Contribution of GIS and remote sensing in evaluating groundwater potential zoning in Mbandjock Open Access

Access to drinking water remains a major challenge for the international community and in particular for developing countries (Emile et al. 2022). Resorting to groundwater, it remains a better alternative because it requires less infrastructure, maintenance, and treatment resources. However, access to this resource remains more problematic given the failures in the choice of collection points (Allafta et al. 2020). In the past, the geological bedrock was considered the main study factor for access to groundwater (Yossa et al. 2023). Nowadays, in view of the failures of underground catchments associated with the study of this unique parameter, several other factors have been highlighted to better understand and model groundwater recharge (Shebl et al. 2022). This is to target areas with high groundwater potential as accurately as possible. As part of this study, we will use the multicriteria analytical hierarchy process (AHP) to assess the groundwater potential in Mbandjock. This approach integrates both geographic information systems (GIS) and remote sensing, which are powerful data analysis and decision-making tools (Ishola et al. 2023). The geospatial analyses mainly contribute to preventing the construction of counterproductive hydraulic structures (Yossa et al. 2023). In Cameroon, the public authorities as well as the decentralized local authorities to meet this need for drinking water in areas not served by the main water distribution network are increasingly using groundwater. However, it is clear how many boreholes have dried up after barely a year of operation. This is potentially due to the exploitation of a water table with a low water content. To better understand this thorny problem, we have resorted to GIS as well as remote sensing via the multicriteria analysis approach (Echogdali et al. 2022). This is based on the factors contributing to the recharge of a groundwater table, which made it possible to identify potential areas favorable to the construction of groundwater catchment works.

The study area as presented in Figure 1 is located in the central region of Cameroon more precisely in the division of Haute Sanaga Mbandjock, which is limited to the north by Yoko and Nanga-Eboko, the south by Edzendouan, Esse, Batchenga, and Lembe Yezoum, the east by Ntui, and West by Nkoteng. It is crossed from the north to the south by the Sanaga River. Mbandjock currently has more than 15 villages, which is considered a rural area of the city of Yaoundé, and till now, it remains partially served by the main drinking water distribution network.

Drainage density is a good indicator for obtaining accurate information on infiltration rates and for establishing a relationship between surface runoff and permeability in a terrain. It is the cumulative length of stream segments of all orders in a basin divided by the area of the basin. In this study, the drainage density was evaluated after generating the hydrographic network through the use of spatial analyst analysis tools in ArcGis 10.8. The entire drainage map is divided into five categories as presented in Table 2.

The study of lineaments has wide applications in different disciplines of geosciences for the identification of potential groundwater areas. Lineaments are straight linear features visible on the surface of the earth, and they are precisely the reflection of discontinuities on the surface of the earth caused by geological and geomorphological phenomena. The lineament density (LD), which is the length of the lineament segments in a particular region by its area, is calculated according to the following formula: LD = ΣLje/A, where ΣLje is the length of the lineament lines and A represents the area. The LD map was obtained by creating hillshades with different altitude limits, allowing us to digitize the lineaments visible from the tools. The map is then classified into five categories as follows and illustrated in Table 3.

The slope is one of the most important parameters representing the local and regional relief and having a significant influence on groundwater recharge in the aquifer. The slope gradient directly controls surface water infiltration and is widely used in the delineation of potential groundwater zones. The slope map was drawn from the corrected DEM (Fill DEM), which consists of filling in the empty parts of the image. Spatial tools were used in Arc GIS 10.8. The entire slope map is divided into five categories (Table 4).

Land use and cover is considered to be another important parameter in identifying potential groundwater areas. Important and necessary information such as infiltration, soil moisture, groundwater, and surface water requirements can be obtained from the LULC map. The Mbandjock study area was previously described, and all the data were selected from the Landsat 8 image with a resolution of 30 m on Earth Explorer. The supervised classification method was used in the production of the LULC map from the Envi 5.5 software. First, we identified the areas of the image, and so we know the real land cover in comparison with the land cover database on Google Earth. These areas are called regions of interest. Subsequently, these zones were used as a reference zone for the generalization of the entire image for the creation of the signature files. A maximum of pixels possible according to the importance of the classes has been selected for the training areas. We then validated our classification to assess the quality. At the end of this treatment, the LULC was classified into seven major classes: water bodies, vegetation, agriculture land, bare soil, wetland, sugar cane plantation, and settlement.

In the aquifer, the soil has an important control over the rates of infiltration and percolation. The shape, size, and arrangement of soil grains and the corresponding pore system are likely to strongly affect the vertical and lateral movement of water. Soil data were obtained from the Geo web portal, which is an FAO site, in ESRI Shapefile format. Soil categories were obtained from SWAT Soil Data. Soil data have been added to ArcGIS 10.8 and georeferenced to the UTM projected coordinate system. We extracted the Mbandjock study area using the analysis tools on ArcGIS. The data were then geocoded for each soil type at different categorical levels in the soil classification according to the classes obtained from the soil SWAT data.

Precipitation plays a big role since it represents the main source of groundwater recharge the annual precipitation for the period from 2012 to 2022 in the district of Mbandjock was obtained from the Climate Research Unit. The rainfall data were converted to a raster layer using the multidimensional tools (Create NetCDF Raster Layer). We then converted the raster layer to points using the conversion tools. These points were interpolated to obtain a map of precipitation distributions.

Geology has a significant impact on the existence and movement of groundwater. It affects recharge by acting on water percolation. Similarly, the type of rock can significantly influence the groundwater recharge potential. By accounting for geology in our analysis, we can minimize uncertainty in estimating drainage densities and lineaments. The geological map of Cameroon was extracted from the Mbandjock study area using analysis tools. The classification of geological categories was carried out according to the model proposed by the FAO.

In this study, AHP was chosen to solve complex problems with a multicriteria decision. The various factors are organized in a structured way while giving a relatively simple solution for the decision-making problems. Seven parameters as presented earlier were used to delineate the GWPZs (Groundwater Potential Zoning) (drainage density, LD, slope, LULC, rainfall, soil characteristics, and geology). The normalized weights were used to evaluate the GWPZs of the Mbandjock district while testing the thematic maps of the different parameters using the AHP method (Achu et al. 2020). The application of the method required the selection of parameters contributing to groundwater recharge, performing a pairwise comparison matrix and evaluating the consistency of the matrix and assigning relative weights.

On the basis of the number of thematic layers selected for GWPZ mapping, we created a pairwise comparison matrix, and each entry in the matrix describes the effect of the row layer versus the column layer. The importance of a layer or the weight of the criteria on the potentiality of groundwater linked to another was noted on the basis of the scale 1–9 presented by Saaty (1990).

This means that low weights correspond to low groundwater potentials, while high weights refer to high groundwater potentials. Based on literature and expert opinion, different parameters imply different impacts on groundwater potential. Thus, precipitation is the main source of water recharge and was therefore selected as the first layer and placed in the first row and the first column in the comparison matrix of our study. Similarly, geology plays a major role in the evolution and flow of groundwater and occupies the second place in the comparison matrix. Slope strongly controls precipitation infiltration into the soil and was selected as the third factor in the hierarchy. The drainage density presents a negative correlation with the permeability, and consequently, the infiltration rate occupies the fourth position. Soil characteristics were considered as the fifth factor in the comparison matrix because permeability and infiltration are directly related to soil grain size and associated pore size characteristics. LULC and LD have a low impact on groundwater recharge and are ranked sixth and seventh, respectively, in the hierarchy. The pairwise comparison matrix based on the AHP method was created to define the rank and priority of the factors. According to a scale of 1–9 points, priorities were assigned to each pair of layers. The order of influence of the parameters on the groundwater potentiality represents the eigenvector.

A pairwise matrix is considered consistent when λmax is greater than or equal to the number of parameters selected (seven parameters in our study). λ max is the principal eigenvalue, and it represents a function of the coherence deviation of the matrix. It was obtained by adding the products of the sum of the columns of the layers of the pairwise comparison matrix and the eigenvectors. For a 7 × 7 matrix, λmax = The value obtained is thus used for the coherence index (IC). The consistency ratio calculation allows us to check the consistency of the normalized weights. If the consistency rate is less than or equal to 10%, the assigned weights are considered consistent. The calculation of the coherence ratio consists in calculating the IC. IC = λmax – n/n – 1, where λmax is the main eigenvalue and n is the number of parameters.

Precipitation being the main source of water recharge has a general characteristic of average annual precipitation of high levels in the southern and western parts, as illustrated in Figure 3. The annual rainfall varies between 1,425 and 1,550 mm. The spatial distribution map of annual precipitation is classified into five classes. Precipitation is the main limiting factor for groundwater recharge. To better understand the importance of this factor, Tolche (2021) shows how the spatial variation of rainfall intensities had a direct impact on groundwater recharge.

The geological constitution of the subsoil gives an idea of the component of the aquifer. Knowing the characteristics of these components, we can have an idea about the permeability level of the aquifer roof. Mbandjock is crossed by six geological structures as illustrated in Figure 4, and the most predominant of which are the pCm in the southern part, the H₂O in the northern part, the Ti in the east, and the MiPi in the west. This structure indicates that the geology of Mbandjock is old and poorly consolidated with potentially permeable free roofs. This is an important indicator for groundwater recharge. This was the case in the study by Aretouyap et al. (2022) in which the author demonstrate that the geological basement is a decisive parameter in the recharge of the aquifer. Gyeltshen et al. (2019) goes further to demonstrate during its work on the contribution of the geology and hydrogeology of discontinuous aquifers of the crystalline basement and describes how this can impact the recharge of the aquifer.

Five slope classes have been determined in the Mbandjock area (Figure 5). A significant portion of the basin falls into the first four categories exhibiting a low slope gradient having the ability to boost the rate of infiltration as it constrains water flow. However, the fifth category in red indicates that the highest percentage of slope is in the South-East part of Mbandjock. This part of the region has a steep gradient and therefore corresponds to a relatively low GWP due to high runoff. The results obtained in the case of our study confirm the hypotheses of Yıldırım (2020), which stipulate that the slope of the land directly impacts the surface runoff process and hence its importance in groundwater recharge. The first three grades of slope (<10°) are gentler slopes with a lower level of topographic elevation, which are classified as very high groundwater potential due to almost nonexistent topography that favors high groundwater rates and high infiltration. Areas with a slope of 10–20° corresponding to the fourth category of our study are classified in the group of potentiality of high groundwater due to their gently rolling terrain and their potential for moderate runoff (Allafta et al. 2020). Areas with a slope range of 20–30° are classified as moderate groundwater potential due to their limited infiltration rates and relatively high runoff. This is the case for the fifth category of slope in our study area. Areas with slopes of 30–40° and over 40° are recognized as having low and very low groundwater potential due to their steeper slope resulting in higher runoff potential. Conclusions drawn from previous studies have shown that areas with low slopes have a high potential for groundwater storage due to the longer residence time promoting water percolation. In contrast, areas with steep slopes have low groundwater potential due to rapid runoff from the terrain. Lower slope values indicate flatter terrain, and higher slope values correspond to a steeper terrain slope (Bhadran et al. 2022).

Table 3 presents the different classes of slope in the study area and the observations of different category of classes.

The drainage density in the Mbandjock area is classified into five classes as presented in Figure 6: 0–0.73 km/km², very low; 0.74–1.13 km/km², low; 1.14–1.52 km/km², low; 1.53–1.96 km/km², low; 1.97–3.2 km/km², moderate. Most of the study area has low to moderate drainage densities and is therefore dominated by a plain area with a penetrable subsoil (permeable) under abundant vegetation (Aykut 2021). They therefore have higher recharge rates. An area with low drainage density is more favorable to high groundwater potential (Shah Porun Ranaa et al. 2022). On the other hand, the zone in blue constituting a high density of drainage extends in a sparse vegetation with a mountainous relief and impermeable soil. They have relatively low recharge rates, promote runoff, and are characterized by poor GWPZ.

Table 4 presents the different classes of drainage in the study area and the observations at different category of classes.

The soils in the Mbandjock area belong to the large set of typical ferralitic soils and subdivided into five subsets as presented in Figure 7: humic clay-brown soils with slightly dark organic matter, yellowish-red clayey soils with fragmentary structure, red sandy-type clayey soils with the presence of gravel, the bright brown slightly clayey soils with the abundance of gravel, and finally the hydromorphic soils.

The LD in the Mbandjock area as presented in Figure 9 varies between 0 and 4.03 km/km². The area is made up of minor lineaments (<3 km) and major lineaments (>3 km). It is classified into five classes. The fifth category (1.95–4.03) km/km² in red has the highest LD, and these are major lineaments. They have spread over most of the region. Zones with high density of lineaments are highly permeable, indicate good porosity, correspond to good recharge zones, and present good GWPZ (Poufone et al. 2022).

Table 5 presents the different classes of LD in the study area and the observations at different categories of classes.

The interpolation and reclassification of the different parameters of the GWPI were done. It was therefore imperative for us to determine the areas with strong potential for groundwater recharge. The production of the map of potentialities reviews the different levels of groundwater recharge by zone is presented in Figure 10.

The water potentiality map here shows that the eastern part of the study area presents a greater recharge susceptibility, given the cumulative nature of the high scores of the various parameters. The soil characteristics (sandy-clayey and hydromorphic) show a moderate level of percolation. The eastern part also has varying slopes from low to moderate, which implies a low movement of rainwater and also a longer infiltration time.

In Mbandjock, 31.35% of the study area has very high recharge potential followed by areas with high potentiality, which cover 15.15% of the study area, 8.81% medium potentiality, 43.12% low potentiality, and finally 1.57% with very low potentiality. Njumbe et al. (2023) did similar research in the determination of groundwater potential zones on the eastern slope of Mount Cameroon using geospatial techniques and seismoelectric method. They came to the same conclusion that groundwater recharge potential behaves in the same way depending on the importance of weights given to each parameter.

At the end of this work there was a question of using GIS as a decision-making tool to determine the potential for groundwater in the locality of Mbandjock in central Cameroon. The study proved to be conclusive and constitutes a capital study to propel the effective and efficient management of groundwater in the locality of Mbandjock. We can see from this study that by applying the AHP method to the different parameters of the GWPZ, the areas with low slopes, high rainfall intensities, high drainage density, and, in the end, sandy-loamy soils had a greater propensity to groundwater recharge. For this reason, local authorities and public authorities will be strongly advised to monitor for the location of future hydraulic structures (wells and boreholes) areas with high potential for groundwater recharge and to carry out more in-depth studies before any installation in areas with medium or low potential for refills.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.