Abstract

Groundwater potential is dynamic and fluctuates with respect to draft and recharge. The purpose of this study is to investigate the potential groundwater zones of the Nambiyar watershed in South India utilizing a probability-based bivariate statistical model frequency ratio (FR) and multi-influencing factor approaches (MIF). For this, spatial relationships between ten factors viz. slope, rainfall, lineament and drainage density, geology, geomorphology, soil texture, land use/land cover, well density, topographic wetness index, and groundwater occurrence were assessed. A total of 162 wells were selected for the study, of which 60% (97 dug wells) were used for training the model, and the remaining 40% (65 dug wells) were used for validating the results. The potential groundwater zones were classified into five categories: very low, low, moderate, high, and very high. Very high zones classified using the FR and MIF models are 186 km² (28.03%) and 97.84 km² (14%), respectively, whereas very low category areas are 63.29 km² (9.50%) and 64.02 km² (9.61%) of the watershed. The results were validated using well data by generating the AUC (area under the curve). The validated results revealed that the AUC for the frequency ratio model was 72%, while the MIF was 62%. This study explores potential groundwater zones using GIS and remote sensing techniques, benefiting government agencies and private sectors for better resource management.

1 Introduction

Groundwater is one of the vital sources of water available to fulfill various needs of mankind. Agricultural, economic development and human health are supported by its quantity and quality (Chaudhry et al. 2019). The easy availability of groundwater near the point of use, which requires little treatment before use, makes it a primary water source, especially in arid regions (Todd and Mays 2005). In a tropical country like India, water scarcity is severe due to its demand created by the growing population, irrigation, industrial growth, and tourism (Sishodia et al. 2016; Etikala et al. 2019; Kumar et al. 2021; Mahanuradha and Pragatheeswaran 2023; Libina et al. 2023). However, man-made influences such as shifting land use patterns and over-extraction of water for various purposes alter underground water levels (Nayak et al. 2015; Kumari et al. 2022). Groundwater systems are dynamic, constantly adapting to changes in climate, groundwater withdrawal, and land use, both in the short and long term. Changes in groundwater levels are also affected by rainfall intensity and by the distribution and volume of runoff. Thus, the scenario shows it is high time to check the quality of water resources and ensure its availability and accessibility to everyone. The study of water level fluctuations helps to assess the gravity of a situation in times of drought, and to take remedial measures (Goyal et al. 2010; Libina et al. 2023).

Similarly, mapping water resources has become more important since demand is increasing daily to sustain the needs of the growing population (Bhattacharya et al. 2021). The delineation of groundwater potential zones utilizing GIS and remote sensing techniques has recently emerged as an excellent technology for obtaining preliminary information about groundwater potential zones before field investigations by interlinking various geo hydrological characteristics of terrain (Fagbohun 2018; Kaewdum and Chotpantarat 2021; Kabeto et al. 2022; Kisiki et al. 2022; Faheem et al. 2023). Researchers have used a variety of statistical models, including multi criteria decision methods, machine and deep learning techniques, bivariate statistics, and multivariate statistics to decipher potential groundwater areas. The combination of geospatial approaches with several other multi-criteria methodologies (MCDM) has proven to be quite beneficial for inexpensive evaluation of potential water locations (Machiwal et al. 2011; Magesh et al. 2012; Patle and Awasthi 2019; Abijith et al. 2020; Anbarasu et al. 2020; Kumar et al. 2021; Masitoh et al. 2022). Several MCDM techniques, such as AHP and fuzzy AHP, are used for groundwater potential mapping (Castillo et al. 2022). Various hybrid models combining AHP and FR methods are being used to rank and weigh each criterion (Aslan and Çelik 2021). The bivariate statistics approach quickly calculates the probabilistic relationship between the dependent and independent variables; hence, this method is widely used in many geospatial studies, such as landslide susceptibility mapping and groundwater potential mapping (Razandi et al. 2015; Trabelsi et al. 2019; Arshad et al. 2020).

Over 73% of the Tamil Nadu State is covered by hard rocks and semi-consolidated and consolidated strata, with the majority occurring in the eastern coastline tract (CGWB 2022). The principal irrigation sources in the watershed are tanks and reservoirs followed by dug wells and bore wells. Excessive groundwater extraction and pumping have also been observed in the Radhapuram block located in the study area, resulting in groundwater depletion. Notably, very few studies have been conducted in the Nambiyar watershed to analyze the potential resource zone by integrating slope, land use/land cover, drainage density, and soil depth, but no in-depth work has been done to explore the groundwater potential zones in the watershed. The watershed is severely vulnerable due to inappropriate land use practices and human activities. Here, both banks of the Nambiyar River follow a mono-cropping pattern where paddy is the major crop during two cropping seasons of the year. The area where a single crop is practiced imposes specific constraints on agricultural activities, particularly those related to the utilization of available water potential and resource depletion (Narmada et al. 2015). In recent years, since no other attempts have been made to understand groundwater potential zones in the watershed, this study was carried out with the objective of exploring and mapping the groundwater potential zones of the Nambiyar watershed using FR and MIF models integrated with GIS and remote sensing techniques. This investigation highlights its scope by investigating potential groundwater sites in the watershed, which assists in understanding the source of readily obtainable groundwater for drinking and irrigation, which is crucial for resource conservation and management.

2 Study area

The area selected for the present study, the Nambiyar watershed, lies in the southern part of the Tirunelveli District of Tamil Nadu, India. It extends from 8°10’N to 8°32’N latitude and 77°28’E to 77°50’E longitude. The watershed covers an area of 665.45 km² (66545.98 ha). The source of the river is in the Kalakkadu Reserved Forest, which is a part of the Western Ghats, with an elevation of 1800 m. The river flows toward the east, enters the plains at Thirukurangudi, and finally drains into the Gulf of Mannar, Indian Ocean. Kodamadi Ar, Paratai Ar, Tamarai Ar, Valliyurankal, and Kallanodai are the major tributaries of the Nambiyar River. The watershed possesses a unique locational setting by spanning across hilly terrain up to coastal plains. The most dominant formation observed in the watershed is metamorphic rocks from the Archean era, (over.5 billion years ago), covered mainly by gneiss, composed of a variety of minerals. Many studies have noted that these rocks were formed from igneous rocks, which are the oldest rocks of the peninsular region. Charnockites, also called black granites, occur in southern Tamil Nadu and extend to Kanyakumari, Western Ghats, and Eastern Ghats (Balasubramanian 2017). The lithology of the coasts is composed of silt, clay, sand, clay sand, alluvium, and calcareous sandstone, which are naturally soft unconsolidated sediments that formed during the Quaternary period. Red aeolian sand, which is locally known as ‘teri’ sand are found along the coasts. The slope of the watershed is categorized as level to nearly level (0–1%), very gently sloped (1–3%), gently sloped (3–8%), moderately sloped (8–15%), moderately steep sloped (15–30%), steeply sloped (30–50%), and very steep sloped (> 50%). The location of the Nambiyar watershed is shown in Figure 1. Previous studies have shown that the coastal strip of the Tirunelveli district in general, notably the Radhapuram Taluk, which is part of the Nambiyar watershed, is composed of less cemented calcareous sediments, which speeds up groundwater circulation. The net groundwater availability in the Kalakkadu and Nanguneri blocks, which are in the northern part of the watershed, decreased from 2375.81 (ham) to 2278.24 (ham), and from 1809.16 (ham) to 1708.55 (ham), respectively (CGWB 2012, 2022). The watershed's land use and land cover categories have changed dramatically over the last two decades, particularly since the construction of the Kodumudiyar Dam in 2003 and the Nambiyar Dam in 2004 (CGWB 2009). These dams are the primary source of irrigation for more than 40 settlements, assisting in the conversion of scrubland to arable land, and increasing agricultural production in the watershed. The watershed falls within Nanguneri, Radhapuram, and Tisayanvilayi Taluks, covering approximately sixty villages, which are the administrative units of the Tirunelveli District.

Figure 1 Location of the Nambiyar watershed.

3 Data and Methods

3.1 Selection of criteria

Various hydrogeological conditions, such as slope, rainfall, soil, geology, drainage density, and land use/land cover, have been considered by researchers in recent years to evaluate areas for groundwater potential zones (Manap et al. 2014; Al-Abadi 2017; Mallick et al. 2019; Bhadran et al. 2022). This section justifies and discusses the different thematic layers selected for groundwater exploration in the present study. Groundwater levels have been used for demarcating potential zones, since shallow groundwater levels indicate high groundwater potential, whereas deeper groundwater levels indicate less potential (Guru et al. 2016; Kim et al. 2018). Transmissivity and storativity help in modeling groundwater productivity since they control the ability of the water-containing layer to transmit and store water (Al-Abadi 2017; Narayanamurthi and Ramasamy 2022). To better understand erosion and runoff processes, the curvature function can be used to define the physical properties of a drainage basin (Manap et al. 2014; Arulbalaji et al. 2019). In addition to the density of drainage and lineaments, the distance to the river (Razandi et al. 2015; Mallick et al. 2019) and distance to faults (Al-Abadi 2017) were also used as criteria for understanding groundwater prospects in a region. The normalized difference water index was taken as a parameter to determine the soil moisture and water body frequency in the area (Sresto et al. 2021). The Topographic Wetness Index (TWI) calculated from the Digital Elevation Model (DEM) defines the tendency for water accumulation in a terrain which allows the identification of a favorable area of groundwater recharge with less runoff (Razandi et al. 2015; Melese and Belay 2022). The density of dug well was taken for delineating potential groundwater zones since wells were constructed in locations where there is abundant water availability in nearby wells. The Topographic Position Index (TPI) is a frequently used technique for measuring topographic slope positions and automating landform identification. A low TPI indicates a flat ground surface, whereas a high TPI indicates undulating topography, and hence has been taken as a criterion for understanding infiltration characteristics (Arulbalaji et al. 2019). The groundwater recharge zones directly indicate potential zones. Water level fluctuations and specific yields of formations are indicators of groundwater recharge zones that show potential zones. The pond frequency also represents the number of water bodies that can hold or accumulate large amounts of water that recharges water (Machiwal et al. 2011; Das and Pal 2019). Hence, the abovementioned parameters were used by many researchers in their studies according to the terrain and hydrological conditions of the selected study region.

3.2 Data sources

This study was carried out using ten influencing factors that control the occurrence of groundwater. The influencing factors are: rainfall (R), slope (Sl), geomorphology (Gm), geology (Gl), lineament density (Ld), drainage density (Dd), soil texture (St), land use/land cover (Lu), topographic wetness index (TWI), and well density (Wd). The sources of the various primary and secondary data used in this study are shown in Table 1. The thematic layers of these influencing criteria were prepared using various primary and secondary data sources. All thematic maps were projected in the Universal Transverse Mercator (UTM) coordinate system and WGS1984 spatial reference. Each thematic layer was resampled and brought into a standard cell size of 30 m × 30 m.

Using the MIF and FR models, weights and ranks were assigned to each criterion selected for the study. Finally, layers of influencing criteria were integrated into GIS software, and a potential groundwater zone was delineated. In total, 162 well locations were taken for study; these locations were randomly divided into training wells (97) and testing wells (65) for validation.

The results were validated with 65 field survey observations. Furthermore, to evaluate the reliability and accuracy of the classified potential map, the area under curve (AUC) was generated to determine the relationship between the field observation wells and classified potential zones. The preparation of thematic layers and spatial analysis work was carried out using software ArcGIS 10.3 and ERDAS Imagine 2014.

Table 1 Data sources for thematic layers.

Dataset	Data Sources	Thematic Layer
Toposheet (1:50,000)	Survey of India (SOI)	Drainage network, watershed boundary, drainage density, well density
DEM (Digital Elevation Model-30m) Cartosat-1	bhuvan.nrsc.gov.in	Slope and Topographic Wetness Index
Daily rainfall data	Directorate of Economics and Statistics, Chennai	Rainfall
District Resource Map	Geological Survey of India	Geology
LISS IV(5m), Resources at 2A	National Remote Sensing Centre, India	Land Use/Land Cover, Geomorphology
Lineaments (1:50,000)	bhuvan.nrsc.gov.in	Lineament density
Soil	Tamil Nadu Agricultural University	Soil texture
Field survey	Nambiyar Watershed (Tirunelveli District)	Training and testing wells

Using the MIF and FR models, weights and ranks were assigned to each criterion selected for the study. Finally, layers of influencing criteria were integrated into GIS software, and a potential groundwater zone was delineated. In total, 162 well locations were taken for study; these locations were randomly divided into training wells (97) and testing wells (65) for validation.

The results were validated with 65 field survey observations. Furthermore, to evaluate the reliability and accuracy of the classified potential map, the area under curve (AUC) was generated to determine the relationship between the field observation wells and classified potential zones. The preparation of thematic layers and spatial analysis work was carried out using software ArcGIS 10.3 and ERDAS Imagine 2014.

3.3 Multiple Influencing Factors (MIF)

The use of Multiple Influencing Factors is a commonly used, simple, reliable, and effective method for mapping the probability of an event, especially when delineating groundwater potential zones, by understanding how influencing factors are related to each other. Major effect represents direct influence of one factor over other factor, whereas minor effect represents its indirect influence (Thapa et al. 2017; Anbarasu et al. 2020; Mahanuradha and Pragatheeswaran 2023). The major influencing factors that have a direct influence compared to others are given a value of 1, whereas the minor or indirect influencing factor is given a value of 0.5. These weights were computed statistically by following a heuristic or knowledge based method for assigning ranks to each sub class. The relationships among factors were analysed to assign ranks and weights based on exensive literature review and author’s expertise.

The sum of the values of both major and minor factors is taken to calculate the weight of each influencing factor (Magesh et al. 2012; Abijith et al. 2020; Duguma and Duguma 2022). The weight of each factor is computed using Equation 1:

(1)

Where:

W	=	weight,
A	=	major influencing factor, and
B	=	minor influencing factor.

The final output map was prepared by integrating all factors using the following formula in the raster calculator tool.

(2)

Where:

GWPI	=	Groundwater Potential Index
SlW	=	slope factor – weight,
SlR	=	slope factor – rank,
Rw	=	rainfall factor – weight,
RR	=	rainfall factor – rank,
LdW	=	lineament density factor - weight,
LdR	=	lineament density factor – rank,
DdW	=	drainage density factor – weight,
DdR	=	drainage density factor – rank,
GlW	=	geology factor – weight,
GlR	=	geology factor – rank,
GmW	=	geomorphology factor – weight,
GmR	=	geomorphology factor – rank,
StW	=	soil texture factor – weight,
StR	=	soil texture factor – rank,
LuW	=	land use factor – weight,
LuR	=	land use factor – rank,
WdW	=	well density factor – weight,
WdR	=	well density factor – rank,
TWIW	=	Topographic Wetness Index factor – weight, and
TWIR	=	Topographic Wetness Index factor – rank.

3.4 Frequency Ratio (FR)

A statistical approach to simulating environmental circumstances is the frequency ratio technique (Kim et al. 2018; Trabelsi et al. 2019; Vishwakarma et al. 2021; Maity et al. 2022). The model also employs factors relevant to the dependent variable. It is the ratio of the chance of an event occurring to the probability of it not occurring under certain conditions (Al-Abadi et al. 2017; Juanico et al. 2020; Olajide et al. 2022). This method attempts to determine the correlation between well location and the factors that influence the occurrence of groundwater, which is computed using Equation 3.

(3)

Where:

	=	number of pixels in each groundwater potential conditioning factor class,
	=	total number of all pixels in the area of interest,
	=	number of well pixels in each factor class, and
	=	number of all well pixels in the area of interest.

Here the ‘FR’ value of each class in a thematic layer was derived, which explains the relation between the groundwater governing factors and the location of training wells considered for delineating groundwater potential zones.

Further, Relative Frequency or ‘RF’ which are normalized ‘FR’ values with range of probability values from 0 to 1 are computed using Equation 4 below (Muavhi et al. 2022). Normalization is achieved by dividing the ‘FR’ of each subclass by the total ‘FR’ value of that groundwater governing factor.

(4)

 Where:

RF	=	relative frequency,
FR	=	frequency ratio of a subclass, and
∑FR	=	sum of the frequency ratio values of all subclasses of a factor.

The next step is to determine the interrelations among factors for which the weight or ‘prediction rate’ needs to be calculated (Razandi et al. 2015; Muavhi et al. 2022).

(5)

Where:

Max RF	=	maximum value of relative frequency, and
Min RF	=	minimum value of relative frequency.

In the final step, the groundwater potential index is computed using Equation 2.

4 Results and Discussion

The thematic layers prepared using various data sources for the assessment of potential groundwater zones are discussed below in detail and shown in Figure 2 (a–j).

Figure 2 (a–­j) Thematic layers generated for delineation of potential groundwater zones.

4.1 Slope

A change in the steepness of a surface determines the hydrogeological nature of a terrain (Manap et al. 2014). As the slope increases, runoff also accelerates, which in turn decreases infiltration, whereas a gentle slope supports more infiltration and percolation of water to the subsurface (Guru et al. 2016). The slope of the percentage of the watershed was calculated and divided into 7 categories: levels to nearly level (0–1%), very gently sloping (1–3%), gently sloping (3–8%), moderately sloping (8–15%), moderately steep sloping (15–30%), steeply sloping (30-50%), and very steeply sloping (> 50%). In this watershed, 583 km² (83%) of the total area comprises nearly level slopes, indicating good to moderate groundwater occurrence in this region, whereas 24.85 km² of the watershed has gently sloping terrain, and 21.43 km² (3.22%) has a moderately sloping terrain.

4.2 Rainfall

Groundwater recharge governs the long-term viability of groundwater withdrawals, of which rainwater is the primary source. When rainfall inputs to the soil exceed evapotranspiration losses, groundwater recharge rates reach a maximum (Kotchoni et al. 2019). The average annual precipitation in the Nambiyar watershed is 1027 mm, and it varies from 639.6 mm to 1881.3 mm. Relief is the most important factor influencing the rainfall distribution in watersheds. The two stations, Balamore and Keeiparai, located in the northwestern part of the watershed and lying in the Western Ghats, receive the highest annual rainfall throughout the year. In general, rainfall shows a decreasing trend from the west to the eastern part of the watershed or, in other words, from the hilly terrain to the plains.

4.3 Lineament density

Lineaments are extensive linear features on Earth’s surface that can be detected when there is a change in topography. They are calculated as the total length of all measured linear features divided by the total area considered (Melese and Belay 2022). They are directly proportional to groundwater occurrence since they are useful for understanding the relationship between porosity and permeability, which induces more water penetration and percolation capability in a terrain. The lineament density (km/km²) was classified into four classes: < 0.2, 0.2- 0.4, 0.4- 0.6, and more than 0.6. A high lineament density is found along the river stretch in the form of faults, joints, and fractures.

4.4 Drainage density

Drainage density is a dominant factor that determines the rate of infiltration of water to the ground. It controls the rate of runoff after a rainfall event. Areas with low drainage density support more groundwater occurrence, as a high drainage density decreases infiltration and increases runoff (Muavhi et al. 2022; Githinji et al. 2022). The drainage density (km/km²) in the Nambiyar watershed was divided into five classes: <0.67, 0.67–1.62, 1.62–2.74, 2.74–4.17, and 4.17–8.91. The drainage density is very high in the western part of the watershed, mainly where the Nambiyar River originates.

4.5 Geology

Various geological and geomorphological elements have significant impacts on the presence and transport of groundwater in any terrain, particularly in hard rock crystalline formations (Krishnamurthy and Srinivas 1995). The dominant rock type found in this region is gneiss, which is composed of a variety of minerals. The garnet biotite gneiss of the Migmatite Complex is dominant in the hilly region toward the western and southern portions of the watershed (245 km²), which are hard foliated and weathered rocks in nature. Khondalite group rocks, mainly garnet-sillimanite gneiss, are found in the northern and eastern parts of the watershed (327 km²). Silt clay kankar sand is found along the mouth of the Nambiyar River near the coast, covering an area of 74.12 km².

4.6 Geomorphology

Understanding different types of landforms, their characteristics, geological structures, and their evolution is necessary for assessing groundwater potential because they have definitive control over the movement and localization of groundwater (Krishnamurthy and Srinivas 1995). The geomorphological features of the watershed have been categorized into four divisions based on the mode of origin: structural, denudational, fluvial, and coastal. The broad classifications of geomorphology in the watershed are structural hills, residual hills, valley fill, inselberg, ridge, pediment, moderately weathered pediplain, shallow weathered pediplain, upper bazada, coastal plain, and dune. The shallow weathered pediplain occupies a major portion of the 409 km² (62%) of the watershed. Structural hills are found in the western part of the watershed and form a part of the Western Ghats.

4.7 Soil texture

Soil texture is one of the primary factors that determine the infiltration capacity of soil. The infiltration rate varies depending on the soil texture. Sand, loamy sand, and sandy loam have very high infiltration rates of 0.3–0.45 in/hr, whereas clay, clay loam, and silty clay loam have very slow infiltration rates of 0.00–0.05 in/hr (Subramanya 2008). Sandy clay loam is the dominant soil found in major parts of the watershed. It makes up 32% of the total watershed area (213.24 km²). Clay soil is the next most predominant soil type found in the Nambiyar River and covers an area of 132.24 km² (18.51%). Sandy clay soil constitutes approximately 137.88 km² (20.71%), which is spread across the entire watershed. Sandy loam is found on either bank of the Nambiyar River and covers an area of 132.24 km² (18.51%). Sandy soil is observed in the coastal region, which covers an area of 4.74 km² (0.712%).

4.8 Land use/Land cover

The land use/land cover of the Earth's surface is rapidly changing, driven by both natural and human forces. This has been the primary cause of shrinking green spaces, soil erosion, and loss of watershed health (Shastri et al. 2020). The watershed land use/land cover categories consist of built-up, cropland, plantations, fallow, forest, land with scrub, land without scrub, sandy area, salt-affected land, barren rock, mining land, reservoirs, tanks/ponds, and rivers. The predominant land use/cover classes found in the watershed are fallow land, which covers 184.62 km² of the total watershed area, and land with scrub occupying an area of 179.57 km². Cropland occupies approximately 75.05 km² of the watershed, which is approximately 11% of the total area. Land without scrub, which is prominent in the northeastern and southwestern parts of the watershed, covers an area of 30.55 km². Tanks and reservoirs are the major sources of irrigation in the watershed and cover a total area of 51.90 km². The water bodies in the watershed are divided into three categories: rivers, tanks/ponds, and reservoirs, which were given higher ranks since they are considered excellent recharge zones.

4.9 Well density

Previously, wells were constructed in locations where water was available in nearby wells. Additionally, more wells are clustered where there is a high groundwater potential (Doke et al. 2021). Hence, well density is considered an important parameter in this study. The density (no. of wells/km²) was classified as < 2, 2–4, 4–6, 6–8, or more than 8 wells per km². The density of wells was higher in the central part of the watershed, especially adjacent to the foothills of the structural hills.

4.10 Topographic Wetness Index

The Topographic Wetness Index (TWI) is a tool for quantifying topographic influences on hydrologic processes (Sorensen et al. 2006). Soil moisture is essential for various environmental functions, but quantifying and interpolating the whole Earth’s surface is difficult; hence, many researchers have developed digital elevation models to measure soil moisture in the form of a wetness index considering catchment area, flow width, and slope gradient (Mallick et al. 2019; Kopecký et al. 2021). The TWI provides a measure of the wetness or moisture of the terrain. A low index value indicates less wetness, which will not cause water accumulation, whereas a high value represents an area near a water body with high wetness that will accumulate water to varying degrees (Arulbalaji et al. 2019). The TWI in the watershed was calculated using the following formula:

(6)

Where:

α	=	upslope area, and
β	=	slope of the terrain.

The wetness indices of the watershed were divided into five classes: < 5, 5–10, 10–15, 15–20 and > 20. A high index value is given a higher weight, and vice versa.

4.11 Potential groundwater zones

The groundwater potential zones delineated using the MIF model were categorized into very low, low, moderate, high, and very high classes. Based on the results, the zones for groundwater occurrence are classified as follows: a very high potential zone covering an area of 97.84 km² (14.70%), a high potential zone covering an area of 163.45 km² (24.55%), a moderate potential zone covering an area of 204 km² (30%), a low potential zone covering 135.5 km² (20.35%), and a very low potential zone covering 64.02 km² (9.61%) of the total area of the watershed. In the study, the soil texture factor was given more weight, followed by geology, with values of 15 and 13, respectively. Both factors are significant in determining the groundwater potential zones. It is observed that the silt clay kankar sand category found downstream of the Nambiyar River had a high potential for groundwater occurrence.

The relationships among the various influencing factors are shown in Figure 3, and their ranks and weights are given in Table 2 and Table 3.

Figure 3 Inter-relationship among influencing factors using MIF model.

Table 2 Weights calculated for each factor using MIF model.

Factors	Major (A)	Minor (B)	Relative weight	Weight
Slope	1	0.5	1.5	7
Rainfall	1+1	0	2	9
Lineament density	1+1	0.5	2.5	11
Drainage density	1+1	0.5	2.5	11
Geology	1+1	0.5+0.5	3	13
Geomorphology	1	0.5	1.5	7
Soil texture	1+1+1	0.5	3.5	15
Land use/Land cover	1	0.5+0.5	2	9
Well density	1+1	0.5	2.5	11
Topographic Wetness Index	1	0.5	1.5	7
Total			22.5

Table 3 Ranks and weights calculated for each factor using an MIF model.

	Class	Rank	Weight		Factors	Class	Rank	Weight
Slope (%)	0-1	10	7		Soil texture	Clay	2	15
	1-3	10				Clayloam	4
	3-8	10				Loamysand	8
	8-15	8				Sand	6
	15-30	2				Sandyclay	6
	30-50	2				Sandyclayloam	6
	>50	2				Sandyloam	10
Rainfall (mm)	633.84-708.50	4	9		Land Use/Cover	Tanks	10	9
	708.508-802.624	6				Reservoirs	10
	802.624-978.35	8				River	10
	978.35-1202.68	8				Urban	4
	1202.68-1487.37	10				Rural	6
Drainage density	<0.67	10	11			Land w/out scrub	4
	0.67-1.62	8				Land with scrub	4
	1.62-2.74	6				Forest	4
	2.74-4.17	4				Plantation	8
Geology	4.17-8.91	2	13			Mining	6
	Garnet biotite sillimanite gneiss	8				Cropland	8
	Garnet biotite gneiss	6				Barren rock	2
	Calcareous sandstone	8				Fallow	6
	Clay sand alluvium	8				Salt affected	4
Geomorphology	Silt clay kankar sand	6	7			Sandy area	6
	Charnockite	4			Well density	<2	4	11
	Residual hills	2				2-4	6
	Inselberg	2				4-6	8
	Ridge	4				6-8	8
	Moderately weathered pediplain	8				>8	10
	Coastal plain	6	7		Topographic Wetness Index	<5	2	7
	Pediment	8				5-10	4
	Structural hills	2				10-15	6
	Sand dune	4				15-20	8
	Shallow weathered pediplain	8				>20	10
	Upper Bazada	10			Lineament	<0.2	4	11
	Valley fill	10			density	0.2-0.4	6
						0.4-0.6	8
						>0.6	10

In the FR approach, after preparing all the controlling parameters, FR values were obtained based on 70% of the training dataset as the ratio of the percentage of available wells in each subclass by area to the percentage of each subclass. The locations of the training and testing wells are shown in Figure 4. The results of the FR model reveal that 186.55 km² is categorized as a very high potential zone for groundwater, comprising 28.03% of the study area, followed by a high potential zone of 327.72 km² (49.24%). The very low potential zones of 63.29 km² (9.50%) are found in the western part of the watershed. The FR was normalized to obtain a probability range value from 0 to 1, which is the relative frequency value. The computed values of the frequency ratio, relative frequency, and prediction rate are presented in Table 4. When the RF values are recorded for the slope factor, 100% of the training wells fall within 0-1% of the slope value (level to nearly level), which supports high infiltration; hence, an RF value of 1 is obtained. The highest RF value for the rainfall factor is 0.445 for the moderate rainfall class (700–900 m), where more than 50% of the wells are found. In the case of the lineament factor, 52% of the wells fall within the class of < 0.2 km/km² of lineament density for which the RF value is 0.23. Since a low drainage density is suitable for groundwater occurrence, 23% of the wells had drainage densities < 0.67 km/km² category with an RF value of 0.30. A moderate drainage density (0.67–1.67 km/km²) had the highest percentage of wells (31%), with an RF of 0.31. Two varieties of hard rocks are dominant in the watershed, viz. garnet biotite sillimanite gneiss and garnet biotite gneiss. The garnet biotite sillimanite gneiss sample has an RF value of 0.19, in which the highest percentage of wells is observed (43%). Calcareous sandstone and clay sand alluvium, which have RF values of 0.29 and 0.20, respectively, are also suitable for groundwater. Both moderately weathered and shallowly weathered pediplains occupy a major part of the Nambiyar watershed suitable for groundwater, which attain RF values of 0.34 and 0.30, respectively. The coastal plains downstream of the River Nambiyar were also found to be suitable for potential groundwater zones and hence achieved a value of 0.29. Variants of clay soil, such as clay loam, sandy clay and sandy clay loam, are the dominant soil textures found in the watershed. Clay loam had a high RF value of 0.29, followed by loamy sand, sandy clay, and sandy clay loam, with RF values of 0.18, 0.17, and 0.14, respectively. Land use categories also determine the rate of infiltration of water in a region. Plantations and croplands obtained the highest RF values of 0.27 and 0.22, respectively. Scrublands are moderately suitable for groundwater occurrence and hence attained an RF value of 0.08. Around 47% of the fallow lands have dug wells, with an RF value of 0.183. Dug well density is an important aspect of determining potential groundwater zones by comparing the positions of previously installed wells. For areas with a density of more than 8 wells/km², a high RF value of 0.40 is created. The topographic wetness index, which measures soil moisture, is an important aspect of groundwater potential research. An index value of 15–20 in the wetness index category achieved a high RF value of 0.61, followed by an RF value of 0.25 in 24% of the wells in the TWI category of 10–15.

Figure 4 Location of training and testing wells selected for FR model.

Table 4 Ranks and prediction rate (weight) assigned to subclasses and factors to determine potential groundwater zones through an RF model.

Factors	Class	Class pixel	%	Well pixel	%	# of wells	FR	RF	RF (%)	RF (INT)	MinRF	MaxRF	Max-Min RF	Min (Max-Min)	PR
Slope (%)	0-1	647860	87.61	69300	100.00	77.00	1.14	1.000	100.00	100
	1-3	21406	2.89	0	0.00		0.00	0.000	0.00	0
	3-8	27434	3.71	0	0.00		0.00	0.000	0.00	0
	8-15	23694	3.20	0	0.00		0.00	0.000	0.00	0
	15-30	13556	1.83	0	0.00		0.00	0.000	0.00	0
	30-50	4774	0.65	0	0.00		0.00	0.000	0.00	0
	>50	766	0.10	0	0.00		0.00	0.000	0.00	0
	Total	739490		69300			1.14				0.00	1.00	1.00	0.27	3.703
Rainfall (mm)	633.84-708.50	195315	26.41	17100	24.68	19.00	0.93	0.327	32.65	32
	708.508-802.624	399464	54.02	47700	68.83		1.27	0.445	44.53	44
	802.624-978.35	73567	9.95	4500	6.49	5.00	0.65	0.228	22.81	22
	978.35-1202.68	39905	5.40	0	0.00	0.00	0.00	0.000	0.00	0
	1202.68-1487.37	31256	4.23	0	0.00	0.00	0.00	0.000	0.00	0
	Total	739507		69300			2.86				0.00	0.45	0.45	0.27	1.666
Lineament density (km/km2)	<0.2	510004	68.97	46800	100.00	52.00	1.00	0.233	23.25	23
	0.2-0.4	62703	8.48	8100	17.31	9.00	1.41	0.327	32.74	32
	0.4-0.6	62430	8.44	5400	11.54	6.00	0.94	0.219	21.92	21
	>0.6	104296	14.10	9000	16.13	10.00	0.95	0.485	48.46	21
	Total	739433		69300			4.30				0.22	0.48	0.27	0.27	1.00
Drainage density (km/ km2)	<0.67	185831	25.13	20700	29.87	23.00	1.19	0.308	30.83	30
	0.67-1.62	247497	33.47	27900	40.26	31.00	1.20	0.312	31.20	31
	1.62-2.74	174868	23.64	17100	24.68	19.00	1.04	0.271	27.07	27
	2.74-4.17	91449	12.37	3600	5.19	4.00	0.42	0.109	10.90	10
	4.17-8.91	39913	5.40	0	0.00	0.00	0.00	0.000	0.00	0
	Total	739558		69300			3.86				0.00	0.31	0.31	0.27	1.148
Geology	Garnet biotite sillimanite gneiss	363550	49.17	38700	55.84	43.00	1.14	0.191	19.10	19
	Garnet biotite gneiss	277169	37.48	20700	29.87	23.00	0.80	0.134	13.40	13
	Calcareous sandstone	5419	0.73	900	1.30	1.00	1.77	0.298	29.80	29
	Clay sand alluvium	8053	1.09	900	1.30	1.00	1.19	0.201	20.05	20
	Silt clay kankar sand	82357	11.14	8100	11.69	9.00	1.05	0.176	17.65	17
	Charnockite	2873	0.39	0	0.00	0.00	0.00	0.000	0.00	0
	Total	739421		69300			5.95				0.00	0.30	0.30	0.27	1.111
Geomorphology	Residual hills	2293	0.31	0	0.00	0.00	0.00	0.000	0.00	0
	Inselberg	1558	0.21	0	0.00	0.00	0.00	0.000	0.00	0
	Ridge	549	0.07	0	0.00	0.00	0.00	0.000	0.00	0
	Moderately weathered pediplain	126838	17.19	16200	23.38	18.00	1.36	0.348	34.79	34
	Coastal plain	16803	2.28	1800	2.60	2.00	1.14	0.292	29.18	29
	Pediment	41777	5.66	900	1.30	1.00	0.23	0.059	5.87	5
	Structural hills	86003	11.65	0	0.00	0.00	0.00	0.000	0.00	0
	Sand dune	704	0.10	0	0.00	0.00	0.00	0.000	0.00	0
	Shallow weathered pediplain	455252	61.69	50400	72.73	56.00	1.18	0.302	30.16	30
	Upper Bazada	3203	0.43	0	0.00	0.00	0.00	0.000	0.00	0
	Valley fill	2928	0.40	0	0.00	0.00	0.00	0.000	0.00	0
	Total	737908		69300			3.91				0.00	0.35	0.35	0.27	1.296
Soil texture	Clay	167564	22.66	16200	23.38	18.00	1.03	0.145	14.51	14
	Clayloam	32119	4.34	6300	9.09	7.00	2.09	0.294	29.43	29
	Loamysand	7398	1.00	900	1.30	1.00	1.30	0.183	18.26	18
	Sand	5264	0.71	0	0.00	0.00	0.00	0.000	0.00	0
	Sandyclay	153196	20.72	18000	25.97	20.00	1.25	0.176	17.63	17
	Sandyclayloam	236863	32.03	22500	32.47	25.00	1.01	0.143	14.25	14
	Sandyloam	137002	18.53	5400	7.79	6.00	0.42	0.059	5.91	5
	Total	739406		69300			7.11				0.00	0.29	0.29	0.27	1.074
Land Use/Cover	Tanks	56774	7.68	1800	2.60	2.00	0.34	0.037	3.66	3
	Reservoirs	870	0.12	0	0.00	0.00	0.00	0.000	0.00	0
	River	4968	0.67	0	0.00	0.00	0.00	0.000	0.00	0
	Urban	6727	0.91	900	1.30	1.00	1.43	0.154	15.45	15
	Rural	24012	3.25	900	1.30	1.00	0.40	0.043	4.33	4
	Land without scrub	33857	4.58	0	0.00	0.00	0.00	0.000	0.00	0
	Land with scrub	199424	26.97	14400	20.78	16.00	0.77	0.083	8.34	8
	Forest	95411	12.90	0	0.00	0.00	0.00	0.000	0.00	0
	Plantation	11324	1.53	2700	3.90	3.00	2.54	0.275	27.53	27
	Mining	5182	0.70	0	0.00	0.00	0.00	0.000	0.00	0
	Cropland	83375	11.28	16200	23.38	18.00	2.07	0.224	22.44	22
	Barren rock	3689	0.50	0	0.00	0.00	0.00	0.000	0.00	0
	Fallow	204992	27.72	32400	46.75	36.00	1.69	0.183	18.25	18
	Salt affected	8271	1.12	0	0.00	0.00	0.00	0.000	0.00	0
	Sandy area	515	0.07	0	0.00	0.00	0.00	0.000	0.00	0
	Total	739391		69300			9.24				0.00	0.28	0.28	0.27	1.037
Well density (No. of wells/km2)	<2	247583	33.49	3600	5.19	4.00	0.16	0.020	1.95	1
	2-4	242447	32.79	18000	25.97	20.00	0.79	0.100	9.96	9
	4-6	136670	18.49	22500	32.47	25.00	1.76	0.221	22.10	22
	6-8	79553	10.76	15300	22.08	17.00	2.05	0.258	25.81	25
	>8	33071	4.47	9900	14.29	11.00	3.19	0.402	40.18	40
	Total	739324		69300			7.95				0.02	0.40	0.38	0.27	1.407
Topographic Wetness Index	<5	30109	4.07	0	0.00	0.00	0.00	0.000	0.00	0
	5-10	585246	79.14	47700	68.83	53.00	0.87	0.135	13.47	13
	10-15	111479	15.07	17100	24.68	19.00	1.64	0.254	25.35	25
	15-20	12156	1.64	4500	6.49	5.00	3.95	0.612	61.18	61
	>20	540	0.07	0	0.00	0.00	0.00	0.000	0.00	0
	Total	739530		69300			6.46				0.00	0.61	0.61	0.27	2.259

Low drainage density, high lineament density, and croplands are highly suitable for groundwater occurrence, which is evident from similar studies carried out using multi-influencing and frequency ratio models (Magesh et al. 2012; Muavhi et al. 2022).

The results of both models show that the groundwater potential zones are clustered in the central part of the watershed, mainly in the banks of the Nambiyar River. The groundwater potentials delineated using the MIF and FR methods are given in Figures 5 (a & b). The results for the groundwater potential classes and area under each class generated using both models are given in Table 5. In high potential groundwater zones, sandy loam soil occurs, which is found along the stretch of the Nambiyar River. Even though highlands located in the north western part of the watershed, which constitute a portion of the Western Ghats, receive more rainfall, they are characterized with a lower potential zone for groundwater due to very steep slopes, high drainage density, and the presence of hard gneissic rocks that are covered by clay, sandy clay, and sandy clay loam.

Figure 5 Potential groundwater zones of the Nambiyar watershed using (a) MIF, and (b) FR methods.

Table 5 Area under each potential groundwater class using FR and MIF models.

Potential groundwater classes	Frequency ratio		Multi-influencing factor
	Area	%	Area	%
Very high	186.55	28.03	97.84	14.70
High	327.72	49.24	163.45	24.56
Moderate	70.02	10.52	204.74	30.76
Low	17.94	2.70	135.50	20.36
Very low	63.29	9.51	64.02	9.62
Total	665.5		665.5

The watershed consists of both porous and fissured formations including:

1. weathered and fractured hard rock formations dating back to the Archaean era, and
2. porous sedimentary formations formed during the Tertiary and Recent periods.

Additionally, isolated deposits of calcareous sandstone and fossiliferous limestone occur in the coastal region, particularly along the southeastern margin. Laterites are exposed as patches along the Radhapuram-Edakkadu, Vijayanarayanam-Kumarapuram, Ittamoli, and Nanguneri areas in the watershed. Exploration in the sedimentary tract has revealed that the depth to basement occurs at 120 mbgl (metres below ground level), and granular zones are found between the depths of 20–92 mbgl. Here, the yield of bore wells varies from 1–4.5 lps (litres per second). The water-bearing properties of crystalline formations, which lack primary porosity, depend on the extent of development of secondary intergranular porosity. These aquifers are highly heterogeneous in nature due to variations in lithology, texture, and structural features, even within short distances (CGWB 2012).

The occurrence and movement of groundwater is restricted to the open system of fractures and joints in un-weathered portions of the watershed, and in the porous zones of weathered formations. Groundwater development in the blocks is categorised as ‘safe’, ‘semi-critical’, ‘critical’, and ‘over-exploited’ areas based on two criteria: stage of groundwater development and long-term trend of pre- and post-monsoon water levels. The ‘over exploited’ and ‘critical’ areas are where there should be intensive monitoring, evaluation, and future groundwater development should be linked with water conservation measures, whereas the ‘semi- critical’ areas are recommended for cautious groundwater development. Radhapuram Block is classified as ‘over exploited’, while Tisayanvilai Block is classified as ‘semi- critical’, both areas are located near the coast. Radhapuram Block is closest to the coast, where a high level of groundwater development is found and might cause seawater intrusion (CGWB 2009).

Key programs implemented for watershed development in the Tamil Nadu State include the Drought Prone Area Program (DPAP), National Watershed Development Project for Rainfed Areas (NWDPRA), Integrated Wasteland Development Program (IWDP), and Desert Development Program (DDP). Additionally, the River Valley Project (RVP) and various initiatives supported by international organizations like DANIDA and DFID are also significant. These watershed development projects aim to improve the productivity of rainfed, degraded land, promote livelihoods, and enhance project effectiveness through integrated management of land and water resources through community-based institutions, and cross-learning mechanisms (CGWB 2012).

In summary, limitations caused by nature, human activity, and the current socioeconomic situation can all be used to analyze groundwater concerns and challenges in the watershed. In terms of the quantitative characteristics of groundwater, more rainfall occurring in short duration results in less recharge and more runoff. This, in turn, causes less groundwater availability because of more extraction than recharging.

Authorities' experimental studies have shown that desilting existing tanks and then creating a percolation pond with recharge wells and recharge shafts are cost-effective ways to boost artificial recharge in the right places and prevent groundwater depletion, particularly in semi-critical and over-exploited blocks. Unused bore and excavated wells can be turned into artificial recharge structures, which are a smart idea for replenishing groundwater (CGWB 2022).

Intensive monitoring of groundwater levels and water quality should be considered in the coastal areas of the district to monitor the interaction of the fresh water/saline water interface. The problems caused due to intensive groundwater extraction, intensive surface water irrigation, intensive mining, and growing urban and industrial establishments are human induced activities prominent in the watershed leading to drastic depletion in groundwater resources. The existing socio-economic setup of farmers here is such that a farmer having more than 5 acres of cropland has less expense than a farmer having one acre. Also, the supply of free electricity to all farmers provides them the ability to extract more groundwater. To avoid this, proper guidance should be given to the farmers for groundwater usage. The pricing policy for using groundwater for irrigation, industrial, and domestic purposes is also an important strategy under consideration for adoption by the State. This will be significant in controlling the illegal extraction of groundwater. In addition, a waste land development program is suggested to increase agricultural output.

4.12 Validation

The assessment of the effectiveness, reliability, robustness, and accuracy of different models used for groundwater mapping is of prime importance. The AUC is often applied to compare and validate probability models (Mas et al. 2013; Abijith et al. 2020; Muavhi et al. 2022). The AUC can ensure the accuracy of groundwater potential models by revealing the relationship between the cumulative percentage of groundwater potential classes and the cumulative percentage of dug well availability (Doke et al. 2020). AUC values ranging from 0 to 1 were used to assess model accuracy. A larger area under curve indicates that the spatial models employed for groundwater mapping are more efficient. The plot analysis showed that the area under curve in the potential maps was 72% for FR and 62% for MIF, as shown in Figure 6. As a result, the method used in this study is good to moderately accurate, and consistent in terms of forecasting prospective zones. Since a larger AUC is obtained using the frequency ratio model compared to the MIF model in this study, it is obvious that the frequency ratio model outperformed the MIF. A similar validation method is used in different studies for similar models where the AUC values ranged from 72.47–75% for the frequency ratio model, and 65–75% for the MIF model (Doke et al. 2020; Muavhi et al. 2022). Also, the accuracy and efficiency of the frequency ratio model is evident from similar results obtained in other studies. While the MIF model generally classifies groundwater zones by analyzing the relationship between each factor, the frequency ratio model offers a more practical solution for delineating groundwater potential zones by considering each governing factor for groundwater occurrence, as well as the influence of these factors on the existing ‘well’ locations to enhance accuracy in the results.

Figure 6 Area under curve generated for MIF and FR models.

A shallow or high groundwater table indicates high groundwater potential, whereas a low water table is an indication of low groundwater potential or occurrence (Guru et al. 2016; Doke et al. 2020).Hence, the groundwater level of observation wells was used to validate the delineated groundwater potential zones. The results of groundwater potential were validated with the water yield capacity of 34 observation wells, and it was found that low groundwater potential zones have a lower water yield capacity, while wells located in high potential zones have a higher water yield capacity (Arulbalaji et al. 2019).

Hence, in addition to the AUC techniques used for validation, groundwater level data collected from the 20 observation wells of the State Ground and Surface Water Resources Data Centre, in Chennai, were also used to validate the potential zones, which are shown in Figure 7. In the present study, groundwater level data shows zones with low groundwater potential have a water table ranging from 7–10 mbgl (metres below ground level) whereas high potential zones for groundwater have a high water table, ranging from 3–5 mbgl. The images shown in Figure 8 (a–e) are some of the field photographs taken during the visit to the Nambiyar watershed.

Figure 7 Groundwater level in the Nambiyar watershed.

Figure 8 a) Upper reaches of the Nambiyar River; b) Nambiyar Dam in Kotaikarungulam; c) River Nambiyar at its mouth; d) Water sample collection from dug well at Veppankulam; and e) Dug well at Valliyur Block.

5 Conclusions

Groundwater supply is critical for meeting additional demand from farmers, industry, and domestic use, which drives the need for groundwater development in a region. Generally, the total irrigated area, average water level, rainfall, cropping pattern, specific yield of formations, and recharge through artificial recharge structures are considered when assessing an area for its groundwater potential. In terms of quantitative factors, rainfall may increase in a short period of time. Therefore, recharge will be lower, runoff will be higher, and groundwater availability will be reduced due to overextension rather than recharge. Probability-based statistical model frequency ratios and multi-influencing factor methods were employed to delineate potential zones for groundwater occurrence in the watershed. The results were categorized into five classes: very low, low, moderate, high, and very high. The very high potential zones determined using the FR model are 186 km² (28.03%) and 97.84 km² (14%) under the MIF technique, whereas 63.29 km² (9.50%) and 64.02 km² (9.61%) were found to have very low potential under the FR and MIF models, respectively. The very low potential zones located in the northwestern part of the watershed are composed of structural hills and ridges covered by reserved forest, which forms part of the Western Ghats. A comparative study of the two models used in this study showed that the FR model produced better and more accurate results than the multi-influencing factor model. The study suggested that because groundwater development is still at a safe stage in many blocks of this area, future groundwater development for the construction of extra irrigation potential must be performed with extreme caution. In the ‘overexploited’ blocks, Radhapuram is in the coastal plain, and necessary steps for regulating groundwater extraction here may be adopted. To track the migration of the fresh water-saline water interface, intensive monitoring of groundwater levels and water quality is required throughout the district's coastal areas. Due to the Tirunelveli district's limited water resources, where the watershed is located, a wasteland development project and micro-irrigation system must be developed to boost agricultural productivity by supplying more food and money per drop of water. The pricing policy for groundwater users is also a significant technique for limiting unlawful groundwater extraction. The utilization of unused dug wells and bore wells as artificial recharge structures will be an effective strategy for replenishing groundwater. As a result, a greater understanding of the nature, flow pattern, quality, and quantity of aquifers achieved by continuous monitoring and maintenance of long-term groundwater data benefits water resource management and increases production, which is helpful for sustaining the livelihood of the population residing in the Nambiyar watershed.

Acknowledgments

The authors would like to express their gratitude for the assistance from the Department of Science and Technology, University Grant Commission, and Indian Council for Social Science Research, New Delhi, India as a Research Fellowship. The authors are highly thankful to the faculty and research scholars of the Department of Geography, Bharathidasan University, Tamil Nadu, for their assistance and guidance.

References

1. Abijith, D., S. Saravanan, L. Singh, J.J. Jesudasan, T. Saranya, and K.S.S. Parthasarathy. 2020. “GIS-Based Multi-Criteria Analysis for Identification of Potential Groundwater Recharge Zones - a Case Study from Ponnaniyaru Watershed, Tamil Nadu, India.” HydroResearch 3, 1–14. https://doi.org/10.1016/j.hydres.2020.02.002
2. Al-Abadi, A.M. 2017. “Modeling of Groundwater Productivity in Northeastern Wasit Governorate, Iraq Using Frequency Ratio and Shannon’s Entropy Models.” Applied Water Science 7, 699–716. https://doi.org/10.1007/s13201-015-0283-1
3. Anbarasu, S., K. Brindha, and L. Elango. 2020. “Multi-Influencing Factor Method for delineation of potential groundwater zones using remote sensing and GIS techniques in the western part of Perambalur District, Southern India.” Earth Science Informatics 13 (2): 317–32. https://doi.org/10.1007/s12145-019-00426-8
4. Arshad, A., Z. Zhang, W. Zhang, and A. Dilawar. 2020. “Mapping Favorable Groundwater Potential Recharge Zones Using a GIS-Based Analytical Hierarchical Process and Probability Frequency Ratio Model: A Case Study from an Agro-Urban Region of Pakistan.” Geoscience Frontiers 11 (5): 1805–1819, https://doi.org/10.1016/j.gsf.2019.12.013
5. Arulbalaji, P., D. Padmalal, and K. Sreelash. 2019. “GIS and AHP techniques-based delineation of potential groundwater zones: A case study from Southern Western Ghats, India.” Scientific Reports 9, 2082. https://doi.org/10.1038/s41598-019-38567-x
6. Aslan, V. and R. Çelik. 2021. “Integrated GIS-Based Multi-Criteria Analysis for Groundwater Potential Mapping in the Euphrates’s Sub-Basin, Harran Basin, Turkey.” Sustainability 13 (13): 7375. https://doi.org/10.3390/su13137375
7. Balasubramanian, A. 2017. “Geological Formations of India.” Report Number 2. University of Mysore, Centre for Advanced Studies in Earth Science, Karnataka, India. https://doi.org/10.13140/RG.2.2.35307.75042
8. Bhadran, A., D. Girishbai, N.P. Jesiya, G. Gopinath, R.G. Krishnan, and V.K. Vijesh. 2022. “A GIS based Fuzzy-AHP for delineating potential groundwater zones in tropical river basin, southern part of India.” Geosystems and Geoenvironment 1 (4): 100093. https://doi.org/10.1016/j.geogeo.2022.100093
9. Bhattacharya, S., S. Das, S. Das, M. Kalashetty, and S.R. Warghat. 2021. “An Integrated Approach for Mapping Groundwater Potential Applying Geospatial and MIF Techniques in the Semiarid Region.” Environment, Development and Sustainability 23, 495–510. https://doi.org/10.1007/s10668-020-00593-5
10. Castillo, J.L.U., D.A.M. Cruz, J.A.R. Leal, J.T. Vargas, S.A.R. Tapia, and A.E.M. Celestino. 2022. “Delineation of Groundwater Potential Zones (GWPZs) in a Semi-Arid Basin through Remote Sensing, GIS, and AHP Approaches.” Water 14 (13): 2138. https://doi.org/10.3390/w14132138
11. Central Ground Water Board (CGWB). 2009. “District Groundwater Brochure Tirunelveli District, Tamil Nadu.” Technical Report Series. Tamil Nadu Central Ground Water Board. pp. 9–11.
12. Central Ground Water Board (CGWB). 2012. “Aquifer Systems of Tamil Nadu & Puducherry.” Southeastern Coastal Region Chennai. pp. 19–25.
13. Central Ground Water Board (CGWB). 2022. “National Compilation on Dynamic Groundwater Resources of India 2022.” Faridabad, India. pp. 79.
14. Chaudhry, A.K., K. Kumar, and M.A. Alam. 2019. “Mapping of Groundwater Potential Zones using the Fuzzy Analytic Hierarchy Process And Geospatial Technique.” Geocarto International 36 (20): 2323–2344. https://doi.org/10.1080/10106049.2019.1695959
15. Das, B. and S.C. Pal. 2019. “Combination of GIS and Fuzzy-AHP for Delineating Groundwater Recharge Potential Zones in the Critical Goghat-II Block of West Bengal, India.” HydroResearch 2, 21–30. https://doi.org/10.1016/j.hydres.2019.10.001
16. Doke, A., S.D. Pardeshi, and S. Das. 2020. “Drainage Morphometry and Groundwater Potential Mapping: Application of Geoinformatics with Frequency Ratio and Influencing Factor Approaches.” Environmental Earth Sciences 79, 393. https://doi.org/10.1007/s12665-020-09137-6
17. Doke, A.B., R.B. Zolekar, H. Patel, S. Das. 2021. “Geospatial Mapping of Potential Groundwater Zones Using Multi-Criteria Decision-Making AHP Approach in a Hardrock Basaltic Terrain in India.” Ecological Indicators 127, 107685. https://doi.org/10.1016/j.ecolind.2021.107685
18. Duguma, T.A. and G.A. Duguma. 2022. “Assessment of Potential Groundwater Zones of Upper Blue Nile River Basin Using Multi-Influencing Factors under GIS and RS Environment: A Case Study on Guder Watersheds, Abay Basin, Oromia Region, Ethiopia.” Geofluids 2022 (1172039): 1–26. https://doi.org/10.1155/2022/1172039
19. Etikala, B., V. Golla, P. Li, and S. Renati. 2019. “Deciphering potential groundwater zones using MIF Technique and GIS: A Study from Tirupati Area, Chittoor District, Andhra Pradesh, India.” HydroResearch 1, 1–7. https://doi.org/10.1016/j.hydres.2019.04.001
20. Fagbohun, B.J. 2018. “Integrating GIS and Multi-Influencing Factor Technique for Delineation of Potential Groundwater Recharge Zones in Parts of Ilesha Schist Belt, Southwestern Nigeria.” Environmental Earth Sciences 77, 69. https://doi.org/10.1007/s12665-018-7229-5
21. Faheem, H., Z. Khattak, F. Islam, R. Ali, R. Khan, I. Khan, and E.T. Eldin. 2023. “Groundwater Potential Zone Mapping Using Geographic Information Systems and Multi-Influencing Factors: A Case Study of the Kohat District, Khyber Pakhtunkhwa,” Frontiers in Earth Science 11, 1097484. https://doi.org/10.3389/feart.2023.1097484
22. Githinji, T.W., E.W. Dindi, Z.N. Kuria, D.O. Olago. 2022. “Application of Analytical Hierarchy Process and Integrated Fuzzy-Analytical Hierarchy Process for Mapping Potential Groundwater Recharge Zone Using GIS in the Arid Areas of Ewaso Ng’iro – Lagh Dera Basin, Kenya.” HydroResearch 5, 22–34. https://doi.org/10.1016/j.hydres.2021.11.001
23. Goyal, S.K., B.S. Chaudhary, O. Singh, G.K. Sethi, and P.K. Thakur. 2010. “Variability Analysis of Groundwater Levels, AGIS-Based Case Study.” Journal of the Indian Society of Remote Sensing 38: 355–364. https://doi.org/10.1007/s12524-010-0024-8
24. Guru, B., K. Seshan, and S. Bera. 2016. “Frequency ratio model for groundwater potential mapping and its sustainable management in cold desert, India.” Journal of King Saud University – Science 29 (3): 333–347. https://doi.org/10.1016/j.jksus.2016.08.003
25. Juanico, D.E., F. Cayson, and C. Patiño. 2020. “Noise-Resilience Horizon of Groundwater Potential Maps Generated by Frequency-Ratio Data Integration.” Frontiers in Environmental Science 8, 523988. https://doi.org/10.3389/fenvs.2020.523988
26. Kabeto, J., D. Adeba, M.S. Regasa, and M.K. Leta. 2022. “Groundwater Potential Assessment using GIS and Remote Sensing Techniques: Case Study of West Arsi Zone, Ethiopia.” Water 14 (12): 1838. https://doi.org/10.3390/w14121838
27. Kaewdum, N. and S. Chotpantarat. 2021. “Mapping Potential Zones for Groundwater Recharge Using a GIS Technique in the Lower Khwae Hanuman Sub-Basin Area, Prachin Buri Province, Thailand.” Frontiers in Earth Science 9, 717313. https://doi.org/10.3389/feart.2021.717313
28. Kim, J-C., H-S. Jung, and S. Lee. 2018. “Groundwater Productivity Potential Mapping Using Frequency Ratio and Evidential Belief Function and Artificial Neural Network Models: Focus on Topographic Factors.” Journal of Hydroinformatics 20 (6): 1436–1451. https://doi.org/10.2166/hydro.2018.120
29. Kisiki, C.P., T.W. Bekele, T. Ayenew, and I.C. Mjemah. 2022. “Geospatial Application on Mapping Groundwater Recharge Zones in Makutupora Basin, Tanzania.” Heliyon 8 (10): e10760. https://doi.org/10.1016/j.heliyon.2022.e10760
30. Kopecký, M., M. Macek, and J. Wild. 2021. “Topographic Wetness Index Calculation Guidelines Based on Measured Soil Moisture and Plant Species Composition.” Science of The Total Environment 757, 143785. https://doi.org/10.1016/j.scitotenv.2020.143785
31. Kotchoni, D.O.V., J.M. Vouillamoz, F.M.A. Lawson, P. Adjomayi, M. Boukari, and R.G. Taylor. 2019. “Relationships between Rainfall and Groundwater Recharge in Seasonally Humid Benin: A Comparative Analysis of Long-Term Hydrographs in Sedimentary and Crystalline Aquifers.” Hydrogeology Journal 27, 447–457. https://doi.org/10.1007/s10040-018-1806-2
32. Krishnamurthy, J. and G. Srinivas. 1995. “Role of Geological and Geomorphological Factors in Ground Water Exploration: A Study Using IRS LISS Data.” International Journal of Remote Sensing 16 (14): 2595–2618. https://doi.org/10.1080/01431169508954579
33. Kumar, M., P. Singh, and P. Singh. 2021. “Fuzzy AHP Based GIS and Remote Sensing Techniques for the Groundwater Potential Zonation for Bundelkhand Craton Region, India.” Geocarto International 37 (22): 6671–6694. https://doi.org/10.1080/10106049.2021.1946170
34. Kumari, S., A. Poddar, N. Kumar, and V. Shankar. 2022. “Delineation of Groundwater Recharge Potential Zones Using the Modeling Based on Remote Sensing, GIS and MIF Techniques: A Study of Hamirpur District, Himachal Pradesh, India.” Modeling Earth Systems and Environment 8, 1759–1770. https://doi.org/10.1007/s40808-021-01181-w
35. Libina, R.S., K. Prakash, R. Jegankumar, S.P. Dhanabalan, and. M.A. Arya. 2023. “Assessment of Groundwater Level and Fluctuations using Geospatial Techniques in Nambiyar Watershed, South India.” In: Rai, P.K. ed. Advances in Water Resource Planning and Sustainability. Springer Nature Singapore, Singapore, p. 203–221. https://doi.org/10.1007/978-981-99-3660-1
36. Machiwal, D., M.K. Jha, and B.C. Mal. 2011. “Assessment of Groundwater Potential in a Semi-Arid Region of India using Remote Sensing, GIS and MCDM Techniques.” Water Resources Management 25, 1359–1386. https://doi.org/10.1007/s11269-010-9749-y
37. Magesh, N.S., N. Chandrasekar, and J.P. Soundranayagam. 2012. “Delineation of Groundwater Potential Zones in Theni District, Tamil Nadu, using Remote Sensing, GIS and MIF Techniques.” Geoscience Frontiers 3 (2): 189–196. https://doi.org/10.1016/j.gsf.2011.10.007
38. Mahanuradha, P. and P. Pragatheeswaran. 2023. “Identification of Potential Groundwater Zones Using Geospatial Techniques in the Varattar Sub-Basin, Dharmapuri District, South India.” Journal of Geology, Geography and Geoecology 31 (4): 668–677. https://doi.org/10.15421/112263
39. Maity, B., S.K. Mallick, P. Das and S. Rudra. 2022. “Comparative analysis of groundwater potentiality zone using fuzzy AHP, frequency ratio and Bayesian weights of evidence methods.” Applied Water Science 12, 63. https://doi.org/10.1007/s13201-022-01591-w
40. Mallick, J., R.A. Khan, M. Ahmed, S.D. Alqadhi, M. Alsubih, I. Falqi, and M.A. Hasan. 2019. “Modeling Groundwater Potential Zone in a Semi-Arid Region of Aseer Using Fuzzy-AHP and Geoinformation Techniques.” Water 11 (12): 2656. https://doi.org/10.3390/w11122656
41. Manap, M.A., H. Nampak, B. Pradhan, S. Lee, W. Nor, A. Sulaiman, and M.F. Ramli. 2014. “Application of Probabilistic-Based Frequency Ratio Model in Groundwater Potential Mapping Using Remote Sensing Data and GIS.” Arabian Journal of Geosciences 7, 711–724. https://doi.org/10.1007/s12517-012-0795-z
42. Mas, J-F., B.S. Filho, R.G. Pontius, M.F. Guitierrez, and H. Rodrigues. 2013. “A Suite of Tools for ROC Analysis of Spatial Models.” ISPRS International Journal of Geo-Information 2 (3): 869–887. https://doi.org/10.3390/ijgi2030869
43. Masitoh, F., F.B. Ar-Rouf, and A.N. Rusydi. 2022. “Identification of Groundwater Potential Zone Using Multi-Influence Factor Technique (Study Case: Brantas Groundwater Basin, East Java, Indonesia).” IOP Conference Series: Earth and Environmental Science 1066, 012004. https://doi.org/10.1088/1755-1315/1066/1/012004
44. Melese, T. and T. Belay. 2022. “Groundwater Potential Zone Mapping Using Analytical Hierarchy Process and GIS in Muga Watershed, Abay Basin, Ethiopia.” Global Challenges 6 (1): 2100068. https://doi.org/10.1002/gch2.202100068
45. Muavhi, N., K.H. Thamaga, and M.I. Mutoti. 2022. “Mapping Potential Groundwater Zones Using Relative Frequency Ratio, Analytic Hierarchy Process and Their Hybrid Models: Case of Nzhelele-Makhado Area in South Africa.” Geocarto International 37 (21): 6311–6330. https://doi.org/10.1080/10106049.2021.1936212
46. Narayanamurthi, V. and A. Ramasamy. 2022. “Groundwater Potential Zoning by Integrating Multi-Criteria Decision and Bivariate Analysis Methods – a Case Study on Cheyyar River Basin, Tamil Nadu, India.” Geocarto International 37 (26): 14570–14595. https://doi.org/10.1080/10106049.2022.2088864
47. Narmada K., K. Gobinath, and G. Bhaskaran. 2015. “An Assessment of Resource Potentials for Sustainable Development of Micro-Watershed in Tirunelveli District Using Geoinformatics– A Case of Nambiyar Micro-Watershed in Tirunelveli District, Tamil Nadu, India.” Aquatic Procedia 4, 1299–1306. https://doi.org/10.1016/j.aqpro.2015.02.169
48. Nayak, T.R., S.K. Gupta, and R. Galkate. 2015. “GIS Based Mapping of Groundwater Fluctuations in Bina Basin,” Aquatic Procedia 4, 1469–76. https://doi.org/10.1016/j.aqpro.2015.02.190
49. Olajide, A., A.A. Akinlalu, and G.O. Omosuyi. 2022. “Application of GIS-Based Frequency Ratio Model to Geoelectric Parameters for Groundwater Potential Zonation in a Basement Complex Terrain.” Indonesian Journal of Earth Sciences 2 (1): 16–32. https://doi.org/10.52562/injoes.v2i1.295
50. Patle, D. and M.K. Awasthi. 2019. “Groundwater Potential Zoning in Tikamgarh District of Bundelkhand Using Remote Sensing and GIS.” International Journal of Agriculture Environment and Biotechnology 12 (4): 311–318. https://doi.org/10.30954/0974-1712.12.2019.3
51. Razandi, Y., H.R. Pourghasemi, N.S. Neisani, and O. Rahmati. 2015. “Application of Analytical Hierarchy Process, Frequency Ratio, and Certainty Factor Models for Groundwater Potential Mapping Using GIS.” Earth Science Informatics 8, 867–83. https://doi.org/10.1007/s12145-015-0220-8
52. Shastri, S., P. Singh, P. Verma, P.K. Rai, A.P. Singh. 2020. “Assessment of Spatial Changes of Land Use/Land Cover Dynamics, Using Multi-Temporal Landsat Data in Dadri Block, Gautam Buddh Nagar, India.” Forum Geografic XIX (1): 72–77. https://doi.org/10.5775/fg.2020.063.i
53. Sishodia, R.P., S. Shukla, W.D. Graham, S.P. Wani, and K.K. Garg. 2016. “Bi-decadal Groundwater Level Trends in a Semi-Arid South Indian Region: Declines, Causes and Management.” Journal of Hydrology: Regional Studies 8, 43–58. https://doi.org/10.1016/j.ejrh.2016.09.005
54. Sorensen, R., U. Zinko, and J. Seibert. 2006. “On the Calculation of the Topographic Wetness Index: Evaluation of Different Methods Based on Field Observations.” Hydrology and Earth System Sciences. 10 (1): 101–112. https://doi.org/10.5194/hess-10-101-2006
55. Sresto, M.A., S. Siddika, M.N. Haque, and M. Saroar. 2021. “Application of Fuzzy Analytic Hierarchy Process and Geospatial Technology to Identify Groundwater Potential Zones in North-West Region of Bangladesh.” Environmental Challenges 5, 100214. https://doi.org/10.1016/j.envc.2021.100214
56. Subramanya, K. 2008. Engineering Hydrology, Third Edition. Tata McGraw-Hill Publishers, Uttar Pradesh, India. pp. 152–159.
57. Thapa, R., S. Gupta, S. Guin, and H. Kaur. 2017. “Assessment of Groundwater Potential Zones using Multi-Influencing Factor (MIF) and GIS: A Case Study from Birbhum District, West Bengal.” Applied Water Science 7, 4117–31. https://doi.org/10.1007/s13201-017-0571-z
58. Todd, D.K. and L.W. Mays. 2005. Groundwater Hydrology, 3rd Edition. John Wiley and Sons, Inc., Hoboken, NJ. p. 652. https://www.academia.edu/1374102/Groundwater_Hydrology_by_D_K_Todd
59. Trabelsi, F., S. Lee, S. Khlifi, and A. Arfaoui. 2019. “Frequency Ratio Model for Mapping Potential groundwater zones Using GIS and Remote Sensing; Medjerda Watershed Tunisia.” In: Advances in Sustainable and Environmental Hydrology, Hydrogeology, Hydrochemistry and Water Resources, ed. H. Chaminé et al. CAJG 2018. Advances in Science, Technology & Innovation. Springer, Cham. pp. 341–45. https://doi.org/10.1007/978-3-030-01572-5_80
60. Vishwakarma, C.A., V. Rena, D. Singh, S. Mukherjee. 2021. “Exploration of Water Resources using Remote Sensing and Geographic Information System.” In: Groundwater Geochemistry, eds. S. Madhav and P. Singh. John Wiley & Sons, Inc. p. 364–78 https://doi.org/10.1002/9781119709732.ch18