals to 10 m, and
the top and bottom boundaries are assumed no-flow bound-
aries. The model is initialized by inserting zero values on the
entire cells in the domain. Figure 8 shows comparison be-
tween the analytical solution and the corresponding numerical
solution from the Excel sheet. The results show perfect agree-
ment between both solutions and provide confidence in the
model.
Fig. 10 Model calibration: profile of the hydraulic head form the data
and the model results at a horizontal section at the center line of the
model. The left part of the graph is sea side
Analysis of model results

The spreadsheet model has been calibrated by changing

the values of the hydraulic conductivity given by the Figure 11 shows the effect of this scenario on the water
data until both the calculated water table elevation and level profile. For this scenario, the water level is reduced by
water table elevation provided by the data match (i.e., about 75 cm on average over the city of Jeddah. The second
the groundwater map in Fig. 1b). Through this calibra- scenario considers the case when there is no action is taken to
tion process, two figures have been produced. Figure 9a solve the problem and there is a population growth that leads
shows a comparison between the schematized ground- to an increase in the leakage there times the current case. The
water map presented in “A” and the calculated ground- results are presented in Fig. 11. The water level increases by
water map from the spreadsheet model presented in 1.65 m. Figure 12 shows the capabilities of the 3-D graphical
“B.” The matching is pretty good; the correlation coef- representation in the spreadsheets. The figure shows the water
ficient between the model results and the observations is table surface in both cases. It might be difficult to recognize
0.96. The groundwater level pattern is pretty similar. the difference in the 3-D maps; however, on the Excel sheet,
The calculated hydraulic head seems to be smoother in you can easily recognize the effects.
comparison with the data which is due to the schema- The presented results show the capabilities of spreadsheet
tization over a relatively large cell sizes (1 km×1 km). to analyze environmental problems and to visualize the effect
Also, the representation of the mountainous part was of any action that may be taken by decision makers for the
not very good. Figure 9b shows the same comparison model output in a quick way and help to understand the
for the hydraulic head distribution superimposed over system behavior under any action.
the satellite image of the study area. This is one of
the good flexibility between Excel outputs and GIS
programs such as SURFER software where the data
can be converted easily and displayed over base maps.
This option is very helpful for decision makers to facil-
itate their decision.
Figure 10 shows the matching between the calculated
hydraulic head and the data from the schematized map
through a profile taking in the middle of the area of interest.
The comparison is also very good and provides more confi-
dence in the model for further elaboration.
Figure 11 shows two proposed scenarios, the first one when
the leakage is reduced by 90 %, i.e., R=0.00012 m/day instead
of R=0.0012 m/day at the current condition. This case corre-
sponds to a performing maintenance of the pipe network
system periodically and preventing leakage from the septic
tanks by constructing a sewerage network system for the city Fig. 11 Comparison of the hydraulic head profiles at two scenarios with
of Jeddah. the current condition
 Arab J Geosci

Fig. 12 3-D view of the hydraulic head profiles for the two considered scenarios: a reduced leakage by 90 % and b increased leakage by three times

Conclusions This scenario is translated as an increase of the recharge three

times of the current situation; under this condition, the ground-
Spreadsheet modeling on microcomputers has been proven to water water level rises up to 1.65 m on average above the
be a powerful tool in analyzing real-world engineering prob- current condition.
lems. The spreadsheet models are capable of handling a fully The results have shown that spreadsheet models are
distributed domain and not just hypothetical cases as in the powerful, flexible, and a user-friendly tool in modeling fully
previous studies. In the current study, an application of spread- distributed real-world engineering problems without
sheet modeling is made to the groundwater level rise in Jeddah performing programming and the results can be transferred
city. A complete development of the model is archived, to GIS software to be displayed over a base map or a satellite
starting from the mathematical formulation of the problem, image. It can also be used to get quick solutions to problems
schematization, introducing model equations into the spread- and provide decision makers with relatively some answers
sheet cells, and the use of the built-in iterative solver to before using a sophisticated modeling approach with specialized
perform numerical solution to the problem. The developed software.
spreadsheet model “GWMAP” is a fully distributed ground-
water model that uses multiple sheets such as layers in terms Acknowledgments The authors would like to thank the graduate student
Mr. Badr Qabalawi for performing some preliminary calculations using the
of GIS technology, namely the following: (1) a sheet for fully
spreadsheet model in his BSc. graduation report. The first author is on
distributed hydraulic conductivity, (2) a sheet for fully distrib- leave of absence from faculty of engineering, Mansoura University,
uted natural recharge or infiltration, (3) a sheet for fully Mansoura, Egypt.
distributed evaporation or exfiltration, (4) a sheet for fully
distributed pumping, (5) a sheet for fully distributed bedrock
elevation topography, (6) a sheet for fully distributed water
surface elevation, (7) a sheet for defining the initial condition, References
and (8) a sheet for the definition of the boundary conditions.
The model has been tested with analytical solution and cali- Abu-Rizaiza OS (1996) Groundwater rise problem in North Jeddah, Final
brated based on observations with a pretty good correlation report (In Arabic) Water and Sanitary Drainage Department of
Makkah province, 2 vol. Saudi Arabia
coefficient of 0.96 for the hydraulic heads. It has been shown Abu-Rizaiza OS (1997) Groundwater rise problem in South Jeddah, Final
how to introduce solution scenarios to the problem and dis- report (In Arabiac) Water and Sanitary Drainage Department of
play the results graphically at once using the interactive Makkah province, Saudi Arabia
graphical interface. Two scenarios are presented. The first Abu-Rizaiza OS (1999) Threats from groundwater table rise in urban
areas in developing countries. IWRA, Water international, Vol.24,
one is the case of reducing the leakage by 90 % (through the no. 1, 46-52
maintenance of the pipe network systems and replacing septic Alquhtani MB and Shehata WM (2003) Vulnerability map of the ground-
tanks with a proper sewerage system) that has lead to a water rise in Jeddah, Saudi Arabia: 12th Asian regional conf. on the
decrease in the groundwater level, with respect to the current soil mechanics and geotechnical engineering, 4–8 August, 2003,
Singapore, pp 357, Chapter 3–3
condition, by about 75 cm on average. However, the second AL-Sefry SA, Sen Z (2006) Groundwater rise problem and risk evalua-
scenarios assumed the case when there is no solution to the tion in major cities of arid lands- Jeddah case in Kingdom of Saudi
problem and there is a growth in the population of the city. Arabia. Water Resour Manag 20:91–108
Arab J Geosci

Aravind H, Rajgopal C, Soman KP (2010) A Simple Approach to Karahan H (2007) Unconditional stable explicit finite difference tech-
Clustering in Excel. Int J Comput Appl 11(7):1–25 nique for the advection-diffusion equation using spreadsheets. Adv
Ayvaz MT, Karahan H (2007) Modeling three-dimensional free-surface Eng Softw 38(2):80–86
flows using multiple spreadsheets an extended pressure application Karahan H (2012) Predicting Muskingum flood routing parameters using
for solving three-dimensional seepage problems with a free-surface. spreadsheets. Comput Appl Eng Educ 20(2):280–286
Comput Geotech 34:112–123 Karahan H, Ayvaz MT (2005a) Transient groundwater modeling using
Basamed AS (2001) Hydrochemical study and bacteriological effects on spreadsheets. Adv Eng Softw 36:374–384
groundwater in the northern part of Jeddah district. Unpublished Karahan H, Ayvaz MT (2005b) Time-dependent groundwater modelling
M.Sc. Thesis, Faculty of Earth Sciences, King Abdulaziz University, using spreadsheets. Comput Appl Eng Educ 13(3):192–199
Jeddah, Saudi Arabia, 234 pp Liengme BV (2003) A guide to Microsoft Excel 2002 for scientists and
Batelaan O, De Smedt F, De Becker P, Huybrechts W (1998) engineers. Elsevier Butterworth-Heinemann, Oxford, 314 pp
Characterization of a regional groundwater discharge area by com- Olsthoorn TN (1985) The power of the electronic worksheet: modeling
bined analysis of hydrochemistry, remote sensing and groundwater without special programs. Groundwater J 23(3):381–390
modeling. In: Dillon P, Simmers I (eds) Shallow Groundwater Olsthoorn TN (1998) Groundwater modelling: calibration and the use of
Systems. IAH publ, Rotterdam, pp 75–86, Chapter 7 spreadsheets. Delft University Press, Delft, ISBN 90-407-1702-8 /
Bhattacharjya RK (2011) Solving groundwater flow inverse problem CIP, about 300 pp
using spreadsheet solver. J Hydrol Eng 16(5):472–477 Ribando RJ (1998) An Excel/Visual Basic for Applications (VBA) prim-
DeCoursey WJ (2003) Statistics and probability for engineering applica- er. Comput Educ J VIII(2):38–43
tions with Microsoft® Excel. Library of Congress Cataloging-in- Ribando RJ, Galbis-Reig V (1998) Convective heat and mass transfer
Publication Data, 396 pp from a runner using some advanced spreadsheet features. Comput
Hancock MC, Heaney JP (1987) Water resources analysis using electron- Educ J VIII(4):22–28
ic spreadsheets. J Water Resour Plan 113(5):639–658 Wang HF, Anderson MP (1982) Introduction to groundwater modeling:
Harbaugh AW (2005) MODFLOW-2005, The U.S. geological survey finite difference and finite element methods. Freeman and Company,
modular ground-water model—the ground-water flow process: U.S. Bedford, 237p
Geological Survey Techniques and Methods 6-A16. http://water. Zheng C, Wang PP (1999) MT3DMS: a modular three-dimensional
usgs.gov/ogw/modflow/MODFLOW.html multispecies transport model for simulation of advection, dispersion
Karahan H (2006) Implicit finite difference techniques for the advection- and chemical reactions of contaminants in groundwater systems,
diffusion equation using spreadsheets. Adv Eng Softw 37(9):601– Documentation and User’s Guide, Contract Report SERDP-99-1.
608 U.S. Army Engineer Research and Development Center, Vicksburg