Groundwater is one of the most natural resources that support human health and ecological diversity, but nowadays it is decreasing and there is an increase in demand for water (Lakshmi and Reddy, 2018). It’s also known as the water that occupies all of the pore space in soils and geologic formations beneath the water table (Wudad et al., 2021). It flows through the aquifer layer towards the discharge point such as wells, springs, rivers, lakes and the oceans (Bekele, 2021). Groundwater is also recognized as one of the most valuable natural resources, an immensely important and dependable source of water supply in all climatic regions all over the world (Debele, 2020).

Ethiopia is one of the countries in Africa known for abundant water resources but minimum utilization of groundwater for irrigation, domestic purposes and livelihood (Berhanu and Hatiye, 2020). The main part of Ethiopian areas has a shortage of rainfall and water scarcity due to a fluctuation in climate (Andualem and Demeke, 2019). The population of Didessa Sub-basins has used waters for drinking and other livestock from ponds and others by going long travel to fetch the water from rivers. The decreased rainfall in the arid zone and increasing population size and demands of water for irrigation and other livelihood requirements urges sustainable exploitation of the groundwater resources in the region (Kindie et al., 2019; Mengistu et al., 2019). In the study area, there is a large demand for groundwater development and utilization. However, there is a lack of in-depth understanding of the groundwater potentials in Ethiopia. In the recent past, the best estimates of the country’s exploitable groundwater potential ranged between 12 and 30 billion cubic meters (Berhanu, 2014). The groundwater potential inside Ethiopia’s territory was estimated to be around 1000 billion cubic meters based on hydrogeological parameters (Kebede, 2013). Other studies have found a different amount for Ethiopia’s groundwater potential about 40 billion cubic meters (Mengistu et al., 2019; Berhanu and Hatiye, 2020). Furthermore, there were just a few studies on the groundwater potentials of smaller catchments. As a result, there is a compelling need to investigate and develop the nation’s groundwater resources so that more people, particularly in rural areas, can benefit from improved water supplies. On the other hand, there is very little research in Ethiopia on the use of groundwater for irrigation (Mengistu et al., 2021). This could be due to a lack of knowledge about the existing groundwater resource and the necessity for big upfront investment in groundwater development for irrigation in the country (Berhanu and Hatiye, 2020). Groundwater potential investigation may thus be a necessary step in ensuring the nation’s long-term usage of groundwater for a variety of needs, including household, agricultural, and industrial uses (Mengistu et al., 2021; Dile et al., 2018). However, extensive use of groundwater based on only limited information may cause unsustainable use of the resource. Subsurface studies are often carried out when there arises a need for local specific development of groundwater exists including borehole, spring, shallow or hand-dug wells for domestic water supply (Wudad et al., 2021). Unsustainable groundwater use is becoming a visible problem and a major source of concern for many developing countries (Hussein et al. 2016). One of these problems is the absence of updated spatial information on the quantity and distribution of groundwater resources (Mengistu et al., 2021). Nowadays, there is a strong need to use groundwater for socio-economic development in Ethiopia, especially in rural areas (Dile et al., 2018). The ministry of water resources, through the water well design and monitoring enterprise, has performed a multi-disciplinary study incorporating hydrogeology and geophysics to investigate Ethiopia’s groundwater potential by employing time-consuming and costly geophysical and hydrogeological procedures, the potential can be realized. Aside from that, the groundwater potential for the Didesa sub-basin has not been characterized in prior research using GIS and RS approaches, separately (Kindie et al., 2019; Mengistu et al., 2019; Bashe, 2017). Since there is a limitation of previously work done in the area of the potentials and recharge zone in detail, the present study fills the gap by applying GIS and remote sensing techniques. These techniques are very easy to access, identify groundwater potential, and recharge zone of large and inaccessible areas (Gidey, 2018; Wudad et al., 2021; Mengistu et al., 2021; Andualem and Demeke, 2019). Considering the problems and the gaps from the other studies in terms of cost and time, an increasing number of population densities and fluctuation of climate on water scarcity in Didessa Sub-basins, this research focused on identifying the potential and mapping the spatial and temporal distribution of groundwater availability in the study area.

The availability, accessibility, movement, and occurrence of groundwater depends on geology, slope aspect, lineaments, drainage density, LULC, Rainfall, surface runoff, and geomorphology of the area (Wudad et al., 2021; Kindie et al., 2019; Bekele, 2021; Debele, 2020). Evaluation, exploitation, exploration, site selection/delineation, and maps of groundwater need serious caution as it is cost and time-effective (Tamiru and Wagari, 2021; Ahmad et al., 2020). It is out of our site improper evaluation of groundwater and site selection is mostly expected to pose the problem.

There are several methods available for the investigation of groundwater exploration or its availability in the sub-surface environment. They range from the oldest traditional water dowsing methods to the recent electromagnetic resonance (EMR) technologies (Wudad et al., 2021). In general, all the methods can be categorized into two broad classes: surface and sub-surface. The most cost-effective and simple methods of groundwater exploration or potential assessment are those that take place on the surface (Gintamo, 2014). Surface geophysical approaches, as well as esoteric geomorphological, geological, soil related, and microbiological remote sensing techniques, are among these technologies. The sub-surface methods mainly consist of test drilling of boreholes and geophysical logging techniques (Bashe, 2017; Tamiru and Wagari, 2021). Although the sub-surface methods are accurate for groundwater exploration, they incur large investment since drilling, completing and development of wells may be needed for the effective application of these methods (Bawoke et al., 2020; Melese and Belay, 2021). Therefore, it is a usual practice to undertake surface investigation methods for locating potential groundwater sources (Gidey, 2018). Remote sensing technique integrated with GIS is becoming a potent tool for the identification and mapping of groundwater potential zones in a time and cost-effective manner (Berhanu and Hatiye, 2020; Kindie et al., 2019; Bekele, 2021). GIS and remote sensing approaches are particularly effective in the identification of groundwater potential of any region and often use spatial data to analyze and characterize required information (Abiy et al., 2014). The geographic information system uses some proxy data from the ground surface to examine subsurface phenomena or objects. Geology, often known as lithology, is the study of rocks and minerals. landforms or geomorphology groundwater potential zones are typically identified using drainage density, rainfall, geological features or lineaments, slope, LULC, and soil properties (Melese and Belay, 2021; Debele, 2020; Kahsay et al., 2019). These data were used as proxy data for groundwater potential mapping in this study. When there are few direct measurements of groundwater, especially in smaller catchments, using such data to assess groundwater potential for Ethiopian conditions is extremely beneficial (Tamiru and Wagari, 2021; Kindie et al., 2019; Mengistu et al., 2021). The Didessa sub-basin is not exceptional to this constraint and there have been no previous investigations on the sub-basin’s groundwater potential. The goal of this work was to use GIS and remote sensing approaches to identify and map the groundwater potential zones in the Didessa sub-basin using groundwater proxy data. The study also aimed to test the veracity of the qualitative analysis GIS and remote sensing techniques yielded. Groundwater potential maps that have been found and organized may provide information on productive wells in the research area. The findings could be utilized as a geo-database to help decision-makers and other stakeholders in the research area appropriately use and manage groundwater resources.

Geographical Information System (GIS) and Remote Sensing (RS) are the optional methods to provide all parameters which influence the groundwater potential of one area and it can access, manipulate, and analyze the spatial and temporal data from satellite images (Yifru et al., 2020; Adeyeye et al., 2019). Besides this, Gupta and Srivastava (2010) explain many decision-making analysis approaches such as Multi-Criteria Decision-Making (MCDM), Analytical Hierarchy Process (AHP) (Wind and Saaty, 1980), and Fuzzy Logic to fill the gap of water scarcity and decision-making on groundwater potential zone for evaluation and mapping. AHP is useful method for complex decision-making. A pairwise comparison matrix is used for checking the consistency ratio (Dinka et al., 2014; Wind and Saaty, 1980). These decision-makers use this to reduce the bias in decision-making (Wind and Saaty, 1980; Kahsay et al., 2019). Therefore, this study was focused on the evaluation of groundwater potential zone in the Didesssa sub-basin by the integrated approach of remote sensing and GIS techniques using the AHP modeling approach. Besides this, the specific objectives of the study were to prepare the thematic maps of the study area such as lithology, LULC, slope, lineaments, soil texture, drainage density, and geomorphology and runoff depth to identify the factors that more affects the groundwater potential and recharge zone to assess, evaluate and delineate groundwater potentials of the area by using remote sensing and geospatial data.

2.1 Study Area

The Didessa sub-basin is the largest among the major sub-basins in Abay Basin in terms of its annual discharge (Kabite and Gessesse, 2018). In the Blue Nile basin, the Didesa sub-basin is the most important both hydrologically and physiographically. The Nile basin, which has supported millions of people since prehistoric times, receives around 60% of its yearly discharge from the blue Nile river, with the Didessa river accounting about 25% of that discharge (Kabite and Gessesse, 2018). The Didesa sub-basin is located in western Ethiopia, and it encompasses the East and West Wollega zones, the Illubabor zone, the Jimma special zone of the Oromiya area, and some parts of the Kamashi zone of Benishangul Gumuz. Geographically the sub-basin is located between 7º and 9ºN and 35º and 36ºE in the western part of Ethiopia. The elevation in the basin ranges between 654m MSL and 3154m MSL. The river drains an area of 28092km², and the Didessa River, which is the blue Nile’s greatest tributary, contributes nearly a quarter of the river’s total flow. In the research region, the average annual rainfall is around 1745mm. The river basin is classified under a humid tropical climate with heavy rainfall with an annual amount that is mainly received when the weather is wet. The highest and lowest temperature ranges between 21 and 31ºC and 10 and 15ºC, respectively. Geologically, Didessa sub-basin is dominated by Jimma Volcanic and the central part with Wellega Basalt and the lower parts are dominated by undifferentiated lower complex, Wellega Basalt and Adigrat Sand Stone.

Figure 1. Study area: Didessa sub-basin in Ethiopia

2.2 Data

In this study, eight thematic layers (drainage density, rainfall, slope steepness, LULC, lineaments density, soil, geomorphology, and lithology) were used. The rainfall map was prepared using Inverse Distance Weighting (IDW) interpolation of long term mean annual rainfall data recorded at 23 stations. The information of lithology and soil are collected from the Ethiopian Geological Survey. These thematic layers were prepared with the help of ArcGIS 10.7 platform. The drainage and lineament density maps have been prepared using the line density analysis tool in Arc GIS 10.7. The slope and drainage density were derived from the Shuttle Radar Terrain Mapping Digital Elevation Model (STRM DEM) of 30m resolution. Landsat-8 Operational Land Imager (OLI) satellite images were obtained from the United States Geological Society (USGS) website and used to generate the lineament density and LULC maps.

2.3 Methodology

GIS and remote sensing techniques were applied to delineate the groundwater potential of the Didesa sub-basins through the AHP. The methodology used in this work includes the stages: i) identification and evaluation of criteria, ii) data collection, iii) preprocessing, iv) input dataset, vi) reclassified input layers, vii) pairwise comparison of criteria and give weight with AHP, and x) overlay analysis with weight sum overlay analysis in ArcGIS tools and ranking the final value (Figure 2). The long-term average point rainfall values were used to generate areal rainfall map. IDW interpolation approach was used to interpolate the areal rainfall map. Twenty-three stations data were employed for the preparation of areal rainfall of the study area. In this study, data sets such as SRTM data, Landsat-8, and geological data of the study area is identified and used as input data sets for further processing. SRTM at 30 X 30m resolution was used to generate digital elevation model of the area. Landsat images have been used for the preparation of LULC, lineament density and geomorphology maps. Then the elevation map of the Didessa sub-basin was prepared using Arc GIS. Further, the elevation data was used to generate the slope raster data in each grid cell. In the elevation raster, slope was measured by the identification of maximum rate of change in value from each cell to neighboring cells. The slope classes in the watershed were identified using slope spatial analyst tool in the ArcGIS package. On the other hand, the data related to drainage density were generated indirectly from the watershed slope data. Drainage density is defined as the closeness of spacing of stream channels. Drainage density (Dd) is the measure of the total length of the stream segment of all orders per unit area. The lineaments of the study area were extracted from the Landsat-8 OLI/TIRS image (path 170, row 51 with zero cloud cover) using automated processing LINE tool in PCI Geomatica 2017. Out of the 11 bands of the Landsat-8 OLI/TIRS images, the lineaments were extracted from band 5 due to its quality and the highest number of lineaments delineated. After the lineaments were extracted, further processes of editing the watershed divide line and road features were done in GIS to ensure the quality of the extracted lineaments. The geomorphological map of the study area was prepared following the Hammond landform classification technique.

Figure 2. Methodology

2.3.1 Weight Assessment and Normalization

Because of all parameters do not have an equal effect on groundwater distribution a ranking of each parameter is needed. Weighting each factor and a pairwise comparison matrix were prepared for each map based on Wind and Saaty (1980) scale. The score for the classes in a layer and weights for each factor was assigned and a pairwise comparison matrix has been prepared for each map by using the Multi-Criteria Decision Analysis technique based on AHP (Wind and Saaty, 1980) by considering seven factors (lithology, lineament density, soil, and drainage, rainfall, slope and LULC). On Saaty’s scale of 1 to 9, each criterion is compared to the other criteria, proportionate to its relevance in matrices (Table 2). The score for classes in a layer and weights for thematic layers of each factor were calculated using the pairwise comparison matrix was carried out by using AHP techniques (Moges et al., 2019). The relative weight of their relevant classes was taken into account while calculating the cumulative weight of the primary criteria. Map categorization and weight assignment for groundwater potential and recharge parameters selected for groundwater potential and nine parameters selected for groundwater recharge potential zone. Normalization of assigned weight using AHP (Wind and Saaty, 1980) based on Saaty’s scale were considered two themes and classes at a time based on their relative importance to determine the groundwater potentials and recharge zone. Following that, pairwise comparison matrices with given weights to various thematic layers and classes created with the help of AHP and weights normalized by the eigenvector approach. To assess the normalized weights of various thematic layers and their particular classes, the consistency ratio (CR) was determined according to Wind and Saaty, (1980). To compute the CR of various thematic layers and their classes, the following steps were carried out:

1. Square matrix: A = a_ij (Equation 1) the element of row i column j was produced and the lower triangular matrix was completed by taking the reciprocal values of the upper diagonal using the formula, a_ij = 1/a_ij

\(A = \begin{bmatrix}\frac{P1}{P1} & \frac{P1}{Pj} & \frac{P1}{Pm}\\[0.3em] \frac{Pi}{P1} & 1 & \frac{Pi}{Pm} \\[0.3em]\frac{Pn}{P1}& \frac{Pn}{Pj} & \frac{Pn}{Pn} \end{bmatrix}\)  (1)    

2. Summation of all columns j in the matrix from (Equation 1) using (Equation 2)

\(\frac{P1}{P1} \ldots \ \ldots+ \frac{Pi}{Pj} \ldots \ \ldots \ \ldots \frac{Pn}{Pn} = \frac{\sum_{i=1}^{n} Pi}{Pi}\)  (2)

3. Divide each element of matrix a_ij = p_i/p_j (Equation 2) by (Equation 3) to get normalized relative weight (Equation 5)

\(\frac{\frac{Pi}{Pj}} {\frac{\sum_{i=1}^{n} Pi}{Pj}}= \frac{Pi}{Pj} \times \frac{Pj}{\sum_{i=1}^{n} Pi} = \frac{Pi}{\sum_{i=1}^{n} Pi} \)  (3)

4. Averaging across the rows (Equation 6) to get the normalized Principal Eigenvector (priority vector) i.e., Rate (Ri) or weight of row ‘i’ (Wi). Since it is normalized, the sum of all elements in the priority vector should be one.

\(\frac{Wi}{Ri}= \frac {Pi} {\sum_{i=1}^{n} Pi} \ldots \ \ldots+ \ \ldots\frac {Pi} {\sum_{i=1}^{n} P1} \times 1/n \)  (4)

The judgments of the pairwise comparison within each thematic layer were checked by the consistency ratio (Wind and Saaty, 1980). The proportion of consistency scholars recommended that for matrices having a CR rating greater than 0.1, the method should be repeated until the desired value of CR = 0.1 is obtained. The value of CR is computed as (Equation 5)

\(RC = \frac{Ci}{R}\)  (5)

where, RI is Saaty’s ratio index; CI is consistency index. Consistency index is a measure of consistency or degree of consistency of the judgment and computed using (Equation 6).

\(CI = \frac{(\lambda max-mn)}{n-1}\)  (6)

where, n is the number criterion, the value of RI for n criteria as indicated in Table 2.

To compute \(\lambda max\) , first multiply the normalized value by the respective weight (Equation 6) and then, the values of the product are added together to get \(\lambda max\)  (Equation 7)

\(\lambda max = \sum_{i=1}^{n} Wi \times \left (\frac {Pi} {\sum_{i=1}^{n} Pi} \right)\)  (7)

where, \(\lambda max\)  is the largest Eigenvalue of the pairwise comparison matrix. \(Pi\)  is the priority of the alternative and \(Wi\)  is the assigned rate/weight.

2.3.2 Overlay Analysis

The final map (Figure 11) was obtained by overlaying all thematic maps using weighted overlay methods after assigning rates for classes in a layer and weights for thematic layers (Kindie et al., 2019; Abiy, et al., 2014; Tamiru and Wagari, 2021), using the spatial analysis tool in are G.I.S 10.7 as shown (Equation 8)

\(GWP= \sum_{i=1}^{n} Wi \times Ri\)  (8)

where, GWP is groundwater potential zone; \(Wi\)  is the weight for each thematic layer and \(Ri\) ; rates for the classes within a thematic layer derived from AHP.

The author declares no potential conflicts of interest with respect to the research, authorship and publication of this article.

The author is highly indebted to the College of Engineering and Technology and Department of Hydraulic and Water Resources Engineering, Wolkite University for providing all kinds of necessary facilities and support during the present study.

AHP: Analytical Hierarchy Process; Dd: Drainage Density; GIS: Geographical Information System; Gm: Geomorphology; GWPZ: Ground Water Potential Zone; IDW: Inverse Distance Weighting; Lith: Lithology; LULC: Land Use Land Cover; Rf: Rainfall; RS: Remote Sensing; Sl: Slope; SRTM: Shuttle Radar Topographic Mission; St: Soil Texture; USGS: United States Geological Society; Wt: Weight.