Environ Earth Sci (2014) 71:3119–3131
DOI 10.1007/s12665-013-2690-7
ORIGINAL ARTICLE
Estimating groundwater vulnerability to pollution using
a modified DRASTIC model in the Kerman agricultural area, Iran
Aminreza Neshat • Biswajeet Pradhan •
Saied Pirasteh • Helmi Zulhaidi Mohd Shafri
Received: 14 February 2013 / Accepted: 20 July 2013 / Published online: 2 August 2013
Ó Springer-Verlag Berlin Heidelberg 2013
Abstract Groundwater contamination from intensive
fertilizer application affects conservation areas in a plain.
The DRASTIC model can be applied in the evaluation of
groundwater vulnerability to such pollution. The main
purpose of using the DRASTIC model is to map ground-
water susceptibility to pollution in different areas. How-
ever, this method has been used in various areas without
modification, thereby disregarding the effects of pollution
types and their characteristics. Thus, this technique must be
standardized and be approved for applications in aquifers
and particular types of pollution. In this study, the potential
for the more accurate assessment of vulnerability to pol-
lution is achieved by correcting the rates of the DRASTIC
parameters. The new rates were calculated by identifying
the relationships among the parameters with respect to the
nitrate concentration in groundwater. The methodology
was implemented in the Kerman plain in the southeastern
region of Iran. The nitrate concentration in water from
underground wells was tested and analyzed in 27 different
locations. The measured nitrate concentrations were used
to associate and correlate the pollution in the aquifer to the
DRASTIC index. The Wilcoxon rank-sum nonparametric
statistical test was applied to determine the relationship
between the index and the measured pollution in Kerman
plain. Also, the weights of the DRASTIC parameters were
A. Neshat  B. Pradhan (&)  H. Z. M. Shafri
Department of Civil Engineering, Faculty of Engineering,
University Putra Malaysia, 43400 Serdang, Selangor, Malaysia
e-mail: biswajeet@mailcity.com; biswajeet24@gmail.com;
biswajeet@lycos.com
S. Pirasteh
Department of Geography and Environmental Management,
Faculty of Environment, University of Waterloo, Waterloo,
Canada
modified through the sensitivity analysis. Subsequently, the
rates and weights were computed. The results of the study
revealed that the modified DRASTIC model performs more
efficiently than the traditional method for nonpoint source
pollution, particularly in agricultural areas. The regression
coefficients showed that the relationship between the vul-
nerability index and the nitrate concentration was 82 %
after modification and 44 % before modification. This
comparison indicated that the results of the modified
DRASTIC of this region are better than those of the ori-
ginal method.
Keywords Modified DRASTIC  GIS 
Groundwater  Hydrogeology  Sensitivity analysis 
Kerman plain  Vulnerability
Introduction
Groundwater is an important and prominent resource in
most countries, particularly for those in arid and semi-arid
areas. Water quality has been given more emphasis in
groundwater management (Pradhan 2009; Ayazi et al.
2010; Manap et al. 2012, 2013). Aquifers are usually
unconfined and highly permeable, thereby causing their
high susceptibility to surface contamination (Javadi et al.
2011a, b). The potential groundwater pollution by human
activities at or near the surface has been considered the
primary basis for the management of this major resource by
implementing preventive policies.
The introduction of potential contaminants to a location
on top of an aquifer at a specified position in an under-
ground system is defined as groundwater vulnerability
(National Research Council 1993). Vulnerability assess-
ment must be based on scientific, accurate, and objective
123
3120
evidence. Various methods have been introduced to esti-
mate groundwater vulnerability with high accuracy (Javadi
et al. 2011a, b). In most cases, these procedures consist of
analytical tools intended to correlate groundwater con-
tamination with land activities. There are three categories
of assessment processes and procedures: the process-based
simulation models, the statistical methods (Harbaugh et al.
2000), and the overlay and index methods (Dixon 2004).
Process-based models generally require a large amount
of primary and secondary data to apply the mathematical
models for creating the principal tool. Such methods seem
more complex and difficult to use on a regional scale.
Statistical methods use data on the known areal contami-
nant distribution and describe the contamination potential
for a specified geographical region using the available data
in the regions of interest (National Research Council 1993).
Overlay and index methods emphasize the combination
of different regional maps by allocating a numerical index.
Both methods are easy to apply in geographic information
systems (GIS), particularly on a regional scale. Therefore,
these techniques are the most popular methods used in
vulnerability evaluation. The most widely used among
these techniques include GOD (Foster 1987), IRISH (Daly
and Drew 1999), AVI (van Stemproot et al. 1993), and
DRASTIC (Aller et al. 1987). DRASTIC is widely applied
in various countries, including the USA (Plymale and
Angle 2002; Fritch et al. 2000; Shukla et al. 2000), China
(Yuan et al. 2006; Huan et al. 2012; Yin et al. 2012),
Jordan (Naqa et al. 2006), Morocco (Ettazarini 2006), Iran
(Javadi et al. 2011a, b), Palestine (Mimi et al. 2012),
Tunisia (Saidi et al. 2010, 2011), and Portugal (Pacheco
and Sanches Fernandes 2012).
Despite its popularity, the DRASTIC method has some
disadvantages. The DRASTIC model consists of seven
hydrogeologic factors: depth of water, net recharge, aquifer
media, soil media, topography (slope), impact of vadose
zone, and hydraulic conductivity of the aquifer.
This approach primarily uses the seven parameters to
compute for the vulnerability index. Each parameter is
allocated specific weights and rating values, as shown in
Table 1 (Aller et al. 1987). The DRASTIC model is con-
sidered one of the best index models for vulnerability map-
ping since it was first introduced by Aller et al. (1985, 1987).
This technique disregards the effects of regional character-
istics. Therefore, uniform weights and rating values are used.
Moreover, this method does not use a standard validation test
for the aquifer. Thus, several researchers have continued to
develop this model using different methods. Some of the
notable studies include those by Secunda et al. (1998),
Melloul and Collin (1998), Zhou et al. (1999), Thirumalai-
vasan et al. (2003), Dixon (2005), Antonakos and Lambrakis
(2007), Bojórquez-Tapia et al. (2009), Ckakraborty et al.
(2007), Denny et al. (2007), Hamza et al. (2007), Leone et al.
123
Environ Earth Sci (2014) 71:3119–3131
(2009), Nobre et al. (2007), Pathak and Hiratsuka (2011),
Remesan and Panda (2008), Saidi et al. (2011), and Hailin
et al. (2011). Several groups attempted to correlate the vul-
nerability index using chemical or contaminant parameters
(Kalinski et al. 1994; Rupert 1999; McLay et al. 2001).
Certain index methods modified the DRASTIC model by
varying the factors and respective weights or by incorpo-
rating alternative data on human activities, such as land use
and contaminant loading. Al-Hanbali and Kondoh (2008)
combined a human activity impact index derived from land
use or cover data using the DRASTIC model. Their work
demonstrated that human activities affect the groundwater
quality and increase the risk of pollution in the Dead Sea
groundwater basin in Jordan.
The existing literature reports a limited number of studies
on groundwater vulnerability from specific contamination
sources based on GIS and index methods. This lack of
information may be due to the previous definition of
groundwater vulnerability for nonpoint sources. However,
the existing index methods may still be useful in evaluating
the risk of groundwater pollution in local and larger scales
when it is combined with site-specific information, such as
the hydrogeology and contamination history (Alexander
et al. 1986). Recent studies have attempted to correlate the
vulnerability index with chemical or contaminant parame-
ters (Kalinski et al. 1994; Rupert 1999; McLay et al. 2001),
whereas others correlate the nearby land to its vulnerability.
However, the rates or weights of the DRASTIC model could
not be used efficiently in these investigations.
Nitrate is not naturally found in surface groundwater. This
compound is considered a good indicator of contaminant
movement from the surface to groundwater, particularly in
land allocated for agricultural use (Javadi et al. 2011a, b).
Carvalho (2009) conducted pollution risk assessment by
integrating DRASTIC results with nitrate concentrations.
Subsequently, other researchers used nitrate to modify
DRASTIC (Panagopoulos et al. 2006; Javadi et al. 2011a, b).
Pacheco and Sanches Fernandes (2012) used nitrates to
perform a correspondence analysis. Panagopoulos et al.
(2006) and Javadi et al. (2011a, b) calibrated the model using
nitrate before a correlation coefficient was obtained to
describe the relationship of the vulnerability index and
nitrate concentration. In the current study, the rate of
DRASTIC parameters was calibrated for the study area by
measuring the nitrate concentration of groundwater. The
relationship between the vulnerability indicators and the
parameters was statistically analyzed to calibrate the rates
using the Wilcoxon rank-sum nonparametric statistical test
(Wilcoxon 1945). In addition, the single-parameter sensi-
tivity analysis was applied to compute the effective weight of
each parameter in the Kerman plain.
The main contribution of this research compared with
previously published literature in Javadi et al. (2011a, b) is
Environ Earth Sci (2014) 71:3119–3131 3121

Table 1 Original DRASTIC weights and rating systems

Depth to water (m) Recharge (mm) Topography Conductivity (m/ Aquifer media Vadose zone material Soil media
(slope %) day)

Range Rating Range Rating Range Rating Range Rating Range Rating Range Rating Range Rating

(0–1.5) 10 (0–50.8) 1 (0–2) 10 (0.04–4.1) 1 Massive shale 2 Confining 1 Thin or 10

layer absent
(1.5–4.6) 9 (50.8–101.6) 3 (2–6) 9 (4.1–12.3) 2 Metamorphic/ 3 Silt/clay 3 Gravel 10
igneous
(4.6–9.1) 7 (101.6–177.8) 6 (6–12) 5 (12.3–28.7) 4 Weathered 4 Shale 3 Sand 9
(9.1–15.2) 5 (177.8–254) 8 (12–18) 3 (28.7–41) 6 metamorphic Limestone 3 Peat 8
igneous
(15.2–22.8) 3 ([254) 9 ([18) 1 (41–82) 8 Glacial till 5 Sandstone 6 Shrinking 7
clay
(22.8–30.4) 2 ([82) 10 Bedded 6 Bedded 6 Sandy loam 6
([30.4) 1 sandstone, limestone, Loam 5
limestone sandstone
Massive 6 Sand and 6 Silty loam 4
sandstone gravel
Massive 8 W. silt Clay loam 3
limestone
Sand and gravel 8 Sand and 8 Muck 2
gravel
Basalt 9 Basalt 9 No 1
shrinking
clay
Karsts 10 Karsts 10
limestone limestone
DRASTIC weight: 5 DRASTIC weight: 4 DRASTIC DRASTIC weight: 3 DRASTIC weight: 3 DRASTIC weight: 5 DRASTIC weight: 2
weight: 1

Source: Aller et al. (1987)

the use of two nitrate samples in the same month, which
makes more accurate correlation. Apart from that, the area
of interest in this study is the Kerman plain which is
located in the southeastern part of Iran. The Kerman plain
is located in arid and semiarid regions, with groundwater as
its only water source because of the scarcity of surface
water. Moreover, the Kerman plain experiences heavy
pumping of groundwater, which becomes a serious prob-
lem because it continuously lowers the water table. The
majority of the study area is covered with agricultural
lands, and the application of fertilizers is a common
practice. The DRASTIC model can be used to demonstrate
the application of the proposed method and to provide a
basis for its environmental management.
in the study area was 108.3 mm in 2011. The vadose zone
consists of silt and clay, sand and gravel, or gravel and sand
with silt. The aquifer medium is composed of marlstone
and shale, silt and clay, massive sandstone, or sand and
gravel. The total net recharge of the study area is 186.06
million cubic meters per year (MCM per year). The max-
imum electrical conductivity (EC) in this study area is
3,880 lmoh/cm (micromhos per centimeter), and the
average EC is 2,700 lmoh/cm. The minimum EC is
1,100 lmoh/cm. In addition, the southeastern regions of
the plain have the maximum EC. The geology of Kerman
plain contains Cretaceous and Eocene conglomerates (PC),
intrusive rocks (gp), Eocene and Neogene volcanism, and
Neogene, or younger, sediments.

Study area
The Kerman plain is an arid and semi-arid area; the
majority of the study area is composed of agricultural land.
Pistachio is the most economically important product in
this region. The Kerman plain has an area of 978 km2 in
the southeastern part of Iran (Fig. 1). The highest ground
elevation in the area is 1,980 m, with the lowest point
being 1,633 m above sea level. The average annual rainfall
Materials and methods
Data and DRASTIC method
The data used to obtain the hydrogeological parameters of
the DRASTIC model are listed in Table 2. This method
was established by the United States Environmental Pro-
tection Agency (USEPA) to classify the pollution potential

123
3122 Environ Earth Sci (2014) 71:3119–3131

Fig. 1 Study area

Table 2 Sources of data used for creation of hydro-geological (Sinkevich et al. 2005; Werz and Hötzl 2007). The index
parameter for DRASTIC method consists of seven parameters with different weighting fac-
No data type Sources tors and is calculated by

1. Hydrogelical data Meteorological Organization of X

7

Kerman
V¼ ðWi  Ri Þ; ð1Þ
i¼1
2. Geology map Geological survey of IRAN
3. Soil map Soil and water research Institute of where V is the index value and Wi is the weighted coeffi-
Kerman cient for parameter i, with an associated rating value of Ri.
4. Topography Water organizations of Kerman The following physical parameters are included in the
5. Wells Water organizations of Kerman DRASTIC method:
6. Hydraulic Water organization of Kerman
conductivity D Depth to water table from the soil surface
7. Geological profile Water organization of Kerman
R net recharge
8. Groundwater balance Water organization of Kerman
A Aquifer media
of Kerman plain S Soil media
9. Sample wells Surveyed in my study area and took two T Topography
times samples using GSP technique I Impact of the vadose zone media
C Conductivity (hydraulics) of the aquifer
of aquifers (Aller et al. 1987). Vulnerability to contami-
nation is defined as a dimensionless index function of Each of these hydrogeological factors is given a rating
hydrogeological factors, contamination sources, and from ‘‘1’’ to ‘‘10’’, and the DRASTIC parameters are
anthropogenic effects in any specific area (Plymale and weighted from ‘‘1’’ to ‘‘5’’ according to their relative
Angle 2002). Groundwater vulnerability commonly refers contribution to the potential contamination (Aller et al.
to the potential contamination from nonpoint sources or 1987). The resulting index is a relative measure of vul-
distributed point sources of pollution, such as pesticides or nerability to contamination. Areas with a higher index
nitrates from fertilizers in agricultural practices. Since the value are more vulnerable, as compared with those with a
development of the DRASTIC model (Aller et al. 1987) for lower index. The rates and weights of the original
groundwater vulnerability assessment by USEPA in the DRASTIC model parameters are presented by Aller et al.
late 1980s, this type of indexing method has become (1987). The seven layers of the DRASTIC model are pre-
popular, and widely used in the US, and worldwide sented in Fig. 2.

123
Environ Earth Sci (2014) 71:3119–3131 3123

Fig. 2 Seven layers of DRASTIC model (a depth of water, b recharge, c aquifer media, d soil, e topography (slope %), f impact of vadose zone,
g hydraulic conductivity)

123
3124
DRASTIC data layers
The assigned layers for the seven DRASTIC parameters
were constructed in raster GIS, based on Table 1.
Depth of water
The water table depths were measured from 28 observation
wells. The ArcGIS Geostatistical Analyst extension by
Krigging interpolation was applied to interpolate the points
and to develop the raster map with a pixel size of 100 m.
Krigging was previously used to obtain significant results
in groundwater level analysis (Kumar 2007; Gundogdu and
Guney 2007; Theodossiou 1999). The methodology for
groundwater vulnerability measures the water table depth
from the surface as the parameter of interest. This param-
eter represents the distance that a contaminant must travel
from the surface to reach the groundwater. A deeper water
level indicates a longer time for contamination (Aller et al.
1987). The depths to the water levels for the Kerman plain
are classified into three classes: 15–23 m, 23–30 m, and
[30 m, with depth to water rates (Dr) of 3, 2, and 1,
respectively. Given that the study area is located in arid and
semi arid regions, two classes less than 30 m were defined
to represent the area of irrigation return flow.
Net recharge (R)
The net recharge is considered the result of rainfall infil-
tration, irrigation return flow, and absorption wells in the
study area. The total net recharge was computed by the
Kerrman Water Authorities.
Aquifer media (A)
Classification of this area was based on the drilling logs for
each well. Two sections of the aquifer rock are basically
composed of marl and conglomerate rocks in the southern
regions and a small area in the northwestern region. The
extensive sand deposits with a very low percentage of fine-
grained material were identified as gravel and sand. Deposits of
fine to coarse sand (fine-medium sand) extended across the
northern and northeastern regions of the study area. Deposits of
silt and clay were located exclusively in the middle region of
the study area. According to Aller et al. (1987) and Rahman
(2008), glacial till is a mixture of gravel, sand, silt, and clay.
Soil media (S)
The soil map of the Soil and Water Institute of Kerman was
used. The soil media layer indicated the recharge rate that
could infiltrate into the pollution. The soil map consists of
clay loam, gravel, loam, non-shrinking and non-aggregated
123
Environ Earth Sci (2014) 71:3119–3131
clay, sand, sandy loam, and silty loam. Aller et al. (1987)
stated that the maximum rate belonged to gravel, sand, and
sandy loam. Based on the observed soil media layer, sand
with high permeability was located in the northern and
southern regions of the study area.
Topography (T)
Topography in the DRASTIC model displays the slope of
the land surface. The topography was derived from the
digital elevation model using a topographic map
(1:25,000). The topography of the area was divided into
five classes (Aller et al. 1987), which were mostly found in
areas with slopes ranging from 0 to 2 % and from 2 to 6 %.
Impact of the vadose zone (I)
The impact of the vadose zone was classified based on the
drilling logs for each well. The most significant part of the
area included gravel and sand with silt and clay (15–30 %)
in the western region. A small section located in the
northeastern region of the study area exclusively contains
gravel and sand.
Hydraulic conductivity (H)
The hydraulic conductivity of the aquifer was computed
according to the following equation: k ¼ Tb , where k is the
hydraulic conductivity of the aquifer (m/s), T is the trans-
missivity (m2/s), and b is the thickness of the aquifer (m).
The hydraulic conductivity distribution map was generated
using pumping test results and a geoelectrical study of the
area. Regions with maximum hydraulic conductivity
exhibited higher chances of distribution contamination.
Hydraulic conductivity was derived by measurement, and
the GIS-ArcView was applied to interpolate the hydraulic
conductivity and create the raster layer. Hydraulic con-
ductivity could be divided into three classes.
All required layers were created; each layer was classified
using the different rating scales The DRASTIC index was then
determined by multiplying the obtained values with the weight
factor. The obtained index was divided into seven groups
(Aller et al. 1987). The vulnerability indices and the corre-
sponding area percentages are presented in Figs. 3 and 4.
Nitrate measurements
The nitrate concentration was selected as the main
parameter of the initial contamination to calibrate the
DRASTIC model. A total of 27 agricultural wells were
chosen for the analysis and sampling, with two nitrate
samples obtained from each well. The first nitrate sample
Environ Earth Sci (2014) 71:3119–3131
Fig. 3 Original vulnerability
map
Fig. 4 The percentage of DRASTIC result
was obtained in May 2010 to calibrate the model, and the
second nitrate sample was obtained in May 2011 to
determine the correlation coefficient between the nitrate
concentration and groundwater vulnerability. The May
2011 samples were normalized before they were used. The
exact location of each well was determined using global
positioning system techniques.
Calibration method
The rates of DRASTIC were initially modified using the Wil-
coxon rank-sum nonparametric statistical test. Subsequently,
3125
the modified DRASTIC was applied for sensitivity analysis.
Nitrate was selected as the primary control parameter in the
study. This parameter was used to change the DRASTIC rates.
Nitrate is not naturally found in groundwater, but it usually
enters via the surface. The Kerman plain is situated in an
agricultural area where fertilizer use is common. Therefore, the
nitrate concentration can be used as an indicator of the vul-
nerability index to reflect the actual situation in the study area.
The following conditions must be met when using nitrate to
optimize the weights and rates: (1) the mean nitrate concen-
tration should be an effect of the agricultural activities on the
surface. (2) The distribution area should be relatively uniform.
(3) Leaching of nitrate occurs because of recharges from the
surface over a long period of time. To ensure the correlation
between contamination and human activities, the vulnerability
index would increase with the increasing nitrate concentration.
Finally, the correlation between nitrate concentration and vul-
nerability would illustrate the higher means vulnerability index.
Agriculture is the primary activity in the selected study area,
thereby ensuring that these basic conditions are satisfied (Pan-
agopoulos et al. 2006; Javadi et al. 2011a, b). The Wilcoxon
rank-sum nonparametric statistical test was used to modify the
rates of the DRASTIC model. Sensitivity analysis was then
applied to optimize the weights of the DRASTIC model.
Sensitivity analysis
According to Babiker et al. (2005), Saidi et al. (2011), the
applied weights for calculating the vulnerability index
123
3126 Environ Earth Sci (2014) 71:3119–3131

Fig. 5 Original vulnerability

map and nitrate (NO3)
concentration for study area

Table 3 Correlation factors between nitrate concentration and ori-

ginal vulnerability index
Pearson’s correlation Number of Factor
coefficient (%) data

100 27 Nitrate concentration

44 DRASTIC index

could differ, depending on the study area. The impact of

the weights of each parameter with their theoretical
weights was compared using the single-parameter sensi-
tivity analysis. In this study, the vulnerability index was
calibrated using rate modification. The influence of the
parameters in the index computation was evaluated using
sensitivity analysis. To investigate for any improvement in
the newly developed DRASTIC map, the nitrate distribu-
tion was compared. The effective weight of each polygon
is defined as
W ¼ ððPr  Pw Þ  V Þ  100; ð2Þ
where W is the effective weight of each parameter and V is Fig. 6 Relationship of DRASTIC intrinsic vulnerability index and
the overall vulnerability index. Pr and Pw are the rating modified DRASTIC to groundwater nitrates concentration for the
value and weight of each parameter, respectively. The Kerman plain
ArcGIS software was used to calculate all combinations of
parameters and their weights. A total of 526 unique suba- displays the lowest effective weights (mean effective
reas were found in the study area; these subareas were weight, 4.12 wt%) compared with the theoretical weights
considered in the statistical analysis of the results. The (21.74 wt%). The net recharge, aquifer media, and
effective weight derived from the single-parameter sensi- hydraulic conductivity had higher effective weights than
tivity analysis is shown in Table 5. The depth to water the theoretical weights assigned by DRASTIC.

123
Environ Earth Sci (2014) 71:3119–3131 3127

Table 4 Original and modified

Factor Range Original rating Mean NO3 Modified rating
DRASTIC weighting rates
based on nitrate concentration Depth to groundwater 15–23 3 No data 3
23–30 2 No data 2
[30 1 No data 1
Recharge (mm) 0–50.8 1 13.4 7.12
50.8–101.6 3 12.8 6.8
101.6–177.8 6 18.8 10
Soil type Clay loam 3 15.13 8.1
Gravel 10 No data 10
Silty loam 4 13.34 7
Loam 5 10 5.3
Sandy loam 6 8.3 4.4
Sand 9 18.83 10
Non-shrinking 1 12 6.4
Topography 0–2 10 13.13 2.5
2–6 9 14.58 2.8
6–12 5 11.59 2
12–18 3 53.2 10
[18 1 No data 1
Impact of vadose Silt/clay 3 19.6 10
zone Sand and gravel 8 7.4 3.8
Sand and gravel with silt and 6 13 6.7
clay
Aquifer media Marlstone 2 15.4 9.1
Depth to water table: nitrate Silt and clay 5 17 10
concentration divided to six Sandstone 6 No data 6
classes, not used for depth of
Gravel sand 8 13.6 7.9
water

Fig. 7 Modified DRASTIC

map

123
3128 Environ Earth Sci (2014) 71:3119–3131

the lowest mean concentration, and the remaining rates were

Fig. 8 Percentage of modified DRASTIC result
Results and discussion
Result of index calibration and evaluation
The difference between the original DRASTIC values and
the nitrate concentration using a sample of 27 wells in May
2010 is presented in Fig. 5. Pearson’s correlation factor
(Table 3) was applied to determine the correlation between
the nitrate concentration and the DRASTIC values, as
illustrated in Fig. 6. Pearson’s correlation value was calcu-
lated at 44 %. This value is relatively low, thereby indicating
that the original vulnerability index must be changed to
obtain a realistic assessment of the potential contamination
in the study area. The highest nitrate concentration was
correlated with the highest rate. Other weighting rates were
linearly changed according to this relation. In the current
method, the rates of the net recharge, aquifer media,
hydraulic conductivity, impact of vadose zone, soil media,
and topography were modified based on the mean nitrate
concentration. The highest rate was assigned to the highest
mean nitrate concentration, the lowest rate was assigned to
linearly modified. This approach was applied to all layers, as
shown in Table 4, and to the new weight of the modified
DRASTIC model with sensitivity analysis. The new
DRASTIC map was computed by applying the modified
system, as shown in Fig. 7. The modified DRASTIC model
concentration and the percentage of its results are shown in
Fig. 8. The Pearson’s correlation factor was computed for
the modified DRASTIC model; its value was increased to
82 % in the new model, as indicated in Table 5 and Fig. 6.
The calibration results suggested that the modified
DRASTIC model significantly affects the study area.
Nitrate is important to obtain better results in the vulner-
ability map, considering that most of the lands in the
Kerman plain are agricultural. The new rates and weights
of the modified DRASTIC map indicated that 41.34 % of
the area belonged to the very high and high vulnerability
class. The percentage for this class was 50.09 % before the
modification. The percentages for the moderate class
before and after modification were 30.81 and 39.85 %,
respectively. The low and very low classes covered 19.09
and 18.81 %, before and after the application of the new
rates, respectively. These results clarified the effect of
modification. In addition, maps were compared to show the
spatial distribution of the index before and after modifi-
cation, as shown in Fig. 9. The result indicated that
45.72 % of the results had a similar class, but 54.28 %
belonged to a different class, thereby verifying the effec-
tiveness of the proposed method. The first nitrate test in
2010 was applied to calibrate the DRASTIC model and to
create the modified DRASTIC model. The second nitrate
test was used to compute for the correlation factor. The
application of the new rates and weights to the layers
clarified the effect of the modification. The region of higher
vulnerability in both maps is situated around the boundary
of the study area, which is mainly composed of highly
permeable sand and gravel. In the modified DRASTIC
model, the vulnerability increases in the center of the
Kerman plain because of agricultural activities (Table 6).

Table 5 Single-parameter
Parameter DRASTIC weight Theoretical weight (%) Effective weight (%)
sensitivity analysis on modified
DRASTIC Minimum Mean Maximum SD

D 5 21.74 2.91 4.12 10.49 1.47

R 4 17.39 14.20 22.47 31.01 4.54
A 3 13.04 12.33 17.52 26.09 2.99
S 2 8.70 4.73 9.49 18.35 3.05
T 1 4.35 0.61 2.34 8.93 2.14
I 5 21.74 10.71 23.37 35.71 6.40
C 3 13.04 10.47 15.83 23.81 3.42

123
Environ Earth Sci (2014) 71:3119–3131
Fig. 9 The percentage of priority between DRASTIC and modified
DRASTIC
Table 6 Correlation factors between nitrate concentration and
modified vulnerability index
Pearson’s correlation Number of data Factor
coefficient (%)
100 27 Nitrate index
82 DRASTIC index
Conclusion
The Kerman plain is located in an arid and semi-arid
region. With groundwater as the only water source in the
area, the evaluation of groundwater quality is crucial. The
increased pumping rates from the water table and the
decreased rainfall in the area accounted for the assessed
groundwater vulnerability in the area.
Applying the DRASTIC model in this region usually
provides a satisfactory assessment of the intrinsic vulner-
ability of groundwater to pollution. In addition, the
majority of the area consists of agricultural land where
inorganic fertilizers are commonly used. Thus, the nitrate
concentrations in the groundwater are primarily due to the
leaching of nitrate from the soil surface layers to the
groundwater. Given the abovementioned considerations,
the original DRASTIC algorithm required calibration and
modification to obtain more accurate results. Thus, we
developed and applied the modified DRASTIC model. The
correlation factor between the nitrate concentrations and
the original vulnerability index was evaluated at 44 %,
whereas the correlation factor between the nitrate con-
centrations and the modified DRASTIC model was calcu-
lated at 82 %. These results indicated that the modified
DRASTIC model could provide better results, as compared
with the original DRASTIC model. An advantage of the
modified DRASTIC model is its flexibility for adjusting the
rates and weights of the said model.
3129
The modified DRASTIC model in this study is recom-
mended for evaluating groundwater vulnerability to pol-
lution in agricultural lands with extensive use of nitrates.
The conditions of a specific area significantly influence the
type of modifications to be applied in the DRASTIC model.
The DRASTIC weights can be varied to improve the model
in this study.
Acknowledgments The authors would like to thank Saman Javadi
and Mohsen Dadras for their valuable contribution in this manuscript.
Thanks to three anonymous reviewers for their helpful reviews which
helped us to improve the quality of the earlier version of the
manuscript.
References
Alexander WJ, Liddle SK, Mason RE, Yeager WB (1986) Ground-
water vulnerability assessment in support of the first stage of the
National Pesticide Survey. Research triangle Park, Research
Triangle Institute, Contract No. EPA 68-01 6646, US Environ-
mental Protection Agency, Washington, DC
Al-Hanbali A, Kondoh A (2008) Groundwater vulnerability assess-
ment and evaluation of human activity impact (HAI) within the
Dead Sea groundwater basin, Jordan. Hydrogeol J 16(3):
499–510
Aller L, Bennet T, Lehr JH, Petty RJ, Hacket G (1985) DRASTIC: A
standardized system for evaluating groundwater pollution using
hydrological settings. Prepared by the National water Well
Association for the US EPA Office of Research and Develop-
ment, Ada
Aller L, Bennet T, Lehr JH, Petty RJ, Hackett G (1987) DRASTIC: a
standardized system for evaluating groundwater pollution
potential using hydrogeological settings. EPA/600/2–87/035.
US Environmental Protection Agency, Ada
Antonakos AK, Lambrakis NL (2007) Development and testing of
three hybrid methods for assessment of aquifer vulnerability to
nitrates, based on the DRASTIC model, an example from NE
Korinthia, Greece. J Hydrol 333(2–4):288–304
Ayazi MH, Pirasteh S, Arvin AKP, Pradhan B, Nikouravan B, Mansor
S (2010) Disasters and risk reduction in groundwater: Zagros
Mountain Southwest Iran using geoinformatics techniques.
Disaster Adv 3(1):51–57
Babiker I, Mohamed M, Hiyama T, Kato K (2005) A GIS based
DRASTIC model for assessing aquifer vulnerability in Kakami-
gahara Heights, Gifu Prefecture, Central Japan. Sci Total
Environ 345(1–3):127–140
Bojórquez-Tapia LA, Cruz-Bello GM, Luna-González L, Juárez L,
Ortiz-Pérez MA (2009) V-DRASTIC: using visualization to
engage policymakers in groundwater vulnerability assessment.
J Hydrol 373:242–255
Carvalho GJP (2009) Vulnerabilidade à contaminação das águas
subterrâneas na bacia hidrográfica do Rio Sordo: comparação de
modelos baseados no método DRASTIC. MSc thesis, Trás-os-
Montes and Alto Douro University, Vila Real
Ckakraborty S, Paul PK, Sikdar PK (2007) Assessing aquifer
vulnerability to arsenic pollution using DRASTIC and GIS of
North Bengal Plain: a case study of English Bazar Block, Malda
District, West Bengal, lndia. JOSH 7(1):101–121
Daly D, Drew D (1999) Irish methodologies for karst aquifer
protection. In: Beek B (ed) Hydrogeology and engineering
geology of sinkholes and karst. Balkema, Rotterdam, pp 267–
327
123
3130
Denny SC, Allen DM, Journea YM (2007) DRASTIC-Fm: a modified
vulnerability mapping method for structurally controlled aquifers
in the southern Gulf lslands, British Columbia, Canada. Hydro-
geol J 15:483–493
Dixon B (2004) Prediction of groundwater vulnerability using an
integrated GIS-based neuro-fuzzy techniques. J Hydrol
4(309):17–38
Dixon B (2005) Groundwater vulnerability mapping: a GIS and fuzzy
rule based integrated tool. Appl Geogr 25(4):327–347
Ettazarini S (2006) Groundwater pollution risk mapping for the
Eocene aquifer of the Oum Er-Rabia basin, Morocco. Environ
Geol 51(3):341–347
Foster SS (1987) Fundamental concepts in aquifer vulnerability,
pollution risk and protection strategy. In: Duijvenbooden W van,
Waegeningh HG van (eds) TNO Committee on Hydrological
Research, The Hague Vulnerability of soil and groundwater to
pollutants. Proceedings and Information, vol 38, pp 69–86
Fritch TG, McKnight CL, Yelderman JC, Arnold JG (2000) An
aquifer vulnerability assessment of the Paluxy aquifer, Central
Texas, USA, using GIS and a modified DRASTIC approach.
Environ Manage 25:337–345
Gundogdu KS, Guney I (2007) Spatial analyses of groundwater level
using universal Kriging. J Earth Syst Sci 116:49–55
Hailin Y, Ligang X, Chang Y, Jiaxing X (2011) Evaluation of
groundwater vulnerability with improved DRASTIC method.
Procedia Environ Sci 10:2690–2695
Hamza MH, Added A, Rodrı́guez R, Abdeljaoued S, Mammou AB
(2007) A GIS-based DRASTIC vulnerability and net recharge
reassessment in an aquifer of a semi-arid region (Metline-Ras
Jebel-Raf Raf aquifer, Northern Tunisia). J Environ Manage
84:12–19
Harbaugh AW, Banta ER, Hill MC, Mcdonald MG (2000) MOD-
FLOW-2000, The US Geological Survey Modular Ground-water
Model- Users guide to modularization concepts and the ground-
water flow process. US Geological Survey Open-File Report
00-92, 121
Huan H, Wang J, Teng Y (2012) Assessment and validation of
groundwater vulnerability to nitrate based on a modified
DRASTIC model: a case study in Jilin City of northeast China.
Sci Total Environ (Article online first available). http://dx.doi.
org/10.1016/j.scitotenv.2012.08.037
Javadi S, Kavehkar N, Mousavizadeh MH, Mohammadi K (2011a)
Modification of DRASTIC model to map groundwater vulner-
ability to pollution using nitrate measurements in agricultural
areas. J Agr Sci Tech 13(2):239–249
Javadi S, Kavehkar N, Mohammadi K, Khodadi A, Kahawita K
(2011b) Calibration DRASTIC using field measurements, sen-
sitivity analysis and statistical method to assess groundwater
vulnerability. Water Int 36(6):719–732
Kalinski RJ, Kelly WE, Bogardi I, Ehrman RL, Yamamoto PO (1994)
Correlation between DRASTIC vulnerabilities and incidents of
VOC contamination of municipal wells in Nebraska. Ground
Water 32(1):31–34
Kumar V (2007) Optimal contour mapping of groundwater levels
using universal Kriging—a case study. Hydrol Sci J
52:1038–1050
Leone A, Ripa MN, Uricchio V, Deák L, Vargay Z (2009)
Vulnerability and risk evaluation of agricultural nitrogen pollu-
tion for Hungary’s main aquifer using DRASTIC and GLEAMS
models. J Environ Manage 90(10):2969–2978
Manap MA, Nampak H, Pradhan B, Lee S, Sulaiman WNA, Ramli
MF (2012) Application of probabilistic-based frequency ratio
model in groundwater potential mapping using remote sensing
data and GIS. Arab J Geosci. http://dx.doi.org/10.1007/s12517-
012-0795-z
123
Environ Earth Sci (2014) 71:3119–3131
Manap MA, Sulaiman WNA, Ramli MF, Pradhan B, Surip N (2013)
A knowledge-driven GIS modelling technique for groundwater
potential mapping at the Upper Langat Basin, Malaysia. Arab J
Geosci 6(5):1621–1637. http://dx.doi.org/10.1007/s12517-011-
0469-2
McLay CDA, Dragden R, Sparling G, Selvarajah N (2001) Predicting
groundwater nitrate concentrations in a region of mixed
agricultural land use: a comparison of three approaches. Environ
Pollut 115:191–204
Melloul AJ, Collin M (1998) A proposed index for aquifer water-
quality assessment: the case of lsrael’s Sharon region. J Environ
Manage 54:131–142
Mimi ZA, Mahmoud N, Madi MA (2012) Modified DRASTIC
assessment for intrinsic vulnerability mapping of karst aquifers:
a case study. Environ Earth Sci 66(2):447–456
Naqa A, Hammouri N, Kuisi M (2006) GIS-based evaluation of
groundwater vulnerability in the Russeifa area, Jordan. Revista
Mexicana de Ciencias Geológicas 23(3):277–287
National Research Council (1993) Groundwater vulnerability assess-
ment: predicting relative contamination potential under condi-
tions of uncertainty. Committee for assessing ground water
vulnerability. National Academy Press, Washington, DC
Nobre RCM, Filho OCR, Mansur WJ, Nobre MMM, Cosenza CAN
(2007) Groundwater vulnerability and risk mapping using GIS,
modeling and a fuzzy logic tool. J Contam Hydrol 94:277–292
Pacheco FAL, Sanches Fernandes LF (2012) The multivariate
structure of DRASTIC model. J Hydrol. doi:http://dx.doi.org/
10.1016/jhydrol.2012.11.20
Panagopoulos GP, Antonakos AK, Lambrakis NJ (2006) Optimiza-
tion of the DRASTIC method for groundwater vulnerability
assessment via the use of simple statistical methods and GIS.
Hydrol J 14:894–911
Pathak DR, Hiratsuka A (2011) An integrated GIS based fuzzy
pattern recognition model to compute groundwater vulnerability
index for decision making. J Hydro-environ Res 5:63–77
Plymale CL, Angle MP (2002) Groundwater pollution potential of
Fulton County, Ohio. Ohio Department of natural resources
division of water, water resources section. Groundwater Pollu-
tion Potential, Report No 45
Pradhan B (2009) Ground water potential zonation for basaltic
watersheds using satellite remote sensing data and GIS tech-
niques. Cent Eur J Geosci 1(1):120–129. http://dx.doi.org/10.
2478/v10085-009-0008-5
Rahman A (2008) A GIS based DRASTIC model for assessing
groundwater vulnerability in shallow aquifer in Aligarh, India.
Appl Geogr 28(2008):32–53
Remesan R, Panda RK (2008) Groundwater vulnerability assessment,
risk mapping, and nitrate evaluation in a small agricultural
watershed: using the DRASTIC model and GIS. Environ Qual
Manage 5(4):53–75
Rupert MG (1999) Improvements to the DRASTIC groundwater
vulnerability mapping method. US Geological Survey Fact Sheet
FS-066-99, USGS, Reston, VA, USA
Saidi S, Bouri S, Dhia HB (2010) Groundwater vulnerability and risk
mapping of the Hajeb-jelma aquifer (Central Tunisia) using a
GIS-based DRASTIC model. Environ Earth Sci 59(7):
1579–1588
Saidi S, Bouria S, Dhiaa HB, Anselmeb B (2011) Assessment of
groundwater risk using intrinsic vulnerability and hazard map-
ping: application to Souassi aquifer, Tunisian Sahel. Agr Water
Manage 98:1671–1682
Secunda S, Collin ML, Melloul AL (1998) Groundwater vulnerability
assessment using a composite model combining DRASTIC with
extensive agricultural land use in Israel’s Sharon region.
J Environ Manage 54:39–57
Environ Earth Sci (2014) 71:3119–3131
Shukla S, Mostaghimi S, Shanholt VO, Collins MC, Ross BB (2000)
A county-level assessment of ground water contamination by
pesticides. Ground Water Monit R 20(1):104–119
Sinkevich MG, Walter MT, Lembo AJ Jr, Richards BK, Peranginan-
gin N, Aburime SA, Steenhuis TS (2005) A GIS-based ground
water contamination risk assessment tool for pesticides. Ground
Water Monit R 25(4):82–91
Theodossiou N (1999) Evaluation of the distribution of hydraulic
head in an aquifer using the Kriging method. Sci J Hellenic
Hydrotech Assoc Hydrotech 9:3–14
Thirumalaivasan D, Karmegam M, Venugopal K (2003) AHP-
DRASTIC: software for specific aquifer vulnerability assessment
using DRASTIC model and GIS. Environ Modell Softw
18:645–656
Van Stemproot D, Evert L, Wassenaar L (1993) Aquifer vulnerability
index: a GIS compatible method for groundwater vulnerability
mapping. Can Water Resour J 18:25–37
3131
Werz H, Hötzl H (2007) Groundwater risk intensity mapping in semi-
arid regions using optical remote sensing data as an additional
tool. Hydrogeol J 15(6):1031–1049
Wilcoxon F (1945) Individual comparisons by ranking methods.
Biometr Bull 1:80–83
Yin L, Zhang E, Wang X, Wenninger J, Dong J, Guo L, Huang J
(2012) A GIS-based DRASTIC model for assessing groundwater
vulnerability in the Ordos Plateau, China. Environ Earth Sci
69(1):171–185
Yuan M, Zhang X, Wang L (2006) Fuzzy pattern recognition method
for assessing groundwater vulnerability to pollution in the
Zhangji area. J Zhejiang Univ Sci A 7(11):1917–1922
Zhou H, Wano O, Yang Q (1999) A multi-objective fuzzy pattern
recognition model for assessing groundwater vulnerability based
on the DRASTIC system. Hydrol Sci J 44:611–618
123