Assessment of Groundwater Potential Zones by Integrating Hydrogeological Data, Geographic Information Systems, Remote Sensing, and Analytical Hierarchical Process Techniques in the Jinan Karst Spring Basin of China

Abstract

Groundwater management in the Jinan Spring basin is hampered by its complex topography, overexploitation, and excessive urbanisation. This has led to springs drying up during dry seasons and a decrease in discharge in recent years. GIS and the AHP were employed to delineate groundwater potential zones using eight thematic layers: slope, geology, lineament density, topographic wetness index (TWI), rainfall, soil, drainage density, and land use/land cover (LULC). The model’s accuracy was assessed by comparing the findings to basin groundwater observation well data. We found that 74% of the observations matched the projected zoning. Further validation utilising the receiver operating characteristic (ROC) curve gave an AUC of 0.736. According to the study, 67.31% of the land has a good GWPZ, 5.60% has a very good one, 27.07% is medium, and 0.03% is low. Heavy rains throughout the rainy season raise water levels. Dry weather lowers water levels. This study’s conclusions will protect groundwater from climate change. Integrating hydrogeological data, GIS, remote sensing, and AHP approaches maximises data use, improves groundwater potential zone delineation, and promotes sustainable groundwater resource management decision making. This integrated method can help land use planners, hydrologists, and policymakers find optimal locations for water supply projects, establish groundwater management techniques, and reduce groundwater risks.

1. Introduction

Water resources are essential for the survival of humans, plants, and animals [1]. The growing global population and economic growth have resulted in an elevated need for water supplies. Presently, more than 40% of the global population is experiencing water scarcity, and if this pattern persists, over 6.3 billion individuals will be afflicted by different levels of water stress by the year 2030. In the last three centuries, around 85% of the Earth’s wetlands have experienced desiccation, while the remaining wetlands have suffered a decline in quality [2].

As a result, the preservation and efficient use of groundwater have gained significant attention in recent years, because surface waters are insufficient to meet human demands [3,4]. Groundwater is not visible to the naked eye, making it challenging to map this resource [3]. Climate change, population growth, and urbanisation exert pressure on groundwater resources, resulting in their overexploitation and the deterioration of water quality [5].

The process of identifying groundwater potential zones (GWPZs) is essential for fulfilling a community’s requirements and maximising the efficient use of available groundwater resources [6]. It is essential to examine the GWPZs in many areas around the country to understand the dynamics and accessibility of groundwater. Conducting a comprehensive analysis of the groundwater potential zones (GWPZs) across various regions in the country is of the utmost importance for gaining insights into the dynamics and availability of groundwater. An in-depth knowledge of the characteristics and behaviour of aquifers, which are geological structures that store and transport water, is necessary for accurately evaluating groundwater supplies and guaranteeing the long-term sustainability of water management techniques [7,8].

Karst groundwater resources play a vital role in global groundwater supplies. Approximately 25% of the global population depends on karst water for drinking purposes, and karst aquifers are present beneath 10–15% of the Earth’s land area [9,10]. Northern China, specifically Shandong, Shanxi, Hebei, Henan, and Shanxi Provinces and Beijing, is home to many essential karst springs [11]. These regions are characterised by a significant presence of Cambrian and Ordovician carbonate rocks, accounting for approximately 61% of the provinces’ areas [12]. Groundwater in karst aquifers is plentiful in these regions and has emerged as the primary source of water for domestic, municipal, industrial, and agricultural irrigation purposes [11,13]. Evaluating the potential of groundwater requires analysing monitoring practices and executing sustainable development strategies for water resources. Several aspects, such as rainfall, geological formations, slope, land use/land cover, geomorphology, and drainage characteristics, greatly influence the boundary of a groundwater potential zone during the planning process. Conventional techniques, such as drilling tests and stratigraphic investigations, are commonly employed to determine the sites of bore wells and the thickness of aquifers in groundwater investigations. Nevertheless, using these techniques across a large geographical area can be laborious and time-consuming [14,15,16].

Recently, due to technological advancements, an increasing number of scientists have been utilising Geographic Information Systems (GISs) and remote sensing (RS) to analyse the availability of groundwater in basins [3,5,17]. RS and GIS technology has demonstrated its effectiveness and cost-efficiency in estimating, analysing, and managing groundwater reserves. This is primarily owing to the abundance of spectral, spatial, and temporal data over large and hard-to-reach areas [18,19]. To find groundwater potential zones, scientists have used a range of GIS and RS methods, including decision trees [20], radial basis function, Bayesian models [21], random forest [22], fuzzy systems [23], weights of evidence [20], and more, to identify and define areas with high potential for groundwater. Satellite data offer useful insights into diverse areas like geology, geomorphology, lineaments, population, land use/land cover, drainage patterns, and soil cover. The data are displayed in thematic maps or layers, which assist in understanding the movement and conditions of water below the surface [24,25].

Various techniques were employed to ascertain the most suitable weight and ranking for these layers, which were subsequently summarised to generate a potential zone. The weights and ranks that were assigned to the parameters were computed via the knowledge-based standard [26]. The Analytical Hierarchical Process (AHP) is one of the key methods within the framework of the multi-criteria decision analysis (MCDA) methodology. This method has been found to be a successful, dependable, and convenient method for identifying groundwater potential zones. It is particularly useful for managing water resources, as it enhances decision making by providing structure, transparency, and firmness [3,16,27]. Moreover, the MCDA technique has demonstrated its efficacy in several applications, including identifying prospective groundwater sources, managing environmental concerns, evaluating agricultural suitability, and other areas. Furthermore, a study conducted by [28] demonstrated the superiority of the AHP technique. Singh conducted a comparative analysis between the Catastrophe technique and the AHP technique in the mapping of groundwater potential zones. The validation of his findings demonstrated that both methods were appropriate for mapping groundwater potential with a relatively high level of precision. The AHP methodology achieved an accuracy rate of 82%, while the Catastrophe technique achieved 74%.

One notable benefit of these methods is their integration with GIS-based techniques, which yield more accurate results, require less processing time, and are cost-effective compared to traditional GWPZ field methods [29]. Therefore, the AHP technique was used in this study.

Although there is available research on Jinan groundwater resources, little is known about the groundwater potential zone delineation within the subject area using GIS and AHP techniques. However, this knowledge is very essential for the proper management of groundwater resources, considering the non-linearity and highly heterogeneous nature of karst water resources. Additionally, with the acceleration of urbanisation, the development of urban rapid rail transit is very important within Jinan City to alleviate problems such as road congestion and the movement of vehicles. This underground construction could interfere with the spring veins and inevitably have an effect on the groundwater environment. To the best of our knowledge, this is the first attempt to delineate the groundwater potential zone in the Jinan Spring Basin using GIS and AHP techniques.

The geological structure of the Jinan Spring Basin is complex, requiring a significant amount of data to accurately evaluate its hydrogeological qualities. A number of researchers have conducted investigations in the Jinan Spring Basin. However, the procedure of hydrogeological exploration and hydrogeological mapping is expensive and requires a significant amount of time. In addition, doing extensive field geological investigations throughout the entire spring basin poses difficulties, and the currently collected data are inadequate for offering a highly precise depiction of the spring basin. Utilising GIS, RS, and AHP methodologies is a cost-effective and efficient method to precisely identify the groundwater potential zones in the research area. These data can be used to carry out hydrogeological surveys more effectively, enabling the early detection of regions that are suitable for exploitation, construction projects, and other human activities without causing harm to the springs.

Owing to the non-linearity and highly heterogeneous nature of the Jinan Karst Spring Basin, the main objectives of this study are (i) to delineate groundwater potential zones (GWPZs) within the study area using GIS-based multi-criteria decision analysis (MCDA), specifically the AHP technique; and (ii) to assess the efficacy of this method by validating it using the receiver operating characteristic (ROC) curve method and observed groundwater level data. The results of this study will serve as a baseline study on groundwater potential zone delineation, as not much research has been conducted in that regard with regard to this area. This will help policymakers and water managers formulate cost-effective, efficient plans and management strategies for sustainable groundwater withdrawal, the development of underground rails, etc.

2. Materials and Methods

2.1. Description of the Study Area and Hydrogeological Conditions

Jinan City is situated in the central and western regions of Shandong Province, China, with Mount Tai to the south and the Yellow River Plain to the north. As a result of this topography, the elevation gradually increases from the northern to the southern part of the basin, as shown in Figure 1. Its geographical position is around Latitude 36°39′59.99″ N and Longitude 116°58′59.99″ E. The Jinan Spring Basin is bounded by the surface divide of Mount Tai to the south, the Carboniferous–Permian igneous rocks to the north, and the Dongwu Fault and the Mashan Fault to the east and west, respectively (please see Figure 1). The Jinan Spring Basin has a warm, temperate continental climate, with an average annual rainfall of around 646.55 mm and a yearly temperature of about 14.2 °C. The distribution of precipitation across the area is uneven, with a general decrease in precipitation from the southeast to the northwest [13,30]. Jinan possesses numerous springs due to its distinctive hydrogeology, topography, and geological composition.

The spring area is situated in the intermediary region between the mountainous terrain in central Shangdong and the gently inclined plain at the foot of the mountain. The Jinan karst aquifer system displays a progressive monoclinic form, with limestone mostly serving as the main aquifer in the groundwater system. Limestone exhibits high solubility, resulting in the development of a complex subterranean network of dissolution fissures, caverns, and pipes in Jinan. The direction of karst groundwater flow is predominantly from the south to the north, in accordance with the inclination of the karst layers and the topography [31,32], as shown in Figure 2. The southern region of the Jinan Spring Basin features the Archean Taishan Group as the underlying bedrock of the aquifer. The basement in the middle of the Jinan Spring Basin consists of Cambrian and Ordovician karst carbonate rock layers. The distinguishing feature of these layers lies in their composition, which includes limestone and shale strata.

The Ordovician is characterised by the predominant presence of extensive deposits of thick limestone, argillaceous limestone, and dolomitic limestone. These rocks have well-developed karst fissures and high permeability, which are conducive for groundwater recharge, runoff, and discharge [13]. The karst network system has a south-to-north flow pattern, and the uniformity of the topography and presence of limestone enable atmospheric precipitation to replenish the karst water (Figure 2). The karst groundwater flow in the north of the Jinan Spring Basin is obstructed by impermeable Mesozoic igneous rocks or Carboniferous–Permian sand and shale. As a result, the groundwater gathers in the contact zone. Due to the elevated water pressure, the enclosed karst groundwater is discharged as upward springs in appropriate topographic and structural regions [33].

The sedimentary basement in this region consists of the Mount Tai group, a succession of paleo-metamorphic rocks dating back to the Archaeozoic era. The cap strata are organised in a sequence from the lowest to the highest, comprising igneous Rocks and Cambrian, Ordovician, and Quaternary formations (Figure 1). These formations play a role in determining the spatial distribution characteristics of karst aquifers in the area. The geological composition of the Jinan Spring Basin can be categorised into three distinct layers: a phreatic aquifer, an aquitard, and a karst aquifer [30,33,34].

2.2. Description of Data Used

This study involves the mapping of different features that influence groundwater potential to different degrees. Different types of thematic layers, such as LULC, slope, geology, lineament density, drainage density, topographic wetness index (TWI), rainfall, and soil, were gathered from different sources and put together to make the GWPZs. Thematic layers were first standardised to ensure consistent projections and cell sizes, and raster layers were transformed into polygon formats.

The LULC map was generated from World LULC ESRI Maps 2022, created with ESA Sentinel-2 imagery with a 10 m spatial resolution, employing ArcGIS 10.8 software. Drainage density and slope shapefiles were generated within ArcGIS using Digital Elevation Model (DEM) data. The drainage density map was generated using ArcGIS’s Line Density tool. The lineament density in km/km² was calculated using the line density tool in ArcGIS. The ArcGIS slope tool was used to analyse slope variations in the study area. The soil map was created utilising data from the Food and Agriculture Organisation’s (FAO) World Soil Data. The topographic wetness index (TWI) is a secondary topographic element that assesses the impact of topography on hydrological processes. The TWI (topographic wetness index) was computed in ArcGIS 10.8 using the following equation:

$$\mathrm{TWI}=\ln (\mathsf{\alpha }/\tan \mathsf{\beta })$$

(1)

where α represents the area of the upslope, and β signifies the slope gradient. Groundwater potential zone mapping was validated using existing well data (i.e., groundwater level data). The process used to identify potential groundwater zones is elaborated in Figure 3 as a flowchart of the study.

2.3. Generation of Thematic Layer Maps

Geospatial techniques were employed to generate thematic layers, including LULC, geology, soil types, slope, rainfall, TWI, drainage density, and lineaments. These thematic layers are very important, as they provide insight into groundwater occurrence, recharge, and movement [29]. The geology and hydrogeological conditions of the region influence the spatial distribution and occurrence of groundwater. Water bodies can also contribute to the replenishment of groundwater. Infiltration rates are significantly low in rural and urban communities, as well as in wasteland areas [35].

2.4. Multi-Criteria Decision Analysis (MCDA) Using Analytical Hierarchical Process (AHP)

Multi-criteria decision analysis (MCDA) with the Analytical Hierarchical Process (AHP) technique is one of the most widely recognised and employed GIS-based approaches for identifying groundwater potential zones. This method enables systematic decision making by establishing comparisons and a numerical scale indicating the relative importance of one parameter over another concerning the criterion that is under examination. The AHP operates through pairwise comparisons, wherein parameters are evaluated against each other employing Saaty’s scale of relative importance [36,37].

2.5. Analytical Hierarchical Process (AHP)

The Analytical Hierarchical Process (AHP), introduced by [27], proved to be a valuable method for making multi-criteria decisions when delineating GWPZs. This study employed the AHP technique to integrate eight thematic layers, recognised as pivotal factors governing water storage and flow within a region. However, it is important to note that not all thematic layers hold the same level of influence. Each thematic layer was assigned a relative weight using Saaty’s scale, which ranges from 1 to 9 (as shown in

Table 1. Saaty’s scale of preference for comparing two parameters in the AHP method.

Scale	Degree of Performance	Description
1	Equal	When both parameters contribute equally to the objective.
2	Intermediate	Signifying a preference lying between 1 and 3.
3	Moderate	The judgment somewhat favours one parameter over the other.
4	Intermediate	Expressing a preference that falls between 3 and 5.
5	Strong	The judgment strongly or significantly favours one parameter.
6	Intermediate	Signifying a preference between 5 and 7.
7	Very Strong	Denoting a very strong preference or high importance.
8	Intermediate	Expressing a preference between 7 and 9.
9	Extreme	Conveying a substantial preference or extreme importance.

), based on their importance in identifying potential zones.

2.6. Multicollinearity Checks

Multicollinearity, a statistical concern in multiple regression models, arises when one input parameter of the model is highly correlated with a combination of other inputs. This correlation suggests that one parameter can be accurately predicted from another, such as slope and slope aspects, leading to significant accuracy in the model output [37]. Thus, validating multicollinearity among input parameters before running the regression model is crucial. This validation involves a linear regression analysis where an input parameter (X_i) serves as the dependent variable, and the R² value (R_i²) is calculated. Using Equations (2) and (3) [37], tolerance (T_i) and variance inflation factor (VIF_i) are computed for each parameter.

$${\mathrm{Tolerance} \mathrm{of} \mathrm{the} \mathrm{ith} \mathrm{predictor} \mathrm{variables} (\mathrm{T}}_{\mathrm{i}})=1 {{\mathrm{R}}_{\mathrm{i}}}^2$$

(2)

$$\mathrm{VIF} \mathrm{of} \mathrm{the} \mathrm{ith}{ \mathrm{predictor} \mathrm{variables} (\mathrm{VIF}}_{\mathrm{i}}) =\raisebox{1ex}{$1$}\!\left/ \!\raisebox{-1ex}{${\mathrm{T}}_{\mathrm{i}}$}\right.$$

(3)

A tolerance below 0.10 or a VIF of 10 or more signals multicollinearity problems [38]. Layers with tolerance levels under 0.10 or VIF values of 10 and above should be excluded from the evaluation, ensuring the accuracy of the assessment. The steps in [37] were used to determine the collinearity statistics of the factors that determine groundwater potential.

2.7. Consistency Analysis

The reliability of the pairwise comparisons that were made among the thematic layers and their subclasses was verified by assessing the consistency ratio (CR) using Equations (4) and (5) [39] below:

$$\mathrm{CI}=(\mathsf{\lambda }\max - \mathrm{n})/(\mathrm{n} - 1)$$

(4)

$$\mathrm{CR}=\mathrm{CI}/\mathrm{RI}$$

(5)

where n is the number of factors, RI is the Random Index, and CI is the Consistency Index.

A CR value of ≤0.10 is deemed acceptable for conducting weighted overlay analysis using the AHP. If the CR > 0.10, it is necessary to re-evaluate the judgements to pinpoint the source of inconsistency and rectify it until the CR is ≤0.10. CR in this study was −0.71 and is considered consistent.

2.8. Groundwater Potential Zone Mapping

The generation of the GWPZ map was achieved in ArcGIS 10.8 software by employing the weighted overlay method after assigning weights to thematic layers and their subclasses [25,36]. This was calculated using Equation (6).

$$\mathrm{G}\mathrm{W}\mathrm{P}\mathrm{Z}={\sum }_{\mathrm{i}=1}^{\mathrm{n}}(\mathrm{W}\mathrm{i}\times \mathrm{R}\mathrm{i})$$

(6)

GWPZ denotes groundwater potential zones, Wi represents the weights allocated to individual thematic layers, and Ri ranks subclasses within each thematic layer. As a final step, the groundwater potential zones were categorised as poor, medium, good, and very good in terms of their potentiality.

2.9. Map Removal Sensitivity Analysis

A sensitivity analysis was conducted to assess the impact of removing individual thematic layers used in the groundwater potential map’s computation. In this analysis, each thematic layer was removed one at a time, and a new GWPZ map was generated with the remaining layers overlaid on each other. The sensitivity index (SI) was calculated using the following equation [40]:

$${\mathrm{DS}}_i^j=\frac{{\mathrm{S}}_{-i}^j-{\mathrm{S}}_{\mathrm{F}}^j}{{\mathrm{S}}_{\mathrm{F}}^j}\times 100$$

(7)

where i represents the parameter number, and j denotes the category of potential zones. ${\mathrm{DS}}_i^j$ signifies the percentage change (±) in the area of the jth type of GWPZ when the ith feature is removed. ${\mathrm{S}}_{-i}^j$ represents the area of the jth type of GWPZ when the ith feature is omitted, while ${\mathrm{S}}_{\mathrm{F}}^j$ represents the area of the jth type of GWPZ when all features are considered.

2.10. Validation

Validation holds significant scientific importance in this modelling process. The GWPZ map was validated using groundwater level data (2021) obtained from observation wells in the study area. Additionally, using the receiver operating characteristics (ROC) curve extended the validation process. This method entails the graphical representation of true positive rates (sensitivity) on the Y-axis and false positive rates (1 specificity) on the X-axis. The resulting curve allows for estimating the area under the curve (AUC), with values ranging between 0.5 and 1. An AUC value closer to 1 indicates excellent model performance, while a value near 0.5 suggests poor prediction accuracy.

3. Results and Discussions

This research provides the basis for integrating hydrogeological data and creating thematic layers related to natural resources using remote sensing (RS), field observation data, and the GIS environment. These integrated datasets are then used to accurately identify groundwater potential zones with the aid of the AHP method. The findings are applied in decision making for development and planning areas. The choice of influential components is contingent upon the hydrological and geological circumstances and the accessibility of relevant data for the study location. To find the groundwater potential zones (GWPZs) in the Jinan Spring Basin, information on eight factors was used. These were drainage density, land use/land cover (LULC), slope, geology, lineament density, topographic wetness index (TWI), rainfall, and soil. The Jinan Spring Basin is located in the mid-western part of Shandong Province and faces significant challenges in the management of springs and developmental projects such as rail construction, which in turn affect the groundwater environment due to the non-linearity and highly heterogeneous nature of the karst environment that is predominant in the basin. The final GWPZ map was developed using the GIS software environment. To validate the model, the GWPZ map was validated using groundwater level data obtained from observation wells in the study area. The validation results revealed that approximately 74% of the groundwater wells’ classifications matched accurately with the zoning that was depicted on the generated groundwater potential map.

The validation process was extended by employing receiver operating characteristic (ROC) analysis by considering the area under the curve (AUC). The validation results demonstrated good prediction accuracy using the Analytical Hierarchical Process (AHP) technique, as the AUC of the GWPZ map was calculated to be 0.736. Furthermore, a sensitivity analysis was conducted to assess the impact of removing individual thematic layers that were used in the groundwater potential map’s computation. The sensitivity analysis tests showed that the drainage density, slope, and lineament density thematic layers had the most significant effect on estimating the groundwater potential. The TWI, soil, and geology had a moderate effect. Notably, the removal of the drainage density thematic layer led to the highest variation in values, followed by lineament density and slope. This suggests that the study area’s surface features have a significant influence on the groundwater potential. The GWPZ map depicted the potential zones of groundwater in the study area, which were categorised into four classes: poor, medium, good, and very good. The good groundwater potential zone covered the majority of the area (roughly 67.31%), and the medium category took up 27.07% of the area. Very good (5.60%) GWPZ zones were observed in small patches in the upper and lower portions of the study area, while the poor zones only occupied 0.02% of the study area. The GWPZ map revealed significant groundwater potential in the good-to-very-good zones, covering 72.91% of the study area.

3.1. Assignment and Normalisation of Weights

Evaluating a thematic layer’s importance in comparison with other layers is a knowledge-driven process, forming the foundation of the AHP. Therefore, relative weights were assigned based on previous studies that were conducted in various geographical regions and informed by field expertise [41]. To determine the relative weights, a pairwise comparison matrix was created, as depicted in

Table 2. Pairwise comparison matrix for the thematic layers and consistency ratio.

Factor	LULC	Slope	Drainage Density	Geology	Lineament Density	TWI	Rainfall	Soil	Weight
LULC	1.00	2.00	4.00	4.00	8.00	2.00	2.00	2.00	0.13
Slope	0.50	1.00	3.00	4.00	6.00	3.00	1.50	1.50	0.12
Drainage Density	0.25	0.33	1.00	2.00	0.50	1.00	0.50	0.50	0.13
Geology	0.25	0.25	0.50	1.00	2.00	3.00	0.50	0.50	0.11
Lineament Density	0.13	0.17	2.00	0.50	1.00	2.00	0.50	0.50	0.10
TWI	0.50	0.33	1.00	0.33	0.50	1.00	0.50	0.50	0.16
Rainfall	0.50	0.67	2.00	2.00	2.00	2.00	1.00	1.50	0.13
Soil	0.50	0.67	2.00	2.00	2.00	2.00	0.67	1.00	0.13
SUM	3.96	6.08	16.00	17.83	24.00	18.00	7.83	8.50	1

Notes: LULC = land use/land cover, TWI = topographic wetness index, CR = −0.71 < 0.1 (consistent).

, which included intensity judgements for the thematic layers. The thematic layers of geology, soil types, and land use/land cover (LULC) were classified based on their respective formations or categories (

Table 3. Assigned and normalised scores of the different features of each factor and theme.

Factors	Classes	Assigned Score	Normalized Score
Drainage Density (km/km2)	4.65–10.4	5	0.33
	10.41–16.15	4	0.27
	16.16–21.9	3	0.20
	21.91–27.66	2	0.13
	27.67–33.41	1	0.07
Land Use/Land Cover	Bare ground	2	0.09
	Crops	4	0.17
	Flooded vegetation	4	0.17
	Trees	3	0.13
	Rangeland	4	0.17
	Built areas	1	0.04
	Water	5	0.22
Slope (degrees)	0–5.74	1	0.07
	5.75–13.12	2	0.13
	13.13–21.05	3	0.20
	21.06–30.07	4	0.27
	30.08–69.71	5	0.33
Geology	Quaternary	4	0.33
	Ordovician	1	0.08
	Cambrian	2	0.17
	Igneous rock	1	0.08
	Archean	4	0.33
Lineament Density (km/km2)	0–0.11	1	0.07
	0.12–0.28	2	0.13
	0.29–0.44	3	0.20
	0.45–0.63	4	0.27
	0.64–1.06	5	0.33
TWI	2.46–5.74	1	0.07
	5.75–7.5	2	0.13
	7.51–9.77	3	0.20
	9.78–13.22	4	0.27
	13.23–23.9	5	0.33
Rainfall	516.05–562.69	1	0.07
	562.70–609.34	2	0.13
	609.35–655.98	3	0.20
	655.99–702.63	4	0.27
	702.64–749.27	5	0.33
Soil	Calcaric cambisols	3	0.08
	Eutric cambisols	4	0.11
	Calcaric fluvisols	4	0.11
	Eutric fluvisols	5	0.14
	Rendzic leptosols	3	0.08
	Gleyic luvisols	2	0.05
	Haplic luvisols	2	0.05
	Calcic luvisols	3	0.08
	Calcaric regosols	2	0.05
	Eutric regosols	3	0.08
	Water	5	0.14
	Urban areas	1	0.03

).

Multicollinearity Analysis

Table 4. Collinearity statistics of the groundwater potential’s determining factors.

#	Thematic Layer	Tolerance	VIF
1	Slope	0.754	1.327
2	LULC	0.896	1.116
3	TWI	0.966	1.036
4	Geology	0.715	1.399
5	Soil	0.691	1.448
6	Lineaments	0.955	1.047
7	Drainage density	0.766	1.306
8	Rainfall	0.844	1.185

illustrates the results of the multicollinearity analysis. The findings show that the variance inflation factor (VIF) values for each thematic layer are below 10, and the tolerance values exceed 0.1, at both significance levels of ρ < 0.01 and ρ < 0.05. This indicates the absence of collinearity among the eight thematic layers that were used in the study. Therefore, the model results remain unaffected by multicollinearity problems, introducing no uncertainties.

3.2. Thematic Maps of Influencing Factors

Eight influencing factors, drainage density, land use/land cover, slope, geology, lineament density, topographic wetness index, rainfall, and soil, were used to identify the groundwater potential zones in the Jinan Spring Basin. The AHP method established the weights for each class in the thematic maps, considering their attributes and capacity for water potential.

3.2.1. Rainfall

Rainfall’s direct impact on groundwater accumulation is due to the percolation and infiltration of rainwater into the subsurface, with longer duration and lower intensity rain causing greater infiltration. This is also in confirmation with research carried out by [24]. Additionally, rainfall significantly influences groundwater potential and serves as a crucial source of groundwater recharge [42]. In the Jinan Spring Basin, annual rainfall (2019) map preparation using the inverse distance weighted (IDW) method showed rainfall ranging from 516 to 749.2 mm (Figure 4A). The distribution of rainfall in the study area was classified into five rainfall zones: 674.3–749.2 mm (very high), 639.5–674.2 mm (high), 606.5–639.4 mm (moderate), 573.7–606.5 mm (low), and 516–573.6 mm (very low). Higher weights were placed on the higher rainfall subclasses, while the lower rainfall subclasses were assigned comparatively lower weights, as depicted in Table 3.

3.2.2. Drainage Density

The drainage density in the study area ranges from 0.01 to 167.5 km/km² (Figure 4B). The drainage density of the study area can be grouped into five classes: (a) ‘very low’ (0–26.3), (b) ‘low’ (26.4–49.9), (c) ‘moderate’ (50–73.6), (d) ‘high’ (73.7–99.9), and (e) ‘very high’ (100–167.5). Drainage densities in the ‘very low’ and ‘low’ categories occupy 20.6% and 24.5%, respectively. The ‘moderate’ density category covers an area of 437.48 km² (representing about 25.4%), while the categories ‘high ‘ and ‘very high’ occupy 30.27% of the study area. Consequently, the drainage density in the research area predominantly falls under the ‘very low’, ‘low’, and moderate classifications, suggesting a higher likelihood of groundwater being present in these regions. This is in accordance with research by [15], which shows a significant inverse correlation existing between the drainage density and the likelihood of groundwater potential. Senapati [15] concluded that the probability of groundwater potential zones decreases as the drainage density increases. When assessing groundwater zones, it is imperative to consider drainage density, an improved measure of permeability. The construction of the drainage density map is based on the existing drainage map of the study area. Locations with a low drainage density are given higher rankings, whereas areas with a high drainage density receive lower rankings.

3.2.3. Land Use/Land Cover (LULC)

Figure 5A shows the spatial distribution of LULC in the study region. The region exhibits seven distinct land use categories. The LULC in the research region consists of the following categories: bare ground (0.02%), crops (15.99%), flooded vegetation (0.20%), trees (8.88%), rangeland (41.37%), built areas (31.52%), and water bodies (2.02%). Clearly, the research region is predominantly covered by rangeland, with built areas, crops, trees, water bodies, flooded vegetation, and bare ground also represented. The presence of crops and rangeland over a significant land area allows for substantial groundwater recharge, resulting in a high potential for groundwater availability. Conversely, built-up regions have a lower potential due to their limited capacity for recharge.

Land use practices significantly influence groundwater’s quality and recharge rates. LULC plays a significant role in affecting the presence of groundwater through various mechanisms such as infiltration, percolation, and surface runoff in different geographical areas. They also represent variables such as soil moisture, surface water availability, groundwater utilisation, and infiltration rates [43]. Land use practices significantly influence groundwater’s quality and recharge rates. Irrigated agriculture has a significant impact on groundwater availability, primarily by increasing recharge rates and enhancing the quality of shallow groundwater [44]. As population growth triggers changes in LULC patterns, understanding and detecting these patterns becomes crucial in delineating GWPZ. LULC encompasses soil types, vegetation density, and housing distribution, which is influenced by human interventions and broader phenomena like agriculture, urban growth, and economic development [45].

3.2.4. Geology

The spatial distribution of major geological classes in the study area is shown in Figure 5B. The formations within the study area are Archaean (29.73%) in the southern part, Cambrian (31.29%) in the middle (west–east direction), igneous rock (4.31%) in the north-western, Ordovician (28.57%) in the north, and Quaternary sediments and volcanic rocks (6.11%) in the north-western direction. Quaternary formations, with materials like alluvium and glacial drift, have high groundwater potential due to their permeable nature. Ordovician rocks, mainly carbonate types such as limestone and dolomite, offer moderate-to-high groundwater potential due to their porosity. Archean formations, like granite and gneiss, generally have limited groundwater potential due to low permeability. Formations with high groundwater potential received greater weights, while those with limited groundwater potential received lower weights. Geology dictates the presence of aquifers where groundwater is stored. Rock porosity is primarily responsible for controlling infiltration and runoff rates. The porosity, hydraulic conductivity, and permeability of formations are what determine aquifers. The geological composition of a location is another pivotal factor regulating groundwater availability [42,46]. Geological layers impact both the presence and movement of groundwater; porous and permeable formations allow for water retention and easy movement.

3.2.5. Slope

Topography, specifically slope, significantly impacts the movement and accumulation of water in the landscape. Slope essentially represents the elevation difference in a particular area. An area’s gradient influences both the runoff and infiltration dynamics. Land inclination acts as a pivotal boundary, dictating water retention and the efficacy of precipitation-driven infiltration [47]. Steeper slopes contribute to elevated runoff rates and reduced water recharge due to diminished percolation and infiltration. This relationship between slope and groundwater is well documented [48,49]. The slopes within the Jinan Spring Basin were categorised into five classes: flat, 0–5° (covering 42.45% of the study area); gentle, 6–12° (20.77%); medium, 13–19° (17.33%); steep, 20–28° (13.94%); and very steep, 29–65° (5.51%) (Figure 6A). Lower weights were assigned to the steep and very steep classes, while the moderate-to-flat slope class received a higher score due to its greater potential for groundwater recharge.

3.2.6. Topographic Wetness Index (TWI)

The TWI is commonly employed to assess the influence of topography on hydrological processes and to gauge the potential for groundwater infiltration [50]. It demonstrates how water that is stored at a particular site is affected by gravity’s pull, guided by the slope. This factor holds significance in evaluating groundwater potential zones and has been established as an indicator of favourable groundwater occurrence. Nevertheless, because of its susceptibility to terrain wetness and topographic gradient (slope), which could result in redundancy within multi-criteria decision making, this thematic layer was excluded as a decisive factor in the AHP model [51]. The TWI is frequently utilised to identify potential groundwater zones by depicting wetness patterns [46]. Ref. [52] investigated the validity of assumptions associated with the use of the topographic wetness index (TWI) in hydrological models. The crucial assumption that was tested was whether groundwater level variations could be adequately approximated by a series of steady-state situations. Their findings indicated a correlation between median groundwater levels and the TWI, but the strength of this correlation was influenced by whether the indices characterized local topography or the topography of the upslope contributing area. The study revealed that the correlation between the TWI and groundwater levels varied over time, decreasing at the start of rainfall events, suggesting significant spatial differences in groundwater responses. However, it increased after peak flow, indicating a more consistent correlation where groundwater levels could be considered spatially in a steady state. In conclusion, the assumptions underlying the TWI are better met when the groundwater levels change slowly. The Jinan Spring Basin’s TWI values range from 2.46 to 23.9, with five categories: very low (−0.6–2.5), low (2.6–4.2), medium (4.3–6.6), high (6.7–10), and very high (10.1–21.5) (Figure 6B). The TWI values vary with the study area, as the topography of the Earth is not uniform. Red [36] reported a TWI ranging from 0.14 to 13.49 in the Burhum district in India, while TWI values ranging from 3.7 to 22 were reported for Komenda-Edina-Eguafo-Abrem (KEEA) Municipality in Ghana [53]. Higher TWI values received greater weight, as higher TWI values correlate with increased groundwater potential.

3.2.7. Soil

In the study area, diverse soil groups exhibit varying groundwater potential. The soil type significantly influences the groundwater recharge and runoff dynamics. The permeability and water-holding capacity of a soil type, which are influenced by its composition and texture, dictate its ability to facilitate infiltration and percolation [54]. The soil found in the study region falls under the following categories: Calcaric cambisols (38.69%), Eutric cambisols (1.74%), Calcaric fluvisols (5.48%), Eutric fluvisols (0.02%), Rendzic leptosols (1.02%), Gleyic luvisols (3.56%), Haplic luvisols (6.64%), Calcic luvisols (0.78%), Calcaric regosols (23.73%), and Eutric regosols (15.16%). These are shown in Figure 7A. Sandy soils (Eutric fluvisols) are well draining with low water retention, which aids in groundwater recharge, and they therefore received the highest weight. Loamy soils (Calcaric cambisols, Eutric cambisols, Calcaric fluvisols, Rendzic leptosols, Gleyic luvisols, Haplic luvisols, and Calcic luvisols) tend to permit higher infiltration due to their greater permeability, while clayey soils (Eutric regosols and Calcaric regosols) hinder infiltration due to lower permeability. The relationship between soil texture, conductivity, permeability, and moisture content dictates groundwater recharge; hence, the clayey soils in this study area received lower weights compared to sandy and loamy soils. In a previous study [36], the author noted that red loamy soils exhibit poor permeability, leading to the assignment of the lowest weight. On the other hand, red sandy soils and laterite soils possess moderate-to-very-high permeability, making them highly conducive to groundwater recharge owing to their increased porosity.

3.2.8. Lineament Density

The groundwater potential within a rock is influenced by both lineament presence and the proximity to drainage systems. These factors impact borehole placements and water yield, alongside the potential for groundwater storage and recharge. Lineaments, representing geological features like faults and fractures, play a pivotal role in groundwater dynamics. Observations indicate that wells on lineaments can yield approximately 14 times more water compared to those that are situated away from lineaments, suggesting superior groundwater potential [55]. High lineament densities enhance the groundwater potential compared to lower lineament densities. Consequently, higher weights are allocated to lineaments with a high density, whereas lineaments with a low density receive lower weights, as can be seen in previous studies [36,53].

The lineament density of the study area (Figure 7B) ranged from 0 to 0.93 km/km² and is categorised as very low (0–0.07 km/km²), low (0.08–0.19 km/km²), moderate (0.2–0.32 km/km²), high (0.33–0.51 km/km²), and very high (0.52–0.93 km/km²), with moderate-to-very-low lineaments covering > 66% of the study area, which could indicate a low potential for groundwater storage and recharge.

3.3. Groundwater Potential Zone (GWPZ) Map

The delineation of groundwater potential zones (GWPZs) is a crucial approach for anticipating future groundwater availability in many regions, especially in arid and semi-arid areas. In such regions, groundwater has been depleting due to factors like over-pumping, urbanisation, and population pressure. By overlaying the relevant thematic layers that are associated with groundwater contribution, potential groundwater areas were delineated. The weighted overlay analysis, conducted using ArcGIS, resulted in the creation of a GWPZ map (Figure 8). The potential groundwater zones were classified as poor, medium, good, and very good based on the assigned weights. The outcomes revealed that of the total area, approximately 5.60% fell into the very good category, 67.31% were categorised as good, 27.07% fell into the medium category, and poor areas constituted about 0.03% of the study area, as shown in

Table 5. Groundwater potential zone map (GWPZ) statistics.

No.	Class	Area (km2)	Area (%)
1	Poor	0.45	0.03
2	Medium	483.29	27.07
3	Good	1201.91	67.31
4	Very good	100.00	5.60

. The map revealed significant groundwater potential in the good-to-very-good zones, covering 72.91% of the study area.

The study region features a low-to-moderate drainage density, which contributes to increased infiltration and recharge rates. The significant coverage of vegetation and agricultural areas across much of the basin corresponds to favourable groundwater potential. Additionally, the study area’s sedimentary rock composition, known for its strong groundwater retention capability, aligns with the identified groundwater potential zones. Furthermore, there is a moderate-to-very-high lineament occurrence in over half of the study area, which suggests a substantial capacity for groundwater storage and recharge.

3.4. Sensitivity Analysis

The outcomes of the sensitivity analysis after removing thematic layers are displayed in

Table 6. Variations in groundwater potential zones resulting from the removal of thematic layers.

Thematic Layer Removed	GWPZs
	Poor (%)	Medium (%)	Good (%)	Very Good (%)
Geology	−0.02	−3.76	+1.82	+1.95
Rainfall	+0.02	−1.19	+1.94	−0.77
Slope	−0.68	−8.84	+6.97	+2.54
Soil	+0.01	+3.52	−0.58	−2.95
Drainage density	−0.02	−8.66	+8.94	−0.40
TWI	+0.02	+4.41	+2.34	−6.78
LULC	−0.01	−0.17	−0.39	+0.57
Lineament density	+0.02	+7.10	+0.84	−7.96
All Layers	0.03	27.07	67.31	5.60

Notes: ‘+’ indicates increase by area and ‘−’ indicates decrease by area.

. These sensitivity assessments revealed that the drainage density, slope, and lineament density thematic layers exerted the most substantial influence on the groundwater potential estimation, while the TWI, soil, and geology had a moderate impact. Notably, the removal of the drainage density thematic layer led to the highest variation in values, followed by lineament density and slope. This suggests that the study area’s surface features have a significant influence on the groundwater potential.

The map removal analysis further demonstrated that eliminating the TWI and soil layers decreased the extent of areas with very good groundwater potential by 6.78 and 2.95%, respectively. Conversely, excluding the geology layer reduced the area with medium groundwater potential by 3.79%, while increasing the extent of areas with very good groundwater potential by 1.95%. The removal of the rainfall thematic layer also resulted in a reduction in the extent of the area with medium groundwater potential and an increase in the area with good groundwater potential by 1.19% and 1.94%, respectively.

The various classified areas displayed significant variations when each thematic layer was excluded. This underscores the importance of thematic layers that contribute to water availability and infiltration in determining groundwater potential. The removal of such layers has a substantial impact on delineating potential groundwater zones.

3.5. Validation of Groundwater Potential Zones

To assess the accuracy of the generated groundwater potential map, it was subjected to cross-validation against groundwater level data, as depicted in Figure 9. This validation process involved analysing a total of 23 groundwater wells distributed across the study area, encompassing diverse geological characteristics, land use/land cover types, and variations in topography. These groundwater wells were classified into four categories based on the measured depth to the water table: shallow (<56.47 m), medium (56.48–85.15 m), deep (85.16–113.83 m), and very deep (>113.83 m). A deeper depth to water table category is indicative of poor groundwater potential.

For the comparative analysis, the locations of the groundwater wells, classified by groundwater level, were overlaid onto the groundwater potential map, as illustrated in Figure 9. The validation results revealed that approximately 74% of the groundwater wells’ classifications matched accurately with the zoning that is depicted on the generated groundwater potential map (refer to

Table 7. Assessment of the groundwater potential map’s precision through a comparison with water level data obtained from drilled wells.

No.	Well ID	Latitude (Decimal Degrees)	Longitude (Decimal Degrees)	Depth to Water Table (m)	Groundwater Level	Location on GWPZ Map	Validation Remark
1	170	36.47	116.92	119.58	Very deep	Good	Disagree
2	S-169	36.52	116.85	51.15	Shallow	Good	Agree
3	S-27	36.53	117.02	101.93	Deep	Good	Disagree
4	358	36.54	117.01	138.42	Very deep	Good	Disagree
5	293	36.55	116.8	29.06	Shallow	Good	Agree
6	S-10	36.55	116.88	49.24	Shallow	Good	Agree
7	257	36.56	116.71	29.72	Shallow	Good	Agree
8	168	36.58	116.77	29.06	Shallow	Good	Agree
9	3	36.58	116.78	29.08	Shallow	Good	Agree
10	288	36.59	116.77	29.36	Shallow	Good	Agree
11	300	36.6	116.74	29.6	Shallow	Good	Agree
12	167	36.6	116.8	28.86	Shallow	Good	Agree
13	286	36.6	116.81	29.45	Shallow	Good	Agree
14	S-7	36.6	116.92	29.02	Shallow	Good	Agree
15	172	36.62	116.8	28.79	Shallow	Good	Agree
16	3A	36.63	116.82	29.01	Shallow	Good	Agree
17	S-9	36.63	116.83	29.12	Shallow	Good	Agree
18	S-7	36.63	116.87	28.09	Shallow	Good	Agree
19	361	36.63	117.12	142.51	Very deep	Good	Disagree
20	6	36.65	116.83	28.92	Shallow	Good	Agree
21	S-8	36.65	116.87	28.96	Shallow	Good	Agree
22	2B	36.67	117.03	29.61	Shallow	Medium	Disagree
23	327	36.69	117.06	27.79	Shallow	Medium	Disagree

). Notably, certain regions within the study area exhibited deeper groundwater levels in the observation wells compared to the GWPZ model’s predictions, which categorised those areas as having moderate-to-good potential. This discrepancy could be attributed to the long-term over-extraction of groundwater for activities like bottled water production, irrigation, or industrial processes [36].

Furthermore, this study conducted a quantitative validation of the predicted groundwater prospect map using the receiver operating characteristic (ROC) curve, as shown in Figure 10. This validation involved comparing the groundwater level data with the generated map, and the ROC tool in the ArcSDM module was employed for this purpose. The ROC curve illustrates the relationship between the true positive rate and the false positive rate across different threshold cutoff points for a given variable. Additionally, the area under the curve (AUC) serves as a measure of how effectively a parameter distinguishes between two diagnostic groups. Every point on the ROC curve indicates a pair of sensitivity values that correlate to a specific decision threshold. AUC values within the range of 0.5–0.6 indicate poor prediction accuracy, while ranges of 0.6–0.7, 0.7–0.8, 0.8–0.9, and 0.9–1 signify average, good, very good, and excellent prediction accuracy, respectively, for the relationship between the AUC and the prediction accuracy [56,57]. The validation results demonstrated good prediction accuracy using the Analytical Hierarchical Process (AHP) technique, as the AUC of the GWPZ map was calculated to be 0.736.

The GWPZ assessment model is robust against the uncertainties associated with multicollinearity issues. According to the sensitivity analysis, all the thematic layers that were taken into account in this analysis are crucial factors influencing the GWPZs of the study area, and the exclusion of any thematic layer would have a significant impact on the achieved accuracy level. Moreover, the GWPZ map was meticulously crafted by refining judgments to resolve inconsistencies among the thematic layers and subclasses within each thematic layer, as indicated by the consistency analysis.

The resulting GWPZ map exhibited a very good accuracy level of 74% when compared with groundwater level data (Figure 9). It also had a very good prediction score of 0.736 using the AUC of the ROC curve (Figure 10). Consequently, this study has achieved a significantly higher level of accuracy by incorporating a greater number of relevant thematic layers through the application of AHP and GIS techniques.

3.6. Practical Applications of the Study

The practical application of this research on the delineation of groundwater potential zones using the Analytical Hierarchical Process (AHP) holds significant implications for effective water resource management. By employing the AHP, decision makers can precisely identify areas with varying degrees of groundwater potential based on diverse factors such as drainage density, land use/land cover, slope, geology, lineament density, topographic wetness index (TWI), soil, and rainfall. This information aids in developing targeted strategies for sustainable groundwater usage and recharge. For instance, regions that are classified as having very good groundwater potential can be prioritized for groundwater extraction, while areas with poor potential may require conservation measures. Such research outcomes empower water resource planners and policymakers 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 utilisation.

4. Limitations and Recommendations

This study is subject to certain limitations due to the inherent nature of the adopted method. The AHP, being a knowledge-driven process, may have introduced some constraints and biases in its predictions. Additionally, the assessment relied on eight thematic layers, soil, rainfall, geology, land use/land cover (LULC), lineament density, drainage density, slope, and topographic wetness index (TWI). Notably, other parameters, such as aquifer thickness, curvature, roughness, pond frequency, topographic position index (TPI), recharge rate, pre- and post-monsoon groundwater depth, Normalised Difference Vegetation Index (NDVI), and water that is pumped for irrigation, industrial, and domestic use, which can also influence the groundwater potentials were not considered in this study owing to the unavailability of data on it.

Moreover, the validation process employed observation wells that were mostly concentrated in the northern part of the study area (since they were the only available data at that time), limiting the representation of the entire region. Although the validation results are good, they may not fully capture the groundwater dynamics in the southern part of the study area. To enhance the accuracy of the groundwater potential zones (GWPZs), it is recommended to conduct additional observations in the lower region to add this to analyses in the future.

In the future, an improved groundwater potential map could be developed by incorporating all these parameters and expanding the dataset with more well data in diverse locations as well as future land use and climate scenarios. Additionally, machine learning models and integration of dynamic data sources, such as real-time satellite data, and considerations of the impact of climate change on groundwater potential could be incorporated for predictive analysis, which may enhance the accuracy and applicability of the study. Despite these limitations, the study results are scientifically valid. The findings of this assessment hold significance for policymakers, providing valuable insights to enhance groundwater management in the study area and potentially serving as a model for other regions.

5. Conclusions

The main aim of this study was to delineate groundwater potential zones (GWPZs) within the Jinan Spring Basin, located in the midwestern Shandong Province of China. Due to the area’s landscape, the elevation reduces from south to north. Jinan has many springs because of its unique topographic features and geological structure. The Jinan karst aquifer system has a gentle monoclinic structure, with limestone serving as the main aquifer in the groundwater system. Limestone is easily dissolvable, forming a complex underground network system of dissolution fissures–cavities–pipes in Jinan. Karst groundwater flows mainly from south to north, which corresponds to the karst strata’s dipping and the topography. The basin has seen persistent difficulties and issues in developing groundwater due to the complex and diverse characteristics of the terrain.

Additionally, due to over-exploitation of the karst groundwater and changes in the recharge conditions, the discharge from the springs has been declining continuously in recent years, and the springs have dried up during dry seasons. To preserve the sustainability of groundwater for current and future generations, it is imperative to comprehend the characteristics and functioning of groundwater, given the rapid urbanisation and excessive exploitation of groundwater that has resulted in the depletion of certain springs. Furthermore, a significant proportion of the population in the region relies extensively on groundwater to fulfil their water needs. The main conclusions are as follows:

(1)

A comprehensive set of eight thematic layers—drainage density, land use/land cover (LULC), slope, geology, lineament density, topographic wetness index (TWI), rainfall, and soil—was developed through the compilation of satellite imagery, topographic maps, and other data sources. GIS technology provided a robust platform for organising and analysing these data layers, allowing for effective visualisation and spatial analysis. Secondly, remote sensing data, including satellite imagery and aerial photography, offer valuable insights into land cover and land use patterns.

(2)

Furthermore, the AHP technique facilitates the integration of various parameters and their relative importance in a consistent and systematic manner. It allows for the establishment of weightings or priority levels for each factor, thus helping to objectively rank and classify the groundwater potential zones. The study categorised the research area into four groundwater potential zones: very good (5.60%), good (67.31%), medium (27.07%), and poor (0.03%).

(3)

The study employed hydrogeological data, specifically from 23 pre-existing observation wells, to validate the accuracy of the groundwater potential zone map. The cross-validation results indicate that the AHP method effectively delineates groundwater potential in the study area, with about 74% of the findings matching the expected zoning. A further validation process achieved a good level of prediction, with an area under the curve (AUC) value of 0.736.

(4)

Secondly, remote sensing data, including satellite imagery and aerial photography, offer valuable insights into land cover and land use patterns. These findings suggest that the method applied in this study produces reliable groundwater potential maps for the Jinan Spring Basin.

(5)

In conclusion, the integration of hydrogeological data, GIS, remote sensing, and Analytical Hierarchical Process (AHP) techniques offers several significant benefits in delineating groundwater potential zones. This study demonstrates that this multi-disciplinary approach provides a more comprehensive understanding of the factors influencing groundwater occurrence and helps in identifying the most suitable areas for groundwater development. Additionally, this study provides valuable insights and tools for water planners and policymakers to better manage groundwater resources in regions facing groundwater scarcity, ultimately aiding in sustainable water resource management. The implementation of effective management strategies is expected to enhance natural groundwater recharge, resulting in increased water availability.