Delineation of groundwater potential zones and identification of artificial recharge sites in the Kinnerasani Watershed, India, using remote sensing-GIS, AHP, and Fuzzy-AHP techniques

Water stands as the most consumed material on the Earth due to its indispensable role in supporting life (Zade et al. 2021; Nguyen et al. 2022; Roy et al. 2022). However, a significant portion (97%) of the Earth's water is salty, rendering it unsuitable for most human purposes (Das et al. 2019). Only a mere 3% of the planet's water comprises freshwater, which is the sole source appropriate for human use (Omar et al. 2020; Goswami & Ghosal 2022). Among these freshwater resources, groundwater, which refers to water stored in underground aquifers (Post & Abarca 2010; Nyakundi et al. 2022), plays a crucial role in safeguarding and preserving water within the Earth's crust (Fitts 2013). Given its wide-ranging applications in all aspects of life, groundwater stands as an important natural resource (Jha et al. 2007). Over time, the utilization of groundwater has significantly increased due to the overuse of surface water and the unpredictability of rainfall. In order to complement surface freshwater sources, households, industries, and agriculture have predominantly relied on groundwater supplies (Andualem & Demeke 2019). Groundwater serves as a crucial water source for approximately 2.5 billion people worldwide, fulfilling their domestic water requirements (UNESCO 2015). Many nations, particularly those situated in semi-arid and arid regions, heavily rely on groundwater for various purposes such as industry, agriculture, and household use (Serele et al. 2020; Ahmed et al. 2021). However, due to the escalating global population and rapid industrialization, there has been a substantial increase in the dependence on groundwater, leading to immense strain on this invaluable resource (Das et al. 2021). Consequently, it is crucial to establish effective delineation techniques and implement sustainable groundwater management practices to meet the rising demand, not only for the present generation but also for future generations (Goswami & Ghosal 2022). Identifying potential zones for groundwater recharge becomes essential for maintaining groundwater quality and ensuring the efficiency of the groundwater management system.

It can be difficult to chart this potential recharge zone because groundwater is typically an inconspicuous source (Allafta et al. 2021). Numerous geoenvironmental factors, including geomorphology, land use and land cover (LULC), lithology, climatic conditions, the extent of fractures, soil, drainage pattern, lineament features, secondary porosity, landform, topography, geological structures, slope, etc., and also their interactions, determine the presence and movement of groundwater in a given area (Greenbaum 1985; Gupta et al. 2020; Dibs et al. 2023).

The many geoenvironmental elements that have an impact on groundwater growth deep inside the Earth's crust make it challenging to identify groundwater. The hunt for groundwater has been answered over time using conventional techniques, such as geophysical examinations (Andualem & Demeke 2019; Agyemang 2022; Shekar & Mathew 2022b). Nevertheless, the drawbacks of these conventional techniques are that they require more money, labour, and time to complete (Gumma & Pavelic 2013; Shyam et al. 2021; Mathew et al. 2022). Geospatial techniques' more recent development has opened the door for creative and effective methods of groundwater exploration (Mandal et al. 2016). These innovative techniques, in contrast to the conventional ones, are quick, economical, and time-saving, and they can even be used in inaccessible locations (Serele et al. 2020; Shekar & Mathew 2022a). In order to analyze and integrate data to provide meaningful outputs for exploring, assessing, and managing essential groundwater resources, the Geographic Information Systems (GIS) and Remote Sensing (RS) technique has proven particularly important (Serele et al. 2020; Ijumulana et al. 2021; Omar et al. 2021; Shekar & Mathew 2022c; Thacker et al. 2022). Over the past few decades, multi-criteria decision analysis (MCDA) approaches, led by remote sensing and GIS, have become more and more popular for managing water resources and the environment (Allafta et al. 2021). Delineating the GWPZs globally is aided by integrated geospatial approaches, using several multi-criteria decision-making techniques. The study conducted by Qadir et al. (2020) aimed to map groundwater potential in a specific region using remote sensing and GIS techniques. The researchers integrated various thematic layers, including lineament density, slope, Topographic Wetness Index (TWI) geology, drainage density, and LULC. Each of these layers was assigned appropriate weights to account for their relative significance in determining groundwater potential using remote sensing and GIS techniques. According to the paper by Arunbose et al. (2021), the authors focused solely on using the GIS-based analytical hierarchy process (AHP) technique method for their investigation of groundwater potential zones (GWPZs).

GIS and AHP integration is a potent tool for combining the evaluation of several criteria (Kumar et al. 2022). The subjective Saaty (1980) technique, known as the AHP, allows users to select the relative weighting of each criterion in a solution to a problem that depends on many criteria. The distribution of impact among the various thematic levels can be determined using AHP, according to Saaty (1996) and Wind and Saaty (1980). It aids in the spatially complex challenge of environmental issues for decision-makers. To find a GWPZ, several characteristics have been taken into consideration (Allafta et al. 2021). Based on their relative weights in the process of deciding where to locate the groundwater zone, each characteristic's weights were assigned. For each problem in the AHP technique, a hierarchical model with goals, criteria, sub-criteria, and options is used (Saaty 2008). The weights or priority of the pairwise comparisons are used to resolve issues with AHP (Ho 2008). When comparing the criteria in pairs, the AHP approach is susceptible to discrepancies to a certain extent.

Fuzzy logic, a branch of artificial intelligence (AI), focuses on approximate reasoning rather than strict precision. It provides a mathematical framework to handle uncertainty and vagueness, acknowledging that things can be partially true or false and allowing for degrees of truth to be expressed through mathematical expressions (Dubois & Prade 2003). Fuzzy logic finds application in various AI domains, including control systems, pattern recognition, decision-making, and data analysis. It proves particularly useful in situations with insufficient or ambiguous information, where conventional logic may not provide clear solutions (Bezdek et al. 1999; Al-Jarrah & Halawani 2001). Recent research highlights the integration of AI, fuzzy logic, and remote sensing techniques for groundwater potential mapping. These methods enable the creation of accurate groundwater potential maps that support resource management and effective groundwater management. To achieve this, it is important to examine the logical consistency of pairwise comparisons (Cheng 1997; Chang et al. 2008). To address these challenges and improve decision-making confidence, Van-Laarhoven & Pedrycz (1983) introduced the Fuzzy-AHP method. Fuzzy-AHP is an extension of the AHP designed for hierarchical fuzzy multi-criteria decision-making. This approach combines alternative strategy selection, systematic weighing, and principles from fuzzy set theory and hierarchical structure analysis. By employing fuzzy numbers to represent weight values, the Fuzzy-AHP approach generates accurate and realistic results (Aryafar et al. 2013; Lee et al. 2013).

Multidisciplinary decision-making has been recognized over the past 20 years as a good method for managing water because it gives decisions more structure, readability, transparency, and firmness (Roberts & Goodwin 2002; Flug et al. 2005). Water resource management requires proper MCDA, which should concentrate on the basis of the decision-making process, including conditional decision-making and potential decisions. Integration with GIS-based techniques, which yield more accurate results and need less computation time than conventional GWPZ field techniques, is a major benefit of these methods (Zolekar & Bhagat 2015).

Groundwater resources play a vital role in the Kinnerasani River basin, located in the Telangana state of India. Encompassing an area of approximately 3,316 km², the basin is predominantly agricultural, relying heavily on groundwater for irrigation of crops such as paddy, vegetables, cotton, and maize. However, the Kinnerasani River faces challenges, including insufficient water flow during dry seasons, which significantly impact both the river's ecosystem and the local communities that depend on it. Given the region's economic dependence on agriculture and groundwater, sustainable management of groundwater resources is crucial for ensuring the socio-economic development of the area. As per the author's knowledge, there has been a lack of investigations regarding the potential of groundwater in the current study area. Furthermore, there have been limited studies that have utilized the AHP and Fuzzy-AHP techniques. In order to bridge these gaps in knowledge and address these challenges, this article endeavours to identify the GWPZs within the Kinnerasani River basin by employing RS and GIS techniques. The objectives of this study are to develop a groundwater potential map in the Kinnerasani River basin by incorporating seven thematic layers: drainage density, elevation, soil, geomorphology, slope, LULC, and rainfall. The application of AHP and Fuzzy-AHP techniques will be employed to analyze and integrate these layers. The accuracy of the GWPZ map will be validated by comparing it with existing well data points using the receiver operating characteristic (ROC) and area under the curve (AUC) methods. Furthermore, this study aims to identify suitable locations for artificial recharge sites in the Kinnerasani River basin based on the generated groundwater potential map. By achieving these objectives, we hope to provide valuable insights into the potential distribution of groundwater resources in the study area and contribute to informed decision-making regarding groundwater management and sustainable development in the Kinnerasani River basin.

Numerous parameters and their subclasses have varied effects on how groundwater behaves, thus they must be assessed and weighed appropriately. This study uses an MCDM approach based on the AHP. AHP, an MCDM tool, is frequently used to manage groundwater (Rahmati & Melesse 2016). Each thematic map's weight was determined using the AHP method. After considerable research and expert advice, theme layers are given priority. The AHP technique, which is based on multi-criteria decision-making, establishes the weight of each thematic layer using Saaty's (1–9) scale. By contrasting the attributes in pairs, the comparison matrix for the seven thematic levels was developed. The eigenvector is computed to normalize the weight assigned to a layer (Saaty 1980). According to the above approach, taking into consideration earlier studies and the expert's knowledge, final weights were established for each thematic layer. GWPZs were created for the Kinnerasani River basin based on AHP by combining all seven theme layers using the weighted overlay analysis technique on the GIS platform. The resultant GWPZs map indicated poor, moderate, and high zones. The effectiveness of AHP approaches was validated using ROC.

In decision-making processes, discrepancies often arise when weighing quantitative and qualitative parameters. To address this, methods such as MCDM techniques play a vital role. These techniques involve assigning weights to different parameters and ranking a subset of parameters based on professional judgement within a specific thematic context (Mohanty & Walling 2008; Quadrio et al. 2015). While the AHP is commonly used in MCDM, the Fuzzy-AHP approach developed by Van-Laarhoven & Pedrycz (1983) provides a subjective assessment of weighing multiple criteria, effectively dealing with differences that may emerge during pairwise comparisons (Chang 1996; Sener et al. 2018). In the present study, the AI-based Fuzzy-AHP approach was utilized to examine the GWPZ, considering seven parameters. To construct the triangular fuzzy judgement matrix, triangular fuzzy numbers were employed as a replacement for crisp numbers in the pairwise comparison matrix (Lyu et al. 2020). This fuzzy approach enables a more nuanced representation of uncertainties and enhances the comprehensiveness of the decision-making process. The geometric mean approach of Buckley (1985) was used to obtain the fuzzy geometric mean and fuzzy weights of each criterion. Defuzzification, or converting a fuzzy number into a crisp number, was ultimately accomplished by first determining the arithmetic mean and then the normalized weight. The combined seven theme layers were processed using the weighted overlay analysis technique in the GIS platform to generate the GWPZ map for the Kinnerasani River basin, employing the Fuzzy-AHP approach. The final GWPZ map identified poor, moderate, and high zones, representing varying levels of groundwater potential. To validate the results of the Fuzzy-AHP technique, the ROC analysis was employed.

In this study, the GWPZ map was developed using GIS by integrating multiple thematic layers that contribute to the occurrence of groundwater. The following is a concise overview of the thematic maps utilized for groundwater potential mapping, along with their corresponding weights assigned through AHP and Fuzzy-AHP, including validation and the resulting final GWPZ map.

Many thematic layers that favour the presence of groundwater have been combined to build a groundwater potential map using GIS. GWPZ mapping thematic maps utilized in this work, their weights determined by AHP, Fuzzy-AHP with validation, and the final GWPZ map are briefly described in the following.

After developing theme layers for groundwater modelling, the next step involved weighing the parameters considering their varying impacts on aquifer replenishment (Serele et al. 2020). The AHP is widely recognized as a reliable method and recommended by many researchers (Roy et al. 2022). Saaty (1980) introduced the AHP as a powerful decision-making approach for determining the relative weights of different variables. In this study, the AHP was employed to evaluate the relative contributions of each factor to groundwater recharge. Saaty (1980) devised a scale of relative relevance ranging from 1 to 9, which was utilized to construct the matrix. The matrix was completed by comparing each pair of criteria and assigning rankings based on their relative importance. The diagonal elements of the matrix were set to 1, indicating equal weights with other diagonal elements. Criteria that exerted a significant influence over others were assigned a rating of 9. The pairwise matrix created for this study is presented in Table 2, whereas Table 3 shows the Saaty scale of relative relevance. According to Saaty (1990), the reciprocal values were utilized to fill the lower triangular matrix and the values were placed in the upper triangle to construct a diagonal matrix. Additionally, Table 4 shows that the relative weights were uniform. By determining the fundamental eigenvalue and examining the consistency ratio (CR), the thoughts' coherence was assessed (Saaty 1990).

Table 2

AHP-based pairwise comparison matrix

.

.

View Large

Table 4

Pairwise comparison matrix with normalized principal eigenvector

.

.

View Large

AHP and the related MCDM approach analyze difficult decisions using numerical and pragmatic principles (Saaty 1990). While AHP is commonly used in MCDM, the Fuzzy-AHP methodology, introduced by Van-Laarhoven and Pedrycz in 1983, allows for a subjective assessment of weighing multiple criteria to address discrepancies that arise during pairwise comparisons (Chang 1996; Tan et al. 2013; Sener et al. 2018). By employing the Fuzzy-AHP technique, factors can be compared and their relative importance can be determined, facilitating the management of challenging decisions. The integration of Fuzzy-AHP into multi-criteria decisions enhances the accuracy of groundwater potential mapping, as suggested by Sener et al. (2018). Fuzzy logic, a branch of AI, deals with approximate reasoning rather than precise calculations, providing a mathematical framework to handle uncertainty and vagueness by acknowledging that things can be partially true or false and allowing degrees of truth to be expressed mathematically. In this study, seven different theme levels were considered. Experts assigned criterion weights using Saaty's scale (1–9) to compare the criteria through a questionnaire. In the context of groundwater potential, a parameter with a low weight has minimal impact, while one with a high weight exerts significant influence. The experts' comments will be taken into account to review the pairwise comparisons and arrive at a conclusion. For example, in AHP, equal importance is denoted as ‘1,’ whereas in Fuzzy-AHP, it is represented as (1,1,1) for the lower, middle, and upper triangle values (Table 7).

Similar to this, Table 8 shows the residual reciprocal values after they have been converted into fuzzy numbers.

Table 9 displays the final values for the fuzzy geometric means in a manner similar to this.

As indicated in Tables 10 and 11, the subsequent step entails multiplying each fuzzy geometric mean value by the reciprocal of the geometric mean summation.

Once weights have been obtained, all weights must then be added together as illustrated in Table 12. 1.07 is an unacceptable number for the sum. To obtain the weight total of 1 as stated in Table 13, the weights are normalized.

The practice of artificial aquifer recharge has gained significance in addressing groundwater degradation and depletion issues in various river basins worldwide (Aju et al. 2021). Artificial recharge involves the deliberate replenishment of groundwater by redirecting excess surface water into underground layers for storage and future use. This process helps replenish the water supply within aquifers. Studies have demonstrated the effectiveness of artificial recharge systems in preserving and replenishing groundwater resources. Capturing and conserving rainwater, which would otherwise seep into the ground or flow through surface streams, is an essential aspect of artificial recharge (Kumar et al. 2021).

Artificial recharge methods, such as spreading, are employed to raise the water table and enhance groundwater levels. The utilization of RS and GIS techniques plays a crucial role in identifying suitable locations for water recharge. Researchers worldwide have successfully employed GIS to delineate suitable areas for various artificial recharge technologies (Ahmadi et al. 2017). By combining scientific methods, including remote sensing, GIS, and artificial recharge techniques, sustainable management of groundwater resources can be achieved. These approaches contribute to the preservation and replenishment of aquifers, ensuring the availability of water for future generations.

For the implementation of artificial recharge, it is recommended to utilize gabion structures as the primary facilities, particularly in lower order streams. These structures enhance the infiltration of water into the subsurface formations, aiding in the recharge process. It is noteworthy that a significant number of identified artificial recharge sites are located in agricultural land, which holds promising implications for fostering higher economic growth. By strategically selecting and implementing artificial recharge sites based on the findings of this study, decision-makers and water users in the research area can actively contribute to sustainable groundwater management. This approach not only improves groundwater recharge potential but also promotes economic development and ecological sustainability within the Kinnerasani Watershed.

The purpose of the current work is to define GWPZs in the Kinnerasani Watershed in India using integrated data, including remote sensing and GIS, in conjunction with the AHP and Fuzzy-AHP techniques. This study utilized seven thematic layers, namely drainage density, slope, elevation, LULC, soil, geomorphology, and rainfall, which were analyzed to delineate the GWPZs. The present study region was divided into three unique areas, each characterized by a different level of groundwater potential, as depicted in the final GWPZ map. Among the factors considered, drainage density emerged as a significant influence on groundwater potential in both the AHP and Fuzzy-AHP models. According to the AHP model, the study area was classified as 17.76% poor potential, 72.79% moderate potential, and 9.45% high potential. In comparison, the Fuzzy-AHP approach classified the area as 18.27% poor potential, 71.07% moderate potential, and 10.69% high potential. Evaluation of predictive capability through ROC calculations revealed that the Fuzzy-AHP technique exhibited better accuracy (AUC = 74%) than the AHP technique (AUC = 73%).

The GWPZ map generated by the Fuzzy-AHP technique serves as a valuable database for groundwater planning and management in the study area. Additionally, this study utilizes the best method derived from both models to identify 26 suitable locations for artificial recharge sites. Given the extensive agricultural practices in the study area, the findings hold considerable potential for improving irrigation facilities, agricultural output, and domestic water use productivity. To maintain sustainable groundwater use in the area, decision-makers can leverage the research findings to develop effective groundwater management strategies and smart development plans. By incorporating the integrated approach and methodologies employed in this study, informed decisions can be made to ensure optimal utilization of groundwater resources in the Kinnerasani Watershed.

Future studies on GWPZs can focus on exploring groundwater–surface water interactions, improving data integration, incorporating socio-economic factors, considering the impacts of climate change, and implementing long-term monitoring and modelling.

The authors would like to thank Telangana Ground Water Department Water Level Data for making public available (https://data.telangana.gov.in) as well as the US Geological Survey (USGS) for making the satellite data available (https://earthexplorer.usgs.gov/). Thank you to the Climatic Research Unit for providing rainfall data on their website at https://crudata.uea.ac.uk/cru/data/hrg/. We would especially like to thank the Geological Survey of India for providing the geomorphology and lineament data on their website at https://bhukosh.gsi.gov.in/Bhukosh/Public.

There was no funding for this project.

All relevant data are available from an online repository or repositories.

The authors declare there is no conflict.