Journal of Environmental Protection, 2011, 2, 56-71
doi:10.4236/jep.2011.21006 Published Online March 2011 (http://www.SciRP.org/journal/jep)
Copyright © 2011 SciRes. JEP
Application of Artificial Neural Networks Model as
Analytical Tool for Groundwater Salinity
Mohamed Seyam1, Yunes Mogheir2
1Civil Engineering Department, Engineering Faculty, Islamic University of Gaza, Gaza, Palestine; 2Environmental Engineering De-
partment, Engineering Faculty, Islamic University of Gaza, Gaza, Palestine.
Email: mohseyam@yahoo.com, ymogheir@iugaza.edu.ps
Received September 1st, 2010; revised October 28th, 2010; accepted December 7th, 2010.
ABSTRACT
The main source of water in Gaza Strip is the shallow coastal aquifer. It is extremely deteriorated in terms of salinity
which influenced by many variables. Studying the relation between these variables and salinity is often a complex and
nonlinear process, making it suitable to model by Artificial Neural Networks (ANN). Initially, it is assumed that the
salinity (represented by chloride concentration, mg/l) may be affected by some variables as: recharge rate, abstraction,
abstraction average rate, life time and aquifer thickness. Data were extracted from 56 municipal wells, covering the
area of Gaza Strip. After a number of modeling trials, the best neural network was determin ed to b e Multila yer Percep-
tron network (MLP) with four layers: an input layer of 6 neurons, first hidden layer with 10 neurons, second hidden
layer with 7 neurons and the output layer with 1 neuron which gives the final chloride concentration. The ANN model
generated very good results depending on the high correlation between the observed and simulated values of chloride
concentration. The correlation coefficient (r) was 0.9848. The high value of (r) showed that the simulated chloride
concentration values using the ANN model were in very good agreement with the observed chloride concentration
which mean that ANN model is useful and applicable for groundwater salinity modeling. ANN model was successfully
utilized as analytical tool to study influence of the input variables on chloride concentration. It proved that chloride
concentration in groundwater is reduced by decreasing abstraction, abstraction average rate and life time. Further-
more, it is reduced by increasing recharge rate and aquifer thickness.
Keywords: Groundwater, Salinity, Artificial Neural Networks, Modeling, Anal yti c al Too l
1. Introduction
The main source of water in Gaza Strip is the shallow
aquifer which is part of the coastal aquifer. The quality of
the groundwater is extremely deteriorated in terms of
salinity and nitrates. Salinity in the Gaza coastal aquifer
is often described by the chloride concentration in
groundwater. Depending on location and hydrochemical
processes, rates of salinization may be gradual or sudden
[1].
Salinization of groundwater may be caused by a num-
ber and/or combination of different processes, including:
seawater intrusion, migration of brines from the deeper
parts of the aquifer, dissolution of soluble salts in the
aquifer (water-rock interaction), and contribution from
discharges from older formations surrounding the coastal
aquifer. In addition, potential man-induced (anthropo-
genic) sources include agricultural return flows, waste-
water seepage, and disposal of industrial wastes [2]. In
addition, water quality (e.g-salinization) is influenced by
many factors such as flow rate, contaminant load, me-
dium of transport, water levels, initial conditions and
other site-specific parameters. The estimation of such
variables is often a complex and nonlinear process, mak-
ing it suitable for Artificial Neural Networks (ANN) ap-
plication [3].
The importance of this article is to develop ANN
model studying th e relation between groundwater salinity
(represented by chloride concentration mg/l) and some
hydrological variables as: recharge rate (R), abstraction
(Q), abstraction average rate (Qr), life time (Lt), and aq-
uifer thickness (Th). Understanding spatial relations be-
tween hydrological variables and salinity of groundwater
can contribute in an integration of water resources man-
agement. Modeling groundwater salin ity usin g tradition al
modeling softwares consume a lot of efforts and required
huge quantity of data while ANN could provide an easy
and efficient tool for modeling and prediction that help in
Application of Artificial Neural Networks Model as Analytical Tool for Groundwater Salinity57
water resources management. This research might be
considered as one of the few contributions in quantita-
tively modeling of the relation between groundwater sa-
linity and the hyd rological v ariables in spatial scale using
ANN.
2. Materials and Methods
2.1. Groundwater Salinity in Gaza Strip
The Gaza Strip is a narrow strip of land on the Mediter-
ranean coast. The area is bounded by the Mediterranean
in the west, the 1948 cease-fire line in the north and east
and by Egypt in the south. The total area of the Gaza
Strip is 365 km2 with approximately 40 km long and the
width varies from 8 km in the north to 14 km in the south
[4]. Figure 1 showed regional and location map of Gaza
Strip.
Gaza Strip is one of the places where the exploitation
level of recourses exceeds the carrying capacity of the
environment. This is especially true for the water and
land resources, which are under high pressure and subject
to sever over exploitation, pollution and degradation.
Quality of the groundwater is a major problem in Gaza
strip. The aquifer is highly vulnerable to pollution. The
domestic water is becoming more saline every year and
average chloride concentrations of 500 mg/l or more is
no longer an exception. Most of the public water supply
wells don’t comply with the drinking water quality stan-
dards and concentrations of chloride and nitrate of the
water exceed the World Health Organization (WHO)
standards in most drinking water wells of the area and
represent the main problem of groundwater quality. Over
pumping of groundwater and salt water intrusion are the
main reasons behind high chloride concentration [2].
It is clearly noticed that the chloride concentration in-
creases significantly over all Gaza Strip especially in
southern east and middle area. The best water quality is
found in the sand dune areas in the north, mainly in the
range of 50 - 250 mg/l. Figure 2 and Figure 3 present
average chloride concentration of pumped Groundwater
of Gaza Strip for the year 2002 and 2007.
2.2. Brief description of Artificial Neural
Networks
ANN refers to computing systems whose central theme is
Figure 1. Regional and location map of Gaza Strip [5].
Copyright © 2011 SciRes. JEP
Application of Artificial Neural Networks Model as Analytical Tool for Groundwater Salinity
58
Figure 2. Average chloride concentration of pumped ground-
water of Gaza Strip for the year 2002 [6].
Figure 3. Average chloride concentration of pumped ground-
water of Gaza Strip for the year 2007 [7].
borrowed from the analogy of biological neural networks.
They represent highly si mplified mathematical models of
biological neural networks. They include the ability to
learn and generalize from examples to produce mean-
ingful solutions to problems even when input data con-
tain errors or are incomplete, and to adapt solutions over
time to compensate for changing circumstances and to
process information rapidly [8].
The brain consists of a large number of neurons, con-
nected with each other by synapses. These networks of
neurons are called neural networks, or natural neural
networks. ANN is a simplified mathematical model of a
natural neural network. ANN are a new information-
processing and computing technique inspired by bio-
logical neuron processing [9] (Lee et al., 1998). The hu-
man brain is a collection of more than 10 billion inter-
connected neurons. Each neuron is a cell that uses bio-
chemical reactions to receive, process, and transmit in-
formation [10]. Figure 4 presented mammalian neuron.
Treelike networks of nerve fibers called dendrites are
connected to the cell body or soma, where the cell nu-
cleus is located. Extending from the cell body is a single
long fiber called the axon, which eventually branches
into strands and sub strands, and is connected to other
neurons through synaptic terminals or synapses. The
transmission of signals from one neuron to another at
synapses is a complex chemical process in which specific
transmitter substances are released from the sending end
of the junction [10].
Artificial neurons connected together form a network.
The structure of ANN is, as rule, layered. Three func-
tional group can be distinguished in the ANN i.e. the
inputs receiving signals from the network’s outside and
introducing them into its inside, the neuron which proc-
ess information and the neurons which generate results.
A model of the artificial neuron is shown in the Figure 5.
ANN is an informational system simulatin g the ability
of a biological neural network by interconnecting many
simple artificial neurons. The neuron accepts inputs from
a single or multiple sources and produces outputs by
simple calculations, processing with a predetermined
non-line ar function [ 1 2].
Most ANN has three layers or more: an input layer,
which is used to present data to the network; an output
layer, which is used to produce an appropriate response
to the given input; and one or more intermediate layers,
which are used to act as a collection of feature detectors.
Determination of appropriate network architecture is one
of the most important, but also one of the most difficult,
tasks in the model-building process. Unless carefully
designed an ANN model can lead to over parameteriza-
tion, resulting in an unnecessarily large network [13].
Figure 6 demonstrated schematic description of a general
Copyright © 2011 SciRes. JEP
Application of Artificial Neural Networks Model as Analytical Tool for Groundwater Salinity59
Figure 4. Mammalian neuron [10].
Figure 5. Model of artificial neurons [11].
Figure 6. Schematic description of a three layer ANN and of
the elements of its (mathematical) neurons [14].
ANN model of three layers.
2.3. Methodology
2.3.1. Construction of Data Matri x for A NN Model
In order to model the groundwater salinity in Gaza strip
using ANN it is necessary to gather data for training
purposes. The training data must include a number of
cases, each containing values for input and output vari-
ables. The first decisions needed are: which are variables
to use, and how many (and which) cases to gather, the
choice of variables (at least initially) is guided by intui-
tion. Understanding and expertise in the problem domain
and conditions give initially idea of which input variables
are likely to be influential. Once in ANN, variables can
be select and deselect, ANN can also experimentally de-
termine useful variables [15]. As a first pass, any vari-
ables which could have an influence on groundwater
salinity should be included on initial studies.
The required data were extracted mainly from the do-
mestic wells in Gaza Strip because it usually have quality
test twice a year in February and October periodically.
The quality test includes the chloride concentration test
which gives us a great chance to monitor groundwater
salinity in Gaza Strip and it’s changes two times per year.
The previous assumed variab les will be gathered, studied,
validated and rearranged to create training data matrix
which should contain many hundreds of cases each con-
taining values for input variables and output.
In this research, it is necessary to deal with regular
time series data to construct data train ing matrix so many
sources of data have been neglected because of the defi-
ciency of complete required data. Since that the detailed
abstraction records have not been obtained for years prior
to 1996, the period of model which include the modelin g
and calibration starts from 1997 to 2006. There are an
estimated 4000 wells within the Gaza Strip, almost all of
these wells are privately owned and used for agricultural
purposes. Approximately 100 wells are owned and oper-
ated by municipalities and are used for domestic supply
[16]. In this research, data were extracted from 56 wells,
most of them are municipal wells and they almost cover
the total area of Gaza Strip as represented in Figure 7
The choice of these wells depends only on the availabil-
ity of required data.
Selection the Variables of ANN Model
Hydrogeologically, the change of chloride concentra-
tion (salinity) was assumed to be depend on many vari-
ables such as infiltration, ab straction, life time of ab strac-
tion from aquifer, groundwater level, aquifer depth, aq-
uifer thickness, and distance from sea shore line. The
variables are described in Table 1.
Time Distribution Phases of ANN Model Data
The model data were extracted mainly from domestic
wells in Gaza Strip because they usually have records of
chloride concentration twice a year in February and Oc-
tober periodically. The time distribution divides the year
in two phases A an d B. The phase A starts from April to
September and the phase B starts from October to March
in next year. For example, time phase 1997-A extends
from April 1997 to September 1997, time phase 1997-B
extends from October 1997 to March 1998 and time
phase 1998-A extends from April 1998 to September
1998, etc. So all other factors were organized according
Copyright © 2011 SciRes. JEP
Application of Artificial Neural Networks Model as Analytical Tool for Groundwater Salinity
Copyright © 2011 SciRes. JEP
60
in the space. As the current data are collected from lim-
ited sources (56 municipal wells), they may constitute
clusters. Therefore, the distribution of each variables
across its range in the database is examined. The mean,
standard deviation and ranges of different variables used
to train the ANN is shown in Table 1. The frequency
distributions of different variables studied across the
range of the 499 cases are represented graphically as
histograms with normal distribution curve in Figure 8.
2.3.3. Bui l di n g AN N M odel
The procedural steps in building and applying for ANN
model varies accord ing to the tool us ed in building ANN
models. Using STATISTICA Neural Networks (SNN),
the procedural steps involves the following procedures:
Data Importation
It include feeding the data matrix for SNN to train the
Network by “importing” or through the data entry proc-
ess. The data must be in acceptable format such as
spreadsheet. The input data is the cases that the network
uses to train itself.
Problem Definition
Problem Definition was achieved by specifying the
inputs (Independent) and the output (Dependent) variable
for the ANN model. Initially, there are nine inputs vari-
ables and one output variables as mentioned in Table 1.
Figure 7. Study wells location in Gaza Strip.
to this time distribution. Extraction of t he Test Set
Organizing of ANN Model Data In SNN, The test set extraction is about 50% of cases
for training. 25% for calibration and 25% for testing and
it is randomly selected and the user can change this per-
centage. Test set provide a means by which the network
knows when to stop training and used for calibration and
Testing.
The organizing of ANN model data are required to
construct some hundreds of data cases of input and out-
put variables. These cases construct data matrix. Data
organizing was carried out using software Ms. Excel and
Access software. The data matrix is considered as row
material to ANN model. Network Design
2.3.2. Analysis of ANN Model Data Determining the appropriate architecture of network
among the available networks based on the type of the
data and the problem. After many trials, Multilayer Per-
ceptron network (MLP) has been chosen because of its
high capabilities to generalize well in problems plagued
with significant heterogeneity and nonlinearity.
Considering only those cases that have complete numeric
values for all variables without any missing data, only
499 cases satisfy the above-mentioned criteria from 1997
to 2004. ANN model might perform well over an entire
space only when the training data are evenly distributed
Table 1. Mean, standard deviation and ranges of variables used to train the ANN model.
Range
Variable Sym. Unit Mean Std. Dev
Min. Max.
Initial chloride concentration Clo mg/l 333.07 253.94 28.00 1412.0
Recharge rate R mm/m2/month 18.19 24.44 0.00 83.07
Abstraction Q m3/hour 105.55 57.99 0.00 254.94
Abstraction average rate Qr mm/m2/month 22.50 5.80 11.37 33.94
Life time Lt y 22.02 13.94 0.00 60.00
Aquifer thickness Th m 64.17 27.25 30.00 124.00
Application of Artificial Neural Networks Model as Analytical Tool for Groundwater Salinity61
(a) (b)
(c) (d)
(e) (f)
Figure 8. Frequency distribution of the variables across the range of 499 cases. (a) Frequency distribution of Clo; (b) Fre-
quency distribution of R; (c) Fre quency distribution of Q; (d) Frequency distribution of Qr; (e) Frequency distri bution of Lt;
h) Frequency distribution of Th. (
Copyright © 2011 SciRes. JEP
Application of Artificial Neural Networks Model as Analytical Tool for Groundwater Salinity
Copyright © 2011 SciRes. JEP
62
Network Training
Once the type of network has been chosen, the condi-
tions to stop training processes were set before the net-
work is trained. Training was controlled by some of con-
ditions as: the maximum number of iterations, target
performance which specifies the tolerance between the
neural network prediction and actual output, the maxi-
mum run time and the minimum allowed gradient and .
Network Cali bration
A trained network was continuously trained in order to
make a model perform best on the training set. However,
after some time, it is very possible for the network to
“memorize” the training set instead of learning it. In or-
der to prevent the possibility of memorization to occur,
calibration is utilized. Calibration is a parameter, which
indicates that the network has trained enough thus stop-
ping the iteration pro cess.
Testing of Network
After the network has been successfully trained well, it
is then tested against a set of cases withheld from it dur-
ing its training session. The ANN is then ready to be ap-
plied to any other values of variables. The results are
then presented in statistical manner. Regression analysis
is utilized to measure the degree of correlation between
the actual output and the network output. Correlation
factor (r) of 1 gives an indication of a perfect model
while an (r) of 0 indicates a very bad model. Mathemati-
cally the values of (r) represented in Equation (1).


2
21
2
1
1
n
ii
in
i
i
actual predicted
R
actual mean

(1)
3. Results and Discussion
3.1. Characteristic of ANN Model
3.1.1. Topology of ANN
After a number of training trials, the best neural network
was determined to be Multilayer Perceptron network
(MLP) with four layers: an input layer of 6 neurons, first
hidden layer with 10 neurons, second hidden layer with 7
neurons and the output layer with 1 neuron as shown in
Figure 9. The six input neurons are: in itial chloride con-
centration (Clo), recharge rate (R), abstraction (Q), ab-
straction average rate of area (Qr), life time (Lt), aquifer
thickness (Th). The output neuron gives the final chloride
concentration (Clf).
3.1.2. Performance of ANN
The progress of the training was checked by plotting the
training, and test mean square errors versus the per-
formed number of iterations, as presented in Figure 10.
Figure 11 presented a comparison of simulated chlo-
ride concentration using ANN and the observed chloride
concentration. The Figure 11 showed a very high corre-
lation between the observed and predicted values of
chloride concentration. The correlation coefficient (r)
between the predicted and observed output values of the
ANN model is 0.9848. The high value of correlation co-
efficient (r) showed that the simulated chloride concen-
tration values using the ANN model were in very good
agreement with the observed chloride concentration
which gave initial impression that ANN model are useful
and applicable. Simulated chloride concentration using
ANN model and observed chlorideconcentration on 1/10/
2000 are presented in Figure 12 .
3.1.3. Regression Statistics of ANN Model
In regression problems, the purpose of the neural net-
work is to learn a mapping from the input variables to a
Figure 9. Topology of final ANN model.
Figure 10. Training progress of ANN.
Application of Artificial Neural Networks Model as Analytical Tool for Groundwater Salinity63
Figure 11. Comparison of simulated chloride concentration
using ANN model and the observed chloride concentration.
Figure 12. Comparison of simulated chloride concentration
using ANN and the observed chloride concentration on
1/10/2000.
continuous output variable. A network is successful at
regression if it makes predictions with accepted accuracy.
SNN automatically calculates correlation coefficient (r)
between the actual and predicted outputs. A perfect pre-
diction will have a correlation coefficient of 1.0. A cor-
relation of 1.0 does not necessarily indicate a perfect
prediction (only a prediction which is perfectly linearly
correlated with the actual outputs), although in practice
the correlation coefficient is a good indicator of per-
formance. It also provides a simple and familiar way to
compare the performance of neural networks with stan-
dard least squares linear fitting procedu res. The degree of
predictive accuracy needed varies from application to
application.
Regression statistics are listed as follows:
Data Mean: Average value of the target output
variable.
Data S.D.: Standard deviation of the target output
variable.
Error Mean: Average error (residual between tar-
get and actua l output va lues) of the output variable.
Abs. E. Mean: Average absolute error (difference
between target and actual output values) of the
output variable.
Error S.D.: Standard deviation of errors for the
output variable.
S.D. Ratio: The error/data standard deviation ratio.
Correlation: The correlation coefficient (r) be-
tween the predicted and observed output values.
The values of regression statistics for the ANN model
are presented in Table 2.
3.2. Response Graph
Response graph shows the effect on the output variable
prediction of ad ju sting inp ut (ind ep end ent) variab les. The
ANN model was utilized to study the influence of the
input variables on output variable which is chloride con-
centration. Figure 13 presen ted a response graph of each
input variables of final ANN model.
Figures 13(a,c-e) indicated that chloride concentration
increases nonlinearly as chloride concentration initial,
abstraction, abstraction average rate and life time in-
crease. Figures13(b,f) indicated that chloride concentra-
tion decreases nonlinearly as recharge rate and aquifer
thickness increase.
3.3. Response Surface
A response surface is a figure shows the effect on the
output variable prediction of adjusting two input (inde-
pendent) variables. Th e ANN model was utilized to study
Table 2. The values of regression statistics for final ANN
model.
Regression sta-
tistics All model
data Training
data set Validation
data set Test data
set
Data Mean 341.105 295.877 345.200 361.427
Data S.D. 260.827 247.433 262.657 263.607
Error Mean 3.242 5.016 8.428 -0.196
Error S.D. 45.371 45.125 47.312 44.204
Abs E. Mean29.798 29.262 32.128 28.911
S.D. Ratio 0.174 0.182 0.180 0.168
Correlation (r)0.9848 0.9832 0.9837 0.9860
Notes: Low value of Error Mean, Abs E. Mean and S.D. Ratio
showed that the error between observed and simulated chloride concen-
tration values using the ANN model are small; High value of correla-
tion coefficient (r) showed that the simulated chloride concentration
values using the ANN model are in good agreement with the observed
hloride concentration. c
Copyright © 2011 SciRes. JEP
Application of Artificial Neural Networks Model as Analytical Tool for Groundwater Salinity
Copyright © 2011 SciRes. JEP
64
(a) (b)
(c) (d)
(e) (f)
Figure 13. Response graph of each input variables of ANN model. (a) Response graph of Qr; (b) Response graph of R; (c)
esponse graph of Q; (d) Response graph of Clo; (e) Response graph of Lt; (f) Response graph of Th. R
Application of Artificial Neural Networks Model as Analytical Tool for Groundwater Salinity65
the influence of each two input variables on chloride
concentration. Figures 14 presented response surface of
each two input variables of final ANN model.
Figures 14(a) indicated that the chloride concentration
increases nonlinearly as recharge decreases and abstrac-
tion increases and the effect of recharge is stronger than
effect of abstraction. Figure 14(b) indicated that the
chloride concentration increases nonlinearly as recharge
decreases and abstraction average rate increases and the
effect of recharge is similar to effect of abstraction aver-
age rate. Figure 14(c) indicated that the chloride concen-
tration increases nonlinearly as life time increases and
recharge decreases and the effect of recharge is stronger
than effect of life time. Figures 14(d) indicated that the
chloride concentration increases nonlinearly as recharge
decrease and aquifer thickness and the effect of aquifer
thickness is stronger than effect of recharge.
Figure 14(e) indicated that the chloride concentration
increases nonlinearly as abstraction and abstraction av-
erage rate increase and the effect of abstraction average
rate is similar to the effect of abstraction. Figure 14(f)
indicated that the chloride concentration increases
nonlinearly as abstraction and life time increase and the
effect of life time is similar to effect of abstraction. Fig-
ure 14(g) indicated that the chloride concentration in-
creases nonlinearly as abstraction increases and aquifer
thickness decrease. In add ition, it was noted that effect of
aquifer thickness is stronger than effect of abstraction.
Figure 14(h) indicated that the chloride concentration
increases nonlinearly as abstraction average rate and life
time increase and the effect of abstr action average r ate is
stronger than effect of life time. Figure 14(i) indicated
that the chloride concentration increases nonlinearly as
abstraction average rate increases and aquifer thickness
decreases and the effect of aquifer thickness is similar to
effect of abstraction average rate.
Figure 14(j) indicated that the chloride concentration
increases nonlinearly as life time increases and aquifer
thickness decreases. In addition, it was noted that effect
of aquifer thickness is similar to effect of life time.
(a) (b)
(c) (d)
Copyright © 2011 SciRes. JEP
Application of Artificial Neural Networks Model as Analytical Tool for Groundwater Salinity
66
(e) (f)
(g) (h)
(i) (j)
Figures 14. Response surface of each two input variables of ANN model. (a) Response surface of R & Q; (b) Response surface
of R & Qr; (c) Response surface of R & Lt; (d) Response surface of R & Th; (e) Response surface of Q & Qr; (f) Response
surface of Q & Lt; (g) Response surface of Q & Th; (h) Response surface of Qr & Lt; (i) Response surface of Th & Qr; (j)
Response surfac e of Th & Lt.
Copyright © 2011 SciRes. JEP
Application of Artificial Neural Networks Model as Analytical Tool for Groundwater Salinity
Copyright © 2011 SciRes. JEP
67
3.4. Utilizing ANN Model as Analytical Tool
The ANN model was utilized to study the influence of
the input variables on chloride concentration. Hypotheti-
cal cases of input variables were assumed to study the
influence of the input variables. Three level of confi-
dence were assumed: the first one was consolidating the
values of input variables on the mean value and changing
the value of studied variable gradually from minimum
value to maximum value in the range of input variable.
The second level of confidence was consolidating the
values of abstraction, abstraction average rate and life
time on the mean plus the value of standard deviation. In
addition it was consolidating the values of recharge rate
and aquifer thickness on the mean subtract the value of
standard deviation which produce conditions lead to in-
crease chloride concentration in groundwater.
The third level of confidence was consolidating the
values of abstraction, abstraction average rate and life
time on the mean subtract the value of standard deviation
and consolidating the values of recharge rate and aquifer
thickness on the mean plus the value of standard devia-
tion which produce conditions lead to decrease chloride
concentration in groundwater.
To obtain the values of gradual changing for input
variable from minimum value to maximum value in the
range of input variable, the range wa s divided to ten steps
and the value gradually was increased from minimum
value to maximum value in the range. Tabl e 3 presented
the hypothetical values of gradual change of input vari-
ables. Hypothetical v alues of inpu t variables fo r the three
analysis conditions were computed as explained above
and they were presented in Table 3.
3.4.1. I n fluence of Re charge Rat e o n Ch loride
Concentration
By application the above mentioned procedure and using
the final ANN model to calculate the valu e of final chlo-
ride concentration for each hypothetical case, the effect
of recharge rate on chloride concentration was studied.
Results of the three conditions (normal, increasing and
decreasing) were presented in Figure 15 and Table 5.
It was noted that increasing recharge rate from 0 to 80
mm/m2/month resulted in a large influence in final chlo-
ride concentration as follows:
In normal condition, when the initial chloride
concentration = 330 mg/l, abstraction =105 m3/hr,
abstraction average rate = 22 mm/m2/month, life
time = 22 years and aquifer thickness = 65 m. Final
chloride concentration decrease from 346.00 mg/l
to 290.91 mg/l. Final ch loride concentration stayed
stable of 330 mg/l on recharge rate of 28 mm/
month.
In increasing condition, when the initial chloride
concentration = 330 mg/l, abstraction = 146 m3/hr,
abstraction average rate = 29 mm/m2/month, life
time = 36 years and aquifer thickness = 40 m. Final
chloride concentration decreased from 363.06 mg/l
to 309.93 mg/l. Final ch loride concentration stayed
stable of 330 mg/l on recharge rate of 52 mm/
m2/month.
In decreasing condition, when initial chloride
concentration = 330 mg/l, abstraction = 47 m3/hr,
abstraction average rate = 16 mm/m2/month, Life
Table 3. Hypothetical values of gr adual change for input variables.
Clo R Q Qr Lt Th
Unit mg/l mm/m2/month m3/hour mm/m2/month y m
Min. 28.00 0.00 0.00 11.37 0.00 30.00
Max. 1412.00 83.07 254.94 33.94 60.00 124.00
1 330.00 0.00 0.00 12.00 0.00 30.00
2 330.00 8.00 25.00 14.30 6.00 39.00
3 330.00 16.00 50.00 16.60 12.00 48.00
4 330.00 24.00 75.00 18.90 18.00 57.00
5 330.00 32.00 100.00 21.20 24.00 66.00
6 330.00 40.00 125.00 23.50 30.00 75.00
7 330.00 48.00 150.00 25.80 36.00 84.00
8 330.00 56.00 175.00 28.10 42.00 93.00
9 330.00 64.00 200.00 30.40 48.00 102.00
10 330.00 72.00 225.00 32.70 54.00 111.00
11 330.00 80.00 250.00 35.00 60.00 120.00
Application of Artificial Neural Networks Model as Analytical Tool for Groundwater Salinity
68
Table 4. Hypothetical values of input var iables for the thr ee analysis conditions.
Clo R Q Qr Lt Th
Min. 28.00 0.00 0.00 11.37 0.00 30.00
Max. 1412.00 83.07 254.94 33.94 60.00 124.00
Mean 333.07 18.19 105.55 22.50 22.02 64.17
S.D 253.94 24.44 57.99 5.80 13.94 27.25
M+S.D 587.01 42.64 163.54 28.30 35.95 91.41
M-S.D 79.13 -6.25 47.56 16.70 8.08 36.92
Normal Condition 330.00 18.00 105.00 22.00 22.00 65.00
Decreasing Condition 330.00 0.00 164.00 29.00 36.00 40.00
Increasing Condition 330.00 43.00 47.00 16.00 8.00 91.00
time = 8 years and aquifer thickness = 91 m. Final
chloride concentration decreased from 316.89 mg/l
to 268.75 mg/l. In this condition final chloride
concentration stayed less than 330 mg/l for all val-
ues of recharge rate even if small values of re-
charge because of very good condition of small
value of abstraction, abstraction average rate life
time and large aquifer thickness.
It is noted that stabilization point of chloride con-
centration for normal condition occurred at re-
charge rate = 22 mm/m2/month and for increasing
condition occurs at recharge rate = 52 mm/m2/
month which mean that increasing condition re-
quired height recharge rate to achieve stabilization
point of chloride concentration. In decreasing con-
dition final chloride concentration stayed less than
330 mg/l with values 316.89 mg/l to 268.75 mg/l
for all values of recharge rate even if small values
of recharge rate were available.
3.4.2. Influence of Abstraction on Chloride
Concentration
By application the above mentioned procedure and using
the final ANN model to calculate the valu e of final chlo-
ride concentration for each hypothetical case, the effect
of abstraction on chloride concentration was studied.
Results of the three conditions (normal, increasing and
decreasing) were presented in Figure 16. It was noted
that increasing abstraction from 0 to 250 m3/hr results in
a small influence in final chloride concen tration.
3.4.3. Impacts of Abstraction Average Rate on
Chloride Concentration
By application the above mentioned procedure and using
the final ANN model to calculate the valu e of final chlo-
ride concentration for each hypothetical case, the effect
of abstraction average rate on chloride concentration was
Figure 15. Impact of recharge rate on chloride concentra-
tion.
Figure 16. Effect of abstraction on chloride concentration.
Copyright © 2011 SciRes. JEP
Application of Artificial Neural Networks Model as Analytical Tool for Groundwater Salinity69
Table 5. Results of ANN model for hypothetical cases stud-
ied the effect of recharge rate on chloride concentration.
Normal Condition
Clo R Q Qr Lt Th Clf
1 330.00 0.00 105.00 22.00 22.00 65.00346.00
2 330.00 8.00 105.00 22.00 22.00 65.00342.11
3 330.00
16.00 105.00 22.00 22.00 65.00337.45
4 330.00
24.00 105.00 22.00 22.00 65.00332.21
5 330.00
32.00 105.00 22.00 22.00 65.00326.58
6 330.00
40.00 105.00 22.00 22.00 65.00320.70
7 330.00
48.00 105.00 22.00 22.00 65.00314.71
8 330.00
56.00 105.00 22.00 22.00 65.00308.67
9 330.00
64.00 105.00 22.00 22.00 65.00302.67
10 330.00 72.00 105.00 22.00 22.00 65.00296.75
11 330.00 80.00 105.00 22.00 22.00 65.00290.91
Increasing Salinity Condition
Clo R Q Qr Lt Th Clf
1 330.00 0.00 164.00 29.00 36.00 40.00363.06
2 330.00 8.00 164.00 29.00 36.00 40.00358.88
3 330.00
16.00 164.00 29.00 36.00 40.00354.22
4 330.00
24.00 164.00 29.00 36.00 40.00349.20
5 330.00
32.00 164.00 29.00 36.00 40.00343.91
6 330.00
40.00 164.00 29.00 36.00 40.00338.45
7 330.00
48.00 164.00 29.00 36.00 40.00332.86
8 330.00
56.00 164.00 29.00 36.00 40.00327.19
9 330.00
64.00 164.00 29.00 36.00 40.00321.47
10 330.00 72.00 164.00 29.00 36.00 40.00315.71
11 330.00 80.00 164.00 29.00 36.00 40.00309.93
Decreasing Salinity Condition
Clo R Q Qr Lt Th Clf
1 330.00 0.00 47.00 16.00 8.00 91.00316.89
2 330.00 8.00 47.00 16.00 8.00 91.00315.78
3 330.00
16.00 47.00 16.00 8.00 91.00 313.21
4 330.00
24.00 47.00 16.00 8.00 91.00 309.44
5 330.00
32.00 47.00 16.00 8.00 91.00 304.73
6 330.00
40.00 47.00 16.00 8.00 91.00 299.33
7 330.00
48.00 47.00 16.00 8.00 91.00 293.47
8 330.00
56.00 47.00 16.00 8.00 91.00 287.35
9 330.00
64.00 47.00 16.00 8.00 91.00 281.12
10 330.00 72.00 47.00 16.00 8.00 91.00 274.89
11 330.00 80.00 47.00 16.00 8.00 91.00 268.75
studied. Results of the three conditions (normal, increas-
ing and decreasing) were presented in Figure 17. It was
noted that increasing abstraction average rate from 12 to
35 mm/m2/month results in a large influence in final
chloride concentration.
3.4.4. Influence of Life Time on Chloride
Concentration
By application the above mentioned procedure and using
the final ANN model to calculate the valu e of final chlo-
ride concentration for each hypothetical case, the effect
of life time on chloride concentration was studied. Re-
sults of the three conditions (normal, increasing and de-
creasing) were presented in Figure 18. It was noted that
increasing Life time from 0 to 60 year results in a large
influence in final chloride concentration.
3.4.5. Influence of Aquifer Thickness on Chloride
Concentration
By application the above mentioned procedure and using
the final ANN model to calculate the valu e of final chlo-
ride concentration for each hypothetical case, the effect
of aquifer thickness on chloride concentration was stud-
ied. Results of the three conditions (normal, increasing
and decreasing) were presented in Figure 19. It was
noted that increasing aquifer thickness from 30 to 120
mm/m2/month results in a large influence in final chlo-
ride concentration.
4. Conclusions
The following conclusions were made based on the re-
sults obtained from the current study:
1) A new approach for Groundwater salinity modelling
in Gaza Strip utilizing ANN was successfully developed
and applied. ANN model was developed to study the
relation between groundwater salinity (represented by
Figure 17. Effect of abstraction average rate on chloride
concentration.
Copyright © 2011 SciRes. JEP
Application of Artificial Neural Networks Model as Analytical Tool for Groundwater Salinity
70
Figure 18. Effect of life time on chloride concentration.
Figure 19. Effect of aquifer thickness on chloride concen-
tration.
chloride concentration in groundwater) and some related
hydrological factors such as recharge rate (R), abstrac-
tion (Q), abstraction average rate (Qr), life time (Lt) and
aquifer thickness (Th).
2) The best neural network was Multilayer Perceptron
network (MLP) with four layers: an input layer of 6 neu-
rons, first hidden layer with 10 neurons, second hidden
layer with 7 neurons and the output layer with 1 neuron
which gives the final chloride concentration (Clf).
3) The new approach generated very good results de-
pending high correlation between the observed and pre-
dicted values of chloride concentration. The correlation
coefficient (r) between the observed and predicted the
output values of the ANN model was 0.9848. The high
value of correlation coefficient (r) showed that the simu-
lated chloride concentration values using the ANN model
were in very good agreement with the observed chloride
concentration which mean that ANN model are useful
and applicable.
4) The ANN model proved that chloride concentration
in groundwater is reduced by decreasing abstraction (Q),
abstraction average rate (Qr) and life time (Lt). Further-
more, it is reduced by increasing recharge rate (R) and
aquifer thickness (Th).
5) Therefore, the current research showed that ANN
model can be used in groundwater quality management
and it is comparable to other used approaches such as
groundwater modelling and statistical modelling.
6) It showed that the strong remedial actions for solv-
ing the groundwater deterioration problem in the aquifer
of Gaza Strip (salinity) are reducing the abstraction rate
and increasing the recharge quantities to the aquifer.
REFERENCES
[1] Metcalf & Eddy, “Costal Aquifer Management Program,”
Final Report: Modeling of Gaza Strip Aquifer, the pro-
gram is funded by US Agency for International Devel-
opment (USAID) and owned by the Palestinian Water
Authority (PWA), Gaza, Palestine, 2000.
[2] Coastal Aquifer Management Plan (CAMP), Gaza
Coastal Aquifer Management Program, Vol. 1, Task 3,
2000.
[3] R. S. Govindaraju, “Artificial Neural Network in Hy-
drology,” Journal of Hydrologic Engineering, Vol. 5, No.
2, 2000, pp. 124-137.
doi:10.1061/(ASCE)1084-0699(2000)5:2(124)
[4] United Nations Environment Programm (UNEP), “Desk
Study on the Environment in the Occupied Palestinian
Territories,” Switzerland, 2003.
[5] United Nations Environment Programm (UNEP), “Envi-
ronmental Assessment of the Gaza Strip Following the
Escalation of Hostilities in December 2008-January
2009,” Gaza, 2009.
[6] Palestinian Water Authority (PWA), (2003). “Groundwa-
ter Levels Decline Phenomena in Gaza Strip Final Re-
port,” Water Resources and Planning Department-Hy-
drology Section, Palestinian National Authority, Gaza,
February 2003.
[7] Coastal Municipality Water Utility (CMWU), “Annual
Report of Wastewater Quality in Gaza Strip for Years
2007 and 2008,” Gaza, Palestine, 2007.
[8] S. K. Jain, V. P. Singh and M. T. van Genuchten,
“Analysis of Soil Water Retention Data Using Artificial
Neural Networks,” Journal Hydrologic Engineering, Vol.
9, No. 5, 2004, pp. 415-420.
doi:10.1061/(ASCE)1084-0699(2004)9:5(415)
[9] S. Lee, S. Cho and M. Wong, “Rainfall Prediction Using
Artificial Neural Networks,” Journal of Geographic In-
formation and Decision Analysis, Vol. 2, No. 2, 1998, pp.
233-242.
[10] A. Ajith, “Artificial Neural Networks Oklahoma State,”
Handbook of Measuring System Design, John Wiley &
Sons, Ltd, University, Stillwater, USA, 2005.
[11] J. Hola and K. Schabowicz, “Application of Artificial
Copyright © 2011 SciRes. JEP
Application of Artificial Neural Networks Model as Analytical Tool for Groundwater Salinity
Copyright © 2011 SciRes. JEP
71
Neural Networks to Determine Concrete Compressive
Strength Based on Non-Destructive Test,” Journal of civil
engineering and management, Vol. 11, No. 1, 2005, pp.
23-32.
[12] D. Jeng and H. Cha, “Application of Neural Network in
Civil Engineering,” School of Engineering, Griffith Uni-
versity Gold Coast Campus, Australia, 2003.
[13] K. P. Sudheer, A. K. Gosain, R. D. Mohana and S. M.
Saheb, “Modeling Evaporation Using an Artificial Neural
Network Algorithm,” Hydrological Processes, Vol. 16,
2002, pp. 3189-3202. doi:10.1002/hyp.1096
[14] M. Claudius and K. Olaf, “Progress Report on Trend
Analysis Methods, Tools and Data Preparation (in Par-
ticular Artificial Neural Networks (ANN)),” Center for
Applied Geoscience, University of Tubingen, Tübingen,
2005.
[15] H. Jiang and W. R. Cotton, “Soil Moisture Estimation
Using an Artificial Neural Network: A Feasibility Study,”
Canadian Journal of Remote Sensing, Vol. 30, No. 5,
2004, pp. 827-839.
[16] Palestinian Water Authority (PWA), “National Water
Plan,” Final Report, The study is funded by United Na-
tion Development Program (UNDP) and owned by the
Palestinian Water Authority (PWA), Gaza, Palestine,
2000.