Introduction

Groundwater or subsurface water is very crucial for sustaining life. It is the largest freshwater reservoir in the world (Makonyo and Msabi 2021). However, it is a renewable and partially non-renewable. They are sometime confined and others unconfined, in addition to perched or artesian (Chatterjee and Purohit Ram 2009). Globally, groundwater is essential to the hydrological water cycle and to methods of water management. Nearly half of all drinking water, a third of industrial water and almost 40% of agricultural water come from underground sources (Lall et al. 2020). However, the availability or distribution of groundwater is unequally distributed over the earth due to local topography, hydrology and human activities. The disappearance of local groundwater resources can have a significant impact on the agricultural value chain, leading to large-scale migrations and economic collapse. India (251 km3/year) had the most groundwater extraction in 2010, subsequently China, the USA, Pakistan, and Iran (63.4 km3/year), which seems to be increased in future (Abdo 2020; Rahaman et al. 2021). Rapid urbanization and global climate change make groundwater an exhausted resource. Identification of groundwater in the hilly mountains is very difficult due to its high terrain (Barman and Biswas 2022). The traditional method of groundwater exploration (drilling) is very time-consuming and manpower intensive. The development of remote sensing and GIS (geographic information system) becomes a suitable tool for groundwater availability mapping by integrating topographic, geological, hydrologic and climatic factors. Recently, statistical data-driven methods like weights of evidence (Al-abadi 2015; Boughariou et al. 2021), frequency ratio (Elmahdy and Mohamed 2014; Doke et al. 2020), evidential belief function (Mogaji and Lim 2016), logistic regression (Ozdemir 2011), artificial neural network model (Profile 2016; Lee et al. 2017), Shannon’s entropy (Khoshtinat et al. 2019) are useful for groundwater availability mapping worldwide.

The identification of groundwater influencing factors in a hilly area is very crucial because of the hidden nature beneath of the ground. Several GIS intergraded models incorporating slope, elevation, aspect, distance to drainage, LULC, rainfall, distance to lineament (Pinto et al. 2017; Namous et al. 2021; Senapati and Das 2021; Rahaman et al. 2022). However, total number of thematic layers for groundwater potentiality may vary place to place based on the topographic, geological and hydrological properties (Roy et al. 2024). Previous study in the Champhai district by (Lalbiakmawia and Kumar 2017) used five thematic factors like lithology, land use land cover (LULC), slope, lineament distances and geomorphology. (Barman et al. 2023b) used a multi-criteria decision-making including slope, aspect, LULC, elevation, rainfall, river distances, precipitation concentration index (PCI), topographic wetness index (TWI), normalized difference vegetation index (NDVI), normalized difference water index (NDWI), geology and LULC. Another study by (Lalbiakmawia 2015) used GIS-based groundwater potential zonation using lithology, land use land cover (LULC), slope, lineament distances and geomorphology in Aizawl district of Mizoram. The objective of the current study is to enhance groundwater potential zonation by incorporating significant groundwater influencing factors, namely elevation, dissection index, slope, aspect, TWI, infiltration number, distance to rivers, rainfall, distance to lineaments and LULC. However, the prediction accuracy of a model may vary due to selection of appropriate conditioning factors (Sutradhar et al. 2023). Therefore, in the study, weights of the factors and domains were estimated using the analytical hierarchy process. After that judgment of the factor’s weights were estimated using single-layer sensitivity analysis and map removal sensitivity analysis. Through the single-layer sensitivity analysis and map removal sensitivity analysis the current study tried to find out most crucial factors for groundwater potentiality mapping based on their role in the study region.

The analytical hierarchy process (AHP) is a knowledge-based semi-quantitative method widely used for natural resource management (Saaty 2003). The significant advantage of the AHP method is the weightage for both factors and domains derived through a pairwise comparison matrix under a considerable consistency ratio (Abdo et al. 2023; Rahaman et al. 2022). Previously, the AHP method was used by a number of scholars (Jothibasu and Anbazhagan 2016; Arefin 2020; Shao et al. 2020; Senapati and Das 2021) for groundwater potentiality delineation, globally. Their result shows a high prediction accuracy as AHP based on the expert opinion and weight for the factors as well as the domains.

The northeaster state like Mizoram entirely covered with hilly terrain. Although rainfall is the major source of groundwater, most of the water is drained as spring water through fractures and joints (Barman and Biswas 2022; Barman et al. 2023c). It was estimated during the monsoons the rate of discharge varies from 0.002 to 0.168 m3/s, while it varies between 0.001 m3/s and 0.084 m3/s in the non-monsoon periods from the springs (Central Ground Water Board 2021). According to an abstract report of (Ground Water Resources Assessment Cell 2020) surface water in the state was found to be corresponding to depths between 8.80 and 13.25 m below ground level. The report also mentioned the groundwater level is declining at a pace of around 0.50 m per decade. The district Champhai is locally known as the ‘rice bowl of Mizoram.’ The district practices their own traditional jhum as well as the wet rice in the Champhai Valley facing water scarcity after the monsoon every year. According to (Central Ground Water Board 2021), the current groundwater extraction in the Champhai city area is 291,800 m3/year and the annual use of groundwater extraction for domestic use was limited to 320,600 m3/year by 2025. No such micro-level work has been published from the study area. Keeping the importance of groundwater potentiality mapping, various conditioning factors were assigned weightage by AHP methods and intergrade by weighted overlay.

Materials and methodology

About study area

Champhai city is the district headquarters of Champhai district which is located middle of the district and eastern part of Mizoram (Fig. 1). The region comes under the Indo-Burma biodiversity hotspot zone. Geographically, the township ranges between 93° 15′E and 93° 24′E longitude and 23° 32′45ʺN and 23° 25′15ʺN latitude. The real extension of the study area is 133.15 sq. km. The township holds a total of 32,734 population with various ethnic diversity (Census of India 2011). The maximum rainfall caused by southwest monsoon ranges between 2226.84 mm/year and 2364.8 mm/year. Although the area received a good amount of precipitation, most of that washed out as surface runoff and spring water (Rao et al. 2023). The significant geomorphological feature ‘The Champhai valley’ which is known as the rice bowl of Mizoram is located in the area. Geologically the study area formed with Jenam, Laisong, and Tikak Parbat formation (Fig. 2). The Jenam formation consists of olive green, splintery shale with minor sandstone; the Laisong formation consists of sandstone, siltstone and shale, and the Laisong formation consists of sandstone, siltstone and shale with conglomerate. The tuipui, tupul and tyao are the only perennial rivers flowing through the study area. As high runoff and no rain in the winter a noticeable water scarcity is reported every year.

Fig. 1
figure 1

Location of the study area, a Mizoram in India, b Champhai district in Mizoram, c Champhai city area within the district, d Champhai city area with spatial pattern of land use land cover

Full size image
Fig. 2
figure 2

(Source Geological society of India)

Geology map of the study area

Full size image

Database and preparation of thematic layers

The largest freshwater storage ‘groundwater’ is available in a limited environment in hilly areas (Barman et al. 2023c). The movement and storage of groundwater is depended on different topographical, hydrological, climatic and other factors. In the current study, the preparation of topographical factors was done using the ASTER digital elevation model (DEM), and hydrological factor including the drainage network was exported from ASTER DEM through threshold value ‘Flow accumulation > 300.’ In case of climatic factors average annual rainfall of 2011 to 2020 by Climate Research Unit (cru) data was used. To gather the tectonic information like ‘lineament’ by Bhukosh was used. Further, land use land cover (LULC) by ESRI 2021 was used to know the human activities over nature (Table 1). The organization of the study was generalized by methodological flow chart in Fig. 3.

Table 1 Description of database
Full size table
Fig. 3
figure 3

Methodological flow chart

Full size image

The elevation is the primary groundwater potentiality controlling factor in a hilly environment (Rahaman et al. 2022). There is a revised relationship between groundwater potentiality and elevation, e.g., groundwater potentiality is maximum in the lowest elevation (Pinto et al. 2017; Namous et al. 2021; Sud et al. 2023). The elevation in the study area was classified into five classes based on the natural breaking classification system (a) 726–986 m, (b) 986.01–1171 m, (c) 1171.01–1323 m, (d) 323.01–1444 m and (e) 1444.01–1672 m. Subsequently, the dissection index (DI) is the ratio of relative relief and absolute relief. The DI ranges between 0 and 1 where 0 denoted no dissected landscape (flat topography) and 1 denoted highly dissected landscape (clip) topography (Nair et al. 2017). The lower value of DI shows minimum groundwater potentiality, while higher DI values show maximum chances of groundwater potentiality (Barman and Biswas 2022). The dissected index was classified into five classes (a) < 0.05, (b) 0.06–0.1, (c) 0.11–0.15, (d) 0.16–0.2 and (e) > 0.20. The slope of a region reflects the inclination or declination of the elevation of two points with respect to horizontal distance (Barman et al. 2023a). There is also a revised relationship between groundwater potentiality and slope value, i.e., chances of groundwater are decreasing with increasing the slope value (Anbarasu et al. 2020). The slope map was classified into five classes based on natural breaking classification as (a) 0–8.74°, (b) 8.75–16.8°, (c) 16.81–24.42°, (d) 24.43–32.93° and (e) 32.94–57.12°. The aspect indicated the direction of slope faces (Biswas et al. 2023). It is dominant controlling factor of vegetation types, solar radiation and precipitation (Barman and Das 2023). The aspect of the study area divided into ten classes like (a) − 1, (b) 0–22.5, (c) 22.5–67.5, (d) 67.5–112.5, (e) 112.5–157.5, (f) 157.5–202.5, (g) 202.5–247.5, (h) 247.5–292.5, (i) 292.5–337.5 and (j) 337.5–360. The TWI is the secondary morphometric index which denoted the hydrological control on topography (Razandi et al. 2015). It has been widely used for groundwater potentiality mapping (Arulbalaji et al. 2019; Ahmad et al. 2020) and representing the spatial flow accumulation. It can be defined as Eq. (1).

$${\text{TWI}} = {\text{ln}}\left( {\frac{{A_s}}{{{\text{Tan}}\;\beta }}} \right)$$
(1)

where \({\text{TWI}}\) is the topographic wetness index, \({A_s}\) is the upslope contributing area, and \(\beta\) is the slope gradient. In the study, TWI was classified into five classes based on the natural breaking classification (a) 2.97–5.25, (b) 5.26–6.71, (c) 6.72–8.86, (d) 6.72–8.86 and (e) 12.04–19.12. The infiltration number is the multiply function of drainage density and stream frequency (Ghosh and Jana 2017). The infiltration number of the study area is classified into five classes (a) 0–121.63, (b) 121.64–334.47, (c) 334.48–623.33, (d) 623.34–1033.82 and (e) 1033.83–1938.42. For the distance to drainage (DTS), the potentiality of groundwater decreases with increasing distance from streams (Rahaman et al. 2022). The DTS map was produced using the Euclidean method in a GIS environment and classified into five classes (a) 0–150 m, (b) 150.01–300 m, (c) 300.01–450, (d) 450.01–600 m and (e) > 600 m. Similarly, rainfall is a significant influencing factor for groundwater (Barman et al. 2023b). The rainfall pattern of the study area was classified into five classes (a) 2226.84–2267.96 mm, (b) 2267.97–2295.01 mm, (c) 2295.02–2320.44 mm, (d) 2320.45–2343.16 mm and (e) 2343.17–2364.8 mm. Again, the lineament distance map was produced by using the Euclidean distance method in a GIS environment and classified into five classes (a) < 300 m, (b) 300.01–600 m, (c) 600.01–900 m, (d) 900.01–1200 m and (e) > 1200 m. For groundwater potentiality land use land cover plays a significant role (Rahmati et al. 2015; Makonyo and Msabi 2021). The LULC map was classified into four major classes, namely water body, vegetation, buildup area and shrub. The thematic view of the groundwater potentiality factors is figured in Fig. 4.

Fig. 4
figure 4figure 4figure 4figure 4

Spatial distribution of subsurface water potentiality factors, a elevation, b dissection index, c slope, d aspect, e TWI (topographic wetness index), f infiltration number, g DTS (distance to streams), h rainfall, i DTL (distance to lineaments) and j LULC (land use land cover)

Full size image

Analytical hierarchy process (AHP)

The AHP is a subjective semi-quantitative approach introduced by (Saaty and Hu 1998). It is a flexible way to determine the relative weightage of a particular output (Senapati and Das 2021). The significant advantage of the approach is the power of inconsistency reduction and estimation of priority vector by pairwise comparison and decomposition (Jothibasu and Anbazhagan 2016). A pairwise comparison matrix was formed as Eq. (2) which is the primary need of the AHP method. In a town planning for groundwater potentiality zonation, ten factors (elevation, dissection index, slope, aspect, TWI, infiltration number, distance to drainage, rainfall, distance to lineament and LULC) were assigned 1 to 9 scale rating (Table 2 and 3) where 1 denotes equal importance and 9 denotes absolute significance (Table 4).

$$A = {\left\{ {{a_{ij}}} \right\}_{n \times n}} = \left[ {\begin{array}{*{20}{c}} 1&{{a_{12}}}& \ldots &{{a_{1n}}} \\ {{a_{21}}}&1& \ldots &{{a_{2n}}} \\ \vdots & \vdots & \ldots & \vdots \\ {{a_{n1}}}&{{a_{n2}}}& \ldots &1 \end{array}} \right] = \left[ {\begin{array}{*{20}{c}} 1&{\frac{{w_1}}{{w_2}} \ldots }&{\frac{{w_1}}{{w_n}}} \\ {w_2}&1&{\frac{{w_2}}{{w_1}}} \\ \vdots & \vdots &{w_n} \\ {\frac{{w_n}}{{w_1}}}&{\frac{{w_n}}{{w_2}}}& \vdots \\ {w_1}&1&{} \end{array}} \right]$$
(2)
Table 2 Comparison matrix for the potentiality factors
Full size table
Table 3 Comparison matrix and weightage for the domains
Full size table
Table 4 Comparison scale after Saaty (1970)
Full size table

In the second step, the matrix was converted into a normalized matrix using Eq. (3) for minimize factor dimension. It is noticeable that the matrix value of a normalized matrix ranges between 0 and 1.

$${a_{i{j_{{\text{nor}}}}}} = \frac{{{a_{ij}}}}{{\mathop \sum \nolimits_{i = 1}^n {a_{ij}}}}\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;i,j = 1,\;2,\; \ldots \ldots n$$
(3)

In the third step, the priority vector which is considered as the factor’s relative weightage was calculated as Eq. (4).

$${w_i} = \left( \frac{1}{n} \right)\mathop \sum \limits_{j = 1}^n {a_{i{j_{{\text{nor}}}}}}\;\;\;\;\;\;\;i = 1,\;2,\; \ldots \ldots \ldots .n$$
(4)

The acceptability of the comparison matrix was estimated by the consistency index (Ci) (Shao et al. 2020). However, consistency ratio (CR) was used to calculate the consistency level. The CR is a ratio between the consistency index (Ci) and random index (Ri) (Eq. 5), where Ri was determined according to the order of the matrix (Table 5). A matrix should be accepted when the CR value is estimated below 0.1 (Biswas et al. 2023).

$${C_R} = \frac{{C_i}}{{R_i}}$$
(5)
Table 5 Random index value
Full size table

Results and discussion

Relationship between influencing factors and groundwater potentiality

The identification of groundwater potentiality zones is very difficult in hilly environments because of the high terrains (Barman et al. 2023c). In the hilly areas of Champhai town area the accessibility of groundwater is very limited due to the environmental condition (Fig. 5). The development of remote sensing and GIS opened a direction for such environmental conditions. In the present study, the relative weightage of ten (10) groundwater water-influencing conditioning factors was estimated through the AHP method. During the rating of the pairwise decision matrix, a subjective experiences and literature survey (Barman and Biswas 2022; Barman et al. 2023b) was done. The (Saaty 2005) rating was used to complete the decision matrix between the aspect and TWI; a rating of 3 was given to the aspect and 1/3 to the TWI because of aspect is more dominant than the TWI for groundwater storage. Based on the normalized weightage of the accepted decision matrix (CR < 0.1), the maximum and minimum weightage was given to the aspect and LULC, respectively, while others played intermediate roles. Further, for the elevation subclasses or domains, the maximum possibility of groundwater potentiality was found below 1,323. Among the five elevational subclasses, the maximum weightage was assigned to 726–986 m domains and minimum to the 1444.01–1672 m(Rahmati et al. 2015; Saranya and Saravanan 2020). For the dissection index, the maximum and minimum weights were assigned to < 0.05 and > 0.20, respectively, as the high dissection index showed a high rate of runoff (Nair et al. 2017). The flow in the gentle slope is a long duration and has larger time for infiltration, while in the steep slope, most of the rainwater flows as surface runoff; thus, the maximum and minimum weightages were given to 0–8.74 degrees and 32.94–57.12 degrees, respectively (Ahmadi et al. 2021). As the study area is located in the southern part of the northern hemisphere, maximum rainfall was done by the influence of the southwest monsoon; thus, the maximum weights were given to south, southwest and flat slope direction (Barman et al. 2023a). In the case of TWI, the maximum and minimum weights were assigned to 2.97–5.25 and 12.04–19.12, respectively, as the higher TWI value indicated more water accumulation (Razandi et al. 2015). Similarly, the maximum weightage was given to 1033.83–1938.42 and the minimum weightage to 0–121.63 for infiltration number as more infiltration denoted more groundwater storage. Regarding the DTS, a revised relationship was seen. Near to the stream has maximum chances to maximum infiltration rate and vice versa (Saranya and Saravanan 2020). So, the maximum and minimum weightages for the DTS’s domains were assigned to 0–150 m and > 600 m, respectively. In the hilly environment rainfall is the major source for groundwater storage; thus, maximum weightage assigned to 2343.17–2364.8 mm, and minimum weightage to 2226.84–2267.96 mm (Saranya and Saravanan 2020; Barman et al. 2023c). In case of DTL, a large quantity of groundwater wash out through lineament for the reason maximum weightage was assigned to larger distance (1200.01–6175.69 m) and minimum weightage to the shorter distance (0–300 m) (Akinwumiju et al. 2016). At least for the LULC, the maximum infiltration has had from water body followed by vegetation cover, buildup area and shrub area; thus, the maximum weightage was given to shrub areas (Senapati and Das 2021).

Fig. 5
figure 5

Field photographs

Full size image

Groundwater potentiality zones

An weighted sum analysis of ten (10) weighted rasters was attempted in order to produce a groundwater potentiality map as Eq. (6). The raster value range between 0.11and 0.29 denoted the upper limit and lower limit of groundwater potentiality.

$${W_{{\text{subsurface}}}} = {\text{Elevatio}}{{\text{n}}_{{\text{AHP}}}} + {\text{D}}{{\text{I}}_{{\text{AHP}}}} + \ldots \; + {\text{LUL}}{{\text{C}}_{{\text{AHP}}}}$$
(6)

To better understand the groundwater potentiality, the study area was zoned into five hierarchical zones based on the natural breaking classification system. The hierarchical zones are, namely very high potential, high potential, moderate potential, low potential and very low potential zones (Fig. 6). As per the model, only 5.08% of the study area had a very high potential for groundwater, whereas the maximum part of the study area came under low potential for groundwater (31.88%). The other zones, namely high potential, moderate potential and very low potential zones, covered 14.73%, 25.34% and 22.98% of the total area, respectively (Fig. 7). The geographical distribution of very high potential zones concentrated mainly along the rivers due to high infiltration, low elevation, high TWI value, low dissection and low slope (Roy et al. 2024). On the contradictory, very low potential zones were found mainly along the ridge due to high elevation and low infiltration (Arulbalaji et al. 2019; Rajesh and Jobin 2019). The low potential area was located mainly on a steep slope due to high runoff (Mukherjee et al. 2012). In the case of moderate and high potential zones, those were concentered 150–450 m from the rivers and between 1000- and 1500-m elevation.

Fig. 6
figure 6

Groundwater potentiality zones of the study area with distribution of wells where larger proportional circle indicated more potential

Full size image
Fig. 7
figure 7

Areal extension of groundwater potentiality zones

Full size image

Validation

A map is not useful without proper validation (Barman et al. 2023a; Biswas et al. 2023). The receiver operating characteristic curve (ROC) and area under curve (AUC) are the scientific validation techniques for groundwater potentiality mapping. The graphical plot of ROC consists of a ‘True Positive Rate’ in the ‘Y-axis’ and a ‘False Positive Rate’ in the ‘X-axis.’ The ‘True Positive Rate’ denoted the proportion of borehole points correctly classified as very and highly potential zones in the prediction map. The ‘False Positive Rate’ rate described the proportion of non-bore hole points correctly classified as low and very low potential zones in the prediction map (Barman et al. 2023a). In the current study, 25 borehole points by Gramin GPS were marked for validation purposes. An ArcGIS extension ‘ArcSDM’ was used for the preparation of the ROC curve (Fig. 8a). The AUC value of 0.737 indicated a good classified groundwater potentiality zone (Al-abadi 2015). As a spatial distribution, 48% (12 points) of the bole hole points fall into the high potentiality zone, 16% (4 points) fall in both very potentiality and moderate potentiality zones, 12% (3 points) fall in low potentiality zones, and only 8% (2 points) fall in very low potentiality zone. Figure 8b shows a scatter plot of groundwater potentiality, water extraction and well depth. In the “Z”-axis the groundwater potentiality was plotted ranging from 1 to 5, in which 1 denoted very low potentiality and 5 denoted very high potentiality. It also indicated that most of the boreholes are located in high and very high potentiality zones. Similarly, the “X-” and “Y”-axis indicated maximum boreholes had a depth of less than 210 ft and weekly groundwater extraction was more than 3500 Liters/week, respectively. It is highlighted that local residences benefit from groundwater in both the winter and summer seasons (Table 6).

Fig. 8
figure 8

Model’s prediction accuracy; a receiver operating characteristic curve and b 3D scatter plot of groundwater potentiality vs amount of groundwater extraction and well depth

Full size image
Table 6 Depth and water extraction from boreholes
Full size table

Sensitivity analysis

The subjective way of weightage selection of the parametric test is an undesirable part (Dadgar et al. 2017). That significantly impacts the model outcome. The sensitivity analysis is a scientific technique to estimate the importance of each influencing factor to groundwater potentiality (Awawdeh et al. 2013). Two well-known tests, namely single-layer sensitivity analysis and map removal sensitivity analysis, were performed for the current study. The single-layer sensitivity shows the effective weight of each influencing factor for the groundwater potentiality. Mathematically, the single-layer sensitivity can be calculated as Eq. (7).

$${\text{S}}{{\text{L}}_{{\text{sensi}}}} = \frac{{{M_{{\text{wi}}}}{R_{{\text{cl}}}}}}{{{W_{{\text{subsurface}}}}}}*100$$
(7)

where \({\text{S}}{{\text{L}}_{{\text{sensi}}}}\) is the single-layer sensitivity index, \({M_{{\text{wi}}}}\) is the empirical weight of factor (s), \({R_{{\text{cl}}}}\) is the rating of the factors, and \({W_{{\text{subsurface}}}}\) is the groundwater potentiality. Based on the sensitivity index, the maximum effective weightage was seen for the slope (mean variation 14.24%) followed by DTS, elevation, aspect and rainfall in the study area (Table 7). It was followed by TWI, dissection index, LULC, infiltration number and DTL (Fig. 9). The empirical weightage for DTS was estimated at 11%, whereas the estimated effective weight was seen at 13.62%. For the aspect, the empirical and effective weightage was calculated at 22% and 11.94%; hence, a big deviation was observed. The difference between empirical and effective weightage was less or more similar for the elevation, dissection index, slope, IN, TWI, DTS, rainfall, DTL and LULC (Table 7).

Table 7 Descriptive statistics of single-layer sensitivity analysis
Full size table
Fig. 9
figure 9

Effective weight of groundwater potentiality factors, a elevation, b dissection index, c slope, d aspect, e TWI (topographic wetness index), f infiltration number, g DTS (distance to streams), h rainfall, i DTL (distance to lineaments) and j LULC (land use land cover)

Full size image

The map removal sensitivity examines the importance of a factor for groundwater potentiality. It can be estimated as Eq. (8).

$${\text{M}}{{\text{R}}_{{\text{sensi}}}} = \frac{{\left| {\left( {\frac{{{W_{{\text{subsurface}}}}}}{N}} \right) - \left( {\frac{{{W_{{\text{subsurfac}}{{\text{e}}^\prime }}}}}{n}} \right)} \right|}}{{{W_{{\text{subsurface}}}}}}*100$$
(8)

where \({\text{M}}{{\text{R}}_{{\text{sensi}}}}\) denotes map removal sensitivity index, \(W_{{{\text{subsurface}}}}\) is the groundwater potentiality map including all factors, and \(W_{{{\text{subsurface}}^{\prime } }}\) is the groundwater potentiality map excluding factor. Further, \(N\) and \(n\) denote the number of factors for \({W_{{\text{subsurface}}}}\) and \({W_{{\text{subsurfac}}{{\text{e}}^\prime }}}\) modeling, respectively. In the current study, no significant changes were seen after removing a single factor. The TWI and aspect were found sensitive factors with the mean sensitivity index (SI) variation of 10.009% and 2.256, respectively. On the contradictory, LULC, rainfall, DTS and slope had less influence in the model (Table 8).

Table 8 Descriptive statistics of map removal sensitivity analysis
Full size table

The groundwater potentiality zone after map removed is figured in Fig. 10. The high potentiality area was increased by map removal after elevation, DI, slope, aspect, TWI, rainfall and LULC. The maximum change was observed after the removal of TWI followed by LULC, rainfall and DTS (Table 9). The area of high potentiality was decreased by the removal of IN and DTL. Similarly, the modeled estimated area of high potentiality was 19.48 Sq. Km, while the maximum area was increased after the removal of TWI followed by LULC, DI, DTS and aspect and the area was decreased by the removal of IN followed by slope, elevation, DTL and rainfall. The moderate potential area was increased by the removal of rainfall, slope and DI and decreased after removal of elevation, aspect, TWI, IN, DTL and LULC. Further, the low potential area was increased than the model estimated area after removal of slope, aspect, IN, DTL and elevation, DI, TWI, DTS, rainfall and LULC. The low potential area was decreased after the removal of DI, slope, TWI, DTS, rainfall and LULC, and the area was increased after the removal of elevation, aspect, In and DTL.

Fig. 10
figure 10

Groundwater potentiality zone after removed single factors, a removed elevation, b removed dissection index, c removed slope, d removed aspect, e removed TWI (topographic wetness index), f removed infiltration number, g removed DTS (distance to streams), h removed rainfall, i DTL (distance to lineaments) and j removed LULC (land use land cover)

Full size image
Table 9 Areal changes of potentiality zones after removed single layers
Full size table

Conclusion

The present study aimed to zonation of groundwater potential zonation of the district headquarters of Champhai district. In this regard, eight factors were assigned weighted using the subjective AHP method under a certain consistency ratio (less than 0.1). The maximum and minimum weightage were assigned to aspect and LULC, respectively. After all, ten weighted rasters were integrated with the ArcGIS environment to produce a groundwater potential map. To better understand, the generated raster groundwater map was classified into five categories very low potentiality, low potentiality, moderate potentiality, high potentiality and very high potentiality using a natural breaking classification system. As a spatial distribution, only 5.08% of the area had a very high potential for groundwater. The result showed that the river valley in the eastern part of the state was excellent for groundwater potentiality. The same was seen for the western part of the study area. The central part of the study area is known heart of the district; the main residential area was good for groundwater potentiality. The other parts of the study area characteristics were low and very low for groundwater potentiality mainly due to the presence of ridge. The AUC value of 0.737 indicated a good classified groundwater potentiality zonation. The sensitivity analysis of single-layer sensitivity highlighted effective weight of slope, elevation, aspect, DTS and rainfall was higher than 10%. Subsequently, the map removal sensitivity analysis showed that TWI and aspect have their maximum influence of 10.009% and 2.256%, respectively.