Introduction

Groundwater is a dynamic resource that plays a vital role in nurturing an ecological balance (Das 2017). The largest freshwater resource in the world is groundwater. Today, the maximum rural population and the total urban population depend on groundwater for their basic needs. Proper management strategies cannot be developed in the context of the groundwater decay scenario due to lack of knowledge in this area. Maximum groundwater is needed for agriculture, drinking water and industry. The absence of groundwater is a common phenomenon in India due to the use of groundwater without proper scientific planning (Rodell et al. 2009). The most valuable resource is groundwater which is unevenly distributed. Distribution of groundwater is controlled according to many criteria, i.e., lithology, geological structure, lineaments, soil, slope, drainage density, distance from the river, vegetation cover, surface water scenario, land use and land cover, rainfall, climatic condition and interrelationship between these factors (Sar et al. 2015; Jothibasu and Anbazhagan2016; Dasho et al.2017).

Six main criteria were used to analyze the groundwater potential of Purba Bardhaman, i.e., soil, land use and land cover, lithology, surface water scenario, water distance and rainfall. In this plain region, they are the most effective. Remote sensing (RS) and GIS techniques (Krishnamurthy et al. 1996; Pothiraj and Rajagopalan 2013; Sar et al. 2015; Anbazhagan and Jothibasu 2016; Jothibasu and Anbazhagan 2017; Choudhari et al. 2018; Das and Pal 2019a, b) are very much useful to delineate the groundwater potential zone. Many researchers in India and aboard have been successfully using this technique to identify groundwater potential zone. Thematic layers have been made through Arc GIS software. Recently, the application of multi-criteria decision making (MCDM) technique in GIS framework plays an important role in scientific investigation in groundwater resource management (Chen et al. 2011). The other areas where MCMD technique and GIS both play a vital role, are natural hazard (Ozturk and Batuk 2011), agriculture (Cisneros et al. 2011) and forestry (Zeng et al. 2007). Analytic hierarchy process (AHP) by Saaty of MCMD is playing a vital role to solve socioeconomic decision-making problems (Jhariya et al. 2016; Pal et al. 2019; Das and Pal 2020). In the AHP model, priority is given not only to the single element, but also to the cluster of elements that are often needed (Saaty 1999). In this paper, the AHP model was successfully used to assign relative importance to parameters that elucidate potential groundwater areas (Machiwal et al. 2015; Jhariya et al. 2016; Dasho et al. 2017; Das et al. 2017; Chakrabortty et al. 2018; Das et al. 2019). The main aim of this work is to identify the groundwater potential area of Purba Bardhaman district through the application AHP method and GIS technique. This study may be helpful in future planning for sustainable groundwater management which may help the people from water crisis.

Study area

The Purba Bardhaman district consisted of an alluvial plain. The study area is enclosed between 22° 15′ 08″ N to 23° 15′ 17″ N latitudes and 87° 13′ 17 E to 88° 7′ 22″ E longitudes (Fig. 1). In this district, the three major rivers are Ajay, Bhagirathi and Damodar, which are situated on the north, east and south side of the district. The undulating lateritic topography is found in the Ausgram region, which is part of the Paschim Bardhaman district. The study area is part of the tropical region. According to the 2011 census, the total population of the district was 48,35,532. Agriculture with a limited number of agricultural industries is a main occupation in this region. In this analysis, there are 11 community blocks, i.e., Ketugram-I, Ketugram-II, Bhatar, Raina-I, Raina-II, Kalna-II, Mongalkote, Katwa-II, Katwa-I, Memari-II, Monteswar were identified as semi-critical blocks (CGWB 2013) out of 23 community blocks, and there is a long-term trend in water level fluctuations (CGWB 2016). The central part of this study area received sufficient amount of rainfall which is categorized as high for this study region, i.e., above 1400 mm. The maximum part of this study region encompases with the distribution of 1200 to 1400 mm rainfall, whereas rest of the region has received below 1200 mm rainfall per year.

Fig. 1
figure 1

Location map

Full size image

Database and methodology

Various types of data are presented here in the tabulated form which has been used for preparing the groundwater potential zone in Purba Bardhaman district. The analyses have been started after the collection of all the required datasets. Numerous types of data (Table 1) and software like ArcGIS 10.3 and ERDAS IMAGINE 2014 have been used in this work. In this work, six criteria, i.e., lithology, soil texture, rainfall, distance from river, surface water body and land use land cover, have been taken into consideration to prepare groundwater potential zone. All the maps have been transformed into Universal Transverse Mercator (UTM) projection northern zone 45 datum WGS 1984 in ArcGIS software. Four thematic layers such as lithology, rainfall, distance from the river and soil texture have been digitized and rectified from various sources which are already mentioned in Table 1. We used the ERDAS 9.0 software for satellite image processing. The maximum likelihood classification method of supervised classification was applied for land use land cover map using the same Landsat 8 OLI satellite imagery of 2019 and software. The surface water body has been extracted using the normalized difference water index (NDWI) method from the same satellite image in ArcGIS 10.3 software. There are various water extraction methods, and one of them is the NDWI method which is used in this paper. Equation of NDWI is as follows:

Table 1 Sources of data
Full size table
$${\text{NDWI}} = \left( {{\text{Green}} - {\text{NIR}}} \right) / \left( {{\text{Green }} + {\text{NIR}}} \right)$$

After preparation of all thematic layers, the AHP method has been applied. Regarding this, the spatial distributions of all geo-environmental variables are from different data sources. AHP is a semiquantitative approach. In the AHP model, score is representing the importance of every individual factor (Saaty 1977, 2000). The preference values are applied for each factor following the importance scale of AHP for the calculation of relative importance in association with the goal (Saaty 1977; Saaty and Vargas 2001).

Generation of groundwater potential zone

Variable factor weights (Vi) and individual factor weights (Fi) for each thematic layers have been assigned to delineate groundwater potential zone. The spatial amalgamation of all map layers has been done using the raster calculator tool of ArcGIS 10.3 software version. Finally, a linear sum combination method has been adopted for getting the groundwater potential zone in RS-GIS environment.

$$\begin{aligned} {\text{GPZ}} & = \left[ {\left( {\text{Liw*Liwi}} \right) + \left( {\text{STw*STwi}} \right) + \left( {\text{DfRw*DfRwi}} \right) } \right. \\ & \quad \left. { + \; \left( {\text{LULCw*LULCwi}} \right) + \left( {\text{SWw*SWwi}} \right) + \left( {\text{Rfw*RFwi}} \right)} \right] \\ \end{aligned}$$

where GPZ = groundwater potential zone, Li = lithology, ST = soil texture, DfR = distance from river, LULC = land use land cover, SW = surface water bodies, Rf = rainfall, the subscripts w and wi refer to the normalized weight of a theme and normalized weight of individual features of a theme, respectively.

Thereafter, the groundwater potential zone map has been validated by the CGWB bore wells data with the help of the interpolation method (Inverse distance weighted (IDW)) in the ArcGIS environment (Fig. 2).

Fig. 2
figure 2

Flowchart of methodology

Full size image

Results and discussion

Lithological unit

Lithology acts as an important role in groundwater potential because the permeability of the rocks directly influences infiltration. The lithological unit determines the movement and prosperity of groundwater (Jhariya et al. 2016). In this study area, four types of lithological unit are found, namely laterite, clay with caliche concretion, clay alternating with silt and sand, sand-silt–clay (Fig. 3). Mainly, 53.36% of the total area in Purba Bardhaman district clay with caliche concretion is higher than another lithological unit. Three units of lithology, i.e., Q1S, Q2K, Q2D, have been made in Pleistocene to Holocene epoch and L units in Cainozoic epoch (Table 2).

Fig. 3
figure 3

Lithological map

Full size image
Table 2 Details of the lithological unit in geological framework
Full size table

Soil texture

The soil potential depends heavily on the soil texture as the soil texture directly affects the infiltration and aquifer conditions. On the basis of porosity and permeability, the infiltration capacity of coarse-grained soil is high compared to fine-grained soil. In this study area, twelve types (Fig. 4) of soil are found which are derived from weathering of laterite soil and transportation of silt soil by the river. Soil types with their percentage of the area are shown in Table 3.

Fig. 4
figure 4

Soil texture map

Full size image
Table 3 Details of soil texture with the percentage of area
Full size table

Mainly, three major types of soil in this region are sandy loam, clayey loam and silty loam. Among those types of soil, sandy loamy soil texture is more permeable than others.

Distances from river

A positive relationship has been found with river and groundwater. Rivers have multifarious interaction with groundwater. Subsurface water flow occurs from river to groundwater zone. This is an uninterrupted flow which increases the groundwater zone. When the groundwater enters the river, it is called base flow. We found a reverse relationship with the river and distance from it for the subsurface and surface flow of the river to the groundwater zone. So groundwater recharge decayed with the increase in distance from the river (Fig. 5).

Fig. 5
figure 5

Buffer zone from River

Full size image

Land use and land cover

Land use and land cover (LULC) is one of the vital factors which directly affect the development of groundwater recharge. Different types of land use act as differently in the runoff, infiltration and groundwater recharge. Generally, forest cover and agricultural land are most suitable for groundwater recharge. On the other hand, the built-up area is not suitable for groundwater recharge. LULC map has been prepared from mosaics of OLI-TRIS Landsat image. In this study area, LULC map has been classified into major five classes, i.e., agricultural land (36.8%), vegetation (19.1%), built-up area (18.97%), fallow or others types of land (17.89%), and water body (7.24) as shown in Fig. 6. Majority of land in this district is under the cover of agriculture and vegetation. Agricultural activities rely on monsoon rainfall, which is why most agricultural land remains vacant during the cold season. The map shows only a few urban settlements and large rural settlements (Fig. 6). In the LULC classification, water bodies, farm property, woodland, other property and built-up areas shall be given the highest to the lowest weight as shown in Table 4.

Fig. 6
figure 6

Land use land cover map

Full size image
Table 4 Pair-wise comparison matrix for AHP and priority and rank of themes
Full size table

Surface water body

There is a complex interconnection between surface water and groundwater. Exchange of surface water and subsurface water is a part of the hydrosphere within the land. This interaction is one of the processes of the natural hydrologic cycle. The interaction between surface water and groundwater is determined by a complex combination of geological, hydrological, geomorphological, climates and landscape factor. Surface water means water reservoirs of this earth such as lakes, swamps, ponds, wetlands, stream, etc. If the permeable rock is present below the surface water penetration stage, it plays a vital role in the recharge of groundwater. The surface water thematic map of the Landsat OLI-TRIS image using the NDWI method has been prepared in this study area as shown in Fig. 7.

Fig. 7
figure 7

Surface water map

Full size image

Rainfall

Rainfall plays a major role in the management of subsurface and surface water sources. Several studies have been carried out using a particular method to calculate the value of rainfall in groundwater variability. The amount of rainfall in this study region is not that high. The precipitation map is divided into three parts shown in Fig. 8.

Fig. 8
figure 8

Rainfall distribution map

Full size image

Groundwater potential zone

Determination of weighting

At first, on the basis of six major criteria, pair-wise matrix has been done with the help of the AHP method to recognize the priority and rank of the themes. Here, pair-wise matrix m × m is a real matrix, and m represents the number of criteria which is considered for evaluation. In this matrix, 1–9 scale has been used. Every criterion is given preference according to Table 4. The comparison matrix of the criterion can be set by this equation, i.e., ajk * akj = 1. The consistency ratio is calculated on the basis of the equation

$${\text{CR}} = {\text{C}}I/{\text{RI}}$$

where CI = consistency index and RI is the random index. Addition or elimination of a feature in the study depends upon the value of consistency ratio. The value of CR should always keep lower than or equal to 0.1 (Saaty 1990). CI reflects the consistency of one’s judgment. CI can be calculated by the following equation

$${\text{CI}} = \lambda \max - n/n$$

where n = order of the matrix, λmax is the largest eigenvalue. CR of each and every individual feature is observed for the inclusion or exclusion in the study.

Calculated priority, rank and AHP weight of individual are presented in Table 4 with the help of above-mentioned procedure and consistency ratio is 2.3% which is below 10% and accepted according to Saaty (1994). The weightage of each class of every parameter has been calculated and presented in Table 5.

Table 5 Weightage of different parameters for groundwater potential zones mapping
Full size table

The groundwater potential zone map is classified in five distinct zones representing very low potential, low potential, medium potential, high potential and very high potential which consist of area of 21.54%, 35.8%, 26.47%, 10.13% and 6.06% of total area. In this district, very high potential and high potential zone areas are found in the side of river Damodar, Ganga and Ajay and dense vegetation areas of the west part of this district, shown in Fig. 9.

Fig. 9
figure 9

Groundwater potential zone map

Full size image

Validation

The groundwater potential zone area was verified with the help of 12 dug well station and 32 tube well data of 44 stations that were collected from CGWB-2018. These 44 stations are distributed throughout the district shown in Fig. 10. Details of tube well location and actual yield are shown in Table 6.

Fig. 10
figure 10

Groundwater validation point

Full size image
Table 6 Detail account of the validation points
Full size table

Table 6 expresses similarity between actual class and expected class. A remark of agreement and disagreement is also prepared in Table 6. In this study area, the total number of tube well and dug well is 44. The number of tube well and dug well under the agreement of coherence between actual and predicted yield range is 39, and the number of tube well and dug well under disagreement of coherence between actual and predicted yield range is 5.

\(\begin{aligned} {\text{The}}\;{\text{equation}}\;{\text{of}}\;{\text{accuracy}}\;{\text{prediction}} & = {\text{No}} .\;{\text{of}}\;{\text{tube}}\;{\text{well}}\;{\text{and}}\;{\text{dug}}\;{\text{well}}\;{\text{under}}\;{\text{the}}\;{\text{agreement}}\;{\text{of}}\;{\text{coherence}} \\ & \quad /{\text{Total}}\;{\text{number}}\;{\text{of}}\;{\text{tube}}\;{\text{well}}\;{\text{and}}\;{\text{dug}}\;{\text{well}}\;{\text{surveyed}} \times 100 \\ & = 39\text{ / }44 \times 100 \\ & = 88.64\% \\ \end{aligned}\) The prediction accuracy value of 88.64 reflects the fact that the AHP model and GIS techniques used in this work are significantly robust and accurate. For validation of this work, groundwater depth zone map (Fig. 11) prepared using the IDW method in ArcGIS is applied and it is compared with the potential zone map (Fig. 10). In order to demonstrate the accuracy, the receiver operating characteristic (ROC) curve was used in this analysis, which compares the tube well data with the groundwater potential area. The ROC curve represents (Fig. 12) that the area under curve (AUC) is 0.868 which means 86.8%. So this ROC result also accepts the AHP model in this work.

Fig. 11
figure 11

Ground water depth zone

Full size image
Fig. 12
figure 12

Receiver operating characteristic (ROC) curve

Full size image

Conclusion

Groundwater is the world’s greatest resource. Groundwater contamination is a big problem for our future climate. To this end, the purpose of this paper is to explore potential groundwater zones using AHP and GIS techniques in the Purba Bardhaman district. In this analysis, six main parameters were taken into account to determine the potential groundwater zone. Low and very low potential zones account for 57.34% of the total area, although the Purba Bardhaman district is a deltaic region. The rest of 43.38% area was covered with medium to very high groundwater potentiality zones. It is obvious that the porous lithological setting, permeable soil texture, location of surface water bodied coupled with rainfall distribution and vegetative cover result in the formation of a high potential area. The central part of the district is mainly characterized by a low and very low field. These areas have low strength, irrespective of the distance from the sea. This work is much more important in the age of drinking water crisis, as this final map would have been useful in the planning and management of water resources. This work has been validated with the results of the 44 tube well station. The empirical study IDW method is used to test and show the precision of the receiver operating characteristic (ROC) curve. The method used is reliable and can be used successfully elsewhere with the required modifications. This study would enable the decision-makers concerned to establish an appropriate study policy.