Arab J Geosci DOI 10.1007/s12517-013-0964-8

ORIGINAL PAPER

Integrated GIS and remote sensing for mapping groundwater potentiality in the Tulul al Ashaqif, Northeast Jordan Muheeb Awawdeh & Mutewekil Obeidat & Mohammad Al-Mohammad & Khaldoon Al-Qudah & Rasheed Jaradat

Received: 23 November 2012 / Accepted: 24 April 2013 # Saudi Society for Geosciences 2013

Abstract Jordan with its limited water resources is currently classified as one of the four water-poor countries worldwide. This study was initiated to explore groundwater potential areas in Tulul al Ashaqif area, Jordan, by integrating remote sensing, geographic information systems (GIS), and multicriteria evaluation techniques. Eight thematic layers were built in a GIS and assigned using multicriteria evaluation techniques suitable weights and ratings regarding their relative contribution in groundwater occurrence. These layers include lithology, geomorphology, lineaments density, drainage density, soil texture, rainfall, elevation, and slope. The final groundwater potentiality map generated by GIS consists of five groundwater potentiality classes: very high, high, moderate, low, and very low. The map showed that the study area is generally of moderate groundwater potentiality (76.35 %). The very high and high potential classes occupy 2.2 and 12.75 % of study area, respectively. The validity of results of this GIS-based model was carried out by superimposing existing hand dug wells on the final map. The single parameter sensitivity test was conducted to assess the influence of the assigned weights on the groundwater potential model, and new effective weights were derived. The resulted groundwater potentiality map showed that the area occupied by each of the groundwater potentiality classes has changed. However, the M. Awawdeh (*) Department of Geography and Geographic Information Systems, King Abdulaziz University, Jeddah, Saudi Arabia e-mail: [email protected] M. Obeidat Faculty of Science and Arts, Jordan University of Science and Technology, 22110, Irbid, Jordan M. Al-Mohammad : K. Al-Qudah : R. Jaradat Deparment of Earth and Environmental Sciences, Faculty of Science, Yarmouk University, 21163, Irbid, Jordan

study area remains generally of moderate groundwater potentiality (70.93 % of the study area). The area occupied by the very high and high potential classes comprises 4.53 and 18.56 % of the study area, respectively. Keywords GIS . Remote sensing . Groundwater potentiality . Multicriteria . Jordan

Introduction Water scarcity is one of the most important environmental challenges facing the world. The demand on groundwater is increasing over the years, and the overexploitation and continuous pollution of this vital resource is threatening both our ecosystems and future generations (Rekha and Thomas 2007). Jordan is currently one of the world’s four water poorest nations, with approximately 80 % of the country receiving <100 mm of precipitation annually. Therefore, proper assessment, planning, and development of water resources are key elements in the overall social and economic development of the country. In Jordan, groundwater is the most important source of water providing about half of the total water demand. Many experts have agreed that the scarcity of water is the gravest environmental challenge facing the country both in the present and in the future (JGE 2001). Whereas water resources in Jordan have fluctuated around a stationary average, the country’s population has continued to increase. Therefore, Jordan needs to expand its water supply to meet the growing needs by exploiting new sources of groundwater. Groundwater occurrence can be explored using a variety of techniques, which are divided into surface and subsurface investigation methods. One of the former techniques for water exploration is the integration of remote sensing and geographic

Arab J Geosci

information systems (GIS). This method has been used by many researchers to prepare groundwater resources maps in many parts of the world (Murthy 2000; Toleti et al. 2000; Dissanayake 2006; Rekha and Thomas 2007). Groundwater exploration has many problems such as the scarcity of existing data, the high cost of data collection, and relatively remote target aquifer. Therefore, remote sensing techniques and geographic information system are considered one of the most appropriate new alternative tools for groundwater exploration (Moore 1982; Dar et al. 2010). Remote sensing technology and geographic information system modeling: An integrated approach towards the mapping of groundwater potential zones in Hardrock terrain, Mamundiyar basin (Dar et al. 2010). One of the greatest advantages of using remote sensing data for hydrogeological investigation and monitoring is its ability to generate information in spatial and temporal domain, which is very crucial for successful analysis, prediction, and validation (Saraf 1999). The large spatial coverage of satellite imagery enables the identification of basin and broad physiographic features, such as stream network character, land use, and watershed coverage. Furthermore, remotely sensed imagery provides useful spatial and spectral information at relatively low cost (Battaglin et al. 1993). Satellite images can be helpful in detecting and mapping an area’s regional structural patterns, including major fracture and fault systems (Chavez et al. 1997), which are major requirements for groundwater resources evaluation and exploration (El-Baz and El-Ashry 1991). The identification and location of groundwater resources using remote sensing data is based on an indirect analysis of some directly observable terrain features like geomorphology, geology, slope, land use/land cover, and hydrologic characteristics (Das et al. 1997; Ravindran and Jeyaram 1997; Srinivas et al. 2000; Sree Devi et al. 2001; Gopinath and Saralathan 2004; Kuria et al. 2012). GIS, on the other hand, permits storing and efficient processing of georeferenced data derived and collected from various sources, maps, satellite imagery, and land surveys (Lillesand and Kiefer 2000). Moreover, GIS allows the overlaying of data layers from different sources to produce user defined maps that contain new output based on criteria previously determined by user. Remote sensing and GIS with their advantages of spatial, spectral and temporal availability, and manipulation of data covering large and inaccessible areas within a short time have become very handy tools in accessing, monitoring, and conserving groundwater resources. With the capabilities of the remotely sensed data and GIS techniques, numerous databases can be integrated to produce conceptual model for delineation and evaluation of groundwater potential zones of an area (Chaterjee and Bhattacharya 1995; Krishnamurthy and Srinivas 1995; Srivasthava and Bhattacharya 2000). The fundamental objective of this study is to delineate groundwater potential zones in the Tulul al

Ashaqif area, northeast Jordan, using multicriteria evaluation techniques (MCET) on thematic maps of a number of factors with the aids of remote sensing and GIS techniques. The word Tulul (Arabic) means highlands.

Description of the study area The study area (Fig. 1) is a part of the Badia region in northeast Jordan, nearly half way between the towns of Safawi and Ruwaishid. It is located between 572193 to 624155 North and 538607 to 599262 East (Jordan Transverse Mercator coordinates). The Amman–Baghdad highway crosses the southern part of the ridge and provides the only major land access into the area. The present climate is arid with cool winters (2–9 °C) and hot summers (may exceed 40 °C). Rainfall occurs in winter and is generally erratic in its spatial and temporal distribution, with average annual rainfall between 60 and 100 mm/year. The study area varies in elevation with the headwater in the west having the highest elevation (1,050 m above sea level), whereas the east side of the study area is much flatter having the lowest elevation (660 m above sea level). The relief of Tulul al Ashaqif is generally moderate, with distinct topographic features defined by volcanic cones and wadis. The highlands separate the Azraq drainage basin in the west from the Hammad drainage basin in the east. The drainage system flows toward the west into Wadi Rajil and toward the east into the playas of Hammad basin (Abu-Jaber et al. 2003). The ridge is a part of the Harrat ash Shaam volcanic rocks (Harrat Ash-Shaam super-group), and it is Neogene in age (Abu-Jaber et al. 2003). According to Ibrahim (1993), much of the ridge is composed of the Safawi Group, whereas the younger Aritain pyroclastics make up the volcanic cones which dot the area. Large-scale stone pavements overlie an aeolian sedimentary mantle. These pavements are believed to affect the surface runoff characteristics and prevent the erosion of the underlying aeolian sediments (Higgitt and Allison 1998). Al-Qudah (2003) classified the pavements into three categories according to percentage of fragments covering the land: pavement 1 (99.60 % stone coverage), pavement 2 (87.80 % stone coverage), and pavement 3 (64.60 % stone coverage). In the wadi system, coarsegrained fluvial sediments are concentrated and include a number of distinct fluvial terraces in some areas. Along the course of the wadi, the width of the channel may increase significantly and result in broad reaches filled with coarse sand and gravel which are known locally as “marabs” (Al Qudah 2003). The marabs typically have a relatively rich vegetative cover, and are sought after by the shepherds in the area (Dottridge and Abu-Jaber 1999). At the termini of these drainage basins, fine-grained playa deposits define the Qa’s, which are almost totally devoid of

Arab J Geosci Fig. 1 Location of the study area

vegetative cover. The geological map of the study area shows several major faults trending NW–SE among which Qitar Al Abid fault. Generally, the regional aquifer system in the study area is classified into three aquifer complexes: the upper aquifer system, which includes Cainozoic aquifers (basalt, Wadi Shallala aquifer, and Umm Rijam aquifer); the middle aquifer system, which is separated from the overlying aquifers by Muwaqqar aquiclude, and involves Mesozoic aquifers (Amman-Wadi As Sir aquifer, Kurnub sandstone aquifer, and Azab-Ramtha aquifer); the lower aquifer system involves the Disi sandstone aquifer, which has been extensively studied in south Jordan; it is separated from the overlying middle aquifer system by the Khreim shales. In the Azraq well-field area, west to the study area, transmissivity of the basalt ranges between 270 to more than 65,000 m2/day (Humphreys 1982). WAJ (1989, unpublished report) reported transmissivity in the range of few to 250 m2/day of Wadi Shalla aquifer. Corresponding permeability is in the range of <0.1 to about 5 m/day. Umm Rijam aquifer has transmissivity ranging between 10 and 5,000 m2/day; it has a permeability of <2 to >56 m/day (WAJ 1995). The hydrochemistry and environmental isotope content of the regional groundwater, mainly Umm Rijam groundwater in Hamad basin, have been extensively investigated by Obeidat (2000) and Zagana et al. (2007). Electrical conductivity of the Umm Rijam aquifer ranges between 970 and 4,900 μS/cm with an average of

2,063 μS/cm. Electrical conductivity of the Amman/Wadi As Sir groundwater ranges between 2,321 and 4,600 μS/cm with an average of 3,158 μS/cm. Electrical conductivityof Ramtha groundwater (Triassic) ranges between 2,916 and 5,000 μS/cm with an average of 3,453 μS/cm. The Tulul al Ashaqif highlands is part of the basaltic plateau of Jordan, which consists of different basaltic flows of Tertiary to Quaternary age (Ibrahim et al. 2001). Most of the basaltic flows are covered with well-developed soil and a veneer of desert pavement rests on the top of the soil surface. Shallow perched aquifers are known in the area and have been used by the local inhabitants for many generations. The water depth in these aquifers ranges from 1.5 to 18 m (Abu-Jaber 2001). Three historical traditional sites are known to contain such water. These are at Wadi Khudari, Wadi Mahdath, and Wadi Ghussein. Simple hand dug pits to depths of few meters accessed water within the shallow alluvium in the reaches. The water is of acceptable quality (total dissolved solids content ranges between 250 and 900 mg/l), and it is used for domestic and livestock purposes (Abu-Jaber 2001). These shallow perched aquifers are distinct from the regional water table, which is over 350 m deep in the area, and of poor quality (Abu-Jaber et al. 1998). Studies have shown that for the most part, infiltration through the pavement surfaces does not contribute to the recharge of the groundwater in the area (Al Qudah 2003). After heavy rainfall events, the runoff water is channeled through the drainage system and partially leaks through the alluvium that makes

Arab J Geosci

up the bottom and the sides of the wadis (Al Qudah 2003). The presence of shallow perched aquifer does not only depends on recharge but also on the prevention of the water from leaking quickly to deeper levels. Interaction of shallow water and carbon dioxide leaking from deeper levels within the local basalt leads to the formation of a clay rich horizon (Kimberley and Abu-Jaber 2005). This horizon is largely impermeable, trapping the water near the surface and keeping it accessible. An attempt was made to locate areas with similar attribute to known well locations. These areas were investigated using remote sensing and geophysical survey (Abu-Jaber et al. 2003). Drilling in the investigated sites yielded two new sites with shallow aquifers. The yield of these aquifers is fairly small and does not exceed few to tens of cubic meters per day (Al Qudah 2003). However, this small amount is very important and fairly sufficient for the local Bedouins for both domestic and livestock use. Spatial analysis study of the location of the traditional well sites and the newly drilled wells shows that the location of the perched aquifers is mainly restricted to the reaches of the main wadis in the area. However, spatial characterization of the location of these aquifers has shown significant factors that influence the development of these aquifers (Al-Qudah and Abu-Jaber 2009). For example, the study shows that perched aquifers are found above 750 m elevation, and they have significant catchment area and stream length above 900 m elevation.

Material and methods Thematic maps of geology, geomorphology, drainage density, slope, rainfall, soil, lineaments density, and elevation were used to generate the final groundwater potentiality map in the study area. Remote sensing data (Landsat Enhanced Thematic Mapper Plus, ETM+) were used to derive and prepare several parameters of the groundwater potential model: drainage network, lineaments, and geomorphology. Remote sensing data preparation included several steps: data acquisition, image mosaicing, stacking, and image subsetting. The images with visible, near-infrared, and mid-infrared bands of the ETM+ imageries were collected from the Landsat digital online database, which is available via the website address: http://www.landsat.org. The ETM+ imageries are composed of eight spectral bands with three different resolutions: 30 m for the visible and near- and mid-infrared (bands 1–5 and 7), 15 m for the panchromatic (band 8), and 60 m for the thermal infrared (band 6). In this study, two individual landsat images ETM+ were acquired for the mapping process. The path and row numbers and coordinates of the center of the imageries are given in Table 1. The images were originally projected to the Universal Transverse Mercator zone 37, which were then reprojected to the Jordan

Table 1 A list of the Landsat Enhanced Thematic Mapper Plus (ETM+) imageries used in the study Image number

1 2

Path

173 173

Row

37 38

Scene Center (JTM) Northing

Easting

672,032 671,279

529,916 790,526

Transverse Mercator (JTM). The landsat data used in this study were acquired in 2000 by Landsat 7 satellite. The images were digitally processed [principle component analysis, band rationing (5/7 and 5/4), edge enhancement] for digitizing lineaments, drainage systems, and geomorphological features. Maps of these features were derived from the images by visual interpretation. Supervised classification was used to derive the land use and geomorphological features of Tulul al Alshaqif. Preparation and analysis of remote sensing data were carried out using several softwares: ERDAS Imagine 8.7, ArcGIS 9.2, ArcView 3.2, and ENVI 4.2. A lithology map (1:250,000 scale) obtained from the Natural Resource Authority (NRA) was digitized and spatially adjusted (rubbersheet method in ArcGIS 9.2) to align well with other spatial data. The geological map was investigated for the rock formations and properties. A digital soil map (1:250,000 scale) obtained from the Ministry of Agriculture was used to generate a thematic map of the soil texture in the study area. Topographic maps (1:50,000 scale) with 20-m contour lines were utilized to create a digital elevation model, from which a thematic map of the slope was generated. The main land use types and geomorphological features (wadies, marabs, mudflats, and volcanic uplands) were digitized with the aid of land use maps produced by the Royal Jordanian Geographic Center. Pavements of wadi el-Ghusein were used as a training area for the supervised classification of the Tulul al Alshaqif area. There are only two weather stations (Tulul al Alshaqif and Ruweishid) within and close to the study area. The two stations were used with the aid of topography to create a spatial rainfall map. Not all the above-mentioned factors have the same importance of contribution of groundwater occurrence and storage. Therefore, an MCET was adopted to prepare the final map of the groundwater potential Table 2 Rating scale used in Saaty’s AHP model Weight

Definition

1 3 5 7 9

Equally likely occurrence Moderately likely occurrence Strongly likely occurrence Very strongly likely occurrence Extremely strongly likely occurrence

The values 2, 4, 6, and 8 can be used to denote intermediary values

Arab J Geosci Table 3 Paired comparison matrix of the used factors Factor

L

G

D

SL

R

S

LD

E

Weight

L G D SL R S LD E

1 1/3 1/3 1/5 1/5 1/7 1/8 1/9

3 1 1 1/3 1/3 1/5 1/7 1/9

3 1 1 1/3 1/3 1/5 1/7 1/9

5 3 3 1 1 1/3 1/5 1/7

5 3 3 1 1 1/3 1/5 1/7

7 5 5 3 3 1 1/3 1/5

8 7 7 5 5 3 1 1/3

9 9 9 7 7 5 3 1

0.3487 0.1917 0.1917 0.0905 0.0905 0.0455 0.0258 0.0156

L Lithology, G geomorpholgy, S soil, LD lineament density, D drainage density, E elevation, SL slope, R annual rainfall

model (GPM). MCET are numerical algorithms that define the suitability of a particular solution on the basis of the input criteria and weights together with some mathematical or logical means of determining trade-offs when conflicts arise (Heywood et al. 2003). In this technique, each data layer (geology, geomorphology, soil, etc.) is assigned a “weight” to reflect its importance relative to other data layers. Moreover, a rating is given to reflect the importance of a class within a data layer. The factors used to prepare the final groundwater potential map include geology, geomorphology, drainage density, slope, rainfall, soil texture, lineament density, and elevation. It is important to understand the control of these factors on the groundwater regime of any area for optimal exploitation and aquifer management. There are two common methods used to assign weights of the parameters: simple additive weighting (SAW) and analytical hierarchy principle (AHP). SAW is probably the best known and very widely used method of multiple attribute decision making. To each of the attributes in SAW, the decision maker assigns importance weights which become the coefficients of the variables. The decision maker can then obtain a total score for each alternative simply by multiplying the scale rating for each attribute value by the importance weight assigned to the attribute and then summing these products over all attributes. This study employs the AHP to assign weights of the factors used. Using the analytical hierarchy principle (AHP) nineTable 4 Average consistency index

Order of matrix

Randomized index (RI)

3 4 5 6

0.90 1.12 1.24 1.32

7 8 9

1.41 1.45 1.51

point scale (Table 2), a paired comparison matrix (Saaty 1980), was prepared for the eight factors selected for Tulul al Ashaqif area, and then maps weights were then developed (Table 3). In AHP model, a matrix A of order n, where n is the Table 5 Weights and ratings of the factors used to map groundwater potentiality Factor

Class

Rating Weight

Elevation (m)

660–750 750–850

50 40

850–950 950–1,050 Soil Silty clay loam to clay Silty clay loam Very stony silty clay loam Often very gravelly, structured silty clay loam Stony and very stony silty clay loam to silty clay Silty clay loam and sandy clay Rainfall (mm) 90 81 75 Geology Alluvium mudflat (lithology) Pleistocene sediments Basaltic dyke Volcanic Alluvium and Wadi sediments Geomorphology Mudflat Pavement 1 Rugged land (volcanic uplands) Pavement 2 Pavement 3 Wadi Marab Drainage density 0–0.5 (km/km2) 0.5–1 1–1.5 1.5–2 2–2.5 2.5–3 3–3.5 3.5–3.8 Lineament density <1 (km/km2) 1–2 2-3 3–4.5 Slope (degrees) >10 5–10 2–5 0.5–2 0–0.5

20 10 15 20 30 30

0.0156

0.0455

30 35 70 60 50 5 25 40 40 70 5 10 20 20 30 50 70 5 10 20 30 40 50 60 70 20 45 50 60 10 25 30 35 40

0.0905

0.3487

0.1917

0.1917

0.0258

0.0905

Arab J Geosci

number of classes in the criterion, is constructed. The matrix A is reciprocal and should be consistent. For each element aij of the matrix, the following condition is satisfied:  aij ¼ 1 aji ð1Þ By solving the matrix, we can find the weight of each parameter. In order to solve for weights, the following equation was used: ðA  I1ÞX ¼ 0

ð2Þ

where I is an identity matrix of order n, and X is the n×1 weight matrix and is the eigenvalue. For a matrix of order n, we get n number of eigenvalues. To calculate the weights, we select the largest value of λ. Using the geometric mean method, the weight (X) of each class is obtained, which are then normalized so that their sum is 1. These are then used to calculate the largest positive eigenvalue λmax. We select the maximum λmax to recalculate the weights. The values in the paired comparison matrix A must be checked for consistency. It is important to check for consistency because there may be inconsistency in judgment while comparing the various classes. The following formulas are used for calculating the consistency index: CI ¼ ðλmax  nÞ=ðn  1Þ

ð3Þ

CR ¼ CI=RI

ð4Þ

where CI=consistency index, CR=consistency ratios, and RI=random index Here, RI is the randomized consistency index for a matrix of order n. The used RI values in this study are shown in (Table 4). For an acceptable level of consistency, we should have CR < 0.1. If the value is found greater, we revise our pairwise comparison matrix until we get the required CR. The eigenvalues and the weights of the different parameter were calculated using a Microsoft Excel macro (Odat, personal communications). Calculations using this algorithm give a very good λmax, which is the largest approximate value of eigenvector. After calculation, we can calculate the consistency ratios using λmax in Eqs. 3 and 4 and Table 4. As discussed earlier, MCET require assigning ratings (scores) of the classes in the thematic layers. Every class in the thematic layers was placed into one of the following categories: (1) very good, (2) good, (3) moderate, (4) low, and (5) poor, depending on their groundwater potential level. After understanding their behavior with respect to groundwater control, the different classes belonging to each factor were given suitable ratings, according to their importance with respect to other classes in the same thematic layer. The ratings were adopted based on field observation, experts’ opinions, and previous similar researches on groundwater potentiality mapping (Brito et al. 2006; Ganapuram et al. 2009; Boru 2012; Kuria et al. 2012). For

Topography

Remote Sensing Data

Lithology

Data Preparation (Reprojection, Moasicking, Clipping, etc)

Data Processing (PCA, Band Rationing, Edge Enhancement)

Lineaments Map

Lineaments Density Map

Digitizing

Supervised Classification

Rainfall

Digitizing

Editing & Spatial Adjustment

Digitizing Contours

Attribute Data Editing

Topography

TIN Map

Drainage System Map Drainage Density Map

Soil

Geomorphology Map

DEM

Slope Map

Lithology Map

Data Classification (Rating) Data Modeling

Groundwater Potential Map Fig. 2 Flow chart showing the methodology used in the study

Soil Map

Rainfall Map

Arab J Geosci

instance, the values assigned to the lithology layer take into account the hydrogeological significance of the rock types. The characteristics considered for lithology are rock/sediment type, degree and type of weathering, fracture density, dykes, grain size, etc. For example, a maximum value of 70 was given for alluvium and wadi sediments (coarse sand) due to their favorable characters for

Fig. 3 Lineaments map of the study area (a) and the rating values of the lineaments density (b)

groundwater accumulation owing to their primary porosities and hydraulic conductivity. A value of 5 was given to alluvium mudflats due to the small particle size and accordingly the very low hydraulic conductivity. Moreover, the slope values were grouped into five classes. The rating value of the slope classes increases as the slope decreases. This implies that the flatter the area, the better are the chances for

Arab J Geosci

groundwater accumulation. The weights of the data layers and the ratings of the classes belonging to each layers are given in Table 5. All data relevant to groundwater potential mapping were collected or created, thoroughly examined, edited, compiled, and assembled in digital format for modeling. Figure 2 shows the different steps involved in preparing the input maps for the model. The GPM was constructed according to the formula: GPM ¼ ðLW  Lr Þ þ ðGw  Gr Þ þ ðSw  Sr Þ þ LDw  LDr Þ þ ðDw  Dr Þ þ ðEw  Er Þ þ ðSLw  SLr Þ þ ðRw  Rr Þ

ð5Þ where L=lithology, G=geomorpholgy, S=soil, LD=lineament density, D=drainage density, E=elevation, SL=slope, R=annual rainfall, W=factor weight, and r = class rating.

Results and discussion Groundwater potentiality mapping can be achieved by investigating the source of water (rainfall), geology, and the structural and geomorphological situation governing its occurrence. Therefore, the controlling factors on groundwater movement, storage, and occurrence were explored and digitally mapped as thematic layers. The layers include geology, geomorphology, soil texture, lineaments density, drainage density, elevation, slope, and rainfall. These factors were integrated using a GIS environment with the aid of ArcGIS 9.2. The factors used vary spatially, and it implies Fig. 4 Rating values of the drainage density

(1) geology (lithology), which determine infiltration, movement, and storage of water; (2) rainfall as a source of water; (3) lineaments (rock fractures), which enhance significantly hydrualic conductivity; (4) elevation; (5) slope, which controls water flow energy; 6) drainage density, which plays a vital role in the runoff distribution and level of infiltration; (7) soil texture, which determine infiltration rate of the soil; and (8) geomorphology, which controls surface runoff and infiltration. Lineament density factor Lineaments analysis for groundwater exploration is a very important step, since structural features such as fractures act as conduits for surface water and rainfall and enhance secondary porosity of rocks. Lineaments provide helpful clues to the movement and storage of groundwater. According to many studies, groundwater potential increases with higher lineament length density (Teeuw 1995; Sener et al. 2005; Shaban 2006). Lineaments of Tulul al Ashaqif were digitized visually from the processed ETM+ satellite images (Fig. 3a). In groundwater exploration studies, lineaments density is used instead of lineaments solely. The lineaments density map was generated in ArcGIS 9.2 using Kernel density method. The density map in the range of 0.009–4.05 (km/km2) was categorized into four classes and assigned ratings according to Table 5 (Fig. 3b). The lineament density factor was assigned a weight value equal to 0.0258 according to the AHP.

Arab J Geosci

Drainage density According to many researchers (Edet et al. 1998; Sener et al. 2005; Shaban 2006), groundwater potentiality is increasing where drainage density is high. Drainage system of the study area was obtained by digitizing ETM+ landsat imagery, and a density map was derived using Kernel density method in ArcGIS 9.2. There are two types of Fig. 5 Rating values of elevations (a) and slope (b)

drainage pattern in Tulul al Ashaqif: The first is dendritic, which dominates the study area, and the second type is radial pattern. The drainage density values varied within the range of 0.03–3.81 km/km2, which was classified into eight classes and assigned suitable ratings (5– 70) with respect to groundwater occurrence (Fig. 4). The drainage density factor was given a weight of 0.1917.

Arab J Geosci

Topography and slope Elevations in the study area (660–1,050 m) were grouped into four classes and given rating in the range of 10–50 (Fig. 5a). In the same way, the slope values (0 and 73°) were given rating values in the range of 10–40 (Fig. 5b). The weights of the elevation and slope factor were 0.0156 and 0.0905, respectively. Previous studies indicated that Fig. 6 Rating values of soil texture (a) and geomorphological units (b)

groundwater potential zones increases with gentle slope and low topographic elevation (Subba 2006; Solomon 2003; Sener et al. 2005; Shaban 2006). Soil data A digital soil map (1:500,000 scale) was obtained from the Ministry of Agriculture in Jordan. Texturally, the soil cover

Arab J Geosci

in the area consists of eight soil units: silty clay loam to clay, silty clay loam, very stony silty clay loam, often very gravelly, structured silty clay loam, stony and very stony silty clay loam to silty clay, and silty clay loam to clay and sandy clay. The eight soil units were given ratings in the range 15–35, depending on its contribution to the groundwater occurrence (Fig. 6a). The soil texture factor was given a weight of 0.0455. Fig. 7 Map of lithology (a) and its rating values (b)

Geomorphology The main land use types and geomorphological features (wadies, marabs, mudflats, and volcanic uplands) were digitized on-screen with the aid of land use maps from the Royal Jordanian Geographic Center. Al-Qudah (2003) categorized the area of wadi el-Ghusein geomorphologically into three classes: pavement 1, pavement 2, and pavement 3,

Arab J Geosci

corresponding to areas having 99.60, 87.80, and 64.60 % of fragments covering the land, respectively. Wadi el-Ghusein was used as a training area for the supervised classification of the Tulul al Alshaqif. Seven geomorphological units were identified in the study area: pavement 1, pavement 2, pavement 3, mudflat, wadi, marab, and rugged land (volcanic uplands). These units were given a rating value ranging between 5 and 70 (Fig. 6b). The weight of the geomorphology factors is 0.1917.

stations (Tulul al Alshaqif and Safawi) were used to generate a detailed rainfall map taking into consideration topography of the area, since the station in Ruwaished was in a different topographical setting. The average annual rainfall in the Swafwi station is 75 mm and that of Tulul al Alshaqif is 90 mm, and the mean annual rainfall in the study area is in the range of 75–90 mm. These values were grouped into three classes and given ratings in the range of 0–100 (Fig. 8). The rainfall factor weight is 0.0905.

Lithology Groundwater potential model and sensitivity analysis A paper lithology map of 1:250,000 scale from the NRA was digitized in ArcGIS 9.2. It was spatially adjusted (rubbersheet method) because it did not align well with other data. The geological map was investigated for the rock formations and properties. The different lithological units (Fig. 7a) were assigned ratings in the range of 5–70 (Fig. 7b). The most dominant class is basaltic dykes and volcanic rocks (rating, 40). Alluvium and wadi sediments ranks second in area coverage (rating, 70), and alluvium mudflat third (rating, 5). The lithology factor was assigned a weight value of 0.3487. Rainfall There is only one weather station within the boundaries of Tulul al Alshaqif. One weather station is located in Safawi town west of the study area, and a second one is located to the east in Ruwaished town. Only two Fig. 8 Rating values of the average annual rainfall

Using the overlay functionality in the ArcGIS 9.2, the above discussed factors were integrated as thematic layers in the GPM using Eq. 5. The output map consists of an index of groundwater potentiality with values ranging between 14 and 70. These values were classified into five categories showing zones of groundwater occurrence potentiality: very high, high, moderate, low, and very low (Fig. 9a). The map shows that most of the study area is occupied by the moderate potential zone (76.35 %), followed by the high potential zone (12.75 %), low potential zone (6.85 %), very high potential zone (2.2 %), and very low potential zone (1.85 %), respectively (Table 6). It can be also seen clearly that the very high and high potential zones are found everywhere in the study area, but less in the western part. Their occurrence and distribution is mainly governed by the factors: lithology and geomorphology.

Arab J Geosci

The very high and high potential zones are elongated in shape representing stream channels and wadi sediments. This also implies that recharge in such areas is mainly indirect (localized). Moreover, it can be concluded that most of the promising areas are found below 800 m asl. To validate the GPM, locations of dug wells by AbuJaber et al. (2003) were superimposed over the map of

Fig. 9 Groundwater potentiality classes using theoretical weights (a) and using effective weights (b). The location of the hand dug wells is also shown

groundwater potential zones (Fig. 9a). The results showed that two successful wells (nos. 1 and 2) were lying within the very high and high potential zones, whereas the other three failed wells (nos.3, 4, and 5) lie within the moderate potential zone; however, two of them are on the margin toward the low potential zone. These results prove that the model is acceptable in

Arab J Geosci Table 6 Classification of groundwater potentialzones in Tulul al Ashaqif

Table 8 GPM classes area and its percent with and after using effective weights

GPM potential class

Area (%)

Very low Low Moderate High Very high

1.85 6.85 76.35 12.75 2.2

GPM potential class

Area (km2)— Area (%)— Area (km2)— Area (%)— theoretical theoretical effective effective weights weights weights weights

Very low Low Moderate High Very high

34 125.93 1403.70 234.33 40.46

delineating zones of groundwater potential. Sensitivity analysis provides valuable information on the influence of ratings and weights assigned to each factor in a GPM. Two sensitivity tests are usually carried out: the map removal sensitivity analysis and the single parameter sensitivity analysis. The first test identifies the sensitivity of GPM by removing one parameter each time the model is run, whereas the single parameter sensitivity test is used to assess the influence of each of the eight parameters of the GPM on the potential value. In this analysis real or “effective” weight of each parameter was compared with its assigned or “theoretical” weight. The effective weight of a parameter in each cell can be calculated using the following equation (Napolitano and Fabbri 1996): W ¼ Xw  Xs =P  100

ð6Þ

where w and s are, respectively, the weight and rating for the parameter X assigned in each cell, and P is the potential value as computed in Eq. 5. For each cell, the sum parameter effective weights are 100 % for all parameters. In the present study, the single parameter sensitivity test is used, and the statistics of the calculated effective weights or variability for each GPM factor are shown in Table 7. The effective weight for most Table 7 Statistics of the effective weights of the eight factors Parameter

E L G LD R SL S D

Theoretical weight

1.56 34.87 19.17 2.58 9.05 9.05 4.55 19.17

Effective weight Min

Max

Mean

SD

0.35 4.56 2.02 0.86 7.44 1.64 1.49 2.23

5.57 65.65 50.13 8.44 38.38 25.84 9.06 48.57

1.75 38.78 12.48 3.11 15.12 9.76 3.11 15.89

0.70 9.06 6.82 0.96 3.12 2.19 0.86 5.94

L Lithology, G geomorpholgy, S soil, LD lineament density, D drainage density, E elevation, SL slope, R mean annual rainfall

1.85 6.85 76.35 12.75 2.2

5.93 104.05 1304.01 341.21 83.21

0.32 5.66 70.93 18.56 4.53

factors deviates clearly from the assigned theoretical weight. The effective weights of rainfall, lithology, slope, lineament density, and elevation increased by 4.69, 3.91, 0.71, 0.53, and 0.19 %, respectively. The weights of the factors geomorphology, drainage density and soil decreased by 6.07, 3.25, and 1.44 %, respectively. Therefore, theoretical weights of the parameters with large deviations should be revised for computing the GPM. The effective weights obtained by the single parameter sensitivity analysis were used to derive a new map showing groundwater potentiality (Fig. 9b). It was discovered that each class changed in areal coverage. The following classes decreased: very low (1.53 %), low (1.19 %), and moderate (5.42 %). The high and very high classes increased by 5.81 and 5.33 %, respectively (Table 8). Again, we compared the locations of dug wells with this newly derived GPM map. The extracted value for the dug water wells changed accordingly; however, four wells (nos. 1, 2, 4, and 5) remained in the same class, and one well (no. 3) moved from the moderate class to the low class.

Conclusions Remote sensing and GIS techniques proved to be a valuable tool to explore new groundwater resources in arid areas in a way such that time and cost can be saved. The aridity of the area and sparseness of vegetation in the study area were optimum conditions for data extraction from satellite images. It is found that lineaments density, geomorphology, drainage density, and annual rainfall were the most effective indicators of the subsurface conditions and therefore aided in deciphering groundwater condition in the area. This means that zones of groundwater exploration are located more accurately when relevant data are combined and modeled. Most of the very high potential zone were found in stream channels and wadi sediments. Sensitivity analysis enabled us to use the effective weights in computing the final GPM, which yielded an increase in the area covered by the high and very high groundwater potential classes. Hand

Arab J Geosci

dug wells from previous projects were very helpful in model validation. The final model can be used for locating targets for further investigations. Acknowledgment The authors are deeply grateful to Deanship of Scientific Research and Graduate Studies, Yarmouk University for funding this research and offering research facilities.

References Abu-Jaber N, Abderahman N, Azaizeh W, Omari H, Haddadin G, Sokhny A, Hamzeh M, Al Qudah K (2003) Investigation of shallow groundwater in the Tulul al Ashaqif highlands, Jordan. Abhath Al Yarmouk. Basic Sci Eng Series 12(2A):381–400 Abu-Jaber N (2001) Geochemical evolution and recharge of the shallow aquifers at Tulul al Ashaqif, NE Jordan. Environl Geol 41:372–383 Abu-Jaber N, Jawad A, Al Qudah K (1998) Use of solute and isotopic composition of groundwater to constrain the groundwater flow system of the Azraq area, Jordan. Groundw 36:361–365 Al Qudah K, Abu-Jaber N (2009) A GIS database for sustainable management of shallow water resources in the Tulul al Ashaqif region, NE Jordan. Water Resour Manag 23:603–615 Al Qudah K (2003) The influence of long-term landscape stability on flood hydrology and geomorphology evoluting of the valley floor in the northeastern badia of Jordan. Ph.D. thesis, University of Nevada, Reno Battaglin W, Ltay L, Parker R, Leavesley G (1993) Applications of a gis for modeling the sensitivity of water resources to alternations in climate in the gunnisan river basin, Colorado. Water Resour Bull, Am Water Res Assoc 25(6):1021–1028 Boru G (2012) Remote sensing and GIS for mapping groundwater potential zones: delineation of groundwater potential zones of Upper Tumet catchment, Western Ethiopia using remote sensing and GIS. MSc thesis, Addis Ababa University Brito M, Costa C, Almeida J, Vendas D, Verdial P (2006) Characterization of maximum infiltration areas using GIS tools. Eng Geol 85:14–18 Chaterjee RS, Bhattacharya AK (1995) Delineation of drainage pattern of coal basin related inference using satellite remote sensing techniques. Asia Pacific Rem Sen J1:107–114 Chavez S, Velasco G, Bowell J, Sides C, Gonzalez R, & Slotesz L (1997) Groundwater resource evaluation. United States Geological Surveys (USGS)-Open Files Report-Of-96-739. Dar I, Sankar K, Dar M (2010) Remote sensing technology and geographic information system modeling: an integrated approach towards the mapping of groundwater potential zones in Hardrock terrain, Mamundiyar basin. J Hydrol 394(3–4):285–295 Das D, B ehara SC, Kar A, N arendra P, Guha S (1997) Hydrogeomorphological mapping in groundwater exploration using remotely sensed data—a case study in Keonjhar District, Orissa. J Indian Soc Remote Sens 25:247–259 Dissanayake D M D O K (2006) Remote sensing and GIS approach for delineating and characterization of groundwater potential zones in hard rock terrain. Conference Proceedings of Map Asia 2006, Bangkok, Thailand, 24 Aug–1 Sept 2006. Dottridge J, Abu-Jaber N (1999) Groundwater resources and quality in northeastern Jordan: Safe yield and sustainability. Appl Geogr 19(4):313–323 Edet AE, Okereke CS, Teme SC, Esu EO (1998) Application of remote sensing data to groundwater exploration : a case study of the cross civer state, southeastern Nigeria. Hydrogeol J 6:21–30

El-Baz F, El-Ashry M (1991) Groundwater for a thirsty earth. Geotimes 36(6):15–18 Ganapuram S, Vijaya Kumar G, Murali Krishna I, Kahya E, Demirel M (2009) Mapping of groundwater potential zones in the Musi basin using remote sensing data and GIS. Adv Eng Softw 40(7):506–518 Gopinath G, Saralathan P (2004) Identification of groundwater prospective zones using IRS- 1D LISS and pump test methods. J Indian Soc Remote Sens 32:329–342 Heywood I, Corneluis S, Carver S (2003) An introduction to geographic information systems, 1st Indian Ed. Pearson Education, Delhi Higgitt D, Allison R (1998) Characteristics of the basalt boulder surfaces. In: Dutton R, Clarke J, Battikhi A (eds) Arid land resources and their management: Jordan’s desert margin. Kegan Paul International, London, pp 171–182 Humphreys H (1982) Azraq well field evaluation: hydrochemistry and monitoring. Water Authority of Jordan, Amman Ibrahim K (1993) The geological framework for the Harrat Ash-Shaam basaltic supergroup and its volcanotectonic evolution. Natural Resources Authority (Jordan), Geology Directorate Bulletin 25 Ibrahim K, Rabba I, Tarawneh K (2001) Geological and mineralogical occurrence of the northern badia region, Jordan. The Higher Council for Science and Technology, Natural Resourse Authority, Geology Directorate, Geological Mapping Division JGE. (2001). Jordan's water shortage. www.kinghussein.gov.jo/geo_ envi4.html. Kimberley M, Abu-Jaber N (2005) Shallow perched groundwater, a flux of deep CO2, and near-surface water-rock interaction in northeastern Jordan: an example of positive feedback and Darwin’s “warm little pond”. Precambrian Res 137:273–280 Krishnamurthy J, Srinivas G (1995) Role of geological and geomorphological factors in ground water exploration: a study using IRS LISS data. Int J Remote Sens 16:2595–2618 Kuria D, Gachari M, Macharia M, Mungai E (2012) Mapping groundwater potential in Kitui District, Kenyausing geospatial technologies. Int J Water Res Environ Eng 4(1):15–22 Lillesand TM, Kiefer RW (2000) Remote sensing and image interpretation. Wiley, New York Moore G (1982) Groundwater applications of remote sensing. Open file report 82-240. US Department of Interior Geological Survey. EROS Data Center, Sioux Falls, South Dakota Murthy KSR (2000) Groundwater potential in a semi-arid region of Andhra Pradesh—geographical information system approach. J Remote Sens 21(9):1864–1884 Napolitano P, & Fabbri AG (1996) Single parameter sensitivity analysis for aquifer vulnerability assessment using DRASTIC and SINTACS. In Kovar K, Nachtnebel HP (eds) Applications of geographic information systems in hydrology and water resource management. Proceedings of the 2nd HydroGIS conference: International Association of Hydrological Sciences, IAHS Publication 235, pp 559–566 Obeidat M (2000) Hydrochemistry and isotope hydrology of groundwater resources in the Hammad basin (Jordan). Hydrogeologie und Umwelt, Heft 26, Wuerzburg, Germany. Ravindran KV, Jeyaram A (1997) Groundwater prospects of Shahbad Tehsil, basan district, eastern Rajasthan. A remote sensing approach. J of Indian Soc Remote Sens 25:239–246 Rekha VB, Thomas AP (2007) Integrated remote sensing and GIS for groundwater potentially mapping in Koduvan Àr-Sub-Watershed of Meenachil river basin, Kottayam District, Kerala. School of Environmental Sciences, Mahatma Gandhi University, Kerala Saaty T L (1980) The analytic hierarchy process. McGaw Hill, New York Saraf A K (1999) A report on landscape modelling in gis for bankura district, project sponsored by DST, NRDMS division Govt. of India Sener E, Davraz A, Ozcelik M (2005) An integration of GIS and remote sensing in groundwater investigation: a case study in Burdur, Turkey. Hydrogeol J 13:826–834

Arab J Geosci Shaban A (2006) Use of remote sensing and GIS to determine recharge potential zones: the case of Occidental Lebanon. Hydrogeol J 14:433–443 Solomon S (2003) Groundwater study using remote sensing and Geographic Information Systems (GIS) in the central highlands of Eritrea. Doctoral dissertation, Environmental and Natural Resources Information Systems Royal Institute of Technology, Stockholm Sree Devi PD, Srinivasalu S, Kesava Raju K (2001) Hydrogeomorphological and groundwater prospects of the Peregu river basin by using remote sensing data. Environ Geol 40:1088–1094 Srinivas Rao Y, Reddy TVK, Nayudu PT (2000) Groundwater targeting in hard rock terrain using fracture battern modelling, Niva river basin, Andhra Pradesh, India. Hydrogeol J 8:494–502 Srivasthava PK, Bhattacharya AK (2000) Delineation of groundwater potential zones in a hard rock terrain of Baragarh district, Orrissa using IRS data. J Indian Soc Remote Sens 28(2&3):129–140 Subba Rao N (2006) Groundwater potential index in a crystalline terrain using remote sensing data. Environ Geology 50:1067–1076

Teeuw RM (1995) Groundwater exploration using remote sensing and a low-cost geographical information system. Hydrogeol J 3(3):21–30 Toleti B V M, Chaudhary B S, Kumar K E, Saroha G P, Yadav M, Singh A, Sharma M P, Pandey A C, & Singh P K (2000) Integrated groundwater resources mapping in Gurgaon district, (Haryana) India using remote sensing and GIS techniques. The 21st Asian Conf. on Remote Sensing, Taipei, Taiwan, December 4–8, 2000. WAJ (1989) Yarmouk Basin water resources study. Unpubl. Rep. North Jordan Water Resources Investigation. Project, Amman, Jordan WAJ (1995) Groundwater investigation in the Hammad and Sirhan Basins, vol 1. Water Authority of Jordan, Amman Zagana E, Obeidat M, Kuells C, Udluft P (2007) Chloride, hydrochemical and isotope methods of groundwater recharge estimation in eastern Mediterranean areas: a case study in Jordan. Hydrol Processes 21(16):2112–2123