Introduction

Groundwater, a critical component of global water resources, accounts for 36%, 42%, and 24% of global drinking, irrigation, and industrial water usage, respectively (Das and Mukhopadhyay 2020; Scanlon et al. 2023). The interplay between population growth, economic development, and the increasing demand for water in the industrial and agricultural sectors has led to overexploitation and subsequent groundwater depletion (Dolan et al. 2021). China, with its substantial groundwater reserves, faces the challenge of uneven distribution exacerbated by swift demographic and economic growth (Fan and Fang 2024). The critical situation in Weifang, Shandong Province, exemplifies this phenomenon. Groundwater overexploitation and utilization in Weifang have resulted in regional overexploitation, posing significant threats to food and water supplies, ecological security, and the capacity for sustainable economic and social development. In response to the pressing need to curb the unsustainable exploitation of groundwater resources, artificial recharge has become a key strategy to ease the growing demand for these resources and ensure sustainable use (Jin et al. 2024). It is important to employ scientific methods to identify and assess potential areas for artificial groundwater recharge. This requires a comprehensive investigation and assessment of the distribution, recharge, and storage characteristics of groundwater resources (Fauzia et al. 2021).

The assessment of artificial recharge suitability zones (ARZs) typically requires an evaluation to determine the effective utilization potential of groundwater resources (Shekar and Mathew 2023). Conventional techniques for groundwater potential assessment depend on technologies including geophysical exploration, geological surveys, and drilling (Bunkar and Nema 2022). Although accurate, these methods are often expensive and time-consuming (Suryawanshi et al. 2023). Owing to technological progress, contemporary methodologies incorporate probabilistic models and machine learning techniques. These include the Frequency Ratio method (Razandi et al. 2015), logistic regression (Nguyen et al. 2020), multiple-criteria decision-making (MCDM) (Ghosh et al. 2022), artificial neural networks (Chowdhuri et al. 2020), and the random forest method (Phong et al. 2021). The integrated application of these modern methods facilitates more effective and expedited identification and evaluation of ARZs. Developing technologies, such as Geographic Information Systems (GIS) and Remote Sensing (RS), is particularly noteworthy (El khalki et al. 2024). GIS provides a comprehensive platform for ARZ assessment by integrating and analyzing various spatial data (Noori et al. 2022).This allows data amalgamation from multiple sources, including topography, soil types, land use, and hydrological conditions, and identifies and comprehends ARZs through spatial analysis capabilities (Xiong et al. 2022). Using RS allows for the non-contact assessment of surface features through satellite observation, providing a valuable approach for groundwater research (Xiao et al. 2023). However, RS data cannot be used directly to access groundwater resource data, highlighting the need to establish correlations between surface features and groundwater conditions (Das et al. 1997; Khan et al. 2021). The application of GIS and RS provides a comprehensive strategy for identifying and understanding ARZs (Yin et al. 2021; Mathenge et al. 2022). However, challenges persist when dealing with intricate groundwater systems.

The MCDM is effective in resolving complex issues by enabling the formulation of more precise decisions (Seif et al. 2024). MCDM enables decision-makers to make well-rounded decisions amidst a multitude of criteria, some of which may be conflicting (Pavelic et al. 2021). The Analytic Hierarchy Process (AHP), a form of MCDM, is widely applied owing to its adaptability, efficiency, dependability, and practicality. The AHP method was designed to simplify complex problems into manageable segments using a hierarchical approach (Chang et al. 2020). This structure allows for comparing elements within each level using a 1–9 scale to determine their relative importance (Tavana et al. 2023). However, the conventional AHP approach may encounter constraints when confronted with ambiguity and subjective assessments (Lipovetsky 2021). The Fuzzy Analytic Hierarchy Process (FAHP) was developed to address this issue. The FAHP methodology introduces the concept of fuzzy numbers, constructs fuzzy complementary judgment matrices, and employs fuzzy operations to calculate the fuzzy weights of each factor (Ahmed and Kilic 2019). This method ensures a more accurate representation of decision-makers' preferences and judgments amidst uncertainty, enabling a more precise evaluation of ARZs and enhancing the congruence between decision outcomes and actual scenarios (Liu et al. 2020). An approach that integrates FAHP, GIS, and RS has proven to be an effective methodology for identifying ARZs (Agarwal et al. 2013; Chen et al. 2018; Singha et al. 2024).

The principal goal of this study is to develop an integrated methodology that leverages RS technology and GIS for data acquisition and synthesis, followed by the development of an artificial recharge suitability analysis model using Fuzzy AHP (FAHP). This model provides a scientific foundation for the judicious development and sustainable management of groundwater resources by identifying suitable zones for artificial recharge within the study area. This study integrated 14 pivotal thematic layers within the watershed and applied FAHP for an exhaustive analysis to assess their impacts on ARZs. Furthermore, the optimal zones for groundwater recharge within the watershed were identified, and the generated ARZ maps were juxtaposed with existing data to assess the model's efficacy. The area under the receiver operating characteristic curve (AUC) served as an evaluation metric, offering a quantitative measure of a model's reliability and precision.

Study area

The Mi River watershed, located in Weifang, Shandong Province, China, is defined by geographical coordinates ranging from 118°28' to 119°7' East and 16°16' to 37°3' North (Fig. 1). The mainstream length of the basin is 193 km and covers a total area of approximately 3,319 km2. The region has a temperate continental climate characterized by significant temperature fluctuations. The region's average annual temperature is approximately 13 °C, with recorded extremes reaching up to 39.7 ºC and as low as –19.3 ºC.

Fig. 1
figure 1

Location of the Mi River watershed

Full size image

The average annual precipitation in the Mi River watershed is approximately 650.8 mm, occurring predominantly during the flood season from July to September. Precipitation during this period represented 71.1% of the total annual, averaging 462.5 mm. Seasonal precipitation patterns have a significant impact on the distribution and utilization of water resources in basins. The river channel in the watershed is primarily a mountain stream rain-flood channel, with discharge varying during the rainy season. Meanwhile, during the dry season, the river flow may be reduced to a point, where it is entirely interrupted. Consequently, the water demand of the downstream area is heavily reliant on the groundwater supply.

Methods and material

Methods for identifying ARZs

The assessment of artificial recharge suitability is a complex multidimensional process that necessitates the integration of various interrelated key factors. The precipitation frequency and distribution are the primary factors affecting groundwater replenishment (Bhadran et al. 2022) and are essential for replenishing and preserving groundwater resources. In addition to precipitation, several other factors significantly impact the suitability of artificial recharge. This encompasses the groundwater depth, soil classification, topographical features, and geological composition (Jin et al. 2024). Together, these factors determine the storage capacity, infiltration routes, and overall water availability. This study integrated a comprehensive set of multidimensional factors, including drainage density (Dd), elevation (E), fault distance (F), geomorphology (Gm), land use/land cover (LULC), rainfall (R), slope (Sl), soil texture (St), groundwater depth (WD), total dissolved solids (TDS), terrain roughness index (TRI), topographic wetness index (TWI), topographic position index (TPI) and curvature (C). The interplay between these factors establishes a multiparameter, multiscale analytical framework for quantifying artificial recharge suitability (Aju et al. 2021).

To assign appropriate weights to thematic parameters, this study employed the FAHP, an improved version of the traditional AHP developed by van Laarhoven and Pedrycz (1983). The FAHP effectively resolves the inherent inconsistencies in traditional comparison matrices (Buckley 1985). The integration of GIS with FAHP technologies enables the categorization of the study area into five distinct levels of artificial recharge suitability, resulting in the creation of a suitability map (Karipoğlu et al. 2023) (Fig. 2). This map not only reveals the spatial distribution characteristics of groundwater resources but also provides a scientific basis for the rational recharge and sustainable management of these resources, thereby ensuring the precision and adaptability of water resource assessments (Karipoğlu et al. 2023).

Fig. 2
figure 2

Flowchart for identifying ARZs

Full size image

Data sources

This study utilized sophisticated spatial data processing methods, leveraging the ASTER GDEM with a 30-m resolution, to generate key topographic parameter maps for the study area. These maps include elevation, slope, drainage density, curvature, TWI, TPI, and TRI, which collectively provide foundational data for topographic analysis of the region (Table 1).

Table 1 Data utilized in this study
Full size table

Integrating RS and GIS is crucial for analyzing groundwater research data (Yin et al. 2021; Paredes-Beltran et al. 2024). Foundational data on precipitation, soil type, and geomorphological structure were obtained from the Institute of Tibetan Plateau Research of the Chinese Academy of Sciences to guarantee the scientific robustness and reliability of the dataset. The LULC data were derived from the annual CLCD at a 30-m resolution, offering insights into the prevailing land use patterns within the study area. Critical fault data, essential for understanding the geological context of groundwater analyses, were supplied by the Geological Survey Bureau. Additionally, data on the groundwater depth and total dissolved solids (TDS) were procured from the Hydrological Center of Shandong Province to complement the dataset.

The pairwise comparison matrix

The pairwise comparison matrix, a cornerstone of AHP, is used to quantify the relative significance of factors within a hierarchical framework (Tavana et al. 2023).This study utilized Saaty's 1–9 scale (9 denotes the highest level of importance, whereas 1 indicates equivalence in importance) to determine the relative priorities among diverse factors (Bhadran et al. 2022). In this context, the relative importance of the factors gradually decreased as the numerical magnitude increased. However, traditional AHP does not fully encapsulate the ambiguity and subjectivity inherent in decision-makers' judgments (Javanbarg et al. 2012). To counter this limitation, fuzzy logic has been integrated into the AHP framework, resulting in FAHP development. FAHP employs triangular fuzzy numbers (TFNs) to represent the results of pairwise comparisons (Reig-Mullor et al. 2020). A TFN is defined by an interval [l, m, and u], where "l" represents the lower bound, "m" denotes the most probable value, and "u" signifies the upper bound (Raja Shekar and Mathew 2023). The use of TFNs allows for a more precise representation of the decision-makers’ preferences, accounting for the potential ambiguity and imprecision from linguistic scales.

To formulate the fuzzy pairwise comparison matrix, expert opinions are solicited and subsequently integrated into the matrix form. Each matrix entry is represented by a TFN that reflects the relative importance of one factor over another. Subsequently, the matrix was normalized to ensure the consistency and comparability of the fuzzy evaluations (Meng and Chen 2017). The normalization process is implemented by calculating the weighted average of the fuzzy numbers, which produces a clear value that represents the relative importance of each factor more precisely manner (Shi and Wang 2019). The FAHP methodology provides a structured and thorough approach to integrating fuzzy logic with the AHP (Khashei-Siuki et al. 2020). This methodology allowed for a more refined and comprehensive assessment of factor prioritization within artificial recharge suitability. The resulting pairwise comparison matrices for the primary parameters, as developed through the FAHP analysis, are presented with their corresponding composite weights (see Table 2).

Table 2 Pariwise comparison matrix that has been FAHP
Full size table

Consistency check

The pairwise comparison process is typically based on a subjective judgment, inevitably leading to inconsistencies. To ensure the consistency of the judgments, the comparison results were verified through consistency checking (Kubler et al. 2016). The Consistency Ratio (CR) was calculated by comparing the Consistency Index (CI) with a random Consistency Index (RI) (Table 3). The CR serves as a vital metric for assessing an acceptable level of consistency within a pairwise comparison matrix.

Table 3 RI for various n scores
Full size table

According to Saaty's criterion, a matrix with a CR value ≤ 0.1 is deemed to have passed the consistency test, reflecting satisfactory consistency, while a CR value of > 0.1 suggests that the pairwise comparison matrix requires revision to guarantee the dependability and accuracy of the analytical results (Saaty 1987, 1980). In this study, the computed CR value was 0.03, which falls below the threshold 0.1, thereby meeting the requirement for consistency verification.

$$\text{CI}=\left({\uplambda }_{\text{max}}-\text{n}\right)/\left(\text{n}-1\right)$$
(1)

where:

\({\uplambda }_{\text{max}}\) represents the maximum eigenvalue and \(\text{n}\) represents the order of the judgment matrix.

$$\text{CR}=\text{CI}/\text{RI}$$
(2)

Weight normalization

Weight normalization is a pivotal step in the evaluation process, ensuring the equilibrium and precision of the influencing factors in the comprehensive analysis (Jiang et al. 2017). In this study, the criteria affecting the suitability of artificial recharge were comprehensively evaluated using a linear combination of techniques. Weights were allocated to each thematic layer, corresponding to the relative significance of each criterion's impact on suitability. These criteria are subsequently divided into sub-criteria, each assigned an independent numerical rating (Saaty 2004). A scale of 1–9 was used to assign ranks to each subclass, as shown in Table 1. The 14 thematic map layers were integrated using the weighted overlay technique in ArcGIS to ensure that the final partition map reflected the combined influence of all factors (Table 4). The total weight of each influencing factor is determined using the following formula:

Table 4 Criteria-assigned ratings and weights for ARZs using FAHP
Full size table
$$\text{ARZ}={\text{Ddw}}_{\text{i}}\,{\text{Ddr}}_{\text{i}}+{\text{Ew}}_{\text{i}}\,{\text{Er}}_{\text{i}}+{\text{DFw}}_{\text{i}}\,{\text{DFr}}_{\text{i}}+{\text{Gmw}}_{\text{i}}\,{\text{Gmr}}_{\text{i}}+{\text{LULCw}}_{\text{i}}\,{\text{LULCr}}_{\text{i}}+{\text{Rw}}_{\text{i}}\,{\text{Rr}}_{\text{i}}+{\text{Slw}}_{\text{i}}\,{\text{Slr}}_{\text{i}}+{\text{Stw}}_{\text{i}}\,{\text{Str}}_{\text{i}}+{\text{WDw}}_{\text{i}}\,{\text{WDr}}_{\text{i}}+{\text{TDSw}}_{\text{i}}\,{\text{TDSr}}_{\text{i}}+{\text{TWIw}}_{\text{i}}\,{\text{TWIr}}_{\text{i}}+{\text{TPIw}}_{\text{i}}\,{\text{TPIr}}_{\text{i}}+{\text{TRIw}}_{\text{i}}\,{\text{TRIr}}_{\text{i}}+{\text{Cw}}_{\text{i}}\,{\text{Cr}}_{\text{i}}$$
(3)

Where w is the feature weight, r is the rank of each subcategory of the feature, and i represents the 14 influencing factors. According to the importance of each factor to the groundwater potential, the weight and grade of the characteristics are expressed by the subscripts w and r, respectively.

Results and discussion

In evaluating artificial recharge suitability, the precise assignment of weights to the selected groundwater contributing factors is of utmost importance to ensure the accuracy and reliability of the evaluation results. The study assigned weights to the selected factors based on expert knowledge gathered and an extensive literature review. The process entailed a comprehensive examination of the influence of each factor on the spatial distribution and migration dynamics of groundwater.

Rainfall

Rainfall is a key contributor to groundwater reserve replenishment. Its spatial and temporal distribution, as well as intensity, exerts a profound impact on the recharge potential of aquifers (Agarwal et al. 2013). Therefore, when assessing the suitability of artificial recharge, areas with high rainfall should be prioritized (Das and Mukhopadhyay 2020). To gain deeper insight into the contribution of rainfall to groundwater replenishment, this study conducted an integrated analysis of the mean annual rainfall data from 2004 to 2023. Interpolation of rain gauge data from 2004 to 2013 and application of the weighted overlay technique enabled the creation of a map depicting the spatial distribution of rainfall in the study area. Considering the natural variability of rainfall, this study used the Natural Breaks method to classify rainfall levels in the wet and dry seasons into five grades to account for natural fluctuations in rainfall patterns. The rainfall levels during the wet season are as follows: (a) < 202.56 mm (3.59%); (b) 202.56–236.96 mm (17.33%); (c) 236.96–258.85 mm (36.59%); (d) 258.85–286.21 mm (28.89%); (e) > 286.21 mm (13.59%) (Fig. 3a). Meanwhile, the dry season rainfall is divided into five grades, as follows: (a) < 34.77 mm (5.78%); (b) 34.77–55.57 mm (6.36%); (c) 55.57–72.21 mm (13.00%); (d) 72.21–85.07 mm (48.92%); (e) > 85.07 mm (25.95%) (Fig. 3B).

Fig.3
figure 3figure 3

Reclassify maps of: A Rainfall distribution of the wet season; B Rainfall distribution of the dry season; C Water depth of the wet season; D Water depth of the dry season; E Soil texture; F LULC; G Slope; H Geomorphology; I Drainage density; J TDS of the wet season; K TDS of the dry season; L Distance from fault; M Elevation; N TWI; O TPI; P TRI; Q Curvature

Full size image

These figures illustrate the distribution of average annual rainfall in the study area under varying rainfall conditions. During the wet season, the overall rainfall was relatively high, with a decreasing trend from the southwest to the northeast and scattered areas of decreased rainfall in the southeast, whereas during the dry season, rainfall was generally lower, and regions with increasingly less rainfall were intermingled in the southern part of the study area, reflecting the spatial heterogeneity of rainfall during the dry season. Our findings align with the seasonal fluctuations and spatial heterogeneity proposed by Magesh et al. (2012), further substantiating the potential for rainfall to recharge groundwater across different regions and seasons. This conclusion is corroborated by the findings of Diriba et al. (2024), highlighting the significant impact of the complexity of rainfall patterns on groundwater recharge.

Water depth

Groundwater depth, a crucial metric for assessing the water storage potential of aquifers, is negatively correlated with the suitability of groundwater recharge (Aloui et al. 2024). In particular, a greater groundwater depth is indicative of a larger storage space, which, in turn, provides a higher potential for replenishment. Conversely, a shallower groundwater depth may impede the efficacy of artificial recharge and increase the risk of surface salinization and other environmental issues (Zhou et al. 2024). This study categorized the research area into five levels based on the depth of groundwater: (a) < 7.58 m; (b) 7.58–10.79 m; (c) 10.79–14.27 m; (d) 14.27–18.01 m; (e) > 18.01 m (Fig. 3c,d). Considering the pronounced seasonal fluctuations in the groundwater table, this study categorized the groundwater depth data into distinct wet and dry seasons for more granular analysis.

The interpolation results indicate that the groundwater depth in the central part of the study area is relatively shallow. The area in question is predominantly plains with flat topography, and the soil and rocks exhibit favorable permeability, which facilitates the accumulation of groundwater in shallow strata, manifested by a lower depth (Yang et al. 2022). In contrast, the Southwest and Northeast regions demonstrated greater groundwater depths due to variations in topography and geological conditions. The southwestern region is predominantly mountainous and hilly, with compact rock structures that impede groundwater accumulation in shallow layers, facilitating deeper infiltration (Chen et al. 2024). The higher degree of soil salinization in Northeast China restricts the capacity of the soil to retain groundwater in the shallow layer, thereby increasing the depth at which the groundwater must be extracted. The groundwater depth was significantly affected by seasonal fluctuations. The groundwater depth is typically deeper during the dry season than during the wet season. This phenomenon is most likely attributed to reduced rainfall and increased evaporation, which collectively cause a seasonal decline in the groundwater table. In the southwest, the groundwater depth is relatively shallow during the wet season but deepens during the dry season. This reflects the sensitivity of groundwater to seasonal hydrological variations.

The consistent shallow groundwater depths observed in the central region during wet and dry seasons indicate that implementing artificial recharge initiatives requires a more prudent and detailed planning approach. Inappropriate recharge activities can result in surface salinization, degradation of wetland ecosystems, and alteration of natural groundwater flow paths. Consequently, a lower weight was assigned to this region during the analysis.

Soil type

Soil texture and hydraulic properties are important for estimating runoff and infiltration rates (Bera et al. 2020). The thematic soil layer of the study area was prepared using the Digital Earth Open Platform. Through analysis and reclassification of the original data, the study area was classified into 37 different soil texture categories, further divided into eight main soil texture structure groups and two types of water bodies. The specific distribution was as follows: Cambisols (22.42%), Lucisols (28.10%), Leptosols (1.09%), Regosols (27.45%), Vertisols (0.44%), Gleysols (1.90%), Fluvisols (13.39%), Solonchaks (4.83%), Water bodies (0.15%), and Fishpond (0.23%) (Fig. 3E).

Soil permeability is a critical factor for groundwater replenishment. Reduced groundwater replenishment is typically associated with low permeability (Döll and Fiedler 2008). The findings are consistent with this view. Fluvisols, commonly found in alluvial and floodplains, are typically distinguished by their higher permeability, which results in a higher weight being assigned to them (Ielpi et al. 2018). The physical properties of these soils facilitate rapid water infiltration, enhancing their groundwater recharge potential. In contrast, Vertisols and Solonchak are frequently associated with human activities, such as cultivation and irrigation, which can affect the structure and salt content of the soil. Owing to their lower permeability, these soils were assigned lower weights. This is in accordance with the findings of Ibrahim-Bathis and Ahmed (2016), which emphasize the impact of human activities on the hydraulic properties of soil. Waterbodies and fishponds are distinct soil texture structures with direct hydraulic connections to groundwater (Patel et al. 2021). Consistent with other studies, the presence of water bodies can significantly enhance the efficiency of groundwater recharge (Raja Shekar and Mathew 2023; Das and Mukhopadhyay 2020). Therefore, they were assigned high weights when assessing the suitability of artificial recharge, as shown in Table 4.

Land use and land cover (LULC)

The availability of groundwater is significantly affected by LULC because it is directly related to percolation processes and surface runoff. These, in turn, affect soil moisture content and groundwater recharge (Gautam et al. 2023). LULC maps were prepared for the study area based on the annual Chinese land cover dataset. This reveals the distribution of different land use types that contribute to the replenishment and protection of groundwater resources. LULC was categorized into six main classes: cropland (55.28%), forest (13.28%), grassland (5.90%), water (2.56%), barren land (0.12%), and impervious land (22.86%) (Fig. 3F).

Areas covered with water bodies and natural vegetation (Forest and Grassland) are conducive to groundwater recharge and thus assigned higher weights (Ravichandran et al. 2022). These areas typically have higher permeability and lower surface runoff, which facilitates the infiltration of water and its replenishment into groundwater. Cropland, as the dominant land use type, has good recharge potential due to agricultural activities and soil management practices and is, therefore, given moderate weighting (Hilal et al. 2024). Conversely, areas with low permeability and high surface runoff, such as barren and impervious areas, contribute only a minimal amount of groundwater recharge. Therefore, these regions were assigned lower weights. In particular, urban and industrial areas, which tend to be impervious, significantly reduce the natural replenishment of groundwater owing to the impermeable nature of the surface cover. This result is consistent with the view of Biswas et al. (2020), who emphasized the adverse effects of urbanization on groundwater recharge.

Slope

The gradient exerts a significant influence on the movement of groundwater, determining its path and speed (Rukundo and Doğan 2019). Studies have demonstrated that steeper slopes are typically associated with higher flow velocities, which result in rapid runoff from rainfall and reduce the opportunities for surface water infiltration (Berhanu and Hatiye 2020). Meanwhile, gentle slopes facilitate seepage, resulting in a more widespread and stable distribution of groundwater, thereby increasing the potential for groundwater recharge. Considering the impact of slope on groundwater seepage, this study assigned greater significance to areas with gentle slopes. The degree of slope was used to divide the study area into five classes: (a) < 3.60° (40.57%), (b) 3.60°–9.12° (30.67%), (c) 9.12°–16.08° (15.90%), (d) 16.08°–25.20° (9.15%), and (e) > 25.20° (3.72%) (Fig. 3G).

By analyzing the diversity of the terrain (as shown in Fig. 3), the gentle slope areas (< 3.60°) were found primarily in the vast plain regions. In these areas, groundwater flow exhibits a high degree of homogeneity and dispersion, which favors the penetration of moisture and the stable replenishment of groundwater. Areas with gentle slopes accounted for 40.57% of the total area of the study region, approximately 1,361.92 km2, indicating that plain terrain occupied a considerable proportion of the study area. As the slope increases, the topography gradually transitions to hilly and mountainous areas, where the conditions for groundwater replenishment and percolation gradually deteriorate (Li et al. 2021). In these areas, groundwater flow pathways may become more complex, and velocity may increase owing to the steeper terrain, which may, in turn, reduce opportunities for moisture infiltration. Steep areas (> 16.08°) constituted 12.87% of the surface of the study area (~ 431.79 km2). The potential for groundwater replenishment in these areas is relatively low primarily because rapid surface runoff reduces the contact time between moisture and soil, thereby limiting natural groundwater replenishment. This result is consistent with the findings of Nazaripour et al. (2024), who highlighted the limited potential for groundwater recharge in areas with steep slopes.

Geomorphology

Assessing groundwater resources relies on the aquifer storage capacity and permeability (Murmu et al. 2019). These parameters are primarily determined by the geomorphological properties of the region in question. As a discipline that studies the forms, processes, and evolution of the Earth's surface, geomorphology provides invaluable tools for understanding and identifying the characteristics of groundwater exploration areas (Senapati and Das 2021). The observed geomorphological features in the study area include submarine accumulation plains (0.09%), flood plains (9.84%), alluvial plains (28.35%), marine deposition plains (10.57%), hills (0.71%), and mountains (50.44%) (Fig. 3H).

These geomorphological features reveal the basic geological features of the area and provide critical information for assessing the groundwater storage capacity (Bhadran et al. 2022). This study assigned different weights to different landforms for assessing groundwater resources by integrating geomorphological features and their potential contributions to groundwater recharge. Due to their high permeability and excellent water storage capacity, submarine accumulation plains contribute significantly to groundwater recharge and are therefore assigned the highest weight (Arulbalaji et al. 2019). These areas typically consist of porous sediments, which promote efficient groundwater flow and storage. The alluvial and floodplains were given more weight due to their good permeability. Long-term river alluviation has resulted in rich layers of sand and gravel, creating favorable conditions for groundwater recharge and movement in these areas. In contrast, marine depositional plains were assigned lower weights due to their relatively low permeability.

These plains are typically composed of marine sediments and are more compact, thereby limiting groundwater infiltration and movement. Finally, due to the complexity and variability of the terrain, hills and mountains were generally assigned the lowest weights.

Drainage density

The drainage density, an essential metric for assessing the hydrological conditions and groundwater potential of a region (Biswas et al. 2020), reflects the number of developed watercourses within a given area and the distance between adjacent watercourses of the same order (Arefin 2020). The magnitude of drainage density affects the distribution of surface water and the capacity for groundwater replenishment. In areas with high drainage density, surface water is rapidly drained owing to high runoff rates, resulting in a lower capacity for groundwater replenishment (Magesh et al. 2012). In contrast, areas with reduced drainage density allow for longer surface water retention, favoring moisture infiltration and, consequently, enhanced groundwater storage and replenishment capacity. Therefore, higher weights were assigned to regions with low drainage densities. Based on the above principle, this study classifies the study area into five classes according to drainage density: (a) < 0.27 km−1 (2.58%); (b) 0.27–0.83 km−1 (7.47%); (c) 0.83–1.34 km−1 (16.53%); (d) > 1.34 km−1 (1.20%); (e) > 1.92 km−1 (72.22%) (Fig. 3I). As shown in the figure, the study area exhibits significant spatial variation in drainage density. Areas of high drainage density are primarily concentrated near central rivers, where the watercourses are well developed, and surface water is rapidly drained, reducing opportunities for groundwater recharge. This result is consistent with the findings of Bhadran et al. (2022), who reported that in areas with high groundwater abstraction, the potential for groundwater recharge is reduced.

Total dissolved solids (TDS)

TDS is a critical measure of the total dissolved solids in water. The TDS value was proportional to the amount of dissolved impurities in the water. Based on the different TDS concentrations, the study area was divided into five categories: (a) < 300 mg/L, (b) 300–500 mg/L; (c) 500–1,000 mg/L; (d) 1,000–2,000 mg/L; (e) > 2,000 mg/L (Fig. 3J and K).

During the wet season, groundwater generally has a lower TDS concentration owing to the dilution of dissolved solids by increased rainfall (Han and Currell 2022). A transition from the wet to the dry season was observed, with an overall trend of increasing TDS concentration. The proportion of areas with extremely low concentrations (< 300 mg/L) decreased from 11.24% to 9.72%, a decrease of 51.17 km2. The area with very high concentrations (> 2,000 mg/L) remained unchanged. The highest TDS concentrations were observed in the northeastern coastal regions, a phenomenon closely related to seawater intrusion. As distance from the coast increased, the TDS concentration gradually decreased. This finding is consistent with the research findings of Han and Currell (2022), who demonstrated that TDS concentrations in coastal areas are directly related to seawater intrusion. High levels of TDS not only limit the suitability of the area for agricultural water use but also affect the physical properties of the groundwater, such as density, which in turn affects its permeability and mobility (Kahal 2024). In addition, in areas with high TDS concentrations, recharge water may chemically react with substances in the groundwater, forming precipitates or altering the chemical composition of the groundwater, which may negatively affect recharge effectiveness. Owing to these potential risks, this study assigned less weight to regions with high TDS levels to ensure rational development and conservation of water resources. This management strategy aligns with the recommendations of El khalki et al. (2024), who proposed cautious water resource management measures in areas with high TDS levels.

Distance from fault

Faults, geologic structures that occur when the crust breaks and shifts under stress, are key features of the topography and geologic framework (Rao et al. 2001) that disrupt the continuity and integrity of subsurface rock formations, potentially creating fractures and fractured rocks in and around fault zones. This disruption significantly affects the pathways and permeability of groundwater flow and tends to create areas of higher permeability. Due to their increased permeability, fault zones tend to be preferred channels for groundwater movement. These channels regulate the groundwater recharge, flow, and discharge patterns.

High-permeability areas near faults may be more suitable for artificial groundwater recharge because they can receive and store recharged water more efficiently. Therefore, regions closer to the fault were assigned higher weights. The study area was divided into five categories: (a) < 38 km (31.80%), (b) 38–78 km (28.28%), (c) 78–122 km (22.58%), (d) 122–177 km (10.25%), and (e) 177 km (7.10%) (Fig. 3L).

Faults were located northwest of the study area, with areas very close to the faults (< 38 km) accounting for 31.80% of the total area, approximately 1,067.50 km2. These areas will likely have higher permeability, significantly affecting groundwater flow and recharge. In contrast, areas farther away from the faults (> 122 km) represented 17.35% of the study area, approximately 582.42 km2. Therefore, the impact of faults on the groundwater systems may be limited. The fault distance distribution map provided an intuitive understanding of groundwater flow and recharge suitability in the study area. For the assessment and management of groundwater resources, particular attention should be paid to areas near faults because of their high groundwater recharge potential and permeability. This conclusion is consistent with the findings of Khosravi et al. (2018), who highlighted the significance of incorporating fault effects in groundwater resource management.

Elevation

Elevation is a key factor influencing the distribution of groundwater potential (Ravichandran et al. 2022), directly affecting topography, groundwater recharge, flow, and storage characteristics. Elevation maps provide essential information about the topography of the study area and are essential tools for identifying areas with high groundwater potential (Andualem and Demeke 2019). Areas at higher elevations, such as mountains and hills, have more complex topographic structures with numerous fractures, possibly resulting in longer and more tortuous groundwater flow paths. In contrast, lower-elevation areas such as plains tend to have flat terrain with thicker, loose sediments, allowing for more direct and continuous groundwater flow and better storage. The study area was divided into five categories based on elevation: (a) < 86 m (36.47%), (b) 86–209 m (23.68%), (c) 209–342 m (18.20%), (d) 342–493 m (15.25%), and (e) > 493 m (6.40%) (Fig. 3M).

In general, the topography of the study area comprises mountains and plains. The highland area in the southwestern corner of the study area is dominated by rocks, with a significant increase in elevation and is characterized by high mountain ranges and plateaus, with areas of very high and high elevation of 214.85 km2 and 511.97 km2, respectively. In contrast, the northeastern region has flat terrain and low elevation, with a very low elevation area (< 86 m) accounting for 36.47% of the study area. The climate of upland areas tends to be cold and dry, with limited precipitation and significant evaporation, making it difficult for moisture to remain in the soil over long periods and reducing the natural replenishment capacity of groundwater. In low elevation areas, climatic conditions favor precipitation infiltration and groundwater recharge. This result is consistent with the findings of Shekar and Mathew (2023), who reported that climate conditions have a positive impact on groundwater recharge in low-altitude regions. When evaluating groundwater resources, low elevation areas were given greater weight due to the effects of elevation on groundwater recharge and flow.

Topographic wetness index (TWI), Topographic position index (TPI) and Terrain roughness index (TRI)

The TWI is a valuable quantitative assessment tool for groundwater recharge, runoff, and drainage (Chen et al. 2018). It incorporates several variables, including slope, flow direction, and topographic undulations. Empirical evidence suggests a positive correlation between the TWI and pore water pressure. Areas with elevated TWI values are frequently distinguished by their relatively flat topographies and larger catchments (Arulbalaji et al. 2019). In areas, where runoff accumulates on the surface, there is increased opportunity and time for infiltration, thereby increasing the amount of groundwater recharge. Therefore, areas with elevated TWI values are given greater consideration when evaluating groundwater resources, reflecting their potential contributions to groundwater recharge. The study area was divided into five classes based on TWI values: (a) < 6.04 (4.02%); (b) 6.04–8.69 (15.97%); (c) 8.69–13.1; the remaining areas are classified as follows: (d) 13.14–17.59 (29.16%); (e) > 17.59 (14.59%) (Fig. 3N). High-TWI areas are mainly concentrated in the northeastern region. As the terrain became more rugged from northeast to southwest, the TWI values decreased.

Meanwhile, TPI is a parameter that describes the relative position of a terrain by comparing the elevation of a target unit with the average elevation of the surrounding area (Singha et al. 2024). Positive TPI values typically indicated the presence of ridges, whereas negative values indicated the presence of valleys or depressions. Areas with lower TPI values, as a consequence of the water-harvesting effect of topography, facilitate the pooling of surface water, thereby increasing the opportunities for infiltration and having a positive impact on groundwater recharge. Consequently, these areas have been given greater consideration in water resource assessments because of their enhanced potential for groundwater replenishment. In turn, areas with elevated TPI values owing to more pronounced slopes may be susceptible to soil erosion and loss. This erosion can disrupt the soil structure, which, in turn, reduces the water retention capacity of the soil and may even exacerbate evaporation and moisture loss. This can have detrimental effects on groundwater recharge. The study area was divided into five grades based on the TPI values to assess the suitability of different topographic locations for groundwater recharge. The grades are as follows: (a) < 4.10 (5.40%)); (b) −4.10– −1.56 (17.57%); (c) −1.56–0.69 (42.53%); (d) 0.69–3.22 (24.62%); (e) > 3.22 (9.87%) (Fig. 3O).

The TRI is a metric that quantifies the complexity and roughness of the terrain (Chowdhuri et al. 2020). The degree of terrain undulation has a direct impact on the water infiltration capacity and efficiency of groundwater recharge. The complexity of the topography increases the diversity and uncertainty of the surface water flow directions, which can lead to dispersion and complexity of groundwater flow paths. Higher TRI values indicate the presence of complex topography in the area, which may be characterized by diverse landform features, such as faults, fractures, and other geological structures (Maxwell and Shobe 2022). These characteristics significantly affect the groundwater flow pathways, potentially causing water to flow along the surface as runoff rather than percolating underground. In assessing groundwater resources, areas with lower TRI values are generally considered more favorable for groundwater recharge since the terrain is flatter, allowing water to penetrate more easily, thereby increasing the amount of groundwater recharge. Consequently, these areas were assigned greater weight in the assessment process to reflect their potential contributions to groundwater resources. The study area was divided into five grades based on TRI values as follows: (a) < 9 (2.09%), (b) 9–17 (7.05%), (c) 17–28 (15.59%), (d) 28–43 (33.76%), and (e) > 43 (41.51%) (Fig. 3P).

Curvature

Curvature is a critical parameter in topographic analysis that has a significant impact on groundwater flow and accumulation dynamics. Curvature is a measure of the convexity or concavity of the surface, which directly affects the percolation of surface water and the groundwater recharge mechanism. Regions with high curvatures are commonly associated with steep ridges or slopes that facilitate the rapid flow of water (Subba Rao et al. 2022). By contrast, low curvature regions are commonly located in valleys and plains. This topography facilitates a uniform distribution of surface water, thereby enhancing the groundwater recharge efficiency. Consequently, regions exhibiting low curvature were given greater significance during the evaluation, reflecting their potential impact on groundwater flow dynamics and accumulation. The study area is divided into five distinct categories based on curvature values to evaluate the suitability of various topographic positions for groundwater replenishment: (a) < 0.51 (43.94%), (b) 0.51–1.07 (38.03%), (c) 1.07–1.54 (9.14%), (d) 1.54–2.16(7.07%), and (e) > 2.16 (1.83%) (Fig. 3Q). This classification method is analogous to those used in previous studies; however, adjustments have been made to accommodate the specific conditions of the study area. For instance, Bera et al. (2020) also utilized curvature values to assess the suitability of groundwater recharge, but their classification thresholds differed from ours. Our classification method incorporates a more detailed consideration of the topographic features of this study area, which may have impacted the evaluation results of groundwater recharge.

Identification of ARZs

This study employed 14 key parameters, namely precipitation, groundwater depth, soil type, LULC, slope, geomorphology, hydrology, TDS, faults, elevation, TWI, TRI, and curvature, to conduct a comprehensive assessment of artificial recharge suitability areas in the Mi River watershed. The FAHP method was employed to assign grades and weights to each parameter, thereby quantifying their contributions to groundwater potential. The weighted index overlay technique in the ArcGIS software was employed to create a zoning map for artificial recharge suitability (Fig. 4). The resulting map divided the study area into five distinct levels based on its suitability for artificial recharge. Grades were assigned as follows: very poor, poor, moderate, good, and very good. This method is similar to the comprehensive evaluation method proposed by Aju et al. (2021), Kadam et al. (2020), and Nazaripour et al. (2024), but has been adapted to the specific conditions of the river basin.

Fig. 4
figure 4

A ARZs map of the wet season; B ARZs map of the dry season

Full size image

The imagery indicates that during the dry season, areas with higher suitability (good, very good) are typically located in the upper and lower reaches of the basin. This distribution can be attributed to higher precipitation, greater groundwater depth, and proximity to faults in these areas. The lower reaches, characterized by flat terrain and predominantly plain geomorphology, are where most of the very good areas are concentrated, which is consistent with the view proposed by Bunkar and Nema (2022) that plain areas are favorable for groundwater recharge. However, despite the favorable topographical conditions in the lower estuary of the study area, the groundwater potential is severely limited owing to the prevalence of saline soils with low percolation capacity and low average annual rainfall (Phong et al. 2021). Conversely, areas with lower suitability (very poor, poor) were primarily situated in the central region of the study area and on the eastern and western sides of the upper reaches. This may be attributed to the elevated topography and steeper slopes, resulting in accelerated surface runoff. In addition, shallow groundwater depth and low precipitation in the central region contributed to low suitability. This finding is consistent with the impact of topography and hydrological conditions on the suitability of artificial recharge as proposed by Arulbalaji et al. (2019).

During the wet season, precipitation levels increase significantly, which improves groundwater recharge conditions. Nevertheless, in comparison to the dry season, there was a notable decline in the number of areas classified as very good and good. Consistent with the findings of Kadam et al. (2020), this study also observed that during the rainy season, the rise in groundwater levels may render certain areas unsuitable for artificial recharge. This suggests that both rainfall and fluctuations in groundwater levels significantly affect the suitability of artificial recharge. The moderate category expanded by 7.57%, indicating that increased rainfall provided more opportunities for groundwater replenishment. The proportions of areas classified as very good and good decreased by 5.22% and 4.57%, respectively. This could be attributed to the increased groundwater table, making some areas unsuitable for artificial recharge. Although overall suitability improved during the wet season, some areas were classified as very poor and poor. This may be attributed to the rugged topography and low soil permeability, which make these areas relatively less suitable for artificial recharge (Aju et al. 2021).

The results of the artificial recharge suitability zoning assessment conducted for both the dry and wet seasons indicate that areas rated as very good and good, which have higher capacities for groundwater recharge and storage, are categorized as highly suitable for artificial recharge. These areas are typically distinguished by favorable hydrogeological conditions, including moderate groundwater depths, highly permeable soils, and low topographic slopes, facilitating efficient infiltration and retention of groundwater (Raja Shekar and Mathew 2023). Moderate areas exhibit a moderate level of suitability for artificial recharge. Further hydrogeological surveys may be required to assess the potential for groundwater recharge and storage and to develop appropriate water management measures to optimize artificial recharge strategies. Areas with a low groundwater potential, classified as poor or very poor, are less suitable for artificial recharge (Kadam et al. 2020). In these areas, it may be necessary to consider the development of alternative water sources or the adoption of special recharge techniques, such as improved land use practices, enhanced soil permeability, or the use of the rainy season for replenishment (Bunkar and Nema 2022). Moreover, the assessment considered the proportion of potential groundwater areas during the wet and dry seasons, reflecting the seasonal variability of groundwater resources. The proportions of wet and dry seasons are shown in Table, which provides a basis for developing seasonal water management strategies.

Verification of ARZs

In this study, we employed FAHP to conduct a comprehensive assessment of ARZs under different seasonal conditions. This approach integrates key hydrogeological parameters, including rainfall, groundwater depth, topographic slope, soil type, and drainage density, to create an accurate and reliable assessment framework. In contrast to the traditional AHP, the FAHP has the advantage of addressing the uncertainties and ambiguities inherent in the evaluation process, which is essential for evaluating groundwater resources. This method is analogous to the comprehensive evaluation method proposed by Singha et al. (2024); however, our study places greater emphasis on the model's applicability across various seasonal conditions.

To determine the model's accuracy, a Receiver Operating Characteristic (ROC) curve analysis was performed using field data from 29 wells (Fig. 5). During the wet season, when groundwater recharge is more abundant, the model yields an AUC value of 73.69%, indicating a high level of accuracy in identifying suitable areas for artificial recharge. This result aligns with the findings of Nazaripour et al. (2024), who also reported high AUC values, which indicate the model's accuracy. Even during the dry season, when hydrological conditions are less favorable, the model exhibited a satisfactory AUC value of 71.29%, indicating its ability to perform well under adverse conditions. A comparison of the AUC values for the two seasons reveals that the FAHP model not only demonstrates robustness across seasons but also exhibits the capacity to adapt to changes in hydrological conditions, thus demonstrating strong adaptation.

Fig. 5
figure 5

A Validation for ARZs of the wet season; B validation for ARZs of the dry season

Full size image

Seasonal assessment is crucial for flexible water management strategies (Yang et al. 2021). This model provides decision-makers with the necessary insights into the availability of groundwater resources in different seasons, thus enabling the optimization of artificial recharge and extraction schemes. Overall, the FAHP model demonstrated high accuracy in assessing suitability across diverse seasonal conditions, providing robust decision-making support for water managers when formulating artificial recharge strategies. A comprehensive seasonal analysis permits a deeper understanding of the intricacies of groundwater systems, thereby providing a scientific basis for the prudent utilization and conservation of water resources.

Conclusion

This study employed GIS and RS technologies to integrate satellite imagery and topographic maps, creating thematic layers that considered the 14 factors influencing groundwater movement and storage. These factors were rainfall, groundwater depth, soil type, LULC, slope, geomorphology, drainage density, TDS, faults, elevation, TWI, TPI, TRI, and curvature. The FAHP method was used to prioritize and weigh these thematic layers to reflect their impact on groundwater.

This study categorized the artificial recharge suitability into five classes: very poor, poor, moderate, good and very good. Field data from 29 wells supported the consistency of the model, with AUC of 73.69% and 71.29% for the wet and dry seasons, respectively. During the wet season, approximately 5.80% of the area exhibited very good artificial recharge suitability, with 35.24% good and 41.96% moderate suitability.

Considering the artificial recharge suitability zoning assessment results for both the dry and wet seasons, it is recommended that more cautious and refined recharge measures be implemented in areas with shallow groundwater in the central region, such as controlling the volume and rate of recharge, to minimize potential impacts on the groundwater system. In contrast, in the upper and lower reaches, where suitability is higher, the storage capacity should be fully utilized by implementing larger-scale recharge measures to facilitate the recovery and replenishment of groundwater resources. In addition, the adaptation of recharge strategies to seasonal variations is recommended to accommodate seasonal variations in groundwater levels.

This study validated the GIS, RS, and FAHP techniques as effective, cost-efficient approaches for assessing artificial recharge suitability in river basins. However, data accuracy influenced the accuracy of the assessment results; therefore, future studies should enhance groundwater monitoring and increase the well density and frequency for more detailed studies on replenishment, water quality, and irrigation suitability to improve the precision and reliability of the assessment.