Application of multi-criteria decision making (MCDM) and

Application of multi-criteria decision making (MCDM)and electrical resistivity tomography (ERT) techniquesfor identiBcation of groundwater recharge zone(s)in granitic hard rock aquifer

ARCHANA KUMARI2, DEWASHISH KUMAR

1,* and PRATIBHA WARWADE2

1CSIR-National Geophysical Research Institute, Hyderabad 500 007, India.

2Central University of Jharkhand, Ranchi 835 205, India.*Corresponding author. e-mail: [email protected]

MS received 9 June 2020; revised 18 January 2021; accepted 19 January 2021

Groundwater recharge is a hydrological process where water Cows from sub-surface layers to the watertable of the aquifer and is the backbone of the hydrogeological system. The present study is carried out ina granitic hard rock aquifer region within and surroundings of the CSIR-NGRI campus, Hyderabad. Theaim of this study to identify the potential groundwater recharge zone(s) using GIS based multi-criteriadecision making (MCDM) along with sub-surface mapping from Electrical Resistivity Tomography(ERT) technique. The assigned weight of the different thematic layers of surface and sub-surfaceparameters and their speciBc characteristics was determined based on their relative contribution to thegroundwater recharge and thus the normalized weight was computed using MCDM technique. Thesethematic layers were integrated with the help of ArcGIS to accurately identify the recharge zones withinthe study region. The resulting recharge map has been categorized into Bve classes viz., very poor, poor,moderate, good and very good. Numerically 23.11% of the study area is in a moderate zone of recharge,4.97% in good and very good zone, while 71.92% falls under the poor and very poor zone, i.e., unsuit-able for groundwater recharge. The recharge zone map of the study area is found to be in agreement with2D inverted resistivity models for two different time periods, which revealed distinct geological featuresand thus identiBed the near surface recharge property, where recharge zone resistivity values lie between*20 and 70 X.m up to 11 m depth. Thus, the integrated results from the present study delineatedgroundwater recharge zone(s) for sustainable groundwater resources in the granitic hard rock system.

Keywords. Multi-criteria decision making (MCDM); electrical resistivity tomography (ERT);groundwater recharge; granite; Hyderabad; India.

1. Introduction

Groundwater is one of the most important com-ponents of a hydrological cycle and found below theearth surface in the cracks and the porous space ofthe soil, sediment and rocks (Al-Garni 2009; Das2019). It is the exhaustible and conBned natural

resources, which is being utilized for farming,industry and domestic purposes. The rapid growthof population and urbanization had enhanced thedemand for groundwater, which led to a globalwater shortage, especially in India, as India is thebiggest user of groundwater in the world that uti-lizes 260 km3 of water every year (Bhanja et al.

J. Earth Syst. Sci. (2021) 130:81 � Indian Academy of Scienceshttps://doi.org/10.1007/s12040-021-01577-3 (0123456789().,-volV)(0123456789().,-volV)

2017). The groundwater intrusion into the sub-surface strata relies largely on the size and shape ofthe weathered/fragmented layer found within thebasement (Kumar et al. 2014a, b) and its Cow canbe measured by the interaction of hydrological,geological, and atmospheric factors (Ghosh et al.2016). Aquifers in the basement rocks are highlylocalized and spatially variable (Satpathy andKanungo 1976; Das and Pardeshi 2018). Recentresearch had shown the increasing groundwaterlevels in South India due to rain and politicalinterventions (Hora et al. 2019). The groundwatertable behaviour is regulated by the intensity, spa-tial distribution and duration of the rainfall. Thedistribution of available groundwater resources inthe region varies unevenly and various fresh wateraquifers are restricted due to naturally contami-nated by the exceeding the permissible limit of PH,TDS, TH, Mg, Na, K, Cl and various other chem-ical parameters of water (Mukherjee et al. 2015;Zolekar et al. 2020). Modelling technique or accu-rate groundwater analysis have a crucial role insustainable growth and development of waterresources. Therefore, researchers are interested inthe mapping of various potential groundwater zone(Das et al. 2018). Currently, remote sensing andGIS techniques play an important role in theassessment of the world’s natural resources (Daset al. 2018). Identifying the potential groundwaterzone is an important part for planning and man-agement of water supply (Garg and Hassan 2007).Groundwater availability depends on many factorssuch as geology, geomorphology, slope, soil struc-ture, drainage density, rainfall and land use(Ganapuram et al. 2009), whereas net groundwaterrecharge depends upon the geology, lineamentdensity, drainage density, type of soil, intensity ofrainfall, slope, vegetative covers, land-use/landcover and water bodies present in the region.Estimates of groundwater storage are frequentlycited in global groundwater assessments (Richeyet al. 2015). The mapping of groundwater resourcesin the hard rock is more complex and difBcult dueto the different geological conditions. For mod-elling and mapping the groundwater resources inthe hard rock region, geospatial and geophysicalmethod is widely used over the globe (Prakash andMishra 1993; Israil et al. 2006; Murasingh 2014).Geophysical resistivity mapping proved to be

reliable and economical in investigations of therecharge source, especially in the hard ruggedgeological scenarios (Rai et al. 2013). Conventionalone-dimensional and modern two-dimensional

electrical resistivity imaging shows its usefulness inthe identiBcation of potential groundwater andrecharge zones (Savita et al. 2018) in the granitichard rock of Andhra Pradesh, India. A one-di-mensional resistivity model is one of the widelyaccepted, used, cheap and most Cexible methodsbecause it injects an electric current into the earthand then resistivity is measured along two elec-trodes (Kumar et al. 2014a, b). Whereas ERTmethod provides a realistic two-dimensionalsub-surface high-resolution resistivity model that,given the complexities of geology and topography,continuously shows changes in resistivity both invertical and horizontal directions (Loke 2000;Kumar et al. 2014a, b, 2016 and 2020). Theinverted model produced using 2D ERT is of highresolution and highly suitable in groundwaterexploration, assessment and sub-surface identiB-cation of saturated and fractured rocks based onthe characteristic resistivity of the sub-surfacegeological formations (Kumar et al. 2014b, 2020).While in geospatial method, Multi-Criteria Deci-sion Making (MCDM) technique is widely appliedin solving different MCDM problems and Analyt-ical Hierarchy Process (AHP) is one of them pro-posed by Saaty (1980); according to Sener et al.(2011), it is deBned as a quantitative techniquecombining objective of a theoretical expertprospective to complete decision-making by ana-lyzing multiple variables of hydrogeological sys-tem. In addition, the integration of RS and GIS is aconvenient way to use water resources andgroundwater maps for a variety of applications.These methods are very eAective and reliable toensure the availability of groundwater in the area(Murthy 2000; Prasad et al. 2008; Rahmati et al.2015; Das et al. 2017; Das and Pardeshi 2018).MCDM is an eAective decision-making tool thatcombines many of the geological and hydrogeo-logical factors that are responsible for delineationof groundwater recharge and potential zone(s)(Chenini et al. 2010). Recently, much research hasbeen done on the application of semi-quantitativemethods such as AHP (Sandoval et al. 2018) andthe frequency ratio model (Arulbalaji et al. 2019).The MCDM method significantly reduces themathematical complexity of the decision-makingprocess by providing estimates based on expertappraisal that compare the judgement of similareAects, and the estimates are an absolute eAect,nine parameters representation of interest. Thecurrent study has been carried out using MCDMand ERT methods to identify the groundwater

81 Page 2 of 17 J. Earth Syst. Sci. (2021) 130:81

recharge source within the surrounding of theNational Geophysical Research Institute (NGRI)campus, Hyderabad. The reliability of MCDM hasbeen validated using geophysical – both VES (1D)model and the 2D ERT model of two different timeperiods in the study area.

2. Geology and hydrogeology of the studyarea

The present study is carried out in a granitic hardrock aquifer, and a micro-watershed (Bgure 1) isfocussed for the detailed study within and sur-rounding the CSIR-NGRI campus, where the totalarea is 53.8 km2, whereas NGRI campus covers anarea of about 147.5 acres (0.60 km2), which is sit-uated in Hyderabad, Telangana, India (Bgure 2).Hyderabad is the largest city of Telangana, whichis situated at about 60 km away on the way ofNH-65, Hyderabad–Suryapet highway and is loca-ted between latitudes 17�24028.800–17�2508.400N andlongitude 78�32049.200–78�33025.200E. It has undu-lating topography with elevation ranges from 460to 560 m above sea level (asl). The area experiencesa hot and dry climate. The maximum averageannual temperature ranges from 39�–43�C and theminimum average annual temperature rangesbetween 13� and 17�C from December to Januaryin a year (Ravi 2012). The mean annual rainfall is

854.6 mm. The study area is drained by a majorriver named Musi River. The pattern of wide valleydrainage is generally dendritic to sub-dendritic innature. Inside the NGRI campus, Bve old largediameter dug wells are located for water supply tothe whole campus whose static water levels varyfrom 7.35 to 9.15 m below ground level (bgl).Water is regularly pumped out from these dugwells to meet the water requirements of CSIR-NGRI campus, both in oDce and in the residentialareas. The exact location of these dug wells isshown in Bgure 2. The entire study area is under-lain by granite rock of the Archaean age, which isone of the known varieties of the crystalline rock.The portion of the sub-surface granitic rocks,which is saturated with water, is an indicator offracture within the host rock and thus it shows lowresistivity formations. The basement rock, which isunderlain in the study area is igneous rock, mostlyundifferentiated soil cover, highly weatheredgranite, semi-weathered granite, fractured andmassive granite thus constituting the entire geo-logical setting of the study region. Geologically, thearea around Hyderabad created an exclusivelandscape magnificent granite structure, which isof Pre-Cambrian age. Figure 3 highlights the localgeology of the region, which falls in the graniticterrain. Most of the area is covered with alkalifeldspars granite. It strikes approximately in NS(north–south) direction and with lineament zone

Figure 1. The micro-watershed of the granitic hard rock aquifer region with different order stream and outlet in the study area.

J. Earth Syst. Sci. (2021) 130:81 Page 3 of 17 81

and dolerite rock. Granite has structural proper-ties such as fractures, joints, cracks and Bssures,which are the major conduits for groundwaterCow. In the upstream region, the area is dendriticand Bnely structured. This type of trough drai-nage is most often observed in shallow horizon-tally stretched basaltic rock, i.e., uniformlyresistant along a gentle regional slope (Horton1945). Commonly the region under the studyshows the hilly topography. Since the topographyis hilly, there are many streams of the Brst orderand second order. Rocks are mostly of Archaeanperiod mainly pink and grey granite with a lessamount of quaternary alluvium. Therefore, thereare two aquifers in the region, such as aquifers ofgranite and alluvium. Alluvial aquifer locatednear the Musi riverbed having medium- to Bne-grained sand, silt deposits at depth up to 5 m inthe region. It is less deep than sedimentary andfractured rock aquifer. While the crystallineaquifer consists of igneous and metamorphic rockshaving insignificant porosity are permeable onlyin fractured zone(s), which is hydrogeologicallycomplex and the movement of water depends onthe fractures and their connectivity present insuch rocks. The water level is usually between 5and 20 m and the average water level is 12 mbelow ground level (bgl). The water level variesbetween 5 and 20 m bgl during pre-monsoon

season and from 2 to 15 m bgl in post-monsoonperiod.

3. Materials and methods

3.1 Review of 1-D VES data interpretation

The aim of the Vertical Electrical Sounding (VES)is to determine the vertical variation of resistivityof the sub-surface rock strata in hard rock aquifer.The VES is an economical geophysical methodwhere current is injected into the ground using twocurrent electrodes for investigation of geologicalsub-surface medium. This is based on the electricalconductivity and/or resistivity of the sub-surfacerock strata. The application of VES has becomevery popular in groundwater research because of itstechnical simplicity and low cost. But VES hasmany limitations as the survey depth is limited bythe distance between the current electrodes. Whileoperating the VES survey, one should concern inmind that the last layer should be of the samethickness otherwise, if it is thinner than the middleand upper layers, the output of the resistance willbe uneven. The resistance of the thin middle layeraAects the measured value. This conventional VESis carried out at one location in the campus areawith a maximum current electrode spacing (i.e.,maximum proBle distance) of 200 m with a station

Figure 2. Location map of the study area with respect to the basin where ERT proBle and dug well locations situated inside theNGRI campus, study area.


spacing (AB/2) that vary from minimum 1.5 m toa maximum 100 m (Kumar et al. 2014a, b). Thepotential can be measured with a resolution of up

to 0.1 mV and acquired the measured apparentresistivity data. The raw VES Beld data was Brstplotted on a log–log graph paper with a modulus of

Figure 3. Showing the different geological setting and hydrogeological scenario within and surrounding of the CSIR-NGRIcampus in a granitic hard rock aquifer.

1 10 100 1000

10

100

1000N G R I 4th G ate

Appa

rent R

esist

ivity

(Ohm

-m)

Spacing (m)

10 100 1000

1

10

100

Depth

(m)

Resistivity (Ohm-m)

N G R I , I ndia

Model

ρ1 = 83.52 Ohm.m; h1 = 1.2m

ρ2 = 46.75 Ohm.m; h2 = 8.5m

ρ3 = 222.8 Ohm.m high RMS Error = 4.84%

Figure 4. Showing the interpretation of VES data (the square box is the Beld result and the sleek line is the calculated curve)along with the model layer parameters resistivity vs. depth model.


62.5 mm and later manual interpretation wasperformed using a conventional master curve(Orellana and Mooney 1996; Rijkswaterstaat1969). Such interpreted results were furtherimproved utilizing a computer-aided system andcurve matching technique (Jupp and VozoA 1975).The process of iteration goes on till the root meansquare (RMS) error is minimized and Bnally dis-play fairly the good match between the Beld andtheoretical curves representing the Bnal sub-sur-face model parameters in terms of true resistivitiesand thickness (qi, hi, i = 1, 2,…,n) for given sub-surface geological formations/layers (Bgure 4). Themodel converges within 5–7 iterations indicatingthat the quality of data is good and it is also seenfrom the RMS error value for the individual modelresults. On analysis and interpretation of VESresults, it revealed three distinct geoelectric layersof the sub-surface formations. Qualitative inter-pretation indicates the VES of H’ type of curves(Bgure 4). The interpreted VES results corrobo-rated with the sub-surface geological formationsof the area. It is found that the resistivity values*84 X.m correspond to weathered granite with athickness of 1.2 m, then follows the highly weath-ered rock of resistivity about 47 X.m of sufBcientthickness of 8.5 m and later follows the highresistivity, which corresponds to the basementhard granite rock as shown in Bgure 4.

3.2 Acquisition and processing of 2D ERT data

Electrical resistivity tomography (ERT) is themost common and precise method among allgroundwater exploration geophysical methods dueto the wide variation of resistivity of the geologicalformations and the changes that occur due to thesaturated conditions of the rock strata. A pair ofcurrent electrodes injected the current into theground and the potential difference between

another pair of potential electrodes are measuredduring each datum acuisition stage, which leads tothe measured apparent resistivity value for eachmeasurement in ERT method (Barker 2001). ERTdata has been acquired from two proBles withinNGRI Campus using three different namelyDipole–Dipole, Wenner and Gradient array inorder to visualize high resolution resistivity data ofnear-surface layers and the characteristics ofhydrogeological scenario. Each of the proBle con-sists of 81 and 40 electrodes with an electrodespacing of 0.5 and 2.0 m, respectively, and coveredthe total proBle distance 40 and 80 m, respectively,on the ground. The measured apparent resistivityBeld dataset was initially processed for eliminatingany noisy or bad data points (e.g., spike type, etc.)in the gathered dataset with respect to variousdepth levels and subsequently on the quality andthe smoothness of resistivity data, damping fac-tors, convergence limit, various mesh parametersand number of iterations to be undertaken have tobe assigned to the individual dataset (Dahlin 1996;Loke 1999) before going for inversion of theindividual data, in order to achieve realistic sub-surface resistivity models of the area.

3.3 Determination of the rank and weightnormalization for the thematic layerintegration using MCDM technique

Multi-criteria decision-making technique is com-monly used for groundwater and analytical hier-archy process (AHP) developed by Saaty (1980),has been used as a method to obtain and Bnalizethe weights assigned to different parameter and thefeature used in the groundwater recharge zonation.According to Sener et al. (2011), it is characterizedas a holistic technology that incorporates realisticand subjective expert opinions to complete deci-sion-making by assessing several factors. For this

Table 1. Pair-wise comparison matrix for the nine themes and calculation of weights by multi-criteria decision technique.

Sl. no. Thematic layer Lithology Lineament GM LULC Soil Slope DD Rainfall DEM

1 Lithology 1 1 2 2 3 1/3 2 1/2 1/3

2 Lineament 1 1 5 3 7 5 2 1 5

3 GM 1/2 1/5 1 1 5 1 3 1/5 1/2

4 LULC 1/2 1/3 1 1 5 1 1/2 1/5 1/3

5 Soil 1/3 1/7 1/5 1/5 1 1 3 1/3 1/2

6 Slope 3 1/5 1 1 1 1 1 1/2 3

7 DD 1/2 1/2 1/3 2 1/3 1 1 1/3 2

8 Rainfall 2 1 5 5 3 2 3 1 1/3

9 DEM 3 1/5 2 3 2 1/3 1/2 3 1


technique, every factor is evaluated based on acombination of elements between one (book value)and nine (significant values), where value 1 repre-sents equal importance between the thematic layerand the value 9 represents the extreme importancebetween the layers (Saaty 1980). This analysisassesses the significance of the parameters aAectinggroundwater availability (Razandi et al. 2015). Theweights assigned to all thematic layers and each oftheir individual feature classes were decidedaccording to the past Beld experience, as well as on

expert opinion. For a pair-wise comparison matrix(table 1), the line follows the mutual significanceof each parameter and the significance of thesecond parameter based on the relative impor-tance of all other parameters (Saaty 1980). It alsogives the percentage inCuence to every thematiclayer. This is done by normalizing the compara-tive matrix, which integrates the two-way com-parison matrix. A thematic layer with highweightage will have higher impact and with lowweightage will have a smaller impact on

Table 2. Normalized and assigned weights for the individual features of parameter for groundwater recharge zone.

Parameter Features classes

Percentage

(%) inCuence

Assigned

weight

Geometric

mean (Gm)

Normalized

weight

Lineament density 0–0.45 24 2 4.79 5.64

0.46–1.2 4

1.3–2.1 5

2.2–3.5 7

3.6–6.1 9

Rainfall 699.9–702.2 17 3 5.97 5.36

702.3–704.2 5

704.3–705.8 7

705.9–707.4 8

707.5–709.3 9

DEM 385–417 13 8 4.58 5.68

418–440 7

441–461 6

462–481 3

482–549 2

Slope 0–2 10 9 4.64 5.81

35 8

69 5

1016 3

17–35 2

Lithology Dolerite 9 2 1.82 3.30

Migmatite gneiss 3

Alkali Feldspar granite 1

Geomorphology Water bodies 8 9 3.02 6.95

Residual hills 1

Residual mound 2

Fluori-older Cood plain 7

Pediment 2

Drainage Density 0–1.1 7 1 3.94 6.35

1.2–2.1 3

2.2–3.2 5

3.3–4.3 7

4.4–5.3 9

LULC Built–up 7 1 3.63 6.61

Water bodies 9

Natural vegetation 5

Wet land/Grass land 7

Open land/Barren land 2

Soil Clayey 5 2 2.00 2.00


groundwater recharge zone. The weight is calcu-lated as the arithmetic mean. The geometric meanwas calculated by the summation of the assignedweights divisible by the number of variables. Thenormalized weight of each parameter has beencalculated using the formula NW=AWN/GM, whereNW is normalized weight, AWN is assigned weightand GM is geometric mean (Ahirwar et al. 2019).The percentages of assignment of normalizedweights and integration of thematic layers andthese weights are presented in table 2. This studyis based on a combination of a remote sensingdatabase and a geographic information systemthat provides a groundwater potential zone map(Bgure 5). Nine types of the parameters are con-sidered, which are of good quality and resolution:lineament density, rainfall, digital elevation modeldata (DEM), slope, lithology, geomorphology(GM), drainage density (DD), land use/land cover(LULC) and soil and these parameters are dis-cussed in detail (table 2). The thematic maps wereanalyzed using MCDM normal weights (table 2)to assess the groundwater recharge zone. Lithol-ogy and geomorphology maps have been

downloaded from the Bhukosh and clipped for thestudy area in ArcGIS by using clip tool. Drainagedensity and slope map are adapted from DEMdata, that are obtained from Earth data. LULCmap and lineament density map obtained fromSentinel 2A satellite data of USGS. Land use/landcover map was prepared from USGS EarthExplorer satellite data Sentinal 2A of 2020. Themaximum likelihood algorithm technique was usedfor generating it. In the preparation, false colourcombination has been processed and supervisedclassiBcation applied to obtain LULC map.Lineament density map was prepared using vari-ous tools present in Geomatica software. A soilmap is prepared from the Bhuvan and extractedusing various tools in ArcGIS and lastly, a rainfallmap is obtained from the Commodities ResearchUnit (CRU) and processed in ArcGIS. Finally, theoutput, i.e., groundwater potential/recharge mapwas evolved through the integration of all layers inArcGIS. The detailed methodology adopted forthe delineation of groundwater recharge zonesin the study region is demonstrated using theworkCow diagram and is shown in Bgure 5.

coCollection of data and maps Generation

Geospatial data Geophysical data

Satellite data Conventional data

Sentinel 2A DEM

Land

Use

/ La

ndco

ver

Slop

e

Dra

inag

e de

nsity

Elev

atio

n

Soil

Geo

mor

phol

ogy

Lith

olog

y

VES-1D 2D ERT

Rai

nfal

l

Weightage Assign & Normalization of Weightage by MCDM Technique

Vector to Raster Conversion

Integration of Thematic Layer Using GIS

Identification of Groundwater Recharge Zone

Schl

umbe

rger

Dip

ole-

Dip

loe

Wen

ner

Gra

dien

t

LS Terrameter

RES2DINV

Methodology

Interpretation

Line

amen

t de

nsity

Figure 5. Showing the collection of data and maps generation for identiBcation of groundwater recharge zone in the methodologyCow chart of the study area.


4. Results

4.1 Description of information obtainedfrom thematic maps

4.1.1 Lineament density

Lineaments are the linear feature that representsthe sub-surface geological structure such as faults,cleavages, fractures, etc. This type of linear fea-tures plays an important part in groundwateroccurrence and movement. Lineament density maphas been prepared which was further divided intoBve classes namely, 0–0.45, 0.46–1.2, 1.3–2.1,2.2–3.5 and 3.6–6.1 (Bgure 6a). High lineamentdensity areas are considered very good prospectzone for groundwater recharge.

4.1.2 Rainfall

Rainfall is one of the most important parametersof groundwater recharge where water is absorbedinto the soil, other rainfall characteristics such asintensity, duration and distribution are alsoimportant to measure the runoA. The annualrainfall map of the area was categorized into Bveclasses which are h699.9–702.2, 702.3–704.2,704.3–705.8, 705.9–707.4 and 707.5–709.3i mm/year as shown in Bgure 6(b).

4.1.3 Digital Elevation Model

The elevation map of the study area represents theground surface undulations available for depressionstorage. The elevation of the study area rangesfrom 385 to 549 m from mean sea level. It is cate-gorized into Bve classes as shown in Bgure 6(c).The elevation of the study area indicates surfacechanges that may persist with depression.

4.1.4 Slope

The slope in the area varies from 0 to 35� as shownin Bgure 6(d). Based on the slope, the study areacan be classiBed into Bve slope classes. The areahaving 0–2� slope falls into the ‘very good’ categorybecause of the nearly Cat terrain. The area with a3–5� slope is considered as good for groundwaterstorage due to slightly undulating topography withsome runoA. The areas with a slope 6–9� causerelatively high runoA and low inBltration to thesub-surface layers and hence are categorized as

‘moderate’. The fourth (10–16�) and Bfth (17–35�)categories are considered as poor due to higherslope values, which cause high runoA.

4.1.5 Lithology

In the study region, three types of geological rocksare found namely, dolerite, migmatite gneiss, andfeldspar granite as shown in Bgure 6(e). Migmatiteis the metamorphic rock, whereas granite anddolerite are igneous rocks. When the migmatitegneiss developed fracturing then it acts as a carrierfor groundwater. The dolerite rock occurs in theform of a dyke across, the area underlain bydolerite dyke acts as a conduit or a barrier forgroundwater Cow, while weathering and fracturingof the granite rock lead to formations of aquifers inthe hard rock region.

4.1.6 Geomorphology

Geomorphology is deBned as the study of thelandscape and plays a significant role in ground-water potential and prospect mapping of a studyarea. Our study area consists of water bodies,residual hill, Cuori-older Cood plain, pediment andresidual mond as shown in Bgure 6(f). Waterbodydeposits have been given the highest weighted,since it is incessant recharge to groundwater.Floodplain deposits were weighted higher thanpediment. The lowest priority was given to residualmond and residual hill.

4.1.7 Drainage density

Drainage density measurements have been madefor all the micro-watershed in the area, whichranges from 0.5 to 24.8 km/km2. Based on thedrainage density of micro-basins, it can be reclas-siBed into Bve classes, (1) 1.1 km/km2, (2) 1.2–2.1km/km2, (3) 2.2–3.2 km/km2, (4) 3.3–4.3 km/km2

and (5) 4.4–5.3 km/km2 as shown in Bgure (6g).

4.1.8 Land use/land cover

The following classes are distinguished: (i) built-uparea, (ii) water bodies, (iii) natural vegetation, (iv)wetland/grassland, and (v) open land/barren land(Bgure 6h). The water bodies were given thehighest weight over other land-use features sinceits incessant recharge to ground, ensured by


Figure 6. Various thematic maps namely (a) lineaments density, (b) rainfall, (c) DEM, (d) slope, (e) lithology area,(f) geomorphology, (g) drainage density, (h) land use/land cover and (i) soil, derived from MCDM-based weightedoverlay technique in the study area.


natural vegetation only, which pull maximumamount of water from the deep root zone for sus-taining, resulting in the presence of water belowthe ground surface. Natural vegetation wasweighted higher than open land/barren land. Openland/barren land is poor in groundwater potentialas compared to wetland/grassland, which is verygood for groundwater potential. The lowest prior-ity was given to the built-up area as it lack vege-tation cover. Built-up or concrete does not allowthe inBltration of water to percolate down theearth.

4.1.9 Soil

Soil emerged from rock fragments as a result of theweathering process. The predominant classes pre-sent in the area are only one clay type of soil(Bgure 6i). Clay soil is a kind of hard soil, beneBtingfrom the high nutrients. These soils are made ofmore than 25% clay and clay soils retain a largeamount of water due to space/pores between clayparticles.

4.2 Groundwater recharge zone

Finally, the groundwater recharge zone (Bgure 7)map has been developed by overlying the

determinant groundwater contributing the ninethematic layers namely, lineament density, rain-fall, digital elevation model data, slope, lithology,geomorphology, drainage density, land use/landcover and soil data using the weightage overlaytool of ArcGIS and MCDM technique as showing inBgure 7. The outcome produced from weightedoverlay analysis for groundwater potential is eval-uated by integrating all the layers in the ArcGISraster calculator. The value of groundwaterpotential index ranges from 2 to 7 and is classiBedinto Bve categories of possible groundwaterpotential zones. They were categorized namelyvery poor (4.32 km2), poor (34.5 km2), moderate(12.4 km2), good (2.21 km2) and very good (0.46km2) covering about 7.87, 64.05, 23.11, 4.10,0.87%, respectively, as well as the same is pre-sented as a pie graph in groundwater potential map(Bgure 7). A huge area covers very poor and poorcategories, which signiBes that groundwater isscarce in the region. Our study region is highlycovered with an urban settlement and rocky hardrock geology. So, groundwater availability is veryless. Only 2.67 km2 area lies in a good and verygood zone of groundwater due to saturated frac-tures present in the region (Bgure 7). Fractures arethe most important source of groundwaterrecharge in the hard rock area.

Figure 7. (a) Showing the groundwater prospect recharge zone map, (b) a percentage (%) of the areal distribution of thegroundwater recharge zones and (c) groundwater potential classes of the study area.


4.3 Interpretation of 2D resistivity models

4.3.1 Interpretation of resistivity modelsduring the post-monsoon period

The two-dimensional electrical resistivity data-sets were acquired during the pre-monsoon periodusing dipole–dipole, Wenner and gradient arraysto study the sub-surface recharge property of thenear-surface layers in a granitic hard rock aquifersystem. The resistivity scenario of the modelsshows that the top layer is high resistivity hardlayer, followed by semi-weathered layers andthen follows the highly weathered granitic rockformations in all the three resistivity models. Thedipole–dipole array is very sensitive to changes inthe geological formation’s resistivity horizontally.The near-surface layer in the model shows thehigh resistivity *80–140 X.m up to 1.5 m depth(Bgure 8a). In this true resistivity model, ano-ther zone with moderate resistivity values, i.e.,25–50 X.m revealed at *2 m depth. Subsequently,resistivity values are very low, i.e., *6–20 X.mbetween depth 5 and 11 m. It suggests thepresence of a highly weathered zone and isinterpreted as the potential groundwater stratawith groundwater recharge. The given 2D sub-surface true resistivity model is constructed byusing the Wenner array. This array covers atotal lateral distance of 40 m and investigateddepth up to 7.26 m. In Bgure 8(b), resistivityranges from *98 to 135 X.m up to 1.3 m depthshowing high resistivity near the top layer. Laterfollowed by a zone with low resistivity values,i.e., below 40 X.m at 3–7 m depth and betweenlateral distance 4 and 37 m. The gradient arrayprovides advantages with a higher signal-to-noiseratio in the structure resolution in quicker dataacquisition (Loke 2004). There is a steadydecrease in resistivity from 2 m to a depth of 6.5m (Bgure 8c), which suggests there is substantialmoisture content within the granitic rock forma-tion and is inferred as highly fractured geologicalstructure present around this depth. The invertedresistivity models obtained from dipole–dipole,Wenner and gradient arrays revealed that theresistivity values are high, i.e., *98–136 X.m atnear-surface depth, right from the top layer withdepths lying between *0.1 and 1.5 m, there is notraces of water as the surface is totally hard,which is shown as high resistivity of the graniticrock at the top layer. The model result is shownin table 3.

4.3.2 Interpretation of resistivity modelsduring the pre-monsoon period

Similarly, the 2D electrical resistivity tomographydata were acquired during the pre-monsoon period.The 2D inverted resistivity models show distinctgeological features as revealed from three differentelectrical arrays namely dipole–dipole, Wenner, andgradient in the pre-monsoon period. In dipole–dipolearray (Bgure 9a), at 13 m lateral distance, there is anear-surface groundwater recharge source up to *9m depth and resistivity value is low\25 X.m. Thereis another water-saturated zone of resistivity*15–25X.m equally visible *10 m depth between lateraldistance 42 and 74 m. At a lateral distance between16 and 48 m indicated weathered zone with higherresistivity *50–90 X.m. Similarly, between lateraldistance 48–80 m another highly weathered zonewith low resistivity of *40–60 X.m up to depth 2.5m. This shows the existence of two saturated zonesinferred as the potential groundwater recharge zone,which are separated by inclined to the vertical geo-logical structure in the study area. Whereas,Bgure 9(b) shows true resistivity model constructedby Wenner array. The Wenner array has a moderateinvestigation depth, i.e., 14.8 m in our case studycompared with other arrays. In this true resistivitysection, a zone with moderate resistivity value, i.e.,25–40 X.m at lateral distance 6–14 m is revealed.Subsequently, resistivity values are high, i.e., 50–90X.m at lateral distance 16–40 m. Later followed by azone with a low resistivity value, i.e., below 35 X.mat 3–8 m depth and located at a lateral distance44–68 m of the resistivity section (Bgure 9b). Itreveals the presence of a saturated weathered zoneand is inferred as the potential groundwater strata.In Bgure 9(c), the given 2D sub-surface true resis-tivity model is constructed by gradient array. Thegradient array model shows that there is a variationin the resistivity values laterally and vertically andthe true resistivity model delineated up to 17.2 mdepth. Initially, the resistivity value is low, i.e.,\40X.m at 2–8 m depth between the lateral distance 42and 68 m. The low resistivity values at these twosites indicated the presence of groundwater rechargezones and are separated by a high resistivity body/barrier between lateral distance of 18 and 42 m, witha resistivity value between 50 and*100 X.m at 2–17m depth. This type of high resistivity signatureindicated a vertical to inclined geological structure,e.g., a dyke or the massive granite between the twolow resistivity zones. The model result is shown intable 3.


5. Discussion

The entire work was accomplished comprising thedetailed hydrogeological study, multi-criteriadecision making (MCDM), remote sensing, GISand high-resolution resistivity tomography in thestudy area for understanding the near-surface

layers for groundwater recharge. The detailedgeological study and hydro-geomorphologicalanalysis provided extensive information about theformation of surface and sub-surface of the studyarea, which is crucial for groundwater analysis.The thematic maps evolved using nine surface andsub-surface geological and hydrogeological param-eters (Bgure 6) were analyzed using MCDM normalweights to assess and delineate the groundwaterrecharge zone. Further, GIS-based multi-criteriaanalysis has been used to identify potentialgroundwater recharge zone (Devanatham et al.2020), which is classiBed into very poor, poor,moderate, good and very good category (Bgure 7)as well as it shows a good integration with bothVES and ERT high-resolution geophysical workand their results (Prakash and Mishra 1993;

Figure 8. Shows the dipole–dipole, Wenner and gradient 2D inverted resistivity models with substantial resistivity values atshallow depth seen during the post-monsoon period.

Table 3. Comparison between resistivity values obtainedthrough 2D inverted resistivity models results at different depthlevels.

Sl no. Formation

Depth

(m)

Resistivity

(X.m)

Recharge

source

1 Top soil 0–2 *80–200 Poor

2 Weathered zone 2–11 *45–70 Moderate

3 Highly weathered

zone

11–17 *20–40 Good to

very good


Murasingh 2014; Kumar et al. 2016). The one-di-mensional and two-dimensional resistivity resultswere compared with the outcome from variousthematic maps of the area. The result of VES datarevealed top-soil and weathered granite resistivityis \100 X.m and highly weathered granitic rockshows resistivity *47 X.m of substantial thicknessof 8.5 m. This highly weathered granitic layer isindicative of water storage repository in theweathered-fractured coupled aquifer system. This1D result is corroborative with respect to the majorsub-surface layers, when compared with two

different time-period results of 2D resistivitymodels. It is understood that the hydrogeologicalcondition within the NGRI campus is varyingabruptly with the depth to the static water levels inthe dug wells vary from 7.35 to 9.15 m bgl. Thisresult coincides with the hydrogeological interpre-tation of the inverted 2D resistivity models espe-cially in terms of the saturated part of the aquifer.This shows that there is a large variation of satu-rated part of the aquifer within a small campusarea, which is totally dependent on the activesource of recharge to the water table of the granitic

Figure 9. Shows the dipole–dipole, Wenner and gradient 2D inverted resistivity models with substantial resistivity values atshallow depth revealed during the pre-monsoon period.


aquifer in this area. Nevertheless, 2D resistivitymodels revealed that the near-surface potentialgroundwater recharge source and is quite visible atshallow depth in high-resolution dipole–dipole,Wenner and gradient array models (Bgures 8 and9). The resistivity ranges between\100 X.m indi-cated weathered zone while 100–200 X.m delin-eated as hard granitic rock, which restricts themovement of groundwater Cow to the watertable of the aquifer. The combined results indi-cated the lithological contact zones between theweathered and highly weathered granitic rock inthe present geological medium, which are the maintarget for groundwater recharge sources in thepresent geological setting of the area.

6. Conclusions

The integrated study utilizing multi-criteria deci-sion making (MCDM), remote sensing, GIS andhigh-resolution resistivity tomography, theirresults and Bndings, etc., exemplify and delineatethe groundwater recharge zones into differentclasses – very poor to very good. The appropriateand successful application of RS data on the GISplatform has helped in achieving a detailed sce-nario of the groundwater scenario in and aroundthe study area. The integration of the geologicaland hydrogeological parameters resulted in Bvedifferent groundwater recharge zones namely, verypoor, poor, moderate, good and very good, which isfurther used in better planning and management ofgroundwater resources. While the outcome of theweighted ranges for various classes of groundwater,recharge zones were categorized such as very good6.1–7, good 5.1–6, moderate 4.1–5, poor 2.1–4, verypoor 0.0–2.0. These Bndings reCected that the poorzone revealed the least favourable area forgroundwater prospect, whereas good and very goodzones show the most favourable area for ground-water prospecting and development. Numerically23.11% of the study area is in a moderate zone ofrecharge, 4.97% in good and very good zone, while71.92% falls under the poor and very poor zone,i.e., unsuitable for groundwater recharge. Thepotential groundwater recharge zone map is vali-dated using 1D and 2D high-resolution geophysicalresults both qualitatively and quantitatively. Theweathered granite resistivity ranges from *45 to70 X.m up to 11 m depth where recharge is mod-erate, while *20–40 X.m up to 17 m depth repre-sents highly weathered granite, which leads to

maximum recharge in the area. The integratedresult based on pie graph reCected that *87% ofthe study region falls in the range of poor tomoderate recharge, and otherwise *28.1% leads tomoderate to very good recharge in the area.Therefore, these delineated recharge areas are usedfor the groundwater recharge strategy, manage-ment as well as development of groundwaterresources. The entire results and Bndings from thepresent study are very useful for prospecting andsustainable management of groundwater resourcesin the similar granitic hard rock system.

Acknowledgements

The authors express their gratitude to Dr V MTiwari, Director and all the respected members ofthe SAC committee of CSIR-National GeophysicalResearch Institute, Hyderabad, India for bestowingthis opportunity to carry out this research work atCSIR-NGRI, Hyderabad, which is a part ofM.Tech dissertation work of the 1st author. Fur-ther 1st author extends her acknowledgment toProf. H P Singh (Head), Department of WaterEngineering and Management, Central Universityof Jharkhand, Ranchi for his kind permission tocarry out this work. The CSIR-NGRI library ref-erence number of the manuscript is NGRI/Lib/2020/Pub-97. The critical review of the manuscriptand the factual comments and suggestions by theanonymous reviewers had significantly improvedthe quality of this research paper.

Author statement

Archana Kumari had done extensive work on themulti-criteria decision making (MCDM) usingremote sensing and GIS dataset and evolved thevarious thematic layers/maps of the study area inthe identiBcation of groundwater recharge zones ingranitic hard rock aquifer. Dewashish Kumar haddone the electrical resistivity tomography (ERT)data acquisition, processing and modelling of theERT dataset and integrated with the preparedmaps of the various thematic layers of the area inorder to achieve the meaning results and Bndingsusing the two different approaches applied in hardrock aquifer. Pratibha Warwade had helped in GISand remote sensing dataset and analyzing them forthe preparation of the various thematic layers andtheir application in groundwater study.


References

Ahirwar R, Malik M S and Shukla J P 2019 Prioritization ofsub-watersheds for soil and water conservation in parts ofNarmada River through morphometric analysis usingremote sensing and GIS; J. Geol. Soc. India 94(5) 515–524.

Al-Garni M A 2009 Geophysical investigation for groundwaterin a complex surface terrain, Wadi Fatima, KSA: A casestudy; Jordan J. Civil Engg. 3(2) 118–136.

Arulbalaji P, Padmala D and Sreelash K 2019 GIS and AHPtechniques-based delineation of groundwater potentialzones: A case study from southern Western Ghats, India;Sci. Rep. 9(1) 1–17.

Barker R D 2001 Imaging fractures in hard rock terrain;University of Birmingham, UK.

Bhanja N S, Mukherjee A, Rodell M, Wada Y, ChattopadhyayS, Veliconga I, Pangaluru K and Famiglietti S J 2017Groundwater rejuvenation in parts of India inCuenced bywater-policy change implementation; Sci. Rep. 7 7453,https://doi.org/10.1038/s41598-017-07058-2.

Chenini I, Mammou A B, and May E I M 2010 Groundwaterrecharge zone mapping using GIS-based multi-criteriaanalysis: A case study in central Tunisia (Maknassy Basin);Water Resour. Manag. 24(5) 921–939.

Dahlin T 1996 2-D Resistivity Surveying for Environmentaland Engineering Applications; First Break 14 275–283.

Das S 2019 Comparison among inCuencing factor, frequencyratio, and analytical hierarchy process techniques forgroundwater potential zonation in Vaitarna basin, Maha-rashtra, India; Groundw. Sust. Dev. 8 617–629, https://doi.org/10.1016/j.gsd.2019.03.003.

Das S, Gupta A and Ghosh S 2017 Exploring groundwaterpotential zones using MIF technique in semi-arid region: Acase study of Hingoli district, Maharashtra; Spat. Inf. Res.25(6) 749–756, https://doi.org/10.1007/s41324-017-0144-0.

Das S, Pardeshi S D, Kulkarni P P and Doke A 2018Extraction of lineaments from different azimuth anglesusing geospatial techniques: A case study of Pravara basin,Maharashtra, India; Arab. J. Geosci. 11 160, https://doi.org/10.1007/s12517-018-3522-6.

Das S and Pardeshi S 2018 Integration of different inCuencingfactors in GIS to delineate groundwater potential areasusing IF and FR techniques: A study of Pravara basin,Maharashtra, India; Appl. Water Sci. 8 197, https://doi.org/10.1007/s13201-018-0848-x.

Devanatham A, Subbarayan S, Singh L, Jennifer J J,Thiyagarajan S and Parthasarthy K S S 2020 GIS-basedmulti-criteria analysis for identiBcation of potential ground-water recharge zones – a case study from Ponnaniyaruwatershed Tamil Nadu, India; Hydro. Res. J. 3 1–14.

GanapuramS,KumarG,Krishna I,KahyaEandDemirelM2009Mapping of groundwater potential zones in the Musi basinusing remote sensing and GIS; Adv. Eng. Softw. 40 506–518.

Garg N K and Hassan Q 2007 Alarming scarcity of water inIndia; Curr. Sci. 93 932–941.

Ghosh P K, Bandyopadhyay S and Jana N C 2016 Mapping ofgroundwater potential zones in hard rock terrain usinggeoinformatics: A case of Kumari watershed in western partof West Bengal; Model. Earth Syst. Environ. 2(1) 1–12.

Hora Tejasvi, Srinivasan Veena and Nandita B Basu 2019 Thegroundwater recovery paradox in south India; https://doi.org/10.1029/2019GL083525.

Horton R E 1945 Erosional development of streams and their

drainage basins: Hydro-physical approach to quantitative

morphology; Bull. Geol. Soc. Am. 56 275–370, Progr. Phys.Geogr.: Earth Environment 19(4) 533–554, https://doi.

org/10.1177/030913339501900406.Israil M, Al-hadithi M and Singhal D C 2006 Application of

resistivity survey and geographical information system(GIS) analysis for hydrogeological zoning of a piedmontarea, Himalayan foothill region, India; Hydrogeol. J. 14753–759.

Jupp D L B and VozoA K 1975 Stable iterative methods for theinversion of geophysical data; Geophys. J. Roy. Astron.Soc. 42 957–976.

Kumar D, Mondal S and Warsi T 2020 Deep insight into the

complex aquifer and its characteristics from high-resolution

electrical resistivity tomography and borehole studies for

groundwater exploration and development; J. Earth Syst.Sci. 129, https://doi.org/10.1007/s12040-019-1336-x.

Kumar D, Mondal S, Nanda M J, Harini P, Soma Sekhar B M

V and Sen M K 2016 Two-dimensional electrical resistivity

tomography (ERT) and time domain induced polarization

(TDIP) study in hard rock for groundwater investigation:

A case study at Choutuppal Telangana, India; Arab.J. Geosci. 9(5) 1–15, https://doi.org/10.1007/s12517-016-

2382-1.Kumar D, Rai S N and Ratnakumari Y 2014a Evaluation of

heterogeneous aquifers in hard rocks from resistivitysounding data in parts of Kalmeshwar Taluk of NagpurDistrict, India; Curr. Sci. 107 1137–1145.

Kumar D, Rao V A and Sarma V S 2014b Hydrogeological and

geophysical study for deeper groundwater resource in

quartzitic hard rock ridge region from 2D resistivity data;

J. Earth Syst. Sci. 123(3) 531–543, https://doi.org/10.

1007/s12040-014-0408-1.Loke M H 1999 A practical guide to 2-D and 3-D Surveys;

Electrical Imaging Surveys for Environmental and Engi-neering Studies, pp. 8–10.

Loke M H 2000 Electrical imaging surveys for environmentaland engineering studies, a practical guide to 2D and 3Dsurveys, 6p.

Loke M H 2004 Tutorial 2-D and 3-D Electrical ImagingSurveys; Geotomo Software, Res2dinv 3.5 software.

Mukherjee A, Dipankar S, Charles F H, Richard G T,KaziMatin A and Bhanja S N 2015 Groundwater systemsof the Indian subcontinent; J. Hydrol. Regional Studies 41–14.

Murasingh S 2014 Analysis of groundwater potential zonesusing electrical resistivity, RS & GIS techniques in a typicalmine area of Odisha; https://doi.org/10.13140/RG.2.2.25059.04641.

Murthy K S R 2000 Groundwater potential in a semi-aridregion of Andhra Pradesh – a geographical informationsystem approach; Int. J. Remote Sens. 21 1867–1884.

Orellena E and Mooney H M 1996 Master tables and curves forVertical Electrical Sounding Over Layered Structures;Interciencia, Madrid, Spain.

Prakash S R and Mishra D 1993 IdentiBcation of groundwaterprospective zones by using remote sensing and geoelectricalmethods in and around Saidnagar area, Dakor block, Jalaundistrict, UP; J. Indian Soc. Remote Sens. 21(4) 217–227.

Prasad R K, Mondal N C, Banerjee P, Nandakumar M V andSingh V S 2008 Deciphering potential groundwater zone in


hard rock through the application of GIS; Environ. Geol.55(3) 467–475.

Rahmati O, Samani A N, Mahdavi M, Pourghasemi H R andZeinivand H 2015 Groundwater potential mapping atKurdistan region of Iran using analytic hierarchy processand GIS; Arab. J. Geosci. 8(9) 7059–7071.

Rai Shivendra Nath, Thiagarajan S, Kumar Dewashish,Dubey Kanchan M, Rai P K, Ramachandran A and NithyaB 2013 Electrical resistivity tomography for groundwaterexploration in a granitic terrain in NGRI campus; Curr.Sci. 105(10) 1410–1418.

Ravi G K 2012–2013 Central Ground Water Board, GroundWater Brochure Hyderabad, Andhra Pradesh, AnnualReport, 20p.

Razandi Y, Pourghasemi H R, Neisani N S and Rahmati O2015 Application of analytical hierarchy process, frequencyratio, and certainity factor models for groundwater poten-tial mapping using GIS; Earth Sci. Inf. 8(4) 867–883.

Richey A S, Thomas B F, Lo M H, Famiglietti J S, Swenson Sand Rodell M 2015 Uncertainty in global groundwaterstorage estimates in a total groundwater stress framework;Water Resour. Res. 51 5198–5216, https://doi.org/10.1002/2015WR017351.

Rijkswaterstaat 1969 Standard graphs for resistivity prospect-ing; European Association of Exploration Geophysicists,Hague, The Netherlands.

Saaty T L 1980 The analytical hierarchy process: Planning,priority setting, resources allocation; McGraw-hill, NewYork, 287p.

Sandoval E, Giacomo V, Jorge N, Jorge P, Jerry P, Fairly HooriA, Jose L A, Evelyn A, Hugo M and Rechard O 2018Groundwater assessment in a rural, arid, mid-mountain basinin north-central Chile;Hydrol. Sci. J. 63(13–14) 1873–1889,https://doi.org/10.1080/02626667.2018.1545095.

Satpathy B N and Kanungo B N 1976 Groundwater explo-ration in Hard rock terrain – a case study; Geophys.Prospect. 24(4) 725–763.

Savita R S,Mittal H K, Satishkumar U, Singh PK, Yadav KK,Jain H K, Mathur S M and Sham Davande 2018 Delineationof groundwater potential zones using remote sensing andGIStechniques in Kanakanala reservoir subwatershed, Kar-nataka, India; Int. J. Curr. Microbiol. Appl. Sci. 7(1)273–288, https://doi.org/10.20546/ijcmas.2018.701.030.

Sener S, Sener E and Karaguzel R 2011 Solid waste disposalsite selection with GIS and AHP methodology: A case studyin Senirkent–Uluborlu (Isparta) Basin, Turkey; Environ.Monit. Assess. 173(1–4) 533–554.

ZolekarRB,TodmalRS,BhagatVS,BhailumeSA,KoradeMSand Das S 2020 Hydro-chemical characterization and geospa-tial analysis of groundwater for drinking and agriculturalusage in Nashik district in Maharashtra, India; Environ. Dev.Sustain., https://doi.org/10.1007/s10668-020-00782-2.

Corresponding editor: ARKOPROVO BISWAS


Documents

Application of multi-criteria decision making (MCDM) and