Delineation of groundwater potential zones using analytical hierarchy process (AHP) and GIS in Weserbi sub-catchment, Oromia, Ethiopia Open Access

Groundwater is a crucial and dynamic resource of nature, with one-third of all fresh water globally sourced from it (Pande et al. 2021). The escalating demand for better quality of water for domestic, industrial and agricultural purposes necessitates a systematic assessment and evaluation of groundwater resources for sustainable development (Nasab et al. 2023; Mojahedimotlagh et al. 2024). Mapping areas with high groundwater potential is crucial for the optimal use and preservation of these resources (Naghibi et al. 2017a, 2017b, 2017c; Zhu et al. 2022). Classifying and mapping potential groundwater zones is a vital stage in understanding resource availability, planning, and implementing effective groundwater strategies. Geospatial techniques hold considerable importance in creating and analyzing thematic layers, including geology, rainfall (RF), lineament density (LD), topography wetness index, drainage density (DD), slope (Sl), soil type, and land use/land cover (LULC), to produce groundwater potential zone (GWPZs) maps (Zarate et al. 2021).

Advancements in remote sensing (RS) technology, combined with the capabilities of geographic information systems (GIS) and high-performance computing, have established these tools as highly effective and precise methods for identifying groundwater resources and delineating GWPZs (Thakur et al. 2017; Sikakwe 2023). These modern approaches offer a more effective and economical alternative to conventional methods, including geological, hydrogeological, and geophysical surveys (Kumari & Singh 2021). The combination of analytic hierarchy process (AHP) and the GIS facilitates the synthesis and normalization of spatial data layers, enabling the generation of actionable information to support informed decision-making (Hasanuzzaman et al. 2022; Kabeto et al. 2022). AHP, a decision-support tool, integrates and transforms spatial data with value judgments, providing actionable insights for decision-making. According to Sulaiman et al. (2021), GIS-based AHP is a flexible and robust methodology capable of handling and synthesizing diverse variables, thereby aiding decision-makers in identifying optimal groundwater locations.

It is well-known that Ethiopia's agricultural activities primarily rely on seasonal rainwater systems and limited small-scale irrigation. This is particularly relevant in the study area of the region, where the agricultural practices depend heavily on RF (Addisu et al. 2015; Alemayehu & Bewket 2017). A significant challenge associated with rain-fed agriculture, both in the region and the country, is the high variability of RF, coupled with the unreliability and depletion of surface water during the dry season. This variability often leads to crop failures during dry periods, resulting in food insecurity that can escalate to famine, especially with slightly adverse climatic changes, which disproportionately affect the livelihoods of rural communities reliant on the rainy season (Moges & Bhat 2021). Additionally, global climate change has distorted seasonal RF patterns, negatively influencing agricultural activities, livestock production, and water supply across various regions.

A significant challenge in Oromia is the scarcity of comprehensive data on groundwater resources. This lack of information hampers effective planning and sustainable management of these resources (Bilal 2024). Groundwater exploration in the region often depends on conventional techniques, which may not accurately capture the spatial variability and complexity of aquifer systems (Ali & Goshu 2017). The region experiences significant fluctuations in RF patterns, leading to challenges in maintaining consistent groundwater levels (Taye et al. 2021). Moreover, activities such as deforestation and unplanned land use have led to reduced groundwater recharge rates, exacerbating water scarcity issues (Ali & Goshu 2017). In this study, by employing the AHP and GIS methodologies, the study offers a more precise and systematic approach to identifying GWPZs. This integration addresses the limitations of traditional methods predominantly on field surveys by considering multiple influencing factors such as lithology, Sl, LULC, and RF simultaneously, providing a holistic assessment of factors affecting groundwater availability. This comprehensive approach enhances the accuracy of groundwater potential mapping. Focusing on the Weserbi sub-catchment, the study delivers detailed insights that are crucial for developing targeted groundwater management strategies in similar arid and semi-arid regions within Oromia.

Given the declining output of rain-fed agriculture and the urgent need to double agricultural output in the upcoming decades, water is considered a critical element in shifting low-yielding rain-fed agriculture to a more sustainable and efficient irrigated farming system (FAO 1994; Gleixner et al. 2017). Previous assessments of GWPZs primarily focus on the regional scales and have limited information on the hydraulic properties of hydrogeological units. The qualitative parameters used for aquifer classification overshadow the available quantitative parameters (Berhe et al. 2021; Kassa et al. 2023). Often, subsurface information is derived indirectly from surface exposures of rocks and soils, such as those found along road cuts, gullies, streams, and riverbanks, to create conventional hydrogeological models. The amount of water withdrawn from an aquifer as well as from surface water, without depletion, largely depends on the rate and amount of recharge. However, there has been insufficient attention paid to quantifying the mechanisms of natural recharge and discharge in the catchment area. Building on these previous findings, this study mainly focuses on assessing and identifying potential groundwater resource areas for domestic water supply and irrigation, ranging from small-scale to household and large-scale applications.

Groundwater mapping in Ethiopia, especially in arid and semi-arid regions, is often limited by a lack of integrated, systematic, and high-resolution methodologies. Many previous studies have relied on traditional field surveys or single-parameter assessments, which may not fully capture the complexity of groundwater occurrence. The application of GIS-based multi-criteria decision analysis (MCDA), particularly the AHP, remains underutilized in Ethiopia's groundwater exploration efforts. There is a need for more refined GWPZs assessments that incorporate multiple hydrogeological, topographical, and climatic factors to improve decision-making. This study applies a robust AHP–GIS approach to delineate GWPZs in the Weserbi sub-catchment, providing a more systematic and quantitative assessment. Unlike many previous studies that focus on limited factors, this research integrates a comprehensive set of thematic layers, including RF, lithology, Sl, soil, LULC, DD, and LD. The study contributes to the growing body of knowledge on groundwater mapping in Ethiopia, demonstrating how AHP–GIS techniques can enhance groundwater resource assessment in arid and semi-arid regions. By focusing on the Weserbi sub-catchment, the study provides localized insights that can be extrapolated to other similar environments, addressing the need for site-specific groundwater exploration strategies.

Studies in the Upper Awash Basin (Ethiopia) have demonstrated the effectiveness of integrating AHP with GIS in groundwater potential mapping, achieving an 82.3% correspondence between high-yield boreholes and areas classified as having ‘very good’ groundwater potential (Adugna & Awoke 2024). Research in the Genale-Dawa Bale sub-basin (Ethiopia) utilized geospatial analysis to map groundwater potential zones, providing essential baseline information for planners and decision-makers in data-scarce regions (Eshetu et al. 2024) and a case study in the Dhungeta-Ramis sub-basin, Ethiopia (Tolche 2021).

This research seeks to identify and delineate GWPZs through the integration of GIS and the AHP methodologies, utilizing various thematic maps. The resultant GWPZ map is subsequently validated using borehole and spring yield data to ensure its reliability and accuracy.

This area is part of the central and northwestern Ethiopian plateau, characterized by deep gorges and diverse landforms resulting from tectonic uplift, volcanism, and ongoing erosion. The elevation within the catchment ranges from 1,250 m in the gorges to a maximum of 3,250 m in the plateau highlands. The identified physiographic features include valleys, low reliefs, gently rolling Sls, linear ridges, and pedi plains. Various drainage patterns have been observed, with dendritic, sub-dendritic, and sub-parallel patterns generated from a digital elevation model (DEM) with a resolution of 30 m × 30 m. The drainage pattern in the sub-basin reflects the geological material's type and degree of weathering. Overall, surface water in the area drains in a north-easterly direction into the Jema River (Figure 1). The mean monthly RF for the study area is 966.5 mm, with significant variations from month to month at different stations. January and February typically experience the lowest RF, while July and August see peak precipitation.

Collection and review of available primary data, field surveys, and integrated professional expert opinions were used to identify the input data for groundwater potential mapping of the research region (Sar et al. 2015). A panel of hydrogeologists, geologists, and GIS specialists in the study area were collected, and identified key thematic layers influencing groundwater potential, and this information performed pairwise comparisons using the Saaty scale (1–9) to assign relative importance to each factor (Saaty 1980). This expert-driven approach enhances the reliability of groundwater assessments, as demonstrated in previous studies (e.g., Jha et al. 2010a, 2010b; Rahmati et al. 2015; Naghibi et al. 2017a, 2017b, 2017c), which confirm the effectiveness of AHP in integrating expert knowledge for groundwater potential mapping. In this study, eight parameters influencing the potential of the groundwater of the area were used and identified as relevant data. These elements included RF, DD, Sl, LULC, LD, topographic wetness index (TWI), lithology, and soil types (S).

The study employs an integrated methodology combining geographic information system (GIS) and AHP tools to assess and delineate potential groundwater zones. The analysis involved collecting diverse datasets, including climatic, geological, hydrogeological, LULC, soil texture, structural and drainage data. Climate data were obtained from the National Meteorological Agency of Ethiopia, well data from regional and zonal water bureaus, geological maps and related information from the Geological Survey of Ethiopia (GSE) (Table 1). Additionally, field observations and verifications were carried out to address data gaps where insufficiency was identified.

Thematic layers were displayed in UTM projection zone 37 using the WGS84 datum and integrated using GIS tools and ASTER images with 30 × 30 m resolution. Polygonal features were converted into raster and point data formats and subsequently classified. Reclassification was performed within the ArcGIS platform, guided by factor prioritization and weighting. The reclassification process employed ArcGIS tools to assign new values to raster cells according to the significance of each parameter influencing groundwater potential. Higher values were assigned to conditions more conducive to groundwater availability, while lower values represented less favorable conditions, following established literature and expert judgment.

The reclassification procedure ensured that all parameters were aligned to a consistent spatial resolution and projection. The process/procedure involved several key steps to ensure the accurate mapping of groundwater potential zones. First, the influential factors were identified by selecting key thematic layers, including lithology, soil types, RF, LD, DD, Sl, LULC, and TWI, based on their impact on groundwater occurrence. Next, data acquisition and preparation were carried out by sourcing each thematic layer from relevant datasets and processing them to align with the study area's spatial extent and resolution. Finally, to facilitate the integration of diverse datasets into a cohesive analysis framework, each thematic layer was reclassified into a standardized scale, typically ranging from 1 to 5, where higher values indicate a greater potential for groundwater occurrence.

AHP is a matrix-based mathematical framework designed to evaluate the comparative importance of several criteria intuitively. It facilitates group decision-making by enabling stakeholders to leverage their experiences, values, and expertise to break down complex problems into hierarchal structures. These structures are solved step-by-step using AHP methodology (Chowdhury et al. 2010). One of AHP's notable strengths is its systematic consistency checks, which distinguish it from other multi-attribute value processes. Consequently, AHP is considered one of the foremost reliable GIS-based methods for delineating groundwater potential zones.

In this study, pairwise comparison matrices were constructed using AHP techniques for determining the total weight of primary criteria. The proportional weight of each criterion's respective classes was also considered. Eight parameters influencing groundwater potential were selected and weighted based on their relevance. The weights were normalized using the AHP framework, considering both criteria and sub-criteria in tandem. Following Saaty's (1980) AHP methodology, pairwise comparison matrices were developed to assign weights to themes and their sub-classes. The consistency ratio (CR) was then calculated to validate the normalized weights of various thematic layers and their classes.

All thematic layers were created and combined using the AHP technique, considering their respective values and impacts on the potential of the groundwater in the area. The important parameters affecting the presence of groundwater, including geology, hydrology, topography, and climatic elements, were systematically evaluated. The study aimed to identify which factor or parameter exerts greater or lesser influence on groundwater occurrence. To identify appropriate locations for potential groundwater zones using GIS-based AHP analysis, the following five procedural steps, as outlined by Machiwal et al. (2011), were employed:

• (a) Criteria selection and weighting: Identification and classification of relevant parameters influencing groundwater potential.

• (b) Pairwise comparison and weight normalization: Application of Saaty's methodology to assign and normalize weights.

• (c) Thematic layer generation: Preparation of thematic maps representing each parameter and their sub-classes.

• (d) Integration and overlay: Combining normalized thematic layers to generate groundwater potential maps.

• (e) Validation and consistency analysis: Calculation of the CR to ensure reliability of judgments and results.

The pairwise comparison method facilitated the weighting of thematic maps using AHP software (Table 2). AHP allows for the identification and adjustment of contradictory judgments through consistency measures, which was crucial in our decision-making process (Saaty 1980). The resulting weight vector was instrumental in pinpointing the key variables that influence groundwater potential within the watershed. Parameter weights were calculated using the nine-point analytical hierarchy principle (AHP) scales (Table 2). According to Saaty (1980), matrices with CR scores exceeding 0.1 require further verification, enabling the identification of inaccuracies in the defined interrelationships.

The process of summing the proportions of all the parameters and dividing the total by the number of parameters yields the principal eigenvalue and the random consistency index (RI) values corresponding to different matrix sizes (n) varies according to the number of layer used for groundwater potential assessment (Table 3). The AHP result shows that the eigenvalue is 8.489 (Table 4). The principal eigenvalue should be equal to or exceeded the total number of parameters within the matrices; if this condition is not met, the matrices require reconstruction (Saaty 1980). In this study, the principal eigenvalue of 8 by 8 matrices was employed to determine the consistency index (CI), as shown in Table 4.

Table 4

View Large

The standardization process involved several key steps to ensure consistency in weighting thematic layers for groundwater potential assessment. First, pairwise comparisons of the thematic layers using the Saaty scale (1–9) (Table 2), assigning relative importance to each factor based on its influence on groundwater occurrence (Table 4). These comparisons were then used to construct a pairwise comparison matrix, from which normalized weights were calculated for each thematic layer (Table 5), ensuring that their sum equaled one. Finally, a CR was computed (Equation (2)) to evaluate the reliability of the expert judgments, with a CR value below 0.1 considered acceptable, thereby confirming the validity of the weighting process.

To identify inconsistencies and derive optimal weights within the pairwise comparison matrix, the CR was computed following the completion of pairwise comparison and the determination of element weights. AHP analysis is considered valid and can proceed if the CR of 0.1 or lower (Saaty 2008). Conversely, if the CR exceeds this threshold, the analysis procedure is regarded as varying, rendering the AHP approach unreliable for producing accurate conclusions (Saaty 2008).

The RI values agreeing to various matrix sizes (n) were utilized, with an RI of 1.41 specifically assigned for n = 8 (Table 4). Subsequently, the CR was computed, yielding a value of 0.05 (5%), which is well below the permissible limit of 0.1 (10%), signifying that the evaluations are reliable and consistent This validation facilitated the combination of the eight thematic layers, together with their respective normalized weights, into ArcGIS software for advanced spatial analysis (Gedam & Dagalo 2020; Raja & Aneesh 2023).

Based on the AHP methodology (Table 4), the assigned weights for the parameters were as follows: lithology (38%), LULC (16%), TWI (14%), LD (10%), Sl (9%), DD (7%), soil (4%), and RF (3%) (Table 4). Lithology emerged as the most influential parameter in the catchment area, as determined by the weighted vector analysis. Conversely, soil and RF were ranked with the lowest priority. The weighted values (Table 5) for each thematic layer class were subsequently applied to reclassify and assign weights to the respective classes using the ARC GIS platform, enabling the overlay of all layers to generate the final groundwater potential index (GWPI) map.

The accuracy of the delineated GWPZ could be validated through the schematic yield data from the study area (Arulbalaji et al. 2019; Bogdan et al. 2024) and by comparing the groundwater well yield data from boreholes with the mapped GWPZ using multiple thematic layers. Therefore, in this study, the validation process involved overlaying observed yield data from 12 points – comprising nine (9) boreholes drilled by various organizations and three (3) springs – onto the categorized groundwater potential map.

Lithology serves as a natural reservoir and is a critical parameter in groundwater studies, significantly influencing the spatial distribution and availability of groundwater (Ramamoorthy & Rammohan 2015; Taye et al. 2021; Eshetu et al. 2024). For this study, the geological map of the Addis Ababa topo-sheet was used, and the lithological features were clipped and converted to raster format using the polygon-to-raster conversion tool for overlay analysis. Lithology plays a crucial role in determining both the quantity and quality of groundwater in a region (Hussein et al. 2017).

Lithology is a primary determinant in GWPZ delineation, as it governs water storage, infiltration, and movement within an aquifer system. Highly permeable formations enhance groundwater potential, whereas impermeable units restrict recharge. Incorporating lithology into GIS and AHP-based assessments significantly improves groundwater potential mapping, providing valuable insights for sustainable water resource management.

Groundwater circulation is significantly affected by lithological characteristics (Chowdhury et al. 2010), as a higher porosity and permeability within lithologically interconnected units lead to increased groundwater storage and enhanced yield. The hydraulic properties of aquifers, such as permeability, hydraulic conductivity, and specific yield, are governed by the physical characteristics of the rock, land feature, overlying unit, and the degree of weathering, among other factors (Singh & Singh 2010; Haile et al. 2024).

LD is defined as the total length of all identified lineaments within a given area divided by the area of the watershed (Pradhan 2009; Eshetu et al. 2024; Adugna & Aweke 2024). Lineaments represent the most prominent structural elements important from a groundwater standpoint and are prominent structural features, including linear arrangements of geological, lithological, topographic, and drainage anomalies. These features may appear straight or curved and have a crucial impact on facilitating the absorption of overland flow into subsurface formation, which is vital for groundwater storage Roy et al. 2020). Areas with fractured bedrocks, which exhibit porosity and permeability, are often associated with increased well yields (Frederiks & Lowry 2022).

The LD map was produced using the spatial analysis tool ArcGIS, employing the line density function. The map was then reclassified using the equal interval method into five categories: very low (0–0.11), low (0.12–0.30), moderate (0.31–0.47), high (0.48–0.67), and very high (0.68–1.2) density (Figure 3(b) and Table 7). Lineaments represent fractures and fault zones that often serve as conduits for groundwater movement. Areas with higher LD were identified as more favorable for groundwater recharge and were associated with increased groundwater storage and flow, as fractures enhance permeability (Saha et al. 2010; Naghibi et al. 2017a, 2017b, 2017c); moreover, regions with a high density of lineaments also exhibit greater well yields due to enhanced secondary porosity (Jha et al. 2010a, 2010b). However, areas with lower LD were less suitable for recharge and discharge (Roy et al. 2020). The map also provided insights into groundwater flow directions within the catchment area.

Sl is a critical factor influencing groundwater recharge systems, as it directly influences overland flow and infiltration dynamics. The Sl, combined with other geomorphological features, serves as a significant determinant of groundwater potential in the area. Regions with low Sl gradients experience reduced surface runoff. Conversely, areas with steep Sls exhibit higher runoff and shorter residence times, limiting infiltration and recharge. In general, low-angle Sls contribute significantly to groundwater recharge. Sl is a critical determinant of groundwater potential (Rajesh et al. 2012) since it regulates the processes of vertical infiltration and runoff from the surface, both of which directly influence groundwater recharge (Adiat et al. 2012; Kumar et al. 2017; Eshetu et al. 2024). Sl and infiltration share an inverse relationship, whereby steeper Sls result in higher runoff and lower infiltration, while gentler Sls facilitate greater infiltration and reduced runoff (Jha et al. 2010a, 2010b; Yeh et al. 2016; Naghibi et al. 2017a, 2017b, 2017c; Adugna & Awoke 2024).

For this study, the Sl map was developed using ArcGIS software and a SPOT DEM with 30 m resolution. The Sl data were categorized into five distinct classes following the FAO classification scheme (FAO 2006). Lower Sl values correspond to gentle terrain, while elevated values correspond to steeper topography. The morphological characteristics of the terrain, including pediplains, low reliefs, gently rolling Sls, linear ridges, and valleys, were used to assign scale values to Sl classes. Steeper Sls are associated with rapid surface drainage, while flatter terrain, such as pediplains, supports prolonged water retention, facilitating precipitation infiltration and recharge.

Soil texture significantly influences groundwater permeability within a region (Kudamnya & Andongma 2017; Taye et al. 2021). Coarse-grained soils with porous structures are particularly conducive to groundwater recharge as they promote the infiltration and percolation of water into the subsurface. The interaction of soil characteristics regulates overland flow and infiltration rates, which consequently dictate soil permeability.

For this study, soil data were obtained from the FAO database, which provides detailed information on soil physical properties, including texture. The infiltration potential and soil moisture properties of the study area were assessed using a soil classification map derived from texture-based factors. A soil-type map produced by Ethiopia's Ministry of Agriculture (MoA) was utilized. Soil features were clipped and converted to raster format using a polygon-to-raster conversion tool in ArcGIS for overlay analysis.

The study area's soil texture and thickness vary significantly across different locations. The river basins are predominantly covered with thick alluvial soils, which are commonly found near rivers, streams, and flat plains and exhibit variable thickness. The catchment area comprises five distinct soil types: Eutric Vertisols, Eutric Cambisols, Haplic Luvisols, Rendzic Leptosols, and Lithic Leptosols (Figure 4(b), Table 7). Eutric Vertisols are the predominant soil type in the study area, mainly covering the highland regions. Lithic Leptosols are the second most common soil type, primarily distributed across the lowland and central areas. In contrast, Eutric Cambisols, Haplic Luvisols, and Rendzic Leptosols occur in isolated pockets, with Eutric Cambisols found in the northwest, Haplic Luvisols in the central region, and Rendzic Leptosols in the northern part of the study area (Figure 4(b) and Table 7).

These soil types were categorized into three texture classes: coarse, medium, and fine based on their grain size. Groundwater compatibility was evaluated following the recommendations of Gumma & Pavelic (2013) and Adugna & Awoke (2024), focusing on the relationship between runoff and infiltration rates, which affect permeability, a critical hydrogeological criterion for assessing groundwater potential.

Sandy soils promote high infiltration rates, whereas clay-rich soils impede percolation and reduce groundwater recharge (Rahmati et al. 2015); sandy soils have high permeability and porosity, allowing rapid infiltration and facilitating groundwater recharge (Jha et al. 2010a, 2010b; Rahmati et al. 2015). Loamy soils, which contain a mix of sand, silt, and clay, provide an optimal balance between water retention and permeability, making them ideal for groundwater recharge (Saha et al. 2010). Hence, regions with predominantly sandy and loamy soils are often classified as high GWPZ due to better infiltration.

Clayey soils were assigned the lowest value due to their poor percolation capacity, attributed to the presence of clay layers, which severely restrict water movement. Conversely, coarse sandy loam soils received the highest value due to their high permeability and rapid infiltration rates. Medium-grained soils, such as those in Haplic Luvisols, indicate moderate groundwater potential.

Fine-grained soils, including Eutric Vertisols and Eutric Cambisols, were associated with low to very low groundwater potential zones, as the lack of interconnected pores restricts water movement. In contrast, coarse-grained soils, including Lithic Leptosols and Rendzic Leptosols, were identified as zones of very high groundwater potential due to their enhanced permeability and infiltration capacity (Hussein et al. 2017).

As a result, Haplic Luvisols, characterized by medium to coarse-grained textures, represent high GWPZ (Figure 4(b)), while Lithic Leptosols and Rendzic Leptosols, being coarse-grained, indicate zones with very high groundwater potential.

LULC are critical components of watershed systems, significantly influencing infiltration, erosion, and evapotranspiration. The LULC map is a key regulatory element for identifying areas with adequate potential for groundwater recharge (Chuma et al. 2013; Ifediegwu 2022; Adugna & Awoke 2024). The development and distribution of groundwater resources are strongly affected by the prevailing LULC types in a region. Surface coverage, such as vegetation, increases surface roughness, promoting infiltration and reducing runoff. Conversely, urban areas exhibit high runoff and low infiltration rates and limit groundwater recharge (Machiwal et al. 2011; Magesh et al. 2012), whereas forested regions typically enhance infiltration and reduce evaporation. Moreover, agricultural lands, depending on irrigation practices, may contribute significantly to groundwater recharge (Rahmati et al. 2015).

The potential for groundwater recharge varies significantly across different LULC types. Vegetation-covered areas, such as forests and croplands, exhibit higher infiltration rates compared to urbanized or barren surfaces, which are less conducive to groundwater recharge (Singh et al. 2010; Li et al. 2018). While some land use practices negatively impact groundwater recharge, others can have positive effects. For instance, urban and residential areas generate substantial runoff, leading to reduced recharge potential (Owuor et al. 2016; Tolche 2021; Kassa et al. 2023). The impact of croplands on groundwater recharge depends on the soil and water conservation practices employed.

Among the identified LULC classes, water bodies, bare lands, and forested areas received the lowest weights for groundwater potential. The prioritization or spatial distribution of these classes of LULC units for groundwater suitability in the study area follows the order: croplands > grasslands > forests > built-up areas > bare lands > water bodies (Figure 5(a) and Table 7).

RF is the main source of groundwater recharge and is integral to all hydrological processes (Agrawal et al. 2013; Merouchi et al. 2024). It is integral to the natural water cycle and significantly influences groundwater potential. Extended periods of precipitation events have a higher probability of contributing significant groundwater recharge compared to short-duration RF, which provides limited recharge opportunities (Kotchoni et al. 2019). Precipitation is a primary source of groundwater recharge. However, its contribution is dependent on other factors such as infiltration capacity and geological conditions (Machiwal et al. 2011). Studies confirm that areas with high RF but poor infiltration characteristics do not necessarily have high groundwater potential (Saha et al. 2010).

For this study, RF data spanning 1998 to 2018 were collected from the National Meteorological Agency of Ethiopia. Annual average RF data points were interpolated using the Inverse Distance Weighting method and categorized into five classes according to the natural breaks in the data (Kumar et al. 2017; Merouchi et al. 2024). Regions receiving higher RF were assigned greater weights, reflecting the increased probability of groundwater recharge in those areas. Conversely, regions with lower RF received lower weights.

RF distribution exhibits both regional and seasonal variability, necessitating detailed analysis to understand its impact on groundwater recharge within a specific area. In the study region, the northeastern areas experience relatively high RF, whereas the southern parts receive comparatively lower precipitation. On average, the watershed receives between 824 and 1,188 mm of RF annually, indicating a relatively high precipitation regime throughout the catchment (Figure 5(b) and Table 7). The spatial distribution of average RF over the past 20 years, as recorded at the region's principal climate stations, is depicted in Figure 5(b).

DD quantifies the closeness of stream channels within a catchment area, expressed as the total length of all stream segments per unit area (Singh et al. 2014). It is a critical factor affecting groundwater recharge, as it reflects the permeability and infiltration capacity of the underlying rock. Higher DD corresponds to increased runoff and reduced infiltration, whereas lower DD indicates enhanced infiltration and recharge potential (Yalcin 2008; Merouchi et al. 2024). Areas with dense drainage networks typically have high surface runoff and low infiltration rates, leading to low groundwater recharge potential. This is often associated with impermeable lithology, steep Sls, and low-porosity soils (Jha et al. 2010a, 2010b; Rahmati et al. 2015). Flat or gently sloping terrain tends to have lower DD, enabling higher groundwater recharge, while steep terrains with high DD promote rapid runoff, reducing infiltration capacity. Consequently, it serves as a key regulating factor for groundwater potential zones.

In this study, watershed drainage was delineated using SPOT DEM data with a 30 m resolution. The ArcGIS spatial analysis line density tool was employed to create a DD map from polyline stream inputs. The catchment area exhibits a dendritic drainage pattern, resembling a tree-like structure. The calculated DD was categorized into five classes using equal intervals, according to Kumar et al. (2017).

The TWI has been extensively employed to analyze the spatial scaling effects of hydrological processes. For the study area, the TWI was classified into five classes (Table 7) of equal intervals: 18.1–29.8 (very high), 14.1–18 (high), 11.9–14 (moderate), 10.2–11.8 (low), and 5.9–10.1 (very low). These classifications provide insights into areas with potential groundwater zones, as illustrated in Figure 6(b). TWI serves as a crucial topographic indicator in groundwater potential zoning, helping to identify high recharge areas in valleys and depressions while distinguishing low recharge zones on ridges and steep Sls.

Determining GWPZ is crucial for effective water resource management, especially in arid regions. Recent studies have demonstrated the efficacy of integrating RS, GIS, and MCDA techniques in this endeavor. For instance, Ajayi et al. (2022) applied these methods in Nigeria, producing a groundwater potential map that was validated with hydrogeophysical data, showing a 68% positive correlation. Similarly, Bogdan et al. (2024) utilized RS and GIS combined with the AHP to delineate GWPZ in Romania, achieving an overall accuracy of approximately 72%. A groundwater potential map evaluates the terrain's physical ability to provide sufficient groundwater resources for specific beneficiaries by analyzing multiple indirect indicators (Gojiya et al. 2018; Ifediegwu 2022). The delineation of GWPZ leverages RS and GIS techniques. To derive the GWPI, eight thematic maps were integrated using the overlay analysis method in ArcGIS software. The eight thematic layers were classified, reclassified, and overlaid to generate the final GWPZ map, which was categorized into three distinct zones: low, moderate, and high groundwater potential.

The analysis indicated that areas with significant groundwater potential are primarily located in the upper catchment. This is attributed to the presence of a Quaternary alluvial sedimentary aquifer in the plains, which exhibits favorable LD conducive to groundwater recharge. High-potential zones encompass approximately 70% (520 km²) of the total Weserbi watershed.

Moderate GWPZ are predominantly distributed in the northern ridge areas, representing a transitional terrain between plains and valleys, characterized by moderately inclined Sls and hill ridges. These zones are also found in parts of the southern, southwestern, and central regions, particularly along areas with higher LD. Overall, moderate-potential zones cover about 24% (184 km²) of the watershed.

In contrast, low GWPZ are concentrated in the eastern regions of the study area, covering approximately 6% (44 km²) of the watershed. These areas are characterized by rugged hill and ravine landscapes, coupled with a lower abundance of water-bearing lithological formations such as mudstone and claystone.

The validation of GWPZ was conducted by analyzing well and spring yields in comparison with the mapped groundwater potential. In this study, yield data from nine (9) boreholes and three (3) springs were assessed, showing a strong correlation with the predicted groundwater potential zones, demonstrating strong agreement between observed data and the developed GWPZ map (Figure 7). This consistency underscores the reliability of the GWPI and the resulting map. Consequently, the validation confirms that the majority of the assessments align well with the collected yield data. The results suggest that the methodology used for groundwater potential mapping is robust and applicable for further hydrogeological assessments. Similar studies have also validated GWPZsusing well and spring yield data, reinforcing the effectiveness of this approach (e.g., Rajasekhar et al. 2020; Jha et al. 2007).

An integrated approach that combines AHP and GIS approaches was used to evaluate the groundwater potential of the research area. The AHP technique was employed to facilitate multi-criteria decision-making for factors influencing groundwater occurrence and movement. Eight thematic layers, including lithology, LD, Sl, TWI, soil, LULC, RF, and DD, were selected, evaluated, and analyzed.

The results demonstrated a strong correlation between the GIS-based groundwater potential analysis and the observed groundwater inventory data. The AHP-derived weights for the thematic layers were as follows: geology/lithology (38%), LULC (16%), TWI (14%), LD (10%), Sl (9%), DD (7%), soil (3%), and RF (3%). The findings showed that the groundwater potential of the area is predominantly influenced by lithology and LULC, with other factors contributing to a lesser extent according to their proportional significance and influence.

Pairwise comparisons were conducted for these thematic layers, both within and across individual factors. These comparisons were incorporated using overlay analysis in ArcGIS to develop the GWPZsmap. Three potential zones were delineated on the final map corresponding to the study area: high (70%), moderate (24%), and low (6%), and validated with borehole well and spring yields in the study area. This study underscores the effectiveness of GIS and AHP as tools for mapping groundwater potential zones, offering a time-efficient and cost-effective approach to groundwater resource assessment. By categorizing the area into zones of high, moderate, and low potential, this study provides valuable insights for planning and managing groundwater resources to meet current and future demands.

In the identified high groundwater potential zones, covering 70% of the study area, can be prioritized by the local decision-maker for well drilling and borehole development to ensure reliable water access for domestic and agricultural use. Besides, to promote sustainable groundwater utilization, integrating these findings with hydrogeochemical assessments will enable the local authorities to implement management strategies that prevent over-extraction and aquifer depletion. Additionally, the moderate and low GWPZs can be targeted for artificial recharge projects, such as check dams and managed aquifer recharge, to enhance water storage and mitigate the effects of droughts and seasonal fluctuations. Additionally, continuous monitoring and hydrogeochemical assessments should be conducted to track groundwater quality and availability, enabling adaptive management approaches. Public awareness and policy interventions should also be strengthened to encourage responsible groundwater use, ensuring long-term water security for domestic, agricultural, and industrial needs.

While this study provides a robust groundwater potential assessment, future research should focus on integrating additional data sources such as geophysical surveys and groundwater quality assessments to enhance model accuracy. Machine learning techniques, including random forest and deep learning algorithms, could also be explored for comparative analysis with the AHP approach. Furthermore, long-term monitoring of groundwater levels and recharge rates would be beneficial for validating and refining the predictive capability of groundwater potential models. Incorporating climate change scenarios into groundwater potential mapping could also help assess the long-term sustainability of groundwater resources in the region.

The first author gratefully acknowledges the financial support provided by Mekelle University, which made this research possible. Their support is sincerely appreciated. Finally, the authors are grateful for the constructive comments and suggestions of the two anonymous reviewers for improving the manuscript.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.