The Use of AHP within GIS in Selecting Potential Sites for Water Harvesting Sites in the Azraq Basin—Jordan

The identification of potential sites for water harvesting is an important step towards maximizing water availability and land productivity in the arid and semi-arid areas. This research aimed to select the optimum sites for water harvesting in the Azraq basin of Jordan through the use of GIS techniques. The Azraq basin is characterized by flash floods that involve large quantities of runoff. The selection criteria in this research were based on six parameters identified based on an extensive literature review. Five experts were then asked to evaluate the importance of each criterion. The consistency ratio between the experts opinions was evaluated using the pairwise comparison method and a final weight was computed for each criterion. A water harvesting suitability map was then generated following the weighted linear combination (WLC) method. The sites that are not suitable for water harvesting within the study area were identified and eliminated following the Boolean method, and final water harvesting suitability map was generated. Finally, the findings of this research can be used to assist in the efficient planning of the water resources management to ensure a sustainable development of the water in Jordan and in other areas suffering from water shortages.

agricultural and domestic users. Jordan is one the poorest four countries in the world in water resource. Jordanians use about 150 m 3 /person/year, compared with a global average at about 1000 m 3 /person/year. According to the Jordan Ministry of Environment [2015], the water scarcity in Jordan is known to be the most important constraints to the Jordan growth and development [1].
Water harvesting techniques can be used to capture and efficiently utilize rainwater runoff to maintain sustainable development of the water in Jordan.
GIS plays a key role in maintaining data and analyzing optimal locations. GIS is a tool that reduces time and cost of the site selection and provides a digital data bank for future monitoring program of the selected sites.
Multi Criteria Decision Analysis (MCDA) in GIS environmental is used to combine layers of spatial data representing the criteria and to specify how the layers are combined. The Analytical Hierarchy Process (AHP) is a method of MCDA that is implemented within GIS, which defines weights for criteria selected.
The use of GIS for identifying the optimum sites for water harvesting schemes has been addressed in many studies [2]- [8].

Research Methodology
The methodology used to select sites for locating suitable water harvesting zones in selected areas of the Azraq basin of Jordan is shown in Figure 1.

Study Area
The Azraq Basin is located in the Northern and central parts of Jordan (Figure 2), with an estimated area of more than 11,052 km 2 . The study area is described to has hot, dry summers and cold winters. Rainfall is in the form of uneven storms of high intensity and short duration; evaporation in this study area is very high.

Selection Criteria Methods
MCDA technique is important for water resources management, which involves choosing criteria and decision options [9]. Several methods of MCDA have been implemented in the GIS environment.  Weighted linear combination (WLC) is a major technique used for site selection within GIS environment. The use of the WLC method for the selection of potential sites for water harvesting has been widely used over the last years. Many of the studies use a WLC method in a GIS environment for the selection of potential sites for water harvesting such as [2] [4] [7] [10]- [13].
Boolean technique, based on the variables are either true or false. The use of Boolean method for identifying the optimum sites for water harvesting projects have been addressed in many studies [2] [3] [14] [15].
AHP is a method of MCDA that is implemented within GIS, which defines weights for criteria. AHP was initially developed by Saaty (1980). Several studies have been carried out for the determination of areas most suitable for groundwater recharge using AHP [16]- [22].
The AHP approach can be used as a set of tools for deriving weights of criteria. The AHP has the ability to deal with inconsistent judgments [23] [24].
The Pairwise Comparison Matrices PCMs involves comparing all the possible pairs of criteria in order to determine which of all the criteria is of a higher priority. The AHP method is based upon the construction of a series of PCMs, which compare all the criteria to one another. Saaty [1980] suggests a scale from 1 to 9 ( Table 1) for PCM elements, where the value of 1 indicates that the criteria are equally important and a value of 9 indicates that the criterion under consideration is extremely important compared to the other criteria. PCM includes a consistency check where judgement errors are identified and a consistency ratio is calculated.
Three main stages to make decisions based on PCM in the AHP method operations [24] [25] are: • The determination of the important criteria in the problem (water harvesting sites). • The assessment of the relative importance of each criterion to each other. This is usually done by experts using a scale from 1 to 9. step includes three main operations [24] [25]: • The assessment of the consistency through pairwise comparisons to assign the Consistency Ratio (CR). This stage involves the following operations [24] [25].
1) Calculating the priority vector for a criterion.
3) Computing the Consistency index (CI). 4) Determining the appropriate value of the random consistency ratio (RI) by

Adopted Selection Criteria
There are many studies concerned with the water harvesting site selection using GIS. These studies were used in this research to define the water harvesting site selection criteria, together with the opinions of local experts.
In addition, seven socio-economic factors are used along with the six criteria in this research to select the optimum sites for water harvesting in the study area. These socio-economic factors have been adopted in several researches [2]-[7] [28]. These factors include, the distances to the urban areas, the farm lands, the roads, the wells, the faults, the wadis and the international borders. These factors represent the major socio-economical activities in the study area. These seven factors were given a zero value for the excluded areas (not suitable) and a value of one for the included ones (suitable). The importance of these factors is:

Selection of the Local Experts
After defining the criteria for selecting sites for the water harvesting, the structured interview was undertaken with local experts. The interviews were carried out in November 2015. The questionnaire, shown in Table 3 was used to identify the relative importance of all the selected criteria. This questionnaire was based on the scale of 1 -9 for the experts to assess the relative importance of each individual criterion. The interview was conducted with 5 experts from Al-al-Bayt university of Jordan that are relevant to the issues of the water harvesting. The background of experts covered the fields of geology, hydrogeology, civil engineering (water resources), groundwater and Geographic Information System (GIS). The experts were selected based on their knowledge of the study area and the water issue in Jordan in general and the study area in particular. Many of the experts had previously published their research [2] [3] [27] [31]- [33]. Interviews were conducted face to face.
Based on the literature review, the criteria were selected and subjected to a review by the local experts to determine the relative importance of each criterion. In this research, an analytic questionnaire survey was carried out to explore the opinions of the experts in the relative importance of the criteria selected for water harvesting using a scale of 1 to 9. The analytic questionnaire was used to check the consistency ratio (CR) and identify the weights for the selected criteria. This involved interviewing the water experts and collecting specific data about the criteria.

Data Collection
The selection sites for the water harvesting require the availability of suitable data; these are both secondary and primary data. the primary data is interviews with experts. The secondary data are collected from various national organizations working in the Jordan. These data are digital maps and other data for different physical and socio-economic aspects of the study area. Table 4 shows the major GIS layers used in this research.

AHP Analysis
Consistency ratio (CR) is calculated using Pairwise Comparison (PWC) technique to assess the consistency between the acquired experts' opinions. PWC was applied to check that the weights for the selection criteria given by the experts are consistent. The traditional implementation of AHP is used in this study [24] [34]- [37]. CR was calculated for all the acquired experts opinions to check if it is less than or equal to 0.1, thereby to check the suitability of each pairwise comparison matrix for the AHP analysis.
The results of the conducted questionnaire are summarised in Table 5. The experts' opinions were selected according to the scale 1 -9, and then the Pairwise Comparison Method (PCM) was applied within the Analytical Hierarchical Process (AHP) to check the CR and to identify the final weights for each criterion. The results of the PWCM (Weights, λ max , CI, RI and CR) is summarised in Table 6.
The CR was less the 0.1 for the pairwise comparison metrics of the experts' evaluation of the site selection criteria represents the consistency ratio of the opinions of the experts. Since the calculated CR is less than or equal to 0.1 for all the experts weightings, this indicates that the acquired experts opinions are consistent, and are suitable for the implementation of the AHP analysis.

Site Selection Criteria (Weights and Rating)
To identify the potential sites for the water harvesting, site selection depends on the rating and the weights of each thematic layer. As stated in this research, opinions of interviewed experts were used to determine the weights of each site selection criterion for the water harvesting and the Analytical Hierarchy Process (AHP) approach used to assess the consistency of the expert opinions by using the consistency ratio(CR), which should be less than or equal to 0.1.
The rating of six physical criteria was selected based on a review of the literature. Using the WLC technique, the rate was assigned to each criterion in the scale of 1 to 4. This is the scale adopted by most of the related literature to date. Table 7 summarizes (weights and rating) the selection criteria for water harvesting projects within the study area.
As discussed in this research, seven implementations representative of socio-economic factors ( Table 8) were taken into account (the international border, wadis, roads, urban areas, faults, wells and farms). These layers were multiplied together after being converted into raster format in ArcGIS. The buffers were then applied on the factors as listed previously. Then, 0 and 1 values were added to the new map containing all the buffered zones. This map was then converted to raster format.

Physical Criteria Analysis
Six physical criteria were used in the research for selecting the suitable sites for the water harvesting recharge are: rainfall, slope, soil clay contents, drainage density, lineament density and Geology. Weights and ratings   were given to each individual criterion as shown in Table 6. The WLC technique was used to integrate these physical criteria. The process of implementing the WLC technique includes standardising the suitability maps, assigning weights of relative importance to the suitability maps, then combining the weights and the standardised suitability maps and obtaining a suitability map. All the generated thematic layers were integrated in ArcGIS® in order to derive a map depicting the suitable areas for the water harvesting of the study area. The total weight of each map of the final integrated layer was computed using Equation 5:

Si= Rw Rr + SLw SLr + SCw SCr + LDw LDr + DDw DDr + Gw Gr
where, "w" represents the weight of each criterion, and "r" represents the rating of each criterion namely: Rainfall (R), Slope (SL), Lineament Density (LD), Soil Clay contents (SC), Drainage Density (DD) and Geology (G). 'Si' is the water harvesting index, which is a dimensionless number that identifies the suitable sites for the water harvesting in the area.  As shown in Table 6, the six GIS layers representing the physical criteria were subjected to a GIS analysis in order to select the optimum sites for the water harvesting in the study area based on these criteria. All maps, the following spatial data techniques were used: • Updating attribute tables according to Raster calculation (integrated to produce the optimum sites for the water harvesting within the study area (Figure 3 and Figure 4)).

Water Harvesting Site Selection Based on the WLC Analysis
According to Equation (1), six thematic layers were added using the WLC method and were classified into five classes of potential water harvesting in the study area, including very low suitability, low suitability, moderate suitability, high suitability, and very high suitability for water harvesting. These thematic layers were integrated to generate a water harvesting suitability map of the study area as shown in Figure 5.

Socio-Economic Criteria
Seven socio-economic criteria (factors) were also integrated into the result shown above. All maps (factors), the following spatial data techniques were used: Updating attribute tables  Raster reclassification  Raster calculation (multiplication). The buffer zones applied for each socio-economic criterion is based on the appropriate buffer distance listed in Table 7. Union in ArcGIS was used to spatially merge the buffer zones of each socio-economic criterion with the study area to incorporate the areas beyond the selected zone distances (Figure 6).
Boolean techniques were applied to the socio-economic factors that cannot be used as sites for water harvesting. The overlay of these factors is illustrated in Figure 7(r). In this figure all areas that are not suitable for harvesting are shown.

Water Harvesting Suitability Mapping
The WLC method was then used to integrate the generated suitability maps of the individual physical criterion in to a one suitability map for water harvesting in the study area. In addition, the Boolean method was used to eliminate the sites that are not suitable for a water harvesting, and generate an unsuitability map. The resultant maps (Figure 8) from the physical criteria, shown in Figure 5 integrated with the socio-economic criteria, shown in Figure 7(r). The study area was classified into five classes based on the minimum and maximum of the criteria maps, the following outcomes for all parts of the study area:  No suitability areas, which represent 7% of the study area.  Very low suitability, which represent 8% of the study area  Moderate suitability, which represent 32% of the study area.  High suitability areas, which represents 26% of the study area.  Very high suitability areas, which represents 27% of the study area Based on [27] [31] [33] [38], it is required to conduct the following steps to validate the outcomes of this research: 1) Surveying the suitable sites to establish if these sites are not occupied with other land uses of high socioeconomic values. This will help in preventing the selection of such sites that have not been known to the researcher when conducting the selection analysis.
2) Geophysical investigation to study the subsurface layers within the study area. This will help in determining wither these layers are suitable to establish a water harvesting schemes above these layers and.
3) Soil sampling from various locations within the study area to test the clay contents. This will help in validating the suitability of soils within the study area to establish water harvesting.    4) Calculate the watershed for the selected sites to determine the amount of runoff that could be utilised in these sites.
In this research, these steps were not conducted due to research fund limitation.

Discussion and Conclusion
In this study, Analytic Hierarchy Process (AHP) was used employed to generate a water harvesting suitability map of a study area located within the Jordan. Six site selection criteria affecting the water harvesting in the study area were defined based on a literature review and discussions with relevant local experts (5 experts). These criteria were the Rainfall, Slope, Soil Clay contents, Drainage Density, Lineament Density and Geology. In addition, seven socio-economic factors that conflict with existing human activities, and thus, affecting the water harvesting were identified based on experts recommendations and literature review. These factors were the international border, wadis, roads, urban areas, faults, wells and farms.
It was found in this study that there was consistency in the experts' opinions by checking the consistency ratio (CR) for experts.
The Weighted Linear Combination (WLC) technique was used to identify the potential sites for water harvesting in the study area. This method is based on the collection of all the criteria after multiplying weights in rating, thereafter determining weights and unifying rating for each criterion. The study area was classified into five classes in terms of the suitability for the water harvesting namely: very low suitability for water harvesting, low suitability for water harvesting, moderately suitable for water harvesting, high suitability for water harvesting and very high suitability for water harvesting. The Boolean technique was then used to eliminate these sites that are not suitable for the water harvesting within the study area including the international border, wadis, roads, urban areas, faults, wells and farms. The Boolean operation resulted in classifying the study area into two Water Harvesting Suitability Based on Boolean and WLC method classes, suitable and not suitable for water harvesting, where the specified sites categories that are not suitable for water harvesting are eliminated. To identify the optimal sites for water harvesting in the study area, the results of the WLC analysis and the results of the Boolean technique were integrated to generate a final water harvesting suitability map of the study area. The study area was classified into no suitability, low suitability, moderate suitability, high suitability, and very high suitability in terms of water harvesting. This research will contribute to the enhancement of the available water resources in the country if the selected sites will be utilised for water harvesting. This will contribute to the sustainable socio-economic development of Jordan. It is recommend to conduct a field work to investigate the selected sites to test the suitability of soil and the sub-surface layers for water harvesting purposes.