Regional overview potential zones for groundwater recharge in Wadi Hodein, south Eastern Desert of Egypt

ABSTRACT Wadi Hodein (WH) is one of the largest arid wadies in the eastern desert, where the groundwater (GW) is the main water resource, to meet the water demand of its residents. Therefore, GW recharge (GWR) potential zones map is to be constructed as an effective tool for GW management in WH. Analytical hierarchy process (AHP) with geospatial analysis using ArcMap was used to delineate the GWR. Six thematic layers (soil map, slope, geology, geomorphology, lineament density, and drainage density) were used to produce the recharge map. The results classified the potential of WH recharging into five categories very low, low, moderate, high, and very high with their corresponding percentage areas of 2.60%, 9.30%, 42.9%, 31.30%, and 14.00%, respectively. Verification analysis using water quality data of the drilled wells inside the study area was conducted to confirm the resulted GWR potential map.


Introduction
Wadi Hodein (WH) is one of the most significant wadis located in the southern part of the eastern desert.It is considered as one of the hyper arid regions, which faces a deficiency of water resources, that could affect different activities such as tourism and mining (Yousif & El Aassar, 2018).It can be assessed that the groundwater (GW) and surface water resources are limited in WH based on the annual precipitation amount over the south of eastern desert of Egypt, which reaches about 20 mm that could be negligible (Ahmed & AbdelMohsen, 2018).It is known that most of the surface runoff does not reach to the outlet of WH due to transport losses (El Bastawesy, White, & Nasr, 2009;Yousef, Salem, Baraka, & Aglan, 2009), so residents depend on GW resources as a main alternative.GW aquifer potential in WH needs to be managed to avoid its depletion.Many authors (Al Gamal, Sayed, Atta, & Nada, 2019;Yousif & El Aassar, 2018) (Embaby, Razack, Porel, & Porel, 2016), (Ismail, El Sayed, & Gomaa, 2006;Sadek, 2007;Yousef, Salem, Baraka, & Aglan, 2009), and (Gomaa, El Fakharany, & Eissa, 2004) concluded that WH has three aquifers: Quaternary, Nubian Sand Stone (NSS), and Basement aquifer.(Al Gamal, Sayed, Atta, & Nada, 2019) classified the GW extraction zones inside WH into three categories: highly depleted zone for Abu -Saafa and El -Dief, moderately depleted zone for Abraq wells, and low depleted zone in El Ghalia and Omeriet wells.(Ismail, El Sayed, & Gomaa, 2006) stated that the basement aquifer has a low GW potential due to weak chance of receiving infiltration from surface runoff while NSS aquifer is the main aquifer which could be detected at Wadi Abu Saafa, El Dief, and Abraq.(Gomaa, El Fakharany, & Eissa, 2004) concluded that salinity of GW in Quaternary aquifer varies from saline to fresh while from brackish to fresh for NSS aquifer.From an unpublished report prepared by The National Water Research Center (NWRC) "Water resources study for some wadies in the eastern desert -Red Sea governorate," it is concluded that the GW is extracted through specified drilled wells concentrated into NSS aquifer and some other hand dug wells distributed randomly in Quaternary and Basement aquifers.The report stated that deterioration for the drilled wells and drying for the hand dug wells are highly expected, which may cause instability for the residents and bedouins.The main objective of the research is to demarcate the GW recharge potential (GWR) zones in WH to categorize the possibility of the recharge (high, moderate, and low) as a part of the strategic plan of the country for achieving the sustainable development of WH and stability for the residents in such strategic region.For achieving the research goal, multicriteria decision-making (MCDM) approach called Analytical Hierarchy Process (AHP) and Geographical Information System (GIS) techniques were integrated.GIS techniques facilitate surface and GW estimation over large areas (Kumar, Mondal, & Ahmed, 2020;Rahmati & Melesse, 2016;Yousif, 2019).Mapping of GWR zones using AHP method was used by many authors among them (Allafta, Opp, & Patra, 2020;Dar, Rai, & Bhat, 2020;Das, Pal, Malik, & Chakraborty, 2018;Saranya & Saravanan, 2020).This method has been proven as an effective and reliable tool by many researchers.Also AHP is recommended to be used in case of lack of data for analysis and validation (Saaty, 2012) (Souissi, Msaddek, Zouhri, & Chenini, Souissi et al., 2018).AHP is a GIS-based MCDM for pairwise comparison of spatial parameters by assigning weights based on expert opinion (Al-Ruzou et al., 2019;Jothibasu & Anbazhagan, 2016).Six thematic layers [Soil type, Slope, Geomorphology, Drainage Density, Lineament Density, and Geology] were utilized for achieving research objective.

Study area
WH is located in the southern part of the eastern desert of Egypt, between latitude 22°00°00°, 24°00° 00° N and longitude 34°00°00°, 36°00°00° E. It covers an area of 11,595 km 2 .WH is considered as one of the Red sea mountains' sub-catchments that drains its runoff water toward Red Sea as shown in Figure 1.It is situated in semi-arid to arid region with maximum rainfall of 10 mm between October and December and almost zero at the rest of the year.The maximum temperature reaches 37.5 C 0 in July and August while the minimum temperature reaches 14 C 0 in December and January.The humidity reaches to its minimum value of 29% in July and August, and the average wind velocity in the study area reaches to about 12 km/h (Rashed & Hassan, 2019).
Geological formations of WH were investigated by many researchers, that is, (Aaglan, 2001;Ghanem, 1972;Ramadan, 1994;Yousef, Salem, Baraka, & Aglan, 2009) and are characterized by three main geological formations: • Pre-Cambrian basement rocks (metamorphic and sedimentary), which cover a large area of WH basin; • Upper Cretaceous Nubian Sandstone (NSS), which covers the north western part of WH and it directly rests on the basement rocks; and • Quaternary deposits, which occupy mostly the eastern part of WH, in addition the main channels.

Material and methodology
Mapping for GWR zones was carried out through spatial analysis for six thematic layers, namely, soil type, geomorphology, geology, drainage density, lineament density, and slope.Figure 2 displays the following methodology to demarcate GW recharge zones in WH.Digital Elevation Model (DEM) was obtained from the Shuttle Radar Topography Mission (SRTM) with the resolution of 90 m.Using ARC-MAP, slope thematic layer was calculated by spatial analysis tool, while drainage density (Dd) was calculated using line density tool.Dd was defined as the total stream length with all order per unit area as given in Equation 1 (Horton, 1932).
where Σ Di is the total stream length in stream order i in Km and A is basin area in Km 2 .Soil and geomorphology maps were extracted as plain rasters from (El Ramady et al., 2019).ARC-MAP was used for digitizing and processing of soil and geomorphology maps.Lineament density and geology maps were obtained from the geological maps of baranis and Jabal Shehab (1:25000).These maps were also digitized and processed using ARC-MAP.

Analytical Hierarchy Process (AHP)
The AHP is considered as one of the MCDM methods that is used in estimating the assigned weight for each different thematic layers (Zghibi et al., 2020).AHP uses a pairwise comparison matrix (PCM) to prioritize each criterion.AHP decomposes the decision problem into four steps: • Determining the factors affecting the decisionmaking; • Constructing the PCM; • Estimating relative weight for each factor; and • Assessing the consistency of the matrix.
In the first step, each factor was given a score from 1 to 9 based on its importance compared to the other factors in the matrix (Saaty, 2008).As shown in Table 1, Score 1 was given for equal relative importance factors while score 9 was given for extremely relative importance (Saaty, 2008).
In the second step, a PCM (m*m) is constructed, where m is the number of factors affecting GWR delineation (Saaty, 2008).Each element in the matrix is assigned a score; 1 when it is compared to itself while its value has a score more or less than 1 based on Saaty's scores when it is compared to another element.Table 2 displays the constructed PCM in a descend order for the six thematic layers influencing GWR in the study area.Soil type parameters were selected as the most important factor influencing GWR compared to the others.Slope parameter was selected as the second most important factor affecting GWR followed by geology, geomorphology, and Dd, and finally lineament density was considered as the least important factor influencing GWR potential.
The third step is to estimate the relative weight of each factor.Normalization for the PCM was performed by dividing the value in each column by the sum value of the same column.A new matrix was produced called a normalized PCM (NPCM) as shown in Table 3.In addition, relative weight for each factor was calculated by the summation of each row.If activity i has one of the above non-zero numbers assigned to it when compared with activity j, then j has the reciprocal value when compared with i The fourth step is to check matrix consistency; consistency vector matrix was calculated by multiplying the PCM with relative weight matrix and then the highest eigen value was calculated by averaging column values of consistency vector matrix as shown in Table 4. Consistency of the PCM was checked based on the consistency index (CI) and consistency ratio (CR).
where CI is the Consistency Index; λ max is the highest matrix Eigen value; n is the number of parameters; CR is the Consistency Ratio; and RI is the Random Index based on number of parameters (Table 5).
The results showed that values of 0.054 for CI and 0.043 for CR with RI of 1.24 were obtained.According to (Saaty, 2008), as long CR is less than 0.1, the PCM of the six thematic layers is consistent.

Producing groundwater recharge zone map
Weighted Overlay Analysis tool in GIS was used to produce GWR zone map.Values of raster pixels were ranked as 1(very low), 2 (Low), 3 (moderate), 4 (high), and 5 (very high) based on their influence on GWR.Pixel value of GWR zone map was calculated by multiplying each pixel rank value by its weighted value as shown in Equation (4).
where GWRP is the pixel value of GWR zone map; Wi is the weighted value of pixel in each raster; Ri is the rank value of pixel in each raster; and i is the raster number.

Verification/Validation analysis
The produced GWR zone map was verified by two different methods: the first method is by comparing the potential of each zone with the existing borewells data.In this research, existing data of the borewells for year 2018 were extracted from (Yousif & El Aassar, 2018).Normally, verification was done based on comparing water level inside the drilled wells (shallow, moderate, and deep) with GWR potential at the same location (Allafta, Opp, & Patra, 2020).In this research, the study area has three different aquifers with no connection between them (Yousif & El Aassar, 2018) and the original water levels map (without discharge) is not available; so, instead, verification will be executed by comparing water quality of wells (fresh, brackish, and saline) with the estimated GWR potential of the study area (Yousif & El Aassar, 2018).The second method is by allocating the recharge ponds executed recently by NWRC for harvesting runoff water and then recharging it to subsoil layers.

Sensitivity analysis
GWR zone map is very sensitive to weight of each raster layer.Sensitivity analysis is performed to assess the impact of each raster ranks and its weights on the GWR zone map (Berhanu & Hatiye, 2020;Rao & Briz-Kishore, 1991).A map removal sensitivity analysis (MRSA) technique was used to check the sensitivity of the output map to the input rasters.

Map Removal Sensitivity Analysis (MRSA)
MRSA was performed to estimate the influence degree of each raster in delineation of GWR zones (Arefin, 2020).The sensitivity index (S) was calculated using equation ( 4): where GWRZ is the output map for GWR zones with all thematic layers; GWRZ* is the output map for GWR zones when removing one of the thematic layers; N is the total number of all thematic layers; and n* is the thematic layers number after removing one of them.

Slope
Topography slope is considered as critical factor that affects the GWR.As the flat and mild slope areas are the most suitable for GWR, the steep slope area has low chance for GWR.Slope raster was extracted from SRTM data using spatial analysis tool of ArcMap as shown in Figure 5. Slope of the study area ranges from 0 to 50%.Slope map has been classified into five categories: (i) plain to flat areas with slopes differs from 0 to 3%.This area covers about 6300 km 2 of study area; (ii) gentle slope areas where slope ranges from 3 to 8% and covers about 3100 km 2 ; (iii) area with moderate slope where slope varies from 8 to 15% and covers about 1300 km 2 ; (iv) steep slope areas where slope varies from 15 to 22% and this area represents very small parts of the study area where area reaches to 540 km 2 ; and (v) very steep slope area with slope more than 22% and this area accounts to 420 km 2 .

Soil
Soil map of WH was obtained from (El Ramady et al., 2019) and then was digitized and processed by ArcMap as shown in Figure 5. WH has many different types of soils such as Mountain, rock out crops, loamy sand, and coarse loamy over sand.Soil type is a dominant factor in GWR process as it controls infiltration potentiality of the study area.Mountain's soil represents about 55% of the total area of WH, which means about one half of WH has very low potentiality for GWR.The rest of soil types has a moderate to high ability of GWR.

Geomorphology
Geomorphology map was digitized and processed by ArcMap to identify the different types of geomorphology in WH as shown in Figure 5. Geomorphology of WH includes Mountains, Outwash Plain, Rock outcrops, Wadi Deposits, Alluvial Deposits, Sand Sheet, Foot Slope, and Delta.Mountains cover about 55% of the study area while wadi deposits account for 17%.Outwash plain and foot slope cover about 12.5% and 9.5%, respectively.While the other geomorphological features account for about less than 6% of WH.

Drainage density
Dd network was extracted from SRTM data as shown in Figure 3 by spatial analysis tool in ArcMap and then Dd map was obtained using line density tool as represented in Figure 5. Dd gives an indication about permeability of the underlying soil as the high Dd indicates for impervious soil with high steep slope.Dd is inversely proportional to GWR (Ahirwar, Malik, Ahirwar, & Shukla, 2020), as the lower Dd value, the higher the GWR potentiality.Dd was classified into five classes, less than 0.15, from 0.15 to 0.25, from 0.25 to 0.35, from 0.35 to 0.45, and more than 0.45 km/km 2 .

Lineament density
Lineament density indirectly represents GWR areas as the existence of these lineaments indicates for a good chance for GWR.Lineaments zones were obtained from the geological maps of baranis and Jabal Shehab (1:25000) and then it is processed by ArcMap and calculated by line density tool.Direction and length of the lineaments were obtained from the geological maps of baranis and Jabal Shehab (1:25000) as shown in Table 6.Lineament density was classified into five classes:    WATER SCIENCE less than 0.03 (Very Low), 0.03-0.08(Low), 0.08-0.12(Moderate), 0.12-0.18(High), and more than 0.18 (V.High) as shown in Figure 5.

Geology
Geological features of WH represent the subsurface layers of the study area.Geological Map of WH was obtained from the geological maps of baranis and Jabal Shehab (1:25000) and then it is digitized and processed by ArcMap as shown in Figure 5.The geological map showed that WH consists of three aquifers: Basement rock, NSS, and Quaternary. Figure 4 shows a conceptual geological cross section for the aquifers of WH.Basement rocks represent about 55% of the geological features of WH.Basement aquifer is composed of igneous rock, which accounts for low chance for GWR; as a result, this aquifer is ranked as very low rank.NSS aquifer represents the western north side of WH.This aquifer is ranked as high chance aquifer for GWR.The third aquifer is the Quaternary aquifer and it is ranked as very high chance aquifer for GWR.

Groundwater recharge potential zone
GWR potential map for WH was created using the ARC-Map as shown in Figure 6 based on weighted overlay concept for the above results of the factors influencing GWR potential.WH was divided into five zones: very low, low, moderate, high, and very high and they cover 2.60%, 9.30%, 42.9%, 31.30%, and 14.00%, respectively.From the aforementioned percentages, it can be concluded that most of WH area is classified as moderate to high GW potential.This can be returned to that about 70% of WH slope is classified as flat to gentle slope.
The very high to high potential zones are located in the southern part of WH with arbitrary distributed zones to the north and at outlet of WH.

Verification analysis
Verification results by the first method as displayed in Figure 7 show a good matching between water quality of the wells (Table 7) and GWR potential zones, as the wells with low total dissolved solids (TDS) were located in very high, high, and moder- ate GWR potentials such as well nos.(7,9,10,11,12).Mismatching was observed in seven wells (Well no. 2,3,4,5,6 and 8).This mismatching could be explained as follows: • For well nos.(2,3,4,5,6), these wells were located near from the coastal line which mean that water quality of these wells is highly affected by sea water salinity.• For well no.(8), well depth reached to 150 m and as mentioned in the drilling report of the well after 120 m, the sandstone is affected by hydrothermal solution which increases the salinity of the GW.(Yousif & El Aassar, 2018).As a result, the produced map could be considered a good estimator for GWR potential zones in WH.
By allocating the recharge ponds recently executed by NWRC, the produced GWR potential map in this research achieved a success prediction for GWR zones as the ponds located in moderate GWR zones as shown in Figure 8.

Sensitivity analysis
Sensitivity analysis results (Table 8) show that GWR potential map is very sensitive for soil type parameter as the coefficient of variation (CV) that equals to the standard deviation (SD) divided by the mean variation index (Avg.)reached to 96.55% followed by geology parameter with CV of 51% followed by slope parameter with CV of 41.63%.These results were logic due to the high theoretical weight given for these layers.It is worth to mention that GWR map was more sensitive for geology than slope despite its lower theoretical weight compared to the slope.
The GWR potential map was found to be moderately sensitive for geomorphology parameter and Dd while it was low sensitive for lineament density with value of 9.35% for the CV.The low sensitivity of the map or lineament density could be referred to the given lower weight.In addition, (Zghibi et al., 2020) stated that the lower lineament density is not necessarily to mean low chance for GWR and the lineament density is considered just as an indicator.

Conclusion and recommendations
GWR potential map was delineated for one of the most important wadies in south of the eastern desert called WH. MCDM using AHP integrated with Arc Map to be used as an efficient tool for achieving the objective of the research.Six thematic layers (soil map, slope, geology, geomorphology, lineament density, and Dd) were considered as input data for the model.The results of the research revealed that WH could be classified into five zones: very low, low, moderate, high, and very high with percentage areas covering 2.60%, 9.30%, 42.9%, 31.30%, and 14.00%, respectively.The model was verified using water quality data for the drilled wells and by allocating the existing recharge ponds inside the study area.The research concluded that WH could be classified into moderate to high GWR potential zone; in addition, the methodology followed in this research proved a good success in demarcating the GWR potential zones in this arid region using a data could be obtained without visiting the study area.As a result, it could be recommended that this methodology could be applied in predicting the GWR potential zone map for the regions having a difficulty in reaching to it.

Figure 1 .
Figure 1.Location map of the study area.

Figure 5 .
Figure 5. Thematic maps used in delineation of GWR potential zone map.

Figure 7 .
Figure 7. Left: Locations of the drilled and hand dug wells in the study area; Right: Zooming to locations of the wells with respect to GWR potential zones.

Figure 8 .
Figure 8. Left: Locations of the existing recharge ponds; Right: Zooming to location the recharge ponds with respect to GWR potential zones.

GWR Potential Zone based on each Thematic layer Published Maps Report of NWRC Estimating Weights of Each thematic layer through AHP method Figure
2. Flow chart of the methodology for mapping GWR zones.

Table 2 .
The pairwise comparison matrix.

Table 3 .
The normalized pairwise comparison matrix .

Table 7 .
Water quality wells of the study area.

Table 8 .
Statistics results of the sensitivity analysis.