An integrated geospatial multi-influencing factor approach to delineate and identify groundwater potential zones in Kabul Province, Afghanistan

This study evaluated the spatial distribution of groundwater potential zones in Kabul Province, Afghanistan using the geospatial multi-influencing factor approach. The influencing parameters employed for the assessment of groundwater potential zones were land slope, geology, soil type, land use/land cover, lineament density, rainfall, and drainage density. The subclasses within each influencing parameter were subdivided based on their influence on groundwater potential as major, minor, and no effect, and were subsequently assigned a score value. The combined score value of these parameters was used for calculating their relative weights. The delineated groundwater potential zones were classified in four groups, i.e., poor, moderate, good, and very good. The study results revealed that zones with a very good groundwater potential covered an area of 355 km2 (2% of the total area), good 1524 km2 (20%), moderate 2251 km2 (73%), and poor 477 km2 (5%). The study concluded that the geospatial-assisted multi-influencing factor approach was very useful and efficient technique for the assessment of groundwater potential zones and can be effectively employed to enhance the conceptual understanding of groundwater resources of Kabul Basin, Afghanistan.


Introduction
Water scarcity and inexpedient management of available water resources is a global concern especially in arid and semi-arid regions of the world. In the current scenario of climate change, unprecedented population growth, rapid urbanization, and increased agricultural and industrial usage, the demand for freshwater is increasing enormously. The situation is further exacerbated by the increasing contamination of surface water, making groundwater more precious and a valuable alternative natural resource (Nicholl 2000;Nedaw and Walraevens 2009).
Almost 1.5 billion of the world population are dependent on groundwater for their daily need. According to the United Nations Educational, Scientific and Cultural Organization (UNESCO) report, by 2025, approximately 1800 million global inhabitants will have to face water scarcity (UNESCO 2006;Murasingh 2014). This is the case in Kabul Province in Afghanistan where inhabitants predominantly rely on groundwater for both domestic use as well as for agricultural irrigation. The population of Kabul has increased from 2.4 to 4.8 million between 2000and 2015(UN 2016. This rapid growth in population and the potential impact of climate change has raised concerns about whether there will be sufficient groundwater available to meet the needs for a substantial proportion of population in Afghanistan in general, and Kabul Province in particular (Mack 2018). In these circumstances, there has been an overwhelming increase in 1 3 453 Page 2 of 13 the demand for groundwater and surface water harvesting (Nasir et al. 2018).
The term groundwater potential refers to the likelihood that water availability below the earth surface in a particular region (Al-Abdi and Shammaa 2014). Incorporation of space-born data and geographic information system (GIS) for groundwater potential is a paradigm shift and significant development in the field of hydrology (Silwal and Pathak 2018). Integration of remote sensing (RS) and GIS for the evaluation of groundwater potential enables the storage, manipulation, analysis, and display of data in various forms and magnitude (Ahmad et al. 2020;Abijith et al. 2020;Raju et al. 2019). Several assessment techniques for delineating groundwater potential zones have been developed recently, and these include single factor analysis (Xin-feng et al. 2012), multifactor analysis (Nag and Kundu 2018), fuzzy-analytical hierarchy process (F-AHP) (Shao, et al. 2020), fuzzy clustering (Ahmad et al. 2020), geographic information fusion systems (Raju et al. 2019), fuzzy-analytical hierarchy process indices (Pinto et al. 2017), the multi-criteria decision making method (Celik 2019), and the multi-influencing approach (Nasir et al. 2018).
For the assessment of groundwater potentiality, various influencing parameters have been used. These include: fault and lineament density; rainfall distribution; altitude; land surface slope; land surface aspect; stream density; land use/ land cover (LULC); geology; geomorphology; physiography; and soil texture (Pinto et al. 2017;Ghorbani-Nejad et al. 2017;Nasir et al. 2018;Mohammadi-Mehzad et al. 2019). Therefore, for the assessment and delineation of groundwater potential zones, the GIS and RS are integrated for creating various thematic parameters' layers with an assigned weight/score in a spatial domain.
A study carried out in 2004 in Kabul Province, Afghanistan, suggested that for an estimated population of 4,089,000 in 2015, the water demand will be around 123.4 million m 3 / year (JICA 2011). The increasing withdrawal has resulted in a declining water table and it is estimated that more than 50% of shallow wells may be dried by 2057. Similarly, the water quality of the well water in urban areas may be degraded due to poor sanitation. The per capita water use in the study area was 110 L/day (Bockh 1971), 50 L/day (Government of Afghanistan 2005 cited by Mack 2018), and 40 L/day (Niard 2007). However, the estimated groundwater availability in the city of Kabul, at about 44 million m 3 /year, can only provide about 2 million people at a modest per capita consumption of 50 L/day (Saffi and Hassan 2011). Several studies have been carried out to monitor the depletion of water table in Kabul Basin. However, the present research is a pioneer study on the assessment of groundwater potentiality of Kabul Basin, and is aimed to delineate groundwater potential zones within Kabul Province with the help of advanced approach of multi-influencing factors integrated in RS and GIS.

Study area
Kabul Province is located in central Afghanistan, and is bounded by Laghman province in the east, Kapisa province in the northeast, Logar province in south, Parwan province in northwest, and Maidan Wardak province in the southwest. It is geographically located between 34° 8΄ 60″ and 34° 54΄ 36″ north latitude and 68° 49΄ 48″ to 69° 57΄ 0″ east longitude. Administratively, the Kabul Province is subdivided into fourteen districts, with Kabul city as the provincial capital. The total area of the province is 4524 Km 2 , sharing only 0.7% of the national land out of 652,225 Km 2 . The province is surrounded by mountains and more than half of the province (56.3%) is mountainous and piedmont, while the remaining (37.7%) is plain area sculptured by river Kabul (JICA 2011). Figure 1 shows a location map of the Kabul Province.
According to the United Nations, Department of Economics and Social Affairs, in 1950, the population of Kabul was 170,784, which increased to 4,221,532 in 2020. It is estimated that since 2015, the population of Kabul has increased by 107,503 persons, with an annual growth rate of 2.61% (UN 2019). According to National Statistics and Information Authority (NISA) estimates, the population of Kabul Province is well over 5 million in 2020 of which 85% are urban inhabitants (Government of Afghanistan 2019).

Methodology
The present study aims to assess and delineate the groundwater potential zones in Kabul Province. Therefore, different influencing parameters' data were used and acquired from various sources. The boundary of the base map was taken from Afghanistan Geodesy and Cartography Head Office (AGCHO), and the slope and stream density layers were generated from Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) Global Digital Elevation Model (GDEM) 2.0, downloaded from NASA website, with 30-m resolution and overall accuracy of 17 m (Tachikawa et al. 2011;ASTER 2011). The LULC and lineament density layers were generated from Landsat OLI, 30-m resolution imagery. The soil map was derived from the existing map of the soil regions of Afghanistan, the geology map from an existing geological and lithological map of Afghanistan, and the rainfall data for past 5 years (2013-2017) for six meteorological stations were collected from the Afghanistan Meteorological Department.
To achieve the study objectives, several influencing parameters were selected and subsequently analyzed by the multi-influencing factor approach. The multi-influencing factor approach is one of the best approaches for the assessment of groundwater potential (Thapa et al. 2017;Nasir et al. 2018;Bhattacharya et al. 2020;Zghibi et al. 2020). It consists of the following steps: The first step was based on literature review for the determination of parameters that influence groundwater potentiality. Geology, soil type, lineament density, slope, rainfall distribution, LULC, and stream density were the seven parameters which were used in the present study.
The second step of the multi-influencing factor approach is to assign scores to selected parameters subclasses and standardization. The subclasses in each influencing parameter were examined for their effectiveness in groundwater aquifer recharge and a suitable score was assigned. Table 1 illustrates the effect of the influencing parameters and their assigned scores. The subclasses which were highly effective (X) in influencing groundwater recharge were assigned a score value of 1, and the subclasses moderately effective (Y) were assigned a score value of 0.5. The subclasses which were not helpful in groundwater recharge were assigned a score value of 0. The cumulative scores (X + Y) of both highly effective (X) and moderately effective (Y) subclasses were considered for calculating the relative effect. The weight of each influencing parameter was calculated using the following formula (Thapa et al. 2017;Zghibi et al. 2020): This step was followed by assigning weight to the selected parameter subclasses. The weight computed for every influencing parameter was divided equally and assigned a rank to every subclass (Gumma and Pavelic 2012). Table 2 shows subclasses of each influencing parameters and their ranks.
The third step was the rasterization and reclassification of all the influencing parameter layers along with the  computed score with the help of the ArcMap 10.5.2 spatial analyst extension. Then, the fourth step was the merging of all influencing parameters and reclassification of the output layers into four groups, viz., very good, good, moderate, and poor groundwater potential zones. Figure 2 shows the methodology adopted for the present study.

Slope
According to Selvam et al. (2015), slope gradient influences the retention of water and percolation capability of the Fig. 2 Flowchart of methodology adopted for delineation of groundwater potential zone of Kabul Province, Afghanistan. GWPZ: Groundwater potential zone; USGS United State Geological Survey; DEM Digital Elevation Model; LULC Land use/land cover surface, which thus is one of the most important parameters for the assessment and delineation of groundwater potential zones. The rate of surface runoff is slow on a gentle slope which increases the rate of infiltration into soils, whereas steep slopes assist rapid runoff and lower infiltration rates (Mumtaz et al. 2019;Sarker et al. 2020). The slope layer was derived from the analysis of ASTER GDEM 2.0, downloaded from https:// search. earth data. nasa. gov with 30-m resolution.
The slope was categorized into four classes, which were: flat slopes (0 • -25°) with low runoff and good infiltration, gentle slopes (26°-50°) can be considered to produce moderate infiltration, moderate slopes (51°-75°) which have poor infiltration rates, and high slopes (76°-89.99°) (very poor) that have very low infiltration rates and a high rate of runoff (Table 2). Figure 3(A) shows the slope classes of the study area.

Geology/rock type
The geology of an area has a strong influence on the amount of recharge an aquifer receives and thus is important in the delineation of groundwater potential zones (Ramu et al. 2014). The rocks storage capacity relies on its porosity and permeability. The water moves from a groundwater recharge zone to a discharge zone under a hydraulic gradient which depends on the hydraulic conductivity or permeability of geological formation (Manikandan et al. 2014). Geological information for Kabul Province was acquired from the geological and lithological map of Afghanistan. The geology map was scanned, imported to ArcMap 10.5.2., and georeferenced. The area of interest was extracted by masking, digitizing, and rasterizing the map. The captured information was subsequently reclassified, and weights were assigned to various subclasses based on their groundwater recharge potential.
In Kabul Province, almost half of the area is underlain by sedimentary rocks. For the present study, Kabul Province was categorized into four geological units which were sedimentary rocks; igneous rocks; igneous and metamorphic rocks; and metamorphic rocks ( Table 2). The highest score was assigned to sedimentary rocks due to their high porosity and permeability. Figure 3(B) shows the geology/rock formation of the study area.

Rainfall
Aquifers are usually recharged by the infiltration of rainfall. Therefore, the amount and frequency of rainfall a region experiences are the important considerations for the assessment and delineation of groundwater potential zones which depends on the environmental condition of an area (Minh et al. 2019;Shao et al. 2020). Typically, groundwater prospects are higher in high rainfall regions, and poorer in low rainfall areas.
In this study, the annual mean rainfall data from 2013 to 2017 for six meteorological stations, including Laghman, Kabul, Mazari Sharif, Jabul Saraj, Jalalabad, and Kunduz were obtained from the Afghanistan Meteorological Department. The rainfall data were interpolated spatially and then reclassified in ArcGIS 10.5.2 software. The study area was categorized into five classes which were: (i) 316-328.6 mm, (ii) 206-315 mm, (iii) 291-205 mm, (iv) 281-290 mm, and (v) 271-280 mm (Table 2). A high score was assigned to high rainfall regions. Figure 3(C) depicts the rainfall regions of the study area.

Lineament density
A lineament is the linear feature on the earth surface which is the surface expression of underlying structural features in bedrock. Lineaments are geological structures (i.e., fractures, faults, joints, and other discontinuity surfaces) which can be detected by remote sensing images (O'Leary et al. 1976). According to Nag (2005), lineament density is a significant parameter for the assessment and delineation of groundwater potential zones.
In the present study, lineaments were generated and extracted from the panchromatic band 8 of Landsat 8 OLI using the software PCI Geomatica 2017. Subsequently, the lineament density was calculated using the density analysis tool of ArcMap 10.5.2. The analysis revealed that Kabul Province is covered by major and minor lineaments that vary in length from 0.15 km to 1.52 km. The computed lineament density layer was then classified into five density zones. The area with the highest lineament density (0.95-1.52) was assigned a high score (very good) due to its effectiveness in groundwater recharge potential, whereas the area with a very low lineament density (0-0.15) was considered to have a poor groundwater potential and was assigned with the lowest score ( Table 2). The lineament density map is illustrated in Fig. 3(D).

Drainage density
According to Strahler (1964), drainage density indicates the total length of the streams per unit area of a watershed. It is expressed as the length of streams per km 2 . Drainage density is also a good indicator of good groundwater potentiality due to its association of the permeability of underlying rocks and runoff (Magesh et al. 2012). Typically, drainage density is lowest in areas where underlying sediments or rocks have a high permeability, and vice versa.
For present study, the drainage network was generated from ASTER GDEM 2.0 (30 m resolution), using the Archydro tool of ArcMap 10.5.2. The drainage network was then used to compute the drainage density using the line density analysis tool of ArcMap 10.5.2. The stream density layer was than rasterized and reclassified into five classes, i.e., 0.07-1.3 (very poor), 1.4-3.3 (poor), 3.4-5.3 (moderate), 5.4-7.8 (good), and 7.9-9.4 (very good) km/ km 2 , respectively ( Table 2). The drainage density is indirectly proportional to the terrain perviousness, and therefore, the high score value was attributed to the poor and very poor drainage density classes, due to the fact that it is a strong indicator of water retention zones, i.e., high infiltration and low runoff (Mahmoud and Alazba 2016). Figure 3(E) illustrates the drainage density of Kabul Province.

Land use/land cover (LULC)
Land use/land cover is the most important humaninduced parameter responsible for the occurrence and groundwater aquifer recharge potential. The LULC change could alter the aquifer recharge rate, which could have negative impact on groundwater potentiality (Chen et al. 2019). The land use refers to the anthropogenic activities relevant to specific piece of land, while land cover referred to the natural covering of earth surface (Lillesand and Kiefer 1979). The LULC of the Kabul Province was derived from a Landsat 8 OLI satellite image downloaded from United State Geological Survey (USGS) web site for the month of April 2019, having a 30-m resolution. The LULC was obtained from this image using a supervised classification algorithm that was undertaken using the spatial analyst tool of ArcMap 10.5.2. The image was classified into six LULC classes, i.e., water bodies, natural vegetation, agriculture land, bush land, built-up areas, and barren land. Figure 3(F) shows the LULC map of the Kabul Province, Afghanistan.

Soil
Soil texture plays a crucial role in the water transport processes and controls the infiltration of surface water and recharge to aquifers (Ahmad et al. 2020). Soil texture refers to the relative percentage of silt, clay, and sand within a soil layer. Soil texture is related to the soil porosity which in turn affects the water holding capacity and infiltration capability of soil. The infiltration rate is comparatively lower in fine textured soil than in coarse-textured soil (Ma et al. 2016).
The textures of soils in Kabul Province were acquired through the existing map of the soil regions of Afghanistan. The map was imported to ArcMap 10.5.2 and georeferenced. The area of interest was extracted, digitized, rasterized, and ultimately reclassified into five classes. The five classes were Haplocambids with Torriorthents (very good prospect), rock outcrops with Lithic cryorthents (good), rock outcrops with Lithic Haplocambids (moderate), rock outcrops with Lithic Haplocryids (poor), and Xerochrepts with Xerorthents (very poor). Figure 3(G) depicts the soil type map of Kabul Province.

Groundwater potential
The groundwater potential of the region was computed using seven weighted thematic parameters layers that were integrated into ArcGIS 10.5.2 software. The ranking and weights were assigned to different thematic layers using the multi-influencing factor approach. The parameters used were geology, rainfall, slope, lineament, soil, LULC, and drainage network. The generated groundwater potential layer was reclassified into four groundwater potential zones which were: (i) very good, (ii) good, (iii) moderate, and (iv) poor groundwater potential. Figure 4 shows the map of groundwater potential zones of Kabul Province.
The map revealed that areas with high prospects of obtaining groundwater were restricted to the central districts of Kabul, Dih Sabz, and Bagrami districts (Wilayat), while the eastern and western districts have the lowest groundwater potential. The eastern and western districts are mostly mountainous. Kabul is surrounded by Koh-e-Paghman mountain and Koh-e-Orough mountain in the east and southwest and Koh-e-Shirdarwaza in the northeast. The spatial extents of  Fig. 4. The analysis revealed that very good groundwater potential zones cover an area of 354.87 km 2 , good 1523.86 km 2 , moderate 2250.99 km 2 , and poor 477.19 km 2 . Figure 5 shows the groundwater potential in the various districts in Kabul Province.
The present study is the first study to delineate the groundwater potential areas of Kabul Province. Previous research was mainly limited to determine the recharge potential of the Kabul Basin (e.g., Akbari et al. 2007Akbari et al. , 2008Mack et al. 2009). Mack et al. (2009) presented the results of their research conducted between 2005 and 2007 regarding the water availability for the growing population and the potential impact of climate change in Kabul Province of Afghanistan. The aquifer recharge in the basin is highly variable both temporally and spatially. The high recharge takes place near irrigated agricultural land, streams, and rivers. In these areas, the recharge rate is estimated to be 1.2 × 10 -3 m/day, whereas at lower altitudes in the areas away from streams and rivers, the recharge may be about 0.7 × 10 -3 m/day. During 2009, the amount of water needed for use in the province was 112,000 cubic meter/day, which is likely to increase to 725,000 cubic meter/day by the year 2057. This research is in line with the present study results. The majority of the groundwater potential zones delineated by the current research are located in the vicinity of rivers and streams. Very good and good groundwater prospects occur in the central districts of Kabul Province where there is a high aquifer recharge capability (Figs. 4,5,6).

Validation of study
To validate the results of this study, well data were acquired from the National Groundwater Monitoring Wells Network for Afghanistan. The monitoring network that was established by the Afghanistan Geological Survey, the Hydrology Group, and the United States Geological Survey has monitored the groundwater table in the Kabul Province since 2004 (Akbari et al. 2007). This network consists of a total of 148 monitoring wells, from the Kabul Basin which are spatially distributed in the five central districts. Of these wells, 61 wells were selected for the validation of results in this study. Most of these wells (44) are concentrated in the Kabul district where the water depth ranges from 9.5 to 90 m. The well location map was derived from USGS Scientific Investigation Report 2009-5262 (Mack, et al. 2009). The map was georeferenced, and well locations were digitized and overlaid on top of the delineated groundwater potential zones computed through multi-influencing factor technique in Arc-GIS 10.5.2 spatial analyst. Figure 7 shows the location and number of well superimposed on top of the groundwater potential zones. The analysis revealed that the majority of well falls in the very good and good groundwater potential zones. None of the well fell in locations where there is a low groundwater potential. Of the 61 total wells, 49 wells (80.30% of total sample wells) fell in areas with a very good or good groundwater potential. Additionally, wells that were located in areas with a very good or good groundwater potential were shallow and had a water depth that ranged from 9.5 to 35 m, which further validates the study results.

Conclusions
Groundwater potential map was developed for Kabul Province in Afghanistan based on seven parameters that were assumed to influence groundwater availability in a region/ satellite images and secondary data acquired from various sources were employed to generate the influencing parameter layers of geology, drainage density, soil, rainfall, land use/land cover, slope, and lineament density. The generated layers were rasterized in ArcMap 10.5.2. using the feature to raster convertor tool. The rasterized thematic layers were reclassified and assigned the appropriate scores and weights depending on their effectiveness in influencing groundwater recharge using the multi-influencing factor approach. Additionally, each weighted parameter layer was finally added together using raster calculator to compute the final Fig. 7 The location and the number of well superimposed on top of the groundwater potential zones delineated through multi-influencing factor techniques in ArcGIS 10.5.2. GWPZ Groundwater potential zone groundwater potential zone map of the study area. This map classified Kabul Province into four groundwater potential zones, namely: very good, good, moderate, and poor. The generated map revealed that most of the area (73%) of Kabul Province has a moderate groundwater potential.
The generated groundwater potential map was verified using the well location map from USGS Scientific Investigation Report 2009-5262. The validation suggested that the geospatial multi-influencing factor approach is an efficient tool for the assessment and delineation of groundwater potential zones in Kabul Province, Afghanistan. It is a cost-and time-effective technique compared to conventional methods. The adopted methodology is empirical in nature and is the most widely used technique for the assessment and delineation of groundwater potential zones. The results of the present study will be valuable for improving groundwater management in the region and could be utilized for future planning of an aquifer storage and recovery (ASR) program in Kabul Province, Afghanistan.