Journal of Groundwater Science and Engineering 11(2023) 317−332

Journal homepage: http://gwse.iheg.org.cn

Research Paper

Using geospatial technologies to delineate Ground Water Potential Zones

(GWPZ) in Mberengwa and Zvishavane District, Zimbabwe
Nyasha Ashleigh Siziba1*, Pepukai Chifamba1
1
Department of Geospatial Science and Earth Observation, Zimbabwe National Geospatial and Space Agency (ZINGSA), 630 Churchill Ave,
Mount Pleasant, Harare Zimbabwe.

Abstract: The main objective of the study was to delineate Ground Water Potential Zones (GWPZ) in
Mberengwa and Zvishavane districts, Zimbabwe, utilizing geospatial technologies and thematic mapping.
Various factors, including geology, soil, rainfall, land use/land cover, drainage density, lineament density,
slope, Terrain Ruggedness Index (TRI), and Terrain Wetness Index (TWI), were incorporated as thematic
layers. The Multi Influencing Factor (MIF) and Analytical Hierarchical Process (AHP) techniques were
employed to assign appropriate weights to these layers based on their relative significance, prioritizing
GWPZ mapping. The integration of these weighted layers resulted in the generation of five GWPZ classes:
Very high, high, moderate, low, and very low. The MIF method identified 3% of the area as having very
high GWPZ, 19% as having high GWPZ, 40% as having moderate GWPZ, 24% as having low GWPZ, and
14% as having very low GWPZ. The AHP method yielded 2% for very high GWPZ, 14% for high GWPZ,
37% for moderate GWPZ, 37% for low GWPZ, and 10% for very low GWPZ. A strong correlation (ρ of
0.91) was observed between the MIF results and groundwater yield. The study successfully identified
regions with abundant groundwater, providing valuable target areas for groundwater exploitation and high-
volume water harvesting initiatives. Accurate identification of these crucial regions is essential for effective
decision-making, planning, and management of groundwater resources to alleviate water shortages.

Keywords: Groundwater resources; Analytical Hierarchical Process; Multi Influence Factor; Lineaments
density; Terrain Wetness Index; Ground Water Potential Zone

Received: 10 Dec 2022/ Accepted: 24 Oct 2023/ Published: 10 Dec 2023

Introduction tion and climate change, aggravate water problems

(Einlo et al. 2023; Yang et al. 2020). Rural and
The demand for groundwater resources has been urban areas in the Mberengwa and Zvishavane
steadily increasing due to erratic rainfall patterns, districts have been significantly affected by water
leading to the depletion of major surface water shortages, as silting of surface water sources, such
sources and resulting in water scarcity for both do- as rivers, streams, ponds, and dams, exacerbates
mestic and agricultural purposes. Growing de- the situation. This water stress has adversely
mand, reduction of water resources and increasing impacted livelihoods, especially for farmers who
pollution of water, driven by an increase in popula- heavily rely on subsistence farming. To address
this pressing issue, it is crucial to identify solu-
*
tions for augmenting surface water supplies. In
Corresponding author: Nyasha Ashleigh Siziba, E-mail address:
[email protected] some regions, groundwater serves as a vital
DOI: 10.26599/JGSE.2023.9280026 resource to supplement the deficits left by surface
Siziba NA, Chifamba P. 2023. Using geospatial technologies to water. Groundwater is a crucial source of water
delineate Ground Water Potential Zones (GWPZ) in Mberengwa supply, especially in regions with erratic rainfalls
and Zvishavane District, Zimbabwe. Journal of Groundwater (Kazerani et al. 2023). In some regions, groundwa-
Science and Engineering, 11(4): 317-332.
ter exploitation has increased to meet the demand
2305-7068/© 2023 Journal of Groundwater Science and Engineering Editorial
Office This is an open access article under the CC BY-NC-ND license (http://
due to surface water quality deterioration and
creativecommons.org/licenses/by-nc-nd/4.0) supply (Raju et al. 2023). Accessing groundwater
317
Journal of Groundwater Science and Engineering 11(2023) 317−332

resources has became more important as water ing elevation, slope, drainage density, flow accu-
demand has increased, particularly in dry regions mulation, geology, geomorphology, soil, land use,
(Emami and Shahamat, 2022). Wells and bore- land cover. An ANN based on the relationship
holes have become indispensable in areas where between groundwater potential data and the above
surface water is limited. This study is vital in meet- factors was implemented to create Ground Water
ing the growing demands for groundwater reso- Potential Zone map. Another machine learning
urces in the Mberengwa and Zvishavane districts. technique that has shown great potential in Ground
Recent research has shown that the mapping of Water Potential Zone mapping is Random Forest
Ground Water Potential Zones (GWPZ) is essen- (RF). In a study by Prasad et al. (2020), Random
tial for identifying high probability areas for ground- Forest (RF) was used to develop a Ground Water
water occurrence and enhancing water resource Potential Zone map along the west coast of India.
management. Geospatial technologies have emer- The study utilized input parameters, including
ged as crucial tools for mapping GWPZs world- elevation, slope, aspect, Slope Length (SL), profile
wide. The use of Geographic Information System curvature, plan curvature, Topographical Wetness
(GIS) has been widely adopted in the research Index (TWI), distance from streams, distance from
methods for ground water potential zone mapping. lineaments, lithology, geomorphology, soil, land
GIS is used for spatial analysis and integration of use, Normalized Difference Vegetation Index
hydrogeological variables such as rainfall, geol- (NDVI), and rainfall, to create a Ground Water Po-
ogy, soil, and terrain data to identify areas with tential Zone map. Similarly, in a study conducted
high groundwater potential (Zhang et al. 2019). by Hanane et al. (2023), Random Forest (RF)
Another remote sensing technique that has been machine learning technique was employed to esti-
utilized to map GWPZs is multispectral data analy- mate groundwater potential zones in the BouSbaa
sis. This method involves the acquisition of satel- area, Marrakech, Morocco. The predictive model
lite data and using it to identify land surface char- was designed using several input variables, includ-
acteristics that indicate groundwater presence ing geology, topography, land cover, slope, and
(Salimi et al. 2019). rainfall, to generate a groundwater potential zone
Various research methods have been employed map. Sarkar et al. (2022) also used Random Forest
to identify and delineate GWPZs accurately. The (RF) and Random Sub-Space (RSS) in the Teesta
Analytic Hierarchy Process (AHP) is a widely- River Basin of Bangladesh.
used research method that employs multiple deci- The integration of machine learning techniques
sion-making criteria to identify GWPZs. AHP such as ANN and Random Forest algorithms has
breaks down research objectives into numerous the potential to enhance the accuracy of groundwa-
decision-making criteria and then compares their ter potential zone mapping and improve predic-
relative importance. Recent studies have shown tions. By incorporating machine learning tech-
that AHP is an effective and powerful tool to anal- niques, experts can identify areas with a high prob-
yze and structure problem-solving and consider ability of groundwater occurrence and help over-
expert judgment on factor selection and weight come the uncertainty and limitations that are asso-
assignment, leading to more accurate identifica-
ciated with traditional geospatial analytical meth-
tion of areas with high groundwater potential
ods. As a result, machine learning tools can be a
(Osinowo and Arowoogun, 2020). The incorpora-
valuable tool in groundwater resource manage-
tion of advanced data processing algorithms and
ment, especially in areas with limited access to
machine learning techniques has the potential to
significantly contribute to improved predictions accurate data and geospatial technologies.
and help overcome uncertainties associated with Recent research findings indicate that the inclu-
geospatial analyses. In recent years, machine learn- sion of meteorological and groundwater yield data
ing has emerged as an important tool in groundwa- in GWPZ mapping leads to more accurate and reli-
ter resource management, especially in identifying able results (Das et al. 2019; Doke, 2019; Gnan-
Ground Water Potential Zones. achandrasamy et al. 2018a; Kabeto et al. 2022;
Artificial Neural Networks (ANNs) can analyze Kumar et al. 2017; Oyedele, 2019; Adeyeye et al.
large datasets from various sources, such as satel- 2019; Senapati and Das, 2022). Notably, different
lite imagery, and effectively predict potential gro- researchers employ distinct sets of factors based on
undwater zones. In a study conducted by Yadav their local environments, which can lead to varia-
and Salui (2021), machine learning models were tions in the final Ground Water Potential Zones
used for identification of Ground Water Potential (GWPZ) results.
Zones in part of Hugli District, West Bengal, India. Overall, a recent review of the relevant research
The study developed a geospatial model that inte- results on groundwater potential zones shows that
grated hydro-geological input parameters, includ- the use of various research methods has proven to
318 http://gwse.iheg.org.cn
 Journal of Groundwater Science and Engineering 11(2023) 317−332

be essential in accurately delineating GWPZs, and 2016). Therefore, recognizing and addressing these
the inclusion of meteorological and groundwater data limitations is crucial for the successful appli-
yield data has further enhanced the accuracy of cation of geospatial technologies in groundwater
these results. resource assessment and management.
The main objective of this study is to utilize To overcome data limitations, this study utilized
geospatial technologies to delineate areas of proba- satellite images to extract information on key land
ble groundwater potential in regions characterized surface features, including land cover, rainfall, and
by heterogeneous terrain and moderate to low topography. By supplementing ground-based data
annual rainfall. The geospatial-based approach collection with these remote sensing techniques,
enables the integration and processing of crucial gaps in spatial datasets were filled and the accu-
factors contributing to groundwater occurrence. racy of groundwater potential zone delineation was
Each factor, in turn, exerts varying degrees of improved.
influence on groundwater availability. In this
study, two techniques, namely Analytical Hierar- 1 Study area
chical Process (AHP) and Multi-Influencing Factor
(MIF), were employed. The MIF method assigns This research is dedicated to explore Agro Eco-
weights to influence factors based on their major or logical Zones (AEZ) 4 and 5a in Zimbabwe,
minor effects on groundwater occurrence, where specifically focusing on areas characterized by an
major effects receive a weight of 1 and minor average annual rainfall of less than 650 mm. The
effects receive a weight of 0.5. The determination scarcity of surface water in these regions makes
of these weights was guided by existing research them particularly suitable for conducting a compre-
(Sohail et al. 2019; Sutradhar et al. 2021; Thapa et hensive groundwater potential mapping study. The
al. 2017; Yeh et al. 2015; Magesh et al. 2012). On research site encompasses two districts, namely,
the other hand, AHP, an effective Multi-Criteria Mberengwa and Zvishavane (Fig. 1). By examin-
Decision-Making (MCDM) technique, has been ing these specific areas, we aim to gain valuable
extensively used by other researchers in GWPZ insights into groundwater resources and their
mapping (Adeyeye et al. 2019; Das et al. 2019; potential implications for sustainable agricultural
Hutti and Nijagunappa, 2011; Melese and Belay, practices in the region. The two districts have a
2022; Muthamilselvan et al. 2022; Zghibi et al. total population of 353,210 (ZIMSTAT, 2022).
2020). The AHP model assigns normalized wei- The study site extends from 19.93°S to 21.05°S
ghts to each thematic layer of groundwater explo- and 29.47°E to 30.53°E. It encompasses an area of
ration, further enhancing the precision of the 7,630.27 km2 and a perimeter of 436,118 km. High
results. elevation values of 1,547 m are found around
As this research was moving forward, several Mberengwa Mountain in the central part of the
aspects must be considered. Firstly, it is essential study area, and low elevation with a minimum of
to validate the accuracy and reliability of the delin- 624 m is in the southeastern part of the Mberengwa
eated groundwater potential zones using field data district. The major activities in the study area are
collection and hydrogeological investigations (Ma- subsistence farming, small-scale gold mining, and
nap et al. 2017). While geospatial technologies platinum mining. The dominant rivers are Ngezi
offer valuable insights, ground-truthing is crucial and Runde, which drain from the northwest to-
to ensuring the practical applicability of the results. wards the southeast. Drainage in the study region
Additionally, the study should explore the poten- exhibits a dendritic drainage pattern. A confor-
tial integration of other geospatial techniques such mable sequence of sedimentary, volcanic, and
as Geographic Information Systems (GIS), Remote older granitic rocks such as gneiss and migmatites
Sensing (RS), and Geostatistics to enhance the ro- with minimum metamorphism is more common in
bustness of the Ground Water Potential Zone map- the study region. Major faults such as the Mber-
ping (Mallick et al. 2021). The incorporation of ad- engwa fault, the Mtshingwe fault, and other sm-
vanced data processing algorithms and machine aller faults contribute a lot to the hydrogeology of
learning techniques may also contribute to imp- the Zvishavane and Mberengwa districts.
roved predictions and help overcome uncertainties
associated with geospatial analyses (Masroor et al.
2023). However, it is essential to acknowledge that 2 Methodology
despite the strengths of geospatial technologies,
their effectiveness can be hindered by data limita- 2.1 Data sources and processing
tions, including inadequate spatial datasets or input
data, which may impact the accuracy of the ground- The procedure adopted for this work comprised
water potential zone delineation (Kumar et al. desktop studies, and Remote Sensing (RS) was the
http://gwse.iheg.org.cn 319
Journal of Groundwater Science and Engineering 11(2023) 317−332

29.5°E 30.0°E 30.5°E

20.0°S

20.0°S
N

W E

Legend
Fault
River
20.5°S

20.5°S
Boundary
Elevation
Band 1 (Gray)
≤620
620-800
800-1000
1000-1300
>1300
21.0°S

21.0°S
0 25 50 km

29.5°E 30.0°E 30.5°E

Fig. 1 Location of the study area

major data source (Table 1). The workflow invo- lite imagery. For lineament extraction, image enhan-
lved is also presented in Fig. 2. Remotely sensed cement and directional filtering were employed
data such as SRTM-DEM and Landsat 8 satellite (Ahmadi and Pekkan, 2021; Al-Nahmi et al. 2016).
imagery were used to drive vital GWP factors. The Principal Component Analysis (PCA) using bands
following data were used in this study. 1–7 was used to enhance the satellite imagery and
the resultant bands were used to extract lineaments
2.2 Methods (Al-Nahmi et al. 2016; Mathew and Ariffin, 2018).
In addition, manual lineament extraction was emp-
In identifying ground water potential zones, the loyed, and lineaments were extracted from Geo-
following procedures were implemented. Geology maps through digitizing techniques. The extracted
and soil shapefiles were converted from vector to lineaments were used to calculate lineament den-
raster format using ILWIS software. These raster sity in ArcMap.
datasets were resampled to 30 m so that they have SRTM DEM was used to extract drainage for
the same spatial resolution as the DEM and Land- the area. The drainage shapefile was used to train
sat images. For rainfall, CHIRPS satellite data and the DEM in QGIS using the r carve algorithm in
meteorology gauging station data were used. A QGIS. Hydro-processing tools in ILWIS were used
Bias correction on satellite rainfall data was app- to extract drainage from the DEM. Overlay Analy-
lied using observed gauging station data. Monthly sis was done for extracted drainage to assess the
CHIRPS data was downloaded from January 1981 performance of the extraction process. A visual
to December 2021. Bias correction was applied agreement existed between extracted drainage and
using the power transformation technique (Gosh- existing drainage networks (Fig. 3). The derived
ime et al. 2020; Gumindoga et al. 2016). drainage was used to calculate drainage density in
Lineaments were extracted from Landsat 8 satel- ArcMap.

Table 1 Data collected, source and use

Data Source Use
SRTM DEM USGS For slope, TWI, TRI, Drainage density maps
Landsat 8 path 170 and row 74 USGS For LULC classification and lineament extraction
Soil Local geodatabase Soil classification
Rainfall Metrological services department and Rainfall map
CHIRPS rainfall data
Geology Geomaps Rock type classification, lineaments

320 http://gwse.iheg.org.cn
 Journal of Groundwater Science and Engineering 11(2023) 317−332

Input data
Meteorological Conventional
Satellite data
data maps and data

Landsat SRTM DEM CHIRPS Geodatabase GWP yield

Image
preprocessing
DEM hydroprocessing Raing guage
Satellite rainfall
drainage extraction rainfall Geology shapefile
PCA
drainage density soil shapefile
lineament
extraction Bias correction
lineament density rainfall Shapefile to raster
Slope
conversion
Band combination
pansharpening TWI
supervised Resampling
classification
accuracy assessment
land use/land cover TRI Geology
soil
Thermatic maps

TWI TRI Slope Geology Soil Rainfall Land use/Land cover Drainage density Lineament density

AHP assigning MIF assigning

weights weights

Reclassification Reclassification

Weighted overlay Weighted overlay

analysis analysis

Groundwater
Groundwater potential zone
potential zone generation
generation
Validation

Final groundwater
potential zones
using AHP and MIF

Fig. 2 Flowchat showing summary of methodology adapted for the research

Terrain and Topographic indices such as slope, neighbouring pixel elevation (Różycka et al. 2017).
TRI, and TWI were derived from DEM. TWI is TRI was calculated in QGIS.
calculated by the following formula Random forest (RF) was used for supervised
TWI = ln(A/tanβ) (1) image classification in QGIS. RF algorithm is a
decision tree-based machine learning algorithms
Where: A is the upslope contributing area and β
for image classification. The area was classified
is the local slope (Karimi-Rizvandi et al. 2021;
into eight land use/land cover classes. Random
Mukherjee et al. 2013; Różycka et al. 2017). Slope
sampling was used to generate 256 points. Acces-
and TWI were calculated in ILWIS. TRI is
computed from elevation value of a central pixel sible points were visited and each point was cate-
which is subtracted from elevation values of neigh- gorised according to relative land use. A simple
bouring pixel. TRI is expressed by the equation Kobo toolbox form was designed which captures
the GPS coordinates of the points and their land
TRI = (Σ(Zc − Zi )2 )2 (2) use type. For navigation to each point, a handheld
Where: Zc is central pixel elevation and Zi is Garmin GPS was used.
http://gwse.iheg.org.cn 321
Journal of Groundwater Science and Engineering 11(2023) 317−332

Normalised Weights for each factor and its rela- reprojected to the same coordinate system, over-
tive classes were calculated in Excel following the laid in a GIS environment, and each factor and
AHP and MIF procedures (Fig. 4). Factors were feature class was assigned its relative weight.

Visual analysis of extracted drainange and

30.1°E 30.2°E 30.3°E
existing drainage N
W E
S
Legend
20.2°S

20.2°S
Extracted drainage
Existing drainage

0 2 4 km

30.29°E 30.30°E 30.31°E

20.34°S

20.34°S
20.3°S

20.3°S
20.35°S

20.35°S
20.36°S

20.36°S
20.4°S

20.4°S
20.37°S

20.37°S
30.1°E 30.2°E 30.3°E 30.29°E 30.30°E 30.31°E

Fig. 3 Visual agreement between existing and extracted drainage

Drainage
density

Geology Soil

Lineament Slope
LULC
density

Rainfall TWI

TRI
Legend
Major effect
Minor effect

Fig. 4 Multi Influencing Factor (MIF) framework

322 http://gwse.iheg.org.cn
 Journal of Groundwater Science and Engineering 11(2023) 317−332

Weights were assigned according to the relevance sis where they have lower weights (Table 2). This
of each factor to the presence and occurrence of highlights the significance of geological analysis in
groundwater. All weighted maps were integrated to comprehending the distribution and availability of
produce a Ground Water Potential Zone map. The groundwater resources.
following Equation was implemented in the raster 3.1.2 Rainfall
calculator to produce a GWPZ map (Chaudhary After geology, rainfall (Fig. 5a) is another crucial
and Kumar, 2018; Das et al. 2019; Gnanachan- factor in determining GWPZ in the current study
drasamy et al. 2018; Kumar et al. 2016; Yeh et al. of agroecological zones 4 and 5. High rainfall
2016). areas are thought to have more potential for water,
GWPI =(Gw ∗ Gwi ) + (Rw ∗ Rwi ) + (S T w ∗ S T wi )+ and vice versa (Das, 2017). Rainfall, a major so-
(S Lw ∗ S Lwi) + (LDw ∗ LDwi ) + (DDw ∗ DDwi )+ urce of groundwater recharge, was analyzed for its
(T WIw ∗ T WIwi ) + (T RIw ∗ T RIwi + spatial distribution in the study area. The region
(LUw ∗ LUwi ) generally experienced moderate to low rainfall,
(3) ranging from less than 450 mm to 750 mm per
Where: GWPI is groundwater potential index, G annum, indicating moderate groundwater recharge
is Geology, R is Rainfall, ST is Soil Type, SL is potential.
Slope, LD is Lineament Density, DD is Drainage 3.1.3 Lineaments
Density, TRI is Terrain Ruggedness Index, TWI A lineament is described as a large-scale linear fea-
is Terrain Wetness Index, LU is Land Use, and the ture that appears in satellite images as relatively
subscript “ w” and “ wi” refer to the normalized straight tonal alignments and the topography of the
weights of layer and the normalized weights in underlying structural elements. Landsat imagery
each thematic layer, respectively. played a significant role in visually interpreting
A normality test was conducted on groundwater and identifying the lineaments in the area for
yield data from boreholes and GWP results from groundwater potential assessment. High lineament
AHP and MIF. Normality testing, followed by density regions (Fig. 5h) were found in the south-
Pearson correlation test was conducted in SPSS. western and northwestern parts of the study area,
indicating potential groundwater concentration and
permeable zones (Gnanachandrasamy et al. 2018a;
3 Results and discussion
Oyedele, 2019). Lineaments are known for their
high permeability and porosity, thus establishing
3.1 Thematic parameters an indirect relationship between lineament density
and groundwater potential zones (Magesh et al.
3.1.1 Geology 2012). The study area consisted of major faults
Geology with high permeability and porosity, such such as the Mberengwa fault, the Mtshingwe fault,
as sandstone and limestone, are more likely to have and other smaller faults. Folding and faulting
groundwater aquifers that can be accessed through contribute a lot to the hydrogeology of the study
wells and springs. It affects the characteristics of region. An increase in fault lines signifies more
the soil. Geology is the most influential factor lineaments, which results in more permeable zones
influencing the occurrence, availability, and distri- and high groundwater potential zones. Findings by
bution of groundwater (Kumar et al. 2017). (Muthamilselvan et al. 2022), states that high
Groundwater occurrence, distribution, and move- lineament density zones indicate areas with high
ment vary with location mainly due to rock forma- groundwater occurrence zones. In this study, the
tion, rock fracturing and the degree of weathering central and northern parts of the study area have
(Raju et al. 2023). The study area featured a higher lineament density and tend to exhibit high
diverse range of geological formations, compris- groundwater potential. Therefore groundwater is
ing around twelve different rock types. Within this associated with fractures, joints and fissures which
diverse range, the Older Gneiss Complex, various exhibits secondary porosity and high groundwater
Gneisses, and Young Intrusive Granite emerged as movement and accumulation.
the dominant rock types (Fig. 5d). However, these 3.1.4 Drainage density
rock types have relatively low porosity, which In hard rock terrains like the study area, drainage
limits their ability to hold water. As a result, the density is recognized as one of the key indications
Older Gneiss Complex, various Gneisses, and for locating groundwater potential zones. It is also
Young Intrusive Granite contribute less to ground- a proximate indicator of a terrain's porosity and
water occurrence, which is reflected in the analy- permeability (Murugesan et al. 2011). Drainage
http://gwse.iheg.org.cn 323
Journal of Groundwater Science and Engineering 11(2023) 317−332

29.7°E 30.0°E 30.3°E 30.6°E 29.7°E 30.0°E 30.3°E 30.6°E 29.7°E 30.0°E 30.3°E
Soils
Rainfall/mm Rainfall/mm Soil Calcic Solonetz Slope Slope
N ≤450 Chromic Cambisols ≤3
Chromic Luvisols
20.1°S

450-550 3-13

20.1°S
Endogleyic Lixisols
Eutric Leptosols
550-650 Ferric Luvisols 13-23
650-750 Haplic Acrisols 23-33
Haplic Lixisols
>750 Haplic Luvisols >33
Hypoluvic Arenosols
20.4°S

20.4°S
Leptosols
Lithic Leptosols
Sodic Solonchaks
20.7°S

20.7°S
21.0°S

21.0°S
0 20 km
(a) 0 20 km (b) 0 20 km (c)
Geology TRI Landuse/Landcover LULC
Geology Alluvium
TRI
Rock Outcrop
≤2 0 20 km
20.1°S

Andesitic

20.1°S
Balsaltic metavolcanics Mine Dumps
Dolerites and gabbros 2-10 Built Up
Felsites and porphyries
Gneisses of various ages 10-25 Crop Lands
Inland water
Metasediments 25-40 Grassland
Norite and gabbro
40-70 Grasslands
20.4°S

Older Gneiss Complex

20.4°S
Serpentinite and pyroxenites
Forest
Serpentinite and ultramafic Water
Ultramafic lavas and intrusions
Younger intrusive granite
20.7°S

20.7°S
21.0°S

21.0°S
0 20 km (d) 0 20 km (e) (f)
Terrain wetness index (TWI) Linearments density Linearments density/km2 Drainage density
TWI 1.2
0 20 km ≤5 0 20 km 0 20 km Drainage density/km2
0
20.1°S

20.1°S
5-12 0
12-18
18-24
20.4°S

24-30

20.4°S
20.7°S

20.7°S
21.0°S

21.0°S

(g) (h) (i)

29.7°E 30.0°E 30.3°E 30.6°E 29.7°E 30.0°E 30.3°E 30.6°E 29.7°E 30.0°E 30.3°E 30.6°E

Fig. 5 Conditioning factors in groundwater potential mapping

a) Rainfall, b) Soil, c) Slope d) Geology, e) Terrain Ruggedness Index (TRI), f) Land use/Land cover (LULC), g)
Terrain Wetness Index (TWI), h) Lineament Density, i) Drainage Density

density (Fig. 5i) also played a crucial role in 3.1.5 Slope

groundwater distribution, affecting runoff and infil- Hilly regions with a slope >33° (Fig. 5c) show
tration (Chaudhary and Kumar, 2018). There was a steeper slopes are not good for groundwater. The
greater visual agreement between extracted drain- slope of the land greatly affects the amount of
age and existing drainage (Fig. 4). The study area water that seeps into the ground and the amount
exhibited high drainage density, particularly with a that runs off. In this study area, slope was catego-
dendritic pattern, associated with low groundwater rized into five classes ranging from very steep to
recharge and permeability. In contrast, areas of low gentle, influenced the groundwater potential thro-
drainage density showed an indirect relationship ugh its control over runoff, infiltration, and perco-
with higher groundwater potential. lation rates (Das et al. 2019c; Chaudhary and
324 http://gwse.iheg.org.cn
 Journal of Groundwater Science and Engineering 11(2023) 317−332

Table 2 Determine the weight of conditioning factors using AHP method

Normalized Normalized theme
Theme Feature class Class weight
theme weight weight (xi)
Geology 0.165 Alluvium 9 0.214
Andesitic and dactic metavolcanic 2 0.048
Basaltic metavolcanic
Dolerites and gabbros 3 0.071
Felsites and porphyries 3 0.071
Gneiss of various ages 2 0.048
Metasediments 3 0.071
Norite and gabro 6 0.143
Older gneiss complex 2 0.046
Serpentinites 4 0.095
Ultramafic lavas 3 0.071
Young intrusive granites 4 0.095
1 0.024
Rainfall 0.146 <450 2 0.100
450–550 3 0.150
550–650 4 0.200
650–750 5 0.250
>750 6 0.300
Lineament density 0.128 0–0.3 2 0.077
/km/km2 0.3–0.6 3 0.115
0.6–0.9 5 0.192
0.9–1.2 7 0.269
>1.2 9 0.346
Drainage density 0.092 0–0.3 9 0.346
/km/km2 0.3–0.6 7 0.269
0.6–0.9 5 0.192
0.9–1.2 3 0.115
>1.2 2 0.077
Slope/o 0.110 0–5 9 0.292
5–10 7 0.250
10–15 5 0.208
15–20 3 0.167
>20 2 0.083
TWI 0.102 0–6 2 0.077
6–12 4 0.154
12–18 5 0.192
18–24 7 0.269
24–30 8 0.308
Soil 0.092 Arenosols 9 0.273
Leptosols 7 0.212
Luvisols 6 0.182
Acrisol 2 0.061
Solonchaks 5 0.152
Lixisols 4 0.121
LULC 0.110 Water 9 0.257
Forest 7 0.200
Shrubs and grassland 6 0.171
Agricultural fields 5 0.143
Bare land 4 0.086
Built up 2 0.057
Mine dumps 2 0.057
Rock outcrops 1 0.029
TRI 0.055 0–2 7 0.300
2–10 6 0.250
10–25 5 0.200
25–40 3 0.150
>40 1 0.100

Kumar, 2018; Zeinolabedini and Esmaeily, 2015). leading to increased groundwater recharge (Adey-
Steeper slopes were associated with higher runoff eye et al. 2019; Doke, 2019).
and lower groundwater potential, while gentler 3.1.6 Topographic wetness index
slopes promoted higher water-holding capacity and By examining the topography of the land surface
more time for water infiltration and percolation, to find areas of high soil moisture and waterlog-
http://gwse.iheg.org.cn 325
Journal of Groundwater Science and Engineering 11(2023) 317−332

ging, which are possible groundwater recharge dumps, and rocky areas. Groundwater potentiality
zones, the Topographic Wetness Index (TWI) is a is minor in mine dumps, built-up and rocky areas
valuable tool for predicting groundwater potential since these regions hinder water from seeping into
in a region. Soil saturation is a characteristic of the ground and facilitate the production of surface
areas with a high TWI value (Fig. 5g), and this is water. Densely vegetated regions promote ground-
thought to be favorable for groundwater recharge water occurrence as roots promote infiltration.
and storage. Based on several factors, including 3.1.9 Terrain Ruggedness Index
topography, soil, and vegetation that affect soil The Terrain Ruggedness Index (TRI) is a quantita-
moisture and waterlogging, the index is calculated tive measure used to assess the roughness or topo-
(Moore et al. 1993). The identification of ground- graphic complexity of a landscape. Its influence on
water recharge sources and efforts to improve groundwater potential can be attributed to its role
groundwater resource management through the in influencing surface water runoff and infiltration
optimization of water use can both benefit from the rates. Studies have shown that areas with higher
estimation of groundwater potential using TWI. TRI values tend to have increased surface runoff
TWI provides critical insights into the distribution, and decreased infiltration, leading to reduced
movement, and recharge potential of groundwater, groundwater recharge (Akbari et al. 2021). The
all of which facilitate the development of sustain- terrain ruggedness index was incorporated as a
able groundwater resource management. factor contributing to groundwater potential. The
3.1.7 Soil study area was dominated by regions of low to
Soil is characterized by its physiography, geology, moderate ruggedness (Fig. 5e), with high rugged-
and climate, all of which are crucial for groundwa- ness primarily observed in mountainous areas with
ter recharge and runoff. The types of soil and their heterogeneity in terrain.
permeability affect the area's ability to hold water.
Different soil types (Fig. 5b) were assessed for 3.2 Weight assignment and ground
their contribution to groundwater potential. Soils water potential mapping
varied in permeability and porosity, with aerono-
sols formed from unconsolidated sands exhibiting Through remote sensing data, GIS and weighted
better permeability, and vertisols formed from overlay analysis, the Analytical Hierarchical Pro-
basalt rock showing very low permeability and low cess (AHP) and Multi-Influencing Factor (MIF)
groundwater potential. techniques (Tables 2 and 3) were employed to
3.1.8 Land use/land cover calculate the weights of different factors contribut-
Land use and land cover (LULC) (Fig. 5f) were ing to groundwater potential. With the MIF tech-
analyzed to understand their impact on infiltration nique, factors with major influence were assigned a
and groundwater recharge. The study area was weight of 1, while those with minor influence were
classified into eight classes, and it was observed assigned a weight of 0.5 (Table 3). The AHP tech-
that vegetation, such as forests, shrubs, and grass- nique further normalized the weights for each
lands, promoted higher infiltration and groundwa- factor and its relative classes (Table 2). The
ter recharge compared to built-up areas, mine combined weights of the nine thematic layers were
Table 3 Determine the weight of conditioning factors using MIF method
Proposed score of each
Proposed relative rates
Factor Major effects (A) Minor effect(B) influencing factor
(A+B)
(A+B) *100/∑(A+B)
Lineaments density 1+1 0.5 2.5 9.80
Drainage density 1+1 0.5+0.5+0.5 3.5 13.73
Land use/Land cover 1+1+1 0.5 3.5 13.73
Geology 1+1+1+1 4 15.69
Soil 1 1 3.92
Rainfall 1+1 0.5+0.5 3 11.76
Slope 1+1+1 0.5+0.5 4 15.69
TWI 1+1 0.5 2.5 9.80
TRI 1 0.5 1.5 5.88
∑25.5 100

326 http://gwse.iheg.org.cn
 Journal of Groundwater Science and Engineering 11(2023) 317−332

used to create the final groundwater potential density, low drainage density, and densely vege-
maps. tated show favourable conditions for groundwater
Finally, the groundwater prospect maps (Figs. 6 infiltration and storage. Moderate to very good
and 7) have been prepared by integrating informa- potential zones exhibit characteristics favourable
tion from all the thematic parameters. The results for groundwater infiltration and storage. These
of both MIF and AHP techniques showed ground- zones are influenced by factors such as gentle
water potential zones with varying proportions slope, high lineament densities, high rainfall, land
across the study area. The area coverage of each surfaces with shrubs to dense vegetation, perme-
GWPZ class is shown in Table 4 and Fig. 8. The able soil. Dense vegetation retards surface runoff,
MIF method indicated very poor, poor, moderate, allowing water to pervade into the soil thereby
good, and very good groundwater potential zones increasing the amount of ground water. Very low
covering 14%, 24%, 40%, 19%, and 3% of the and low GWP is more dominant in the western,
study area, respectively. On the other hand, the eastern and south eastern region, these are charac-
AHP method identified groundwater potential terised by communal lands, rainfall less than 550
zones as very poor, poor, moderate, good, and very m, lineament density less than 0.3 km/km2, Gnei-
good, covering 10%, 37%, 37%, 14%, and 2% of sses and lixisols. These regions are dominated by
the study area, respectively. Using both methods, it factors which retards groundwater occurrence and
is clear that more than 50% of the study region has replenishment. Areas with high TWI values are
moderate to very high groundwater potential. This strongly related to groundwater potential (Sarkar et
can be explained by high lineament density in the al. 2022). In this research waterways and rivers
study area. Lineaments, such as faults and joints, exhibited moderate to very good groundwater po-
play a crucial in groundwater accumulation and tential and these regions had high TWI values. The
movement (Raju et al. 2023). The AHP and MIF study revealed that soil type (arenosols), geology,
results indicated that the central and northern parts land use (mainly forest), and high lineament den-
have well to very good GWP. These areas are sity were key factors contributing to good to very
dominated factors such as sedimentary rocks, good groundwater prospects.
higher lineaments densities, and rainfall above 650 A correlation analysis was conducted (Fig. 9a
mm per annum which promote groundwater accu- and 9b) to compare the groundwater potential
mulation. This research is in tandem with the find- results from MIF and AHP techniques with bore-
ings of (Muthamilselvan et al. 2022), where areas hole yield data from 330 sites. The analysis
gentle slope of lower than 3°, high lineament showed a very strong agreement (ρ=0.911) bet-

29.5°E 30.0°E 30.5°E 31.0°E

20.0°S

20.0°S

N
W E

S
GWPZ
Very poor
Poor
Moderate
Good
Very good
Area in hectares
20.5°S

20.5°S

2%
10%

14%

37%

Very good
Very low 37%
Good
Moderate
Low
21.0°S

21.0°S

0 25 km

29.5°E 30.0°E 30.5°E 31.0°E

Fig. 6 Ground water potential zones derived from AHP method

http://gwse.iheg.org.cn 327
Journal of Groundwater Science and Engineering 11(2023) 317−332

29.5°E 30.0°E 30.5°E 31.0°E

20.0°S

20.0°S
N
W E

S
GWPZ
Very poor
Poor
Moderate
Good
Very good
20.5°S

20.5°S
Area in hectares
3%
14%

40%

Very good 19%

Very low
Good
Low
24%
Moderate
21.0°S

21.0°S
0 25 km

29.5°E 30.0°E 30.5°E 31.0°E

Fig. 7 Ground water potential zones derived from MIF method

Table 4 The percentage and area covered by Ground Water Potential Zones (GWPZ)
Classification AHP area /Ha AHP area /% MIF area /Ha MIF area /%
Very good 13,942 2 24,748 3
Good 101,436 14 141,217 19
Moderate 281,693 37 300,451 40
Low 282,259 37 184,238 24
Very low 75,195 10 103,871 14

can be attributed to the different approaches, geol-

Area/km2
3,500 ogy and topography data used to derive groundwa-
3,004.5
3,000 2,822.6 2,816.9 ter potential zones. The MIF method is a data-
2,500 driven approach that considers multiple factors that
2,000 1,842.4
1,500 1,412.2 influence groundwater potential, including topog-
1,038.7 1,014.4 raphy, geology, land use/land cover, and rainfall.
1,000 752.0
500 247.5 139.4 On the other hand, the AHP method is a subjective
0 approach that requires the expert weighting of
Very poor Poor Moderate Good Very good
MIF AHP criteria based on their perceived importance.
Therefore, the reliability of the MIF method in this
Fig. 8 Bar chart representing area covered by GWPZ study can be attributed to its ability to consider a
wide range of factors that influence groundwater
ween the MIF results and borehole yield data, indi-
potential. The conclusions of the study can be
cating the reliability of the MIF method. The AHP applicable to other areas with similar hydrogeolog-
results exhibited moderate agreement (ρ=0.696) ical characteristics as the study area, although the
with the borehole yield data. Fig. 10a and 10b methodology may need to be tailored to the site-
shows scatter plots for MIF and AHP with their R2 specific conditions and data inputs available.
of 0.83 and 0.47, respectively. Correlation between Groundwater resource management requires a
the two methods shows that the derived groundwa- comprehensive and integrated approach that add-
ter potential zones are reasonably in line with the resses countless physical and socio-economic di-
actual results. The strong agreement between the mensions (Einlo et al. 2023). MIF and AHP pro-
MIF method and borehole yield data, as well as the vide integrated techniques involving various the-
moderate agreement shown by the AHP method, matic factors to delineate groundwater potential.
328 http://gwse.iheg.org.cn
 Journal of Groundwater Science and Engineering 11(2023) 317−332

Correlations Correlations
Yield MIF Yield AHP
Yield Pearson correlation 1 0.911** Yield Pearson correlation 1 0.696**
Sig. (1-tailed) <0.001 Sig. (1-tailed) <0.001
N 303 303 N 303 303
MIF Pearson correlation 0.911** 1 AHP Pearson correlation 0.696** 1
Sig. (1-tailed) <0.001 Sig. (1-tailed) <0.001
N 303 303 N 303 303
**.Correlation is significant at the 0.01 level (1-tailed). **.Correlation is significant at the 0.01 level (1-tailed).

Fig. 9 Correlation between MIF and Borehole yield

Borehole yield vs MIF Borehole yield vs AHP

7 7
6 6
5 5
4 4

AHP
MIF

3 y=0.69x+1.2912 3 y=0.6034x+1.8705
2 R2=0.8294 2 R2=0.4844
1 1
0 0
0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7
Borehole yield Borehole yield

Fig. 10 Scatter plot for MIF derived GWPZ

Through the use of various factors different envi- tively contributed to understanding the distribu-
ronments were classified according to their poten- tion of groundwater potential. The results revealed
tial in groundwater exploration. Each of the two distinct zones with varying groundwater potential,
techniques applied different weights on each fac- offering valuable insights for sustainable water
tor, thereby yielding different GWP and this led to resource management and informed decision-
variation in the correlation of each technique. In making. The study classified the study area into
this research the two applied techniques had differ- five categories of groundwater potential zones:
ent weights for thematic layers and this difference Very poor, poor, moderate, good, and very good.
in weights yielded different GWP results and The AHP method assigned these zones the follow-
different levels of agreement between GWPZ and ing proportions: Very poor (10%), poor (37%),
Borehole yield data. The two implied techniques moderate (37%), good (14%), and very good (2%),
comprehensively and integrate factors to address while the MIF method yielded: Very poor (14%),
the issue of groundwater exploration so as to curb poor (24%), moderate (40%), good (19%), and
water shortages in Mberengwa and Zvishavane. very good (3%). Notably, a strong correlation was
These regions are suffering from surface water is observed between MIF-derived groundwater pot-
depleting due to siltation, erratic rainfall and ground- ential results and borehole yield data, as well as a
moderate correlation between AHP-derived ground-
water pose a viable solution for water supply.
water potential results and borehole yield data. As
per the validations results AHP method (ρ=0.696)
4 Conclusion is shown more poor and moderate areas in the
watershed area as compared to MIF method (ρ=
In conclusion, the integration of geospatial tech- 0.911). The comparison analysis showed that the
nologies and the utilization of Multi-Influencing combined approach MIF method are performed
Factor (MIF) and Analytical Hierarchical Process acceptably for demarcating groundwater potential
(AHP) techniques proved to be highly effective in areas in the study area. The outcomes of this
conducting a comprehensive assessment of ground- research underscore the significance of geospatial
water potential zones in the study area. The analy- technologies as indispensable tools for exploring
sis encompassed nine crucial factors, namely geol- groundwater in suitable regions. Moreover, this
ogy, soil, rainfall, land use/land cover, drainage study has demonstrated the efficacy, time effi-
density, lineament density, terrain wetness index, ciency, and cost-effectiveness of Geographic Infor-
terrain ruggedness index, and slope, which collec- mation System (GIS) and Remote Sensing (RS) in
http://gwse.iheg.org.cn 329
Journal of Groundwater Science and Engineering 11(2023) 317−332

groundwater potential exploration. By narrowing DOI：10.1007/s40808-017-0396-7.

down the search areas for planners and decision- Doke A. 2019. Delineation of the groundwater
makers, this research offers valuable guidance in potential using remote sensing and GIS: A
mitigating water shortages in the study area. case study of Ulhas Basin, Maharashtra,
India. Archives of Photogrammetry, Cartogra-
References phy and Remote Sensing, 31(1): 49−64. DOI：
10.2478/apcrs-2019-0004.
Adeyeye OA, Ikpokonte EA, Arabi SA. 2019. GIS- Einlo F, Ekhtesasi MR, Ghorbani M, et al. 2023.
based groundwater potential mapping within Determine the most appropriate strategy for
Dengi area, North Central Nigeria. The Egyp- groundwater management in arid and semi-
tian Journal of Remote Sensing and Space arid regions, Abhar Plain, Iran. Journal of
Science, 22(2): 175−181. DOI：10.1016/j. Groundwater Science and Engineering, 11(2):
ejrs.2018.04.003. 97−115. DOI：10.26599/JGSE.2023.92800
Ahmadi H, Pekkan E. 2021. Fault-based geologi- 10.
cal lineaments extraction using Remote Emami H, Shahamat A. 2022. Geospatial ground-
Sensing and GIS—A review. Geosciences, water potential zone mapping using geologi-
11(5): 183. DOI：10.3390/geosciences11050 cal, remote sensing data and multi criteria
183. analysis techniques. Earth Observation and
Al-Nahmi F, Alami OB, Baidder L, et al. 2016. Geomatics Engineering, 6(1): 43−57. DOI：
Using remote sensing for lineament extrac- 10.22059/EOGE.2022.344474.1116.
tion in Al MaghrAbah area–Hajjah, Yemen. Gnanachandrasamy G, Zhou Y, Bagyaraj M, et al.
The International Archives of the Photogram- 2018. Remote sensing and GIS based ground-
metry, Remote Sensing and Spatial Informa- water potential zone mapping in Ariyalur
tion Sciences, XLII-2/W1: 137–142. DOI： District, Tamil Nadu. Journal of the Geologi-
10.5194/isprs-archives-XLII-2-W1-137-2016. cal Society of India, 92(4): 484−490. DOI：
Akbari M, Meshram SG, Krishna RS, et al. 2021. 10.1007/s12594-018-1046-z.
Identification of the groundwater potential Goshime DW, Absi R, Haile AT, et al. 2020. Bias-
recharge zones using MCDM Models: Full corrected CHIRP satellite rainfall for water
Consistency Method (FUCOM), Best Worst level simulation, Lake Ziway, Ethiopia. Jour-
Method (BWM) and Analytic Hierarchy nal of Hydrologic Engineering, 25(9):
Process (AHP). Water Resources Manage- 05020024. DOI：10.1061/(ASCE) HE.1943-
ment, 35(14): 4727–4745. DOI：10.1007/ 5584.0001965.
s11269-021-02924-1. Gumindoga W, Rientjes THM, Haile AT, et al.
Chaudhary BS, Kumar S. 2018. Identification of 2016. Bias correction schemes for CMORPH
groundwater potential zones using Remote satellite rainfall estimates in the Zambezi
Sensing and GIS of K-J Watershed, India. River Basin. Hydrology and Earth System
Journal of the Geological Society of India, Sciences Discussions, 1−36. DOI：10.5194/
91(6): 717−721. DOI：10.1007/s12594-018- hess-2016-33.
0929-3. Hanane T, Ahmed A, Abdellah A. 2023. Ground-
Das B, Pal SC, Malik S, et al. 2019. Modeling water potentiality mapping using machine
groundwater potential zones of Puruliya learning algorithms BouSbaa area, Marrak-
district, West Bengal, India using remote ech, Morocco. In Review. Research Square, 1-
sensing and GIS techniques. Geology, Ecol- 21. DOI：10.21203/rs.3.rs-2645081/v1.
ogy, and Landscapes, 3(3): 223−237. DOI： Hutti B, Nijagunappa R. 2011. Identification of
10.1080/24749508.2018.1555740. groundwater potential zone using Geoinfor-
Das S. 2017. Delineation of groundwater potential matics in Ghataprabha Basin, North Karna-
zone in hard rock terrain in Gangajalghati taka, India. International Journal of Geomat-
block, Bankura district, India using remote ics and Geosciences, 2(1): 91−109.
sensing and GIS techniques. Modeling Earth Kabeto J, Adeba D, Regasa MS, et al. 2022.
Systems and Environment, 3(4): 1589–1599. Groundwater potential assessment using GIS
330 http://gwse.iheg.org.cn
 Journal of Groundwater Science and Engineering 11(2023) 317−332

and Remote Sensing Techniques: Case study Mathew TG, Ariffin KS. 2018. Remote sensing
of West Arsi Zone, Ethiopia. Water, 14(12): technique for lineament extraction in associa-
1838. DOI：10.3390/w14121838. tion with mineralization pattern in Central
Karimi-Rizvandi S, Goodarzi HV, Afkoueieh JH, Belt Peninsular Malaysia. Journal of Physics:
et al. 2021. Groundwater-potential mapping Conference Series, 1082: 012092. DOI：10.
using a self-learning bayesian network model: 1088/1742-6596/1082/1/012092.
A comparison among metaheuristic algo- Melese T, Belay T. 2022. Groundwater potential
rithms. Water, 13(5): 658. DOI：10.3390/ zone mapping using Analytical Hierarchy
w13050658. Process and GIS in Muga Watershed, Abay
Kazerani P, Ziaei AN, Davari K. 2023. Determin- Basin, Ethiopia. Global Challenges, 6(1):
ing safe yield and mapping water level zoning 2100068. DOI：10.1002/gch2.202100068.
in groundwater resources of the Neishabour Moore ID, Gessler PE, Neilsen GA, et al. 1993.
Plain. Journal of Groundwater Science and Soil attribute predicted using terrain analysis.
Engineering, 11(1): 47−54. DOI：10.26599/ Soil Science Society of America Journal, 57:
JGSE.2023.9280005. 443−452. DOI：10.2136/sssaj1993.03615995
Kumar P, Herath S, Avtar R, et al. 2016. Mapping 005700020026x.
of groundwater potential zones in Killinochi Mukherjee S, Joshi PK, Mukherjee S, et al. 2013.
area, Sri Lanka, using GIS and remote sens- Evaluation of vertical accuracy of open source
ing techniques. Sustainable Water Resources Digital Elevation Model (DEM). Interna-
Management, 2(4): 419−430. DOI：10.1007/ tional Journal of Applied Earth Observation
s40899-016-0072-5. and Geoinformation, 21: 205−217. DOI：10.
Kumar YY, Moorthy DVS, Srinivas GS. 2017. 1016/j.jag.2012.09.004.
Identification of groundwater potential zones Murugesan V, Krishnaraj S, Kannusamy V, et al.
using Remote Sensing and Geographical 2011. Groundwater potential zoning in Thiru-
information System. International Journal of manimuttar sub-basin Tamilnadu, India-A
Civil Engineering and Technology, 8(3): GIS and remote sensing approach. Geo-
01−10. Spatial Information Science, 14(1): 17−26.
Magesh NS, Chandrasekar N, Soundranayagam JP. DOI：10.1007/s11806-011-0422-2.
2012. Delineation of groundwater potential Muthamilselvan A, Anamika S, Emmanuel I.
zones in Theni district, Tamil Nadu, using 2022. Identification of groundwater potential
remote sensing, GIS and MIF techniques. in hard rock aquifer systems using Remote
Geoscience Frontiers, 3(2): 189−196. DOI： Sensing, GIS and Magnetic Survey in
10.1016/j.gsf.2011.10.007. Veppanthattai, Peram-balur, Tamilnadu. Jour-
Mallick J, Singh CK, AlMesfer MK, et al. 2021. nal of Groundwater Science and Engineering,
Groundwater quality studies in the Kingdom 10(4): 367−380. DOI：10.19637/j.cnki.2305-
of Saudi Arabia: Prevalent research and 7068.2022.04.005.
management dimensions. Water, 13(9): 1266. Osinowo OO, Arowoogun KI. 2020. A multi-crite-
DOI：10.3390/w13091266. ria decision analysis for groundwater poten-
Manap MA, Sulaiman WNA, Ramli MF, et al. tial evaluation in parts of Ibadan, southwest-
2013. A knowledge-driven GIS modeling ern Nigeria. Applied Water Science, 10(11):
technique for groundwater potential mapping 228. DOI：10.1007/s13201-020-01311-2.
at the Upper Langat Basin, Malaysia. Arabian Oyedele AA. 2019. Use of remote sensing and GIS
Journal of Geosciences, 6(5): 1621−1637. techniques for groundwater exploration in the
DOI：10.1007/s12517-011-0469-2. basement complex terrain of Ado-Ekiti, SW
Masroor M, Sajjad H, Kumar P, et al. 2023. Novel Nigeria. Applied Water Science, 9(3): 1−13.
ensemble machine learning modeling DOI：10.1007/s13201-019-0917-9.
approach for groundwater potential mapping Prasad P, Loveson VJ, Kotha M, et al. 2020.
in Parbhani District of Maharashtra, India. Application of machine learning techniques in
Water, 15(3): 419. DOI：10.3390/w1503 groundwater potential mapping along the west
0419. coast of India. GIScience & Remote Sensing,
http://gwse.iheg.org.cn 331
Journal of Groundwater Science and Engineering 11(2023) 317−332

57(6): 735−752. DOI：10.1080/15481603. ment of groundwater potential zones using

2020.1794104. multi-influencing factor (MIF) and GIS: A
Raju BA, Rao PV, Subrahmanyam M. 2023. Esti- case study from Birbhum district, West
mating aquifer transmissivity using Dar- Bengal. Applied Water Science, 7(7):
Zarrouk parameters to delineate groundwater 4117−4131. DOI：10.1007/s13201-017-0571-
potential zones in Alluri Seetharama Raju z.
District, Andhra Pradesh, India. Journal of Yadav S, Salui CL. 2021. Artificial neural network
Groundwater Science and Engineering, 11(2): for identification of groundwater potential
116−132. DOI：10.26599/JGSE.2023. zones in part of Hugli District, West Bengal,
9280011. India. In Shit PK, Bhunia GS, Adhikary PP,
Różycka M, Migoń P, Michniewicz A. 2017. (Eds. ), Groundwater and Society, 247–257.
Topographic wetness index and terrain Springer International Publishing. DOI：10.
ruggedness index in geomorphic characterisa- 1007/978-3-030-64136-8_11.
tion of landslide terrains, on examples from Yang L, Zhang YP, Wen XR, et al. 2020. Charac-
the Sudetes, SW Poland. Zeitschrift Für teristics of groundwater and urban emer-
Geomorphologie, Supplementary Issues, gency water sources optimazation in Luo-
61(2): 61−80. DOI：10.1127/zfg_suppl/2016/ yang, China. Journal of Groundwater Science
0328. and Engineering, 8(3): 298−304. DOI：10.
Salimi AH, Masoompour Samakosh J, Sharifi E, et
19637/j.cnki.2305-7068.2020.03.010.
al. 2019. Optimized artificial neural networks-
Yeh HF, Cheng YS, Lin HI, et al. 2016. Mapping
based methods for statistical downscaling of
groundwater recharge potential zone using a
gridded precipitation data. Water, 11(8):
GIS approach in Hualian River, Taiwan.
1653. DOI：10.3390/w11081653.
Sustainable Environment Research, 26(1):
Sarkar SK, Talukdar S, Rahman AS, et al. 2022.
33−43. DOI：10.1016/j.serj.2015.09.005.
Groundwater potentiality mapping using
Zghibi A, Mirchi A, Msaddek MH, et al. 2020.
ensemble machine learning algorithms for
Using analytical hierarchy process and multi-
sustainable groundwater management. Fron-
influencing factors to map groundwater
tiers in Engineering and Built Environment,
recharge zones in a semi-arid Mediterranean
2(1): 43−54. DOI：10.1108/FEBE-09-2021-
Coastal Aquifer. Water, 12(9): 2525. DOI：
0044.
10.3390/w12092525.
Senapati U, Das TK. 2022. GIS-based compara-
tive assessment of groundwater potential zone Zhang K, Gann D, Ross M, et al. 2019. Accuracy
using MIF and AHP techniques in Cooch assessment of ASTER, SRTM, ALOS, and
Behar district, West Bengal. Applied Water TDX DEMs for hispaniola and implications
Science, 12(3): 43. DOI：10.1007/s13201- for mapping vulnerability to coastal flooding.
021-01509-y. Remote Sensing of Environment, 225:
Sohail AT, Singh SK, Kanga S. 2019. A geospa- 290−306. DOI：10.1016/j.rse.2019.02.028.
tial approach for groundwater potential ZIMSTAT. 2022. Population and Housing Census.
assessment using Multi Influence Factor https://www.zimstat.co.zw/wp-content/
(MIF) technique. International Journal on uploads/2022/07/Census2022_Preliminary_R
Emerging Technologies, 10(1): 183−196. eport.pdf
Sutradhar S, Mondal P, Das N. 2021. Delineation Zeinolabedini M, Esmaeily A. 2015. Groundwater
of groundwater potential zones using MIF and potential assessment using Geographic Infor-
AHP models: A micro-level study on Suri mation Systems and AHP method (Case
Sadar Sub-Division, Birbhum District, West study: Baft City, Kerman, Iran). The Interna-
Bengal, India. Groundwater for Sustainable tional Archives of the Photogrammetry,
Development, 12: 100547. DOI：10.1016/j. Remote Sensing and Spatial Information
gsd.2021.100547. Sciences, XL-1/W5: 769–774. DOI：10.5194/
Thapa R, Gupta S, Guin S, et al. 2017. Assess- isprsarchives-XL-1-W5-769-2015.

332 http://gwse.iheg.org.cn