Assessing Groundwater Potential in the Kabul River Basin of Pakistan: A GIS and Analytical Hierarchy Process Approach for Sustainable Water Management

Abstract

The rapid urbanization in the Kabul River Basin has increased the demand for water for both drinking and commercial purposes, leading to domestic and industrial water insecurity. Assessing the groundwater potential of the Kabul River Basin is highly crucial for effective water management. The aim of this paper is to identify potential zones for groundwater by employing a Geographic Information System and an Analytical Hierarchy Process approach to formulate a cumulative score based on seven thematic images—rainfall, geology, lineament density, drainage density, land use/land cover, soil type, and slope—within the Kabul River, with assigned weightages of 32%, 27%, 12%, 10%, 8%, 6%, and 5%, respectively, with a consistency ratio of 0.053 (5%), demonstrating the reliability of the results. The study shows that the first three factors contribute more to the percentages of Groundwater Potential Zones. The identified groundwater potential is classified into very good, good, medium, poor, and very poor zones, covering 35.45% (19,989 km²), 37.2% (20,978 km²), 23.16% (13,063 km²), 4.13% (2332 km²), and 0.06% (19 km²), respectively. Groundwater potential in the basin is predominantly classified as good to medium; however, there are notable variations across sub-basins. The Swat sub-basin and western parts of the Kabul River Basin, encompassing the Panjshir and Parwan districts, exhibit exceptionally high groundwater potential. In contrast, the Panjkora sub-basin (Dir district) and southwestern areas of the Kabul River Basin, covering parts of the Ghazni and Wardak districts, have very limited groundwater potential.

1. Introduction

Groundwater is the most valuable natural resource that protects human health and supports economic development. In many urban areas, groundwater is an important water supply source due to its continuous supply and excellent quality [1]. This demand becomes more prominent owing to the economic development, growing population, and land use changes made with no consideration of the balance of groundwater [2]. Globally, groundwater extraction used for domestic, agricultural, and industrial use accounts for 36%, 42%, and 27%, respectively [3]. In Pakistan, about 90% of groundwater is used in rural areas for domestic purposes and 50% for agricultural purposes. Due to the lack of a sustainable groundwater management program, the groundwater remains in a deteriorating condition [4]. In addition to rising domestic demand, socio-economic growth and an increase in population growth have created a threat to the inhabitants of different provinces in Pakistan [5,6]. Therefore, it is necessary to identify Groundwater Potential Zones to counteract the demand for freshwater, industrial use, and other agricultural use [3]. Groundwater plays a key role in the socio-economic growth of the Kabul River Basin, both in Pakistan and Afghanistan. As groundwater in the basin has not been properly managed, sustainable management is needed to assess Groundwater Potential Zones [7]. Groundwater in the subsurface is explored through a range of methods, from the oldest water dowsing method to modern tools and electromagnetic resonance technologies. In general, there are two classes of groundwater exploration, that is, surface and subsurface. Surface groundwater exploration is the least costly, and no labor work is required. These methods include geomorphological, soil and microbiological, surface geophysical, remote sensing, and esoteric methods. On the other hand, subsurface explorations intended for groundwater exploration within aquifers need investments in drilling wells and effective application of the subsurface method. However, it should be made clear that not all subsurface water is groundwater. For example, soil moisture and water in the unsaturated zone do not qualify as groundwater. Therefore, it is a useful practice to explore Groundwater Potential Zones through the surface investigation method [8].

Remote sensing and GIS-based investigations have become the most powerful, rapid, and cost-effective methods for the exploration of groundwater mapping. The incorporation of these tools is used for the evaluation, preservation, and monitoring of groundwater resources [9,10,11,12,13]. Several methods have been employed to investigate the Groundwater Potential Zones, i.e., the Analytical Hierarchy Process (AHP) [14,15,16], multifactor analysis [17], fuzzy modeling [18,19], and data envelopment analysis [20]. Among these methods, the combination of GIS and AHP has received much attention due to its strong decision-making capabilities [11,21,22,23,24,25]. The method was developed by hierarchically evaluating, planning, and weighting the criteria based on similar research studies and expert opinions [26]. In groundwater potential zoning, the decision-making criteria employed is based on multi-criteria decision analysis (MCDA), incorporating many factors (e.g., geomorphology, lithology, slope, soil texture, land use/land cover, drainage density, lineament density, and rainfall), reflecting the hydrologic, geologic, hydrogeological, and metrological features of the basin [13,27,28,29]. The goal of MCDA is to determine an option for spatially predicting potential zones for groundwater resources [30]. Ref. [31] also delineated GWPZ by applying the Fuzzy-AHP method with RS and GIS for the semi-arid Shanxi Province, China. This study concluded that the Fuzzy-AHP model generated findings that aligned well with ground-truth data, demonstrating that this method is proficient and efficient for the identification of GWPZ. Ref. [15] identified GWPZ using the RS, GIS, and AHP approaches in the transboundary Shatt Al-Arab Basin. Consequently, this study was developed to help authorities in developing policies and future plans for managing the groundwater.

The Kabul River Basin, a transboundary basin shared by Pakistan and Afghanistan, is situated in an arid water-scarce zone [32]. The Kabul River Basin comprises a total of 13 provinces, whose dwellers face groundwater scarcity due to numerous factors, including increased irrigation water demand, growth in urbanization and population, lack of future planning and management of water resources, climate change, and inadequate awareness. In most of its sub-basins, a very large number of new bores are constructed without following any strategies offered by the government, which enhances the threat to the availability of freshwater [7]. Additionally, due to the increased number of bores, the level of groundwater has depleted by 1.7 to 3 m from 2008 to 2016 and as a result, about 33% of boreholes are not functional [33]. Recent studies on groundwater depletion in the regions have labeled this basin as “alarming”. Ref. [34] found that groundwater is declining by around 4 m per year in certain parts of central Kabul City, while averaging 1.5 m per year throughout the whole urban area of Kabul City. The paper concluded that the decline is alarmingly larger than other regions in the country. Ref. [35] concluded that the northern Kabul River Basin is one of the regions with the highest rate of groundwater depletion. The paper emphasized the stress on underground water resources in the Kabul River Basin. Additionally, surface water in the Kabul River is contaminated due to discharges of untreated sewage from the settled areas in Kabul City and cities in Khyber Pakhtunkhwa, i.e., Peshawar, Nowshera, Pabbi, and Bara. This contamination affects the surface water directly and the groundwater indirectly [36]. Hence, it is very important to assess the basin-wide groundwater situation both qualitatively and quantitatively in the Kabul River Basin.

The Kabul River Basin is one of the three major river basins in Afghanistan, along with the Amu Darya River Basin and the Helmand River Basin. These basins have been readily used to draw water from underground due to the lack of adequate surface water supply projects. As agriculture is the primary source of earning across these basins, including the upper part of the Kabul River Basin, a lot of groundwater is consumed for irrigation purposes, without proper management. These basins have not been studied well for ground and surface water exploration due to the Afghan war (1978 to 2021); however, international agencies have investigated aspects such as climatology, agricultural management, hydrology, and land use of the region using advanced satellite observations [37,38]. Consequently, due to the availability of new datasets, certain studies [39,40] were conducted in order to address the water-scarcity situation in Afghanistan. These studies focused on Afghanistan and were based on groundwater recharge and modeling of selective rivers, i.e., the Helmand River Basin. However, the Kabul River Basin received little attention. Despite KRB being strategically more important, as it encompasses major urban cities in Afghanistan, such as Kabul and Jalalabad, it remains less studied compared to its downstream counterpart, i.e., the Indus River Basin, which is very intensively studied due to its geopolitical and economic importance and international research programs, such as those by the World Bank, IWMI, and NASA. Even being part of the Indus River Basin, the KRB has received less attention due to border instability between Pakistan and Afghanistan, a lack of data acquisition, and the American war on terror. Consequently, researchers have focused on localized studies. One such study, Ref. [41], studied the Indus River Basin for ground and surface water interactions, especially in Punjab. But this study potentially neglected the Kabul River Basin, which is a tributary of the Indus River Basin. Most recently, another study, Ref. [42], targeted other basin in Afghanistan, i.e., Amu Darya Basin, which is the northwestern basin; the authors studied Groundwater Potential Zones in the Amu Darya River Basin. Additionally, the authors of [43] have conducted groundwater potential zoning for Kabul province only, and not Kabul River Basin as a whole. Ref. [7] assessed the GWPZ using the GIS-AHP method for the Kabul River Basin (KRB) only for Afghanistan and did not take into account the downstream consumption in the basin. During the research, it was found that only 18% of the total annual average rainfall contributes to the recharge of groundwater. Ref. [44] studied the small area of the Swat River, i.e., the Swat district only, which is another sub-basin of the Kabul River Basin. Although there are a number of localized groundwater studies, many scholars [45,46,47,48,49] have reinforced the need for basin-scale assessments for sustainable water management. Subsequently, the Kabul River Basin has been identified as a critical region for integrated basin-wide groundwater evaluation to address water scarcity, groundwater depletion, and management challenges. Therefore, it is necessary to delineate the Groundwater Potential Zones using remote sensing and GIS to assess and create a guide map of groundwater exploration for future generations and to ensure that this natural resource will be sustainably and optimally operated and controlled [7]. This study focuses on basin-wide analysis because the localized studies cannot address the complexity of the interconnected dynamics of the basin system.

2. Materials and Methods

2.1. Description of Study Area

The Kabul River Basin (KRB), spanning Afghanistan and Pakistan, stretches from 68.8392° east to 34.357° north, with a total watershed area of 87,499 km², out of which 40,064 km² is located in Afghanistan and the remaining is located in Pakistan. The Kabul River starts from the Hindu Kush Mountains and flows about 700 km eastward, and drains into the Indus River in Punjab, Pakistan [50]. The Kabul Basin, located in Afghanistan, is divided into the following sub-basins: Paghman River Basin, Logar River Basin, Kunar River Basin, Salang, Ghorband, Panjshir River Basin, Alishang, and Alingar River Basin [51,52], while the part of the KRB located in Pakistan is divided into four sub-basins, namely, the Kabul River Basin, the Chitral River Basin, the Bara River Basin, and the Swat River Basin [50].

In the KRB, 72% of runoff is from seasonal snowmelt and serves around 9 million people in Pakistan and Afghanistan [53]. The average precipitation observed in the winter months is 110 mm, with an average maximum temperature of 28 °C and an average minimum temperature of −6 °C [51]. According to census records from 1951 to 2017, the total population increase within the KRB is approximately 85.7%. The population increase plays a key role in the increase in anthropogenic activities, leading to surface water contamination and wild use of groundwater, causing a depletion in groundwater levels [50].

The geographical location of the Kabul River Basin and the 30-m digital elevation model from the USGS’s Shuttle Radar Topography Mission (DEM_SRTM), along with all hydrological stations, are illustrated in Figure 1, while the district-wise distribution of the Kabul River Basin is given in Figure 2.

2.2. Factors Influencing the Groundwater Potential Zones

Many factors affect the recharge and occurrence of groundwater in any specific region [55]. These factors are represented in maps using equal scales and are categorized into suitable classes using a multi-criteria decision-making method. The same approach is used in the present study; seven thematic layers are chosen to delineate Groundwater Potential Zones. These factors are soil type, land use/land cover (LULC), rainfall, lineament density, geology, drainage density, and slope. These factors are listed as conventional and remote sensing maps in Figure 3. All these factors play a major role in classifying Groundwater Potential Zones [32]. The data were collected from different reliable sources, which are presented in

Table 1. Required data and sources.

S. No.	Data Collected	Source	Description
1	Soil map	Food and Agriculture Organization of the United Nations	Scale: 1:5,000,000
2	Land use/land cover	USGS, Landsat	30 m grid
3	Rainfall	NASA, LaRC Power Project	---
4	Lineament density	USGS, Landsat	30 m grid
5	Geology	USGS, CERT Mapper	30 m grid
6	Drainage density	USGS, Landsat	30 m grid
7	Slope	USGS, SRTM	30 m grid

.

The annual rainfall data were obtained for 12 stations (7 from Pakistan, 5 from Afghanistan) from 2000 to 2022 from the online NASA Langley Research Centre (LaRC) POWER Project funded through the NASA Earth Science/Applied Science Program in

Table 2. Average precipitation recorded at different stations in KRB.

Gauging Station	Latitude	Longitude	Avg_Rainfall
Cherat	34	72	427
Chitral	36	72	415
Dir	35	72	127
Drosh	36	72	619
Saidu Sharif	35	72	1474
Kalam	35	73	639
Malam Jabba	35	73	728
Peshawar	34	72	817
North Salang	35	69	990
South Salang	35	69	1036
Paghman	34	69	437
Kabul	34	69	299

. The digital elevation model with 30 m resolution for the study area, downloaded from USGS, was used to reveal the slope and drainage density of the area. The authors of [56] showed that a 30 m resolution is well suited for the broader groundwater potential trends and not sufficient for fine-scale site-specific variability. The researchers of [57] concluded that the use of a resolution of 10 m or less increases the computational load, severely affecting data processing in such basin-level studies. Refs. [9,58] successfully demonstrated that 30 m resolution satellite imagery can sufficiently capture topographic and land-use variations in both complex and simple terrains at the watershed/basin level with acceptable accuracy. For this particular study, LULC data were obtained from the USGS Landsat collection.

The mentioned thematic layers play an important role in the analysis of GWPZ. Lineament density represents the topography of subsurface faults and fractures in the region. As these fractures increase the rate of infiltration, areas with higher lineament density have higher potential for groundwater [59]. Lineament density is the total length of lineaments in a unit area of the watershed [60] and is expressed as follows:

$$Lineament Density=\sum _{i=1}^{i=n}{\displaystyle \frac{L_i}{A}}$$

(1)

where Li denotes the total length of lineaments, and A represents the total area of the watershed.

Drainage density is another thematic layer that can provide information about both surface and subsurface formations. The lower the drainage density, the greater the ability of the surface to absorb surface runoff, because the drainage density of an area depends on the type of vegetation, structure, nature of the bedrock, soil infiltration rate, and slope [61]. The drainage density can be defined as the total length of streams to the area of the grid, and is expressed as follows:

$$Drainage Density=\sum _{i=1}^{i=n}{\displaystyle \frac{D_i}{A}}$$

(2)

where D_i is the total length of streams present in the mesh and A is the total area of the grid (km²).

The lithology of a region also has a strong impact on groundwater recharge [62]. Sandstone, lime stone, dolomite, colluvium, fan alluvium, basalt, volcanic, and sedimentary rocks are considered the best rocks for groundwater recharge. Therefore, these types of stones have a high weight value.

Rainfall is another important thematic layer used in delineating Groundwater Potential Zones. Rainfall is a major component of the water cycle and causes groundwater recharge. The higher the rainfall in the region, the greater the groundwater recharge [3]. The rainfall map was produced by spatially distributing point precipitation data using the Kriging method in ArcGIS 10.3, as the inverse weighing distance method can be problematic if sampling points are limited.

The slope of any region is also an important feature affecting water movement on the surface and subsurface topography, directly affecting the groundwater recharge. The steeper the slope, the larger the surface runoff and the lower the infiltration rate [63]. The most favorable surface for higher groundwater recharge will be the flat land.

The land use/land cover map provides detailed information about the nature of the land. Land surfaces with water bodies, snow covers, crop lands, and other unpaved bare land are considered the most favorable surfaces for groundwater recharge and generation, while paved surfaces are the least favorable surfaces for groundwater generation [23].

The last thematic layer for the AHP method is the type of soil, which plays a key role in groundwater recharge [64].

All these layers are geo-referenced using the Universal Transverse Mercator (UTM), Zone 36 N projection system, and the World Geodetic System (WGS 1984). After the necessary changes are made to the symbology of each layer, the obtained layer is converted to a raster format.

2.3. Analytical Hierarchy Process (AHP)

The Analytical Hierarchy Process is a sub-technique of the multi-criteria decision method (MCDM) established to analyze the qualitative measures in a systematic and scientific approach. The method is responsible for comparing the thematic layer parameters in a hierarchical structure to assess the comparative weightages. The AHP method was developed by Thomas L. Saaty, also known as Saaty’s method [65,66], commonly employed in the last few decades for the assessment of the groundwater potential. After the generation of thematic layers, as shown in Figure 3, the first step of the AHP method is to create a pair-wise comparison matrix by selecting the important criteria based on judgment and expert opinions [26]. Using Saaty’s scale from 1 to 9 (

Table 3. Saaty’s scale of relative importance or pair-wise comparison scale [66].

Saaty’s Scale Values	Description
1	Equally important
2	Equally to moderately important
3	Moderately important
4	Moderately to strongly important
5	Strongly important
6	Strongly to very strongly important
7	Very Strongly important
8	Very strongly to extremely important
9	Extremely important

), the criteria of the complex decision-making process are reduced to a single level, where 1 indicates equally important and 9 is extremely important.

$$P=\left[\begin{array}{c}\begin{array}{ccc}{A}_{11}& A_{12}& {\dots A}_{1n}\\ A_{21}& A_{22}& \dots A_{2n}\\ {\begin{array}{c}⋮\\ A\end{array}}_{1n}& {\begin{array}{c}⋮\\ A\end{array}}_{2n}& \dots {\begin{array}{c}⋮\\ A\end{array}}_{nn}\end{array}\end{array}\right]$$

(3)

where $A_{ij}$ is an element of the criteria comparison matrix

The second step is to calculate the normalized weights (

Table 4. Comparison matrix of 7 thematic layers for normalized weights.

Parameters	Rainfall	Geology	Lineament Density	Drainage Density	Slope	Soil Type	LULC	$\mathbf{Normalized} {\mathit{W}}_{\mathit{n}}$
Rainfall	1	1	5	4	6	5	2	31.86%
Geology	1	1	3	3	4	6	3	27.71%
Lineament density	1/5	1/3	1	1	3	3	2	11.65%
Drainage density	1/4	1/3	1	1	3	1	2	10.25%
Slope	1/6	1/4	1/3	1/3	1	1	1	5.17%
Soil type	1/5	1/6	1/3	1	1	1	1	5.93%
LULC	1/2	1/3	1/2	½	1	1	1	7.43%
Sum	3.32	3.4	11.67	10.83	19	18	12

) by calculating the geometric mean of each criterion, as shown in Equation (4).

$$W_n={\displaystyle \frac{{Gm}_i}{{\sum }_{n-1}^{n}Gm}}$$

(4)

where $W_n$ denotes the normalized weights, called the eigenvector, and Gm denotes the geometric mean of each row in the comparison matrix.

The last step is to calculate the consistency ratio (CR) using Equation (5). The calculated CR value is 0.053. The CR value is less than 0.10 or 10% for consistent weights. If CR > 0.10, the comparison matrix and, consequently, the weightages (Table 4), should be revised. From

Table 5. Random consistency index (RCI) values for various evaluation criterion numbers (n) [67].

n	10	9	8	7	6	5	4	3	2	1
RCI	1.49	1.45	1.41	1.32	1.24	1.12	0.89	0.58	0	0

, the random consistency index value for the seven thematic layers is given as 1.32, and the principal eigenvalue is calculated as 7.42.

$$Consistency Ratio\left(CR\right)={\displaystyle \frac{Consistency Index \left(CI\right)}{Random Consistency Index \left(RCI\right)}}$$

(5)

where the consistency index (CI = 0.07) is calculated using the following equation:

$$CI={\displaystyle \frac{{\lambda }_{max}-N_i}{N_i-1}}$$

(6)

where N_i denotes the number of factors, and ${ \lambda }_{max}$ is the maximum value of normalized weights or the eigenvalue of the comparison matrix, as shown in

Table 6. Calculation of the (${\lambda }_{max}$) principal eigenvalue.

Layers	Sum	${\mathit{W}}_{\mathit{n}}$	$\mathbf{Sum} \times {\mathit{W}}_{\mathit{n}}$
Rainfall	3.32	0.318	1.06
Lithology	3.4	0.28	0.95
Lineament density	11.67	0.12	1.4
Drainage density	10.28	0.10	1.08
Slope	19	0.051	0.97
Soil type	18	0.06	1.08
LULC	12	0.73	0.88
${\mathsf{\lambda }}_{\mathit{m}\mathit{a}\mathit{x}}$			Sum = 7.42

. This can be calculated using the following equation:

$${\lambda }_{max}={\displaystyle \frac{1}{N_i}}\sum _{n=1}^{N_i}{\displaystyle \frac{A.W_n}{W_n}}$$

(7)

where W_n is the normalized weighted vector.

Finally, we prepare the groundwater potential index (GWPI) using Equation (8) if all criteria are found consistent (CR < 0.1) after assigning the weightage values as shown in

Table 7. Weights and classifications of different thematic layers and their sub-classes for KRB.

Factors/Layers	Classes	Potential for Groundwater	Class Rank	Normalized Weights	Weights of Each Rank
Rainfall	127–396	Very poor	1	32	32
	396–665	Poor	2		64
	665–934	Medium	3		96
	934–1203	Good	4		128
	1203–1472	Very good	5		160
Geology	Cambrian metamorphic intrusive rocks (CMI)	Very good	2	27	54
	Carboniferous sedimentary rocks (Cs)	Very good	3		81
	Jurassic and Triassic rocks (Jms)	Very good	4		108
	Cretaceous and Jurassic sedimentary rocks (KJs).	Very good	4		108
	Cretaceous sedimentary rocks (Ks)	Medium	3		81
	Metamorphic intrusive rocks (Mi)	Good	2		54
	Mesozoic intrusive rocks (Mz)	Good	2		54
	Neogene sedimentary rocks (N)	Good	3		81
	Permian rocks (Pr)	Medium	2		54
	Permian intrusive metamorphic rocks (Prim)	Poor	2		54
	Undivided Paleozoic rocks (Pz)	Very good	3		81
	Paleozoic igneous rocks (Pzi)	Poor	2		54
	Lower Paleozoic rocks (Pzl)	Poor	2		54
	Paleozoic Precambrian rocks (PzPc)	Very good	3		81
	Quaternary sediments (Q)	Very good	4		108
	Undivided Silurian rocks (S)	Good	3		81
	Tertiary igneous rocks (Ti)	Good	2		54
	Triassic rocks (Tr)	Poor	3		81
	Triassic metamorphic and sedimentary rocks (Trms)	Medium	3		81
	Tertiary sedimentary rocks (Ts)	Poor	4		108
	Undivided igneous rocks (Pc)	Good	3		81
Lineament density	0–3	Poor	2	12	24
	3–9	Medium	3		36
	9–15	Good	4		48
	15–26	Very good	5		60
	26–51	Very good	5		60
Drainage density	0–4	Poor	2	10	20
	4–12	Medium	3		30
	12–21	Good	4		40
	21–30	Very good	5		50
	30–47	Very good	5		50
Slope	0–2,375,614	Very good	5	5	25
	2,375,614–5,279,143	Good	4		20
	5,279,143–8,182,672	Medium	3		15
	81,826,72–12,933,901	Poor	2		10
	12,933,901–67,309,080	Very Poor	1		5
Soil	Rocky land with Lithic Cryorthents (Ro_Li_Cr)	Poor	2	6	12
	Rocky land with ice-capped bare rock (Ro_Ic)	Good	4		24
	Xerochrepts with Xerorthents (Xe_Xe)	Poor	2		12
	Haplocambids with Torriorthents (Ha_To)	Poor	2		12
	Rocky land with Lithic Haplocryids (Ro_Li_Ha_Cr)	Medium	3		18
	Calcixeralfs with Xerochrepts (Ca_Xe)	Very poor	1		6
	Mountain rock outcrops with very shallow loamy soils (Lo_Sh_So)	Poor	2		12
	Rocky land with Lithic Haplocambids (Ro_Li_Ha_ca)	Poor	2		12
	Torriorthents with Torrifluvents (To_Th_To_Fl)	Very good	5		30
	Torrifluvents with Torripsamments (To_To)	Very poor	1		6
	Loamy and clayey, mainly non-calcareous soils (Lo_Cl_Nc)	Poor	2		12
	Loamy and clayey partly non-calcareous soils (Lo_Cl)	Medium	3		18
	Mountainous area with mainly loamy, shallow soils (Ro_Lo_Sh)	Poor	2		12
LULC	Rivers	Very good	5	8	40
	Trees	Poor	2		16
	Water Body	Very good	5		40
	Agriculture land	Medium	3		24
	Buildings	Poor	2		16
	Mountains	Good	4		32
	Snow	Very good	5		40
	Bare land	Medium	3		24

.

$$GWPI={\sum }_{i=1}^n{W}_{n}x{R}_i$$

(8)

where R_i is the rating of each weighted criterion.

3. Results

The soil type, land use/land cover, rainfall, lineament density, geology, drainage density, and slope data were obtained from different sources and were processed using ArcGIS 10.3 software to obtain the presented thematic layers. Geological features, land use/land cover, soil types, and slope were categorized into classes according to data variance. Each specific type within these categories was analyzed for its properties related to groundwater potential, with its contribution to Groundwater Potential Zones determined by weightage, enhancing spatial result accuracy. Rainfall, drainage density, and lineament density were classified into five classes to ensure compatibility with groundwater potentiality classes, i.e., good, very good, medium, poor, and very poor, as depicted in Table 7.

The thematic images and their performance in the overall groundwater potential zoning are presented below:

3.1. Rainfall

Rainfall is the primary source of groundwater recharge; this is why its weightage in general should be higher, and this is how it is projected in the comparison matrix. A major part of the rainwater percolates deeply, contributing to groundwater recharge. High rainfall means high groundwater recharge if accompanied by fractured geological formations and lower drainage density. The spatiotemporal distribution of rainfall is a key factor in determining the Groundwater Potential Zones. Figure 4 shows the distribution of stations from which data are obtained and the distribution of the rainfall over the Kabul River Basin based on the data collected from an online open-access source. Based on the data, the Kabul River Basin was divided into five classes, i.e., very low, low, medium, high, and very high. It can be established that most of the basin receives medium, low, or very low rainfall. Very high rainfall occurs in Pakistan in the basin of the Swat River. The average rainfall values recorded in various sub-basins of the Kabul River Basin, like Dir, Kabul, Chitral, Cherat, Paghman, Drosh, Kalam, Malam Jabba, Peshawar, North Salang, South Salang, and Saidu Sharif (Swat), are 127, 298, 414, 427, 436, 619, 639, 728, 817, 990, 1035 and 1474 mm, respectively. The above rainfall values have been divided into five categories; the lowest value is 127, which is considered very low, while the highest value is 1474, which is considered very high.

3.2. Geology

Geology is an important indicator for determining zones of groundwater recharge because infiltration will not occur even if all other conditions are met and the geology is not ideal for replenishing groundwater. That is why geology is anticipated to be the second most significant element in groundwater potential zoning, behind rainfall. Rocks are classified into very good, good, medium, poor, and very poor based on their porosity and permeability. Porosity and permeability are mostly conceived as similar terms for rocks, but there is a logical difference—a rock may not be porous enough, but it might be permeable for water due to its tectonic history and the geophysics of the rock (i.e., a CMI rock is labeled as very good for GWP). This rock is formed under high temperature and pressure and is found in tectonically active zones, i.e., the Himalayas. Even though CMI rocks are low in porosity, the fractures, joints, and faults induced in these rocks make them very permeable, and this is why they are classified as very good for GWP [68]. Cs are very porous and permeable. They are found mainly in layered structures like KJs, which can serve as potential aquifers for groundwater; hence, these rocks are also very good in terms of GWP. A similar situation is found with Jms, which exhibits very high porosity and permeability, leading to good water storage. These formations contain sandstone, limestone, and shale. Sandstone and limestone are very good for storage and recharge. However, shales do reduce permeability, but these rocks are the recipe for an aquifer [69]. Moreover, Pzi and Pzl are labeled as poor in terms of contribution to GWPZ. Granite is the most common member of the Pzi formation. These rocks are crystalline in nature and are very low in porosity and permeability, with very little water movement. However, certain fractures can develop, which can act as aquifers, but these features are very limited in such type of rock. Slate and phyllites are the most common rock types in the formation of Pzl rocks. These rocks are compact, fine-grained, and are very low in porosity and permeability, yielding very low or no infiltration at all, which makes a perfect recipe for an aquitard [70]. Hence, they are labeled as poor in terms of contribution toward GWPZ. Based on data gathered from the USGS website, a geological map of the Kabul River Basin has been created. The basin is classified according to the underlying areas of geology as shown in Figure 5.

3.3. Lineament Density

The underlying geological structures are good indicators of groundwater potential as they facilitate groundwater movements. Lineaments show the potential layers of the bedrock, which can act as aquifers. Higher lineament density favors high water storage potential—if accompanied by adequate land use, permeable soil, and adequate slope—so that the rainfall could potentially lead to underground water storage after it infiltrates the ground. This is why it is the third most important parameter in determining GWP, as depicted in the comparison matrix, because underground fractures serve as pathways for rainfall [70]. Based on the data produced from the online USGS source, the lineament density map is divided into five classes, i.e., very high, high, medium, low, and very low. The areas of the lower slopes and the lower part of the basin are dominated by zones of very low lineament density, while the higher altitudes of the Kabul River are dominated by high lineament density, as shown in Figure 6.

3.4. Drainage Density

The drainage density map plays an important role in groundwater potential zoning because areas with high drainage density have higher runoff, and vice versa; this means reduced infiltration. Drainage density is very much related to the slope, soil type, and land use of the area. The drainage density map is classified into five classes, i.e., very high, high, medium, low, and very low. All the areas adjacent to the rivers have higher drainage density, while most of the mountainous areas have low and very low drainage densities. The higher drainage density lies mostly in areas of lower elevation. A higher drainage density area is assigned a lower score, and vice versa, because the greater the drainage density, the lower the infiltration and the lower the groundwater recharge. So, in this research work, the SRTM digital elevation model downloaded from the USGS online source was integrated into the ArcGIS 10.3 software to draw the drainage density map for the Kabul River Basin. The drainage density map is shown in Figure 7.

3.5. Slope

Slope is another influencing factor that is associated with the percolation of the rainwater into the underground water table. From the literature review, it is evident that the slope must be incorporated in the assessment of groundwater potential for a particular basin. In steep slopes, water runs off quickly compared to gentle and flat slopes, which means very low groundwater storage. It is the least important factor among all. Even if the slope is flat, but the land use is not favorable to infiltration, it would lead to infiltration and groundwater recharge. It only works when other factors are favorable. Hence, we can see that not all places with groundwater storage potential have a flat slope. Slopes alone cannot work if there is very low rainfall.

In this work, the SRTM DEM was downloaded from the USGS online source and integrated into ArcGIS 10.3 software to draw the slope map for the Kabul River Basin. The slope map is given in Figure 8.

3.6. Soil Type

The thematic map of soil properties was derived with the help of data obtained online from the source of Food and Agriculture Organization of the United Nations. The classification of soil types in the Kabul River Basin shows that most of the basin consists of either rocky land or mountainous bed with shallow loamy soil, as shown in Figure 9.

Soil types were investigated based on certain criteria to judge their contribution to the GWPZ. These criteria include the sandy loam nature of the soil, its location with respect to the river, infiltration rate, soil texture, topography, and water yield. To_Th_To_Fl is labeled as the only very good soil type in terms of groundwater potential contribution because it is often associated with floodplains, and is highly porous, enhancing infiltration. The typical texture of Torrifluvents is sandy and silty loam, which causes increased vertical water movement and has a relatively flat topography, enhancing infiltration and reducing runoff; however, this soil type is only available in the southwesternmost part of the KRB and, hence, the overall contribution of this particular soil is limited [71]. However, Ro_Li_Ha_Cr is classified as a medium contributor to the GWP, but the study area has an abundance of this soil type. Infiltration in this type of soil is very limited because these rocks are mostly available in colder regions. Their infiltration is limited only to porosity and fractures induced by the freeze–thaw cycle within the rock; hence, they have limited storage potential [72]. However, Xe_Xe is classified as poor in the GWP. These soils have very low water-holding capacity due to poor development and poor structure. Their coarse structure promotes fast drainage, making it very difficult to accumulate water that can be yielded [73]. In Figure 9, it can be seen that this type of soil can be found more readily along with other soil types like Ro_Li_Cr and Ro_Lo_Sh, which have more of the same physical properties. Due to the abundance of poor soil in terms of GWP in the KRB, the contribution of the soil to the overall GWPZ is limited.

In Figure 9, it can be seen that the lowest downstream part (confluence of the Kabul River and Swat River) of the Kabul River Basin, which has an abundance of clayey non-calcareous soil, is classified as a soil type with good groundwater potential. The upper part of the basin mainly contains rocky soils, which are less rechargeable and contribute very little to the total weight of the soil type. Consequently, these areas are marked as poor and very poor in the overall groundwater potential map. The clayey and loamy soil has greater water potential, which covers a major part of the lower Swat River basin upstream of the junction of the Panjkora River and the Swat River.

3.7. Land Use/Land Cover

The land use/land cover is responsible for the information required about the nature and extent of the land. From the literature, we know that land use plays a vital role in the discharge calculations of the particular area, whereas it also has a strong impact on the rate of recharge. Land use represents the first interaction between rainfall and the ground, making it very important for local recharge. But when determining the groundwater recharge of the whole basin, it is not as important as the drainage density, lineament density, rainfall, and geology, which incorporate the distribution of rainfall at a larger scale across the basin. However, rivers, water bodies, and snowmelt are considered to be the primary sources of groundwater recharge, subject to certain gradients, i.e., the GWT is below the riverbed, the riverbed is permeable, and fractured terrains are available to guide river water deep into the aquifers. Consequently, ponds and agricultural lands are considered more pervious than mountains and other impervious surfaces like buildings. Additionally, in this analysis, water bodies like rivers and lakes have very good GWP, while agricultural land and bare land fall under the medium GWP, and buildings and mountains are labeled as having poor GWP. The land use/land cover of the basin area is classified into nine classes based on the data collected from the online USGS source. The land cover classes are agricultural land, bare land, buildings, green land, mountains, ponds, rivers, snow, and trees. The agricultural land, rivers, and water bodies are mostly responsible for groundwater recharge, which covers a major part of the basin, as shown in Figure 10. It can be seen that almost all of the very good GWP in Figure 3 can be found adjacent to the river bodies.

Integrating these thematic images (Figure 4, Figure 5, Figure 6, Figure 7, Figure 8, Figure 9 and Figure 10), Groundwater Potential Zones for the study area were developed using ArcGIS 10.3 based on their weighted factors to classify the given basin area into five zones, i.e., very good, good, medium, poor, and very poor, covering 35.45% (19,989 km²), 37.2% (20,978 km²), 23.16% (13,063 km²), and 0.06% (19 km²), respectively. The Groundwater Potential Zones in Figure 11 show that most of the region lies in the good class of groundwater potentiality, and 19 km² of the total area is classified as very poor, very good, or good. The Groundwater Potential Zones are distributed along the course of the Kabul River and its tributaries, i.e., the Swat River, Panjkora River, and Kunar River. However, a major part of the Kabul River Basin is classified as having medium groundwater potential.

The multi-criteria decision-making (MCDM) approach operates under the assumption that different criteria hold varying degrees of importance in assessing groundwater potential. Consequently, the results demonstrate that each thematic layer is allocated a specific weight based on the comparison matrix. It is evident from the results that three factors, i.e., rainfall, geology, and lineament density, contribute more to the groundwater potential of the study area. As a result, they have increased weights of 32%, 27%, and 12%, respectively. The highest average rainfall value occurs in Saidu Sharif, Swat, and as rainfall receives the highest weightage in the groundwater potential zone, most of the Swat Basin falls within the good and very good Groundwater Potential Zones. Similarly, the zone with very low annual rainfall, i.e., Dir, falls in the poor GWP region. Generally, the areas with greater rainfall indicate good water potentiality if all other conditions are favorable. As the geology of the Kabul River Basin contains a total of 22 rock types, only 5 are marked as poor, with no rock marked as very poor in terms of groundwater potentiality. A major part of the basin is covered with CMi, N, PTr, Pr, Prim, Ks, Pz, Pc, Ti, and Trms; out of these rock types, Ks are classified as medium and Prim are classified as low, while the other rock types are all classified as very good and good in terms of GWP. Although most of the basin is covered by low rainfall, it is classified as good and very good in terms of GWP; this can be credited to the geology of the basin, which is the second most influential factor after rainfall in terms of GWP. Lineament density is another important factor in the analysis of GWPZs in the Kabul River Basin. Most of the Kabul River Basin is dominated by very low to low lineament density, but the eastern part of the Swat sub-basin contains areas of medium lineament density, while the southwesternmost part of the Parwan district’s sub-basin contains very high, high, and medium lineament density. Consequently, these areas are classified as good and very good in GWPZs. The Swat sub-basin receives more rainfall compared to the Panjshir and Parwan sub-basins, as shown in the rainfall map in Figure 4, but Panjshir and Parwan receive their fair share of GWP weight due to their higher lineament density compared to the other sub-basins, which means channelized runoff concentration. The basin contains a total of 13 types of soil, as shown in Figure 9. The elevated portion of the basin is mostly covered by rocky land with Lithic Cryorthents soil, which has lower water-holding properties compared to clay and loamy soils. Elevated parts of the Parwan sub-basin are also covered by this type of rock, but the determining factor in characterizing this basin with higher Groundwater Potential Zones is its geology and lineament density. The upper part of the Swat River basin is covered by Ro_Lo_Sh, and the lower part is partly covered by Lo_Cl_Nc and Lo_Sh_So soil types, which have low water-holding capacity, as assessed subjectively through a literature study. Consequently, this means that the factors that contribute more to the high groundwater potentiality of the Swat River basin are mostly rainfall and lineament density. The clayey and loamy soil has greater water potential, which covers a major part of the lower Swat River Basin upstream of the junction of the Panjkora River and the Swat River. The existing area consists of 70.07%, 16.83%, 4.8%, 4.46%, 3.44%, 0.36%, and 0.004% of bare land, mountains, agricultural land, buildings, snow cover, rivers, and lakes, respectively. The western portion of the Kabul River Basin lies in the very good and good zones of groundwater potential, as most of the area consists of agricultural land, rivers, and lakes. The lowest-elevation parts of the Kabul River Basin to the east consist of metropolitan areas; therefore, groundwater recharge is very low due to the land cover characteristics, and the zones lie within the medium and poor GWPZs.

3.8. Validation of Qualitative Results

To validate the results against existing groundwater levels, data from 354 boreholes were collected from the Tehsil Municipal Administration (TMA), the Public Health Engineering Departments of Pakistan, and the Meteorological Department in Kabul. Figure 12 shows the distribution of existing groundwater level data throughout the basin; the data are projected on the GWPZ map to compare the current groundwater levels.

Table 8. Comparison of the actual groundwater level with the GWPZ.

GWPZ	Distribution of Different Depths of Tube Well
	1–45	46–90	91–115	116–135	>135
Very good	16	130	67	1	4
Good	6	27	14	6	5
Medium	4	23	14	2	3
Poor	0	6	9	0	0
Very poor	0	12	4	1	0
No. of tube wells in different zones	26	198	108	10	12
No. of tube wells in medium or above GWPZ	26	180	95	9	12
% of tube wells in medium or above GWPZ	8.07	55.90	29.50	2.80	3.73

shows the comparison between the actual groundwater levels to GWPZs. The actual groundwater depth is distributed across five different classes, as follows: <45 m, 46–90 m, 91–115 m, 116–135 m, and >135, based on very good, good, medium, poor, and very poor groundwater depth. To compare the actual groundwater level to the GWPZ map, the groundwater table data were overlapped on the GWPZ map based on their latitude and longitude, and a comparison matrix was developed. Figure 12 shows different classes of groundwater table levels and the distribution of borehole data across GWPZs, falling into the very good, good, medium, poor, and very poor classes, which are 8.07%, 55.9%, 29.5%, 2.8%, and 3.73%, with different classes of groundwater tables in bores, respectively. The comparison shows that more than 93.47% of the tube well data classified as medium or above coincide with zones marked as medium, good, and very good in the GWPZ map (Figure 11). The results are presented in Table 8.

4. Conclusions

This study successfully assessed the Kabul River Basin for groundwater potential zones, employing remote sensing and the AHP method. The groundwater potential identified was classified into very good, good, medium, poor, and very poor zones, covering 35.45% (19,989 km²), 37.2% (20,978 km²), 23.16% (13,063 km²), 4.13% (2332 km²), and 0.06% (19 km²), respectively. This study successfully delineated the given study area for GWPZs based on the relative importance of seven thematic factors, that is, rainfall, geology, lineament density, drainage density, land use/land cover, soil type, and slope of the Kabul River with assigned weightages of 32%, 27%, 12%, 10%, 8%, 6%, and 5%, respectively, with a consistency index of 0.07 and a consistency ratio of 0.053. This means that rainfall, geology, and lineament density played a vital role in classifying the study area into GWPZs.

A major part of the study area is classified into medium and good GWPZs. The study shows that certain parts of the districts of Panjshir, Parwan, Mohmand, and Swat are classified as very good GWPZs. The study area shows the movement of groundwater from the northwestern part to the eastern part of the basin. The areas situated at higher elevations, such as the Dir sub-basin and the southernmost part of the Logar sub-basin (Ghazni), show depletion in the water table, and most of the perennial streams are observed to have dried up. This study shows that these sub-basins fall under poor GWPZs and are areas of higher drainage densities, which means that these areas need proper water conservation strategies. Urban development and lower rainfall in the eastern part of the Kabul River Basin (Peshawar District) put the area at risk of groundwater depletion, as this region also falls under zones with higher drainage density and very low lineament density, resulting in low groundwater recharge; it is marked as a poor GWPZ.

Only seven factors were considered in this study and their effects in terms of groundwater potentiality were studied; however, there can be many factors, such as the curvature of the surface, aquifer thickness, topographic wetness index, and topographic position index, which can be important in any particular case study, but were not taken into account in this research work. Although MCDM is a numerical method that quantifies the contribution of each factor to the desired outcome, assigning thresholds to different numerical values of the dataset for parameters is contingent upon one’s own expertise and subjectivity, which can vary according to the range, variance, and standard deviation of the data available. The GWP assessment was validated using tube well data obtained from 354 local and administrative sources, dispersed mainly in Peshawar, Mardan, Swat, Chitral, and some parts of the Kabul region. The validation shows a significant overlap of results, demonstrating accuracy in identifying GWPZs.

This assessment of groundwater potential is essential for different stakeholders to assess water resources in the Kabul River Basin, both in Pakistan and Afghanistan. This study is beneficial in providing a surface investigation framework for the case study basin based on the MCDM using the AHP method. This study can be a benchmark for further groundwater investigative studies in the basin.

Consequently, it is very important for the implementation of developmental schemes, particularly in the mentioned areas, where the management and proper utilization of groundwater is very important due to the prevalent urbanization of the basin. This study offers a deeper comprehension of irrigation water demand, offering helpful recommendations for sustainable irrigation practices.

The present research directly contributes to the Afghanistan Emergency Food Security Project’s (OSRO/AFG/213/WBK) objectives, aiding in informed irrigation and water management strategies. The GWP map was compared with the irrigation water demand map generated by the Food and Agriculture organization of the United Nations for the summer season (https://openknowledge.fao.org/server/api/core/bitstreams/c59d46e0-77d7-4430-83ae-03b84d326e86/content (accessed on 23 February 2023)) and winter season (https://openknowledge.fao.org/server/api/core/bitstreams/c59d46e0-77d7-4430-83ae-03b84d326e86/content (accessed on 23 February 2023)). According to FAO maps on irrigation water demand for Kabul, the potential evapotranspiration in the Kabul sub-basin ranges from a mean value of 358.49 mm in the winter to 1210 mm in the summer season, where the cropland extent ranges from a mean value of 7.78% in the winter to 9% (1227 km²) in the summer. The situation with the Kabul sub-basin in the winter is critical, and the area requires irrigation facilities compared to other neighboring areas. However, the estimated water loss during the winter season in the Kabul River Basin (https://openknowledge.fao.org/server/api/core/bitstreams/77c9f440-2ac3-4092-a9d7-f38188249f23/content (accessed on 23 February 2023)) shows that the Panjshir and Ghorband sub-basins also require irrigation infrastructure. This research study demonstrates that there is a potential for groundwater exploration in these sub-basins.

It is also demonstrated that most regions within the Kabul and Panjshir sub-basins are situated in good and medium GWPZs, with some areas classified as very good GWPZs; there is sufficient potential for groundwater exploration to meet the irrigation demands of the community. Refs. [74,75,76] show that groundwater is depleting in the Peshawar district, and particular recharge and proper conservation measures are necessary. In this study, Peshawar is classified as a medium GWPZ, and rapid urbanization is going to make it worse.

Most of the factors necessary for the recharge study are covered in this research work. Using this study, decision-makers can accurately pinpoint areas with varying levels of groundwater potential, considering factors like drainage density, land use/land cover, slope, geology, lineament density, soil, and rainfall. However, other factors can also be included other than these factors. Ref. [77] incorporated the aquifer resistivity and aquifer thickness in its analysis in order to improve the accuracy of the analysis, but rainfall was anticipated to be the most important factor. Ref. [78] considered the topographic wetness index (TWI) but found that the most important factors were geology, slope, and lineament density, while TWI received the lowest weightage of three percent in the normalized pairwise comparison matrix. Refs. [79,80,81] also found that geology, lineament density, drainage density, and slope were the most important factors. Consequently, this study utilizes these seven factors to delineate the KRB for the GWPZ.

This study successfully demonstrates that AHP and remote sensing can be applied effectively to delineate the KRB into GWPZ. However, this research work [82] also demonstrates that remote sensing is very effective and economical for sites where data are scarce and physical measurements of the parameters are not possible. Furthermore, this study provides the basis for further advanced studies, such as the identification of recharge locations [83]. Additionally, this research also provides a benchmark for advanced studies, like the incorporation of other methods in order to improve the accuracy of the existing method. Refs. [84,85] combined AHP, RS, and GIS with geospatial artificial intelligence (Geo-AI) and machine learning, respectively, in order to improve the precision in the identification of GWPZ. This information is invaluable for developing targeted strategies for sustainable groundwater usage and recharge. For example, regions identified as having very good groundwater potential can be prioritized for groundwater extraction, while areas with poor or very poor potential may necessitate conservation measures. Such research outcomes provide water resource planners and policymakers with the data needed to make informed decisions, optimize resource allocation, and implement tailored groundwater management practices. Ultimately, the practical application of this study contributes to efficient and sustainable water resource utilization in the case study area.

5. Limitations

While this study provides a good framework for the assessment of the groundwater situation in the KRB and also provides validation using the actual tube well data of 356 boreholes, certain limitations to this study should be acknowledged.

First, it is widely recognized that AHP is subjective, and this may result in an influenced final result of the GWPZ. However, to ensure that the pairwise comparison is logically coherent, consistency ratios were computed. Still, there is a potential bias in the judgment of the assigned weights by the experts [86,87]. Several studies based on AHP, while selecting the same layers, have different results concerning the weightages of different layers due to the expert judgment involvement in assigning the weightage in the comparison matrix. Ref. [88] concluded that geology is the most important factor. However, some researchers [89] did not even consider geology in thematic maps, which is considered very important for GWPZ. This implies that different authors considered different layers to be important in determining the GWPZ in their studies due to different requirements and characteristics of the study area under consideration. However, the general rule is that whichever factors are considered, the selection should be logical and scientifically important for the study area.

Furthermore, there are many key hydrogeological parameters, i.e., surface curvature, sunshine, transmissivity, aquifer thickness, and groundwater fluctuation, which were not accounted for in this study due to the fact that these datasets have very low spatial resolution. Many of these factors can also be obtained from remotely sensed data but certain factors can only be estimated using a physical measurement method in reliable form, i.e., aquifer thickness; due to the long-held conflict in the Kabul River Basin, the data collection was seriously disturbed, and many hydrological and meteorological stations were destroyed [90]. An illustrative example of this is the climate record, which was stopped in 1980 and then restarted in 2003 [91]. Due to this lack of data, many methods to date have to rely only on remotely sensed data. However, the seven factors considered in this study are very critical for a complete picture of the status of the GWP [92]. Furthermore, it is acknowledged that future studies are needed in order to incorporate a variety of these factors in a variety of combinations and assigned weightages, and the results should be validated quantitatively using the groundwater data from tube wells [93].

Another important aspect of the assessment is its validation process. Even though a wide range of borehole data was obtained, the data were spatially very limited; they were collected mainly for metropolitan regions, which may have potentially led to biased interpolation. While it was physically observed that the water table can vary, even within a close range, in this case study—particularly in the Afghanistan region—due to the lack of availability of borehole data, the available data were averaged across a larger area, which might misrepresent the actual underground water level in the upstream and central regions, where there are less or even no borehole data. Hence, the strength of validation of the assessment in these areas is weak [24]. Future assessments in the study area can be enhanced by collecting further borehole data in these particular areas.

Lastly, the results obtained from this study are, by and large, qualitative and do not present quantifiable underground availability or recharge rates; hence, there is a limitation on the objectivity and reproducibility of the outcome. Further research work is needed to develop a basin scale methodology in order to somehow quantitatively delineate the basin using certain in situ hydrological measurements (e.g., groundwater level fluctuations, well yield data, soil infiltration tests) with high-resolution remotely sensed datasets (e.g., satellite-based soil moisture estimates, land surface temperature, vegetation indices, and evapotranspiration products) coupled with advanced machine learning algorithms, hydrological modeling and physics-based groundwater simulation models. These models will enhance the accuracy and precision of future research work within this domain. Developing an accurate quantitative method with good spatial and temporal precision can be very beneficial for better decision-making in the future.