A combined GIS, remote sensing and geophysical methods for groundwater potential assessment of Ilora, Oyo central, Nigeria

The unavailability of water has been a major threat to agricultural and commercial activities in Ilora, Oyo central, Nigeria. This study combined GIS, remote sensing, and geophysical methods for groundwater potential zone mapping within the area in an attempt to identify prospective zones for borehole drilling activities. Landsat 8 OLI, ASTER DEM, geological, geophysical, and soil data of the research area were acquired for this study. To map the groundwater potential of the area, eight thematic layers that influence groundwater occurrences and movement controlling factors such as, geology, elevation, slope, land use land cover (LULC), fault proximity, soil, lineament density, and drainage density were mapped out from the acquired data of the area. The influence of each theme and subunit/class on groundwater recharge based on previous studies were evaluated using Analytical Hierarchical Process (AHP). The groundwater potential of the area of study was qualitatively classified into five classes, namely: very high, high, moderate, low, and very low which account for 0.3%, 7.8%, 54.8%, 35.6%, and 1.5% of the total area, respectively. The results were cross validated using well data from the area and 88.9% correlation and 0.74 kappa coefficient were achieved. The groundwater potential map generated in this research could be used as a preliminary reference in selecting suitable sites for groundwater resource exploitation in the area to ameliorate the current scarcity of water in this region.


Introduction
Water, for many years, has been one of the most important resources that humans need to survive. While it exists in abundance in some areas, it remains a scarce commodity in others due to various geogenic and anthropogenic factors. Nigeria, like many West African countries, has two seasons; the dry season in which rainfall is scarce and the rainy season in which rainfall is abundant. However, the duration of these seasons varies across cities in the country. Lack of rainfall results in exhaustion of surface water and consequently water scarcity. This has led to overreliance on groundwater for many domestic and agricultural activities in the country.
The geology of Nigeria is made up of three major geological components, they are; the basement complex, younger granites, and the sedimentary basins (Obaje 2009). The presence of favorable porosity and permeability within the sedimentary basins of Nigeria makes it easy for groundwater recharge and accumulation, hence groundwater occurs in appreciable quantity in areas located within the sedimentary basins. The basement complex geology is a bit more complex compared to the sedimentary terrains. Groundwater is hard to come by in these areas due to the impervious nature of the basement rocks.
Within the Basement complex terrain of Nigeria, groundwater is said to be abundant in areas where thick overburden overlies a fractured basement (Olasehinde 1999). Some of the problems encountered by groundwater development in many parts of Nigeria include the non-uniform, discontinuity, and complexities of the aquifer systems in the crystalline basement rocks which result in well/borehole dry ups (Hazell et al. 1992). Other obstacles encountered include lack of data, equipment, methods, and manpower which would have aided the identification of the promising high yield well/borehole sites (Boeckh 1992;Offodile 2002;Badamasi et al. 2016). Groundwater exploration in Nigeria is still traditional and improvement in the needed technology is slow-paced, therefore, there is a need for costeffective, accurate, and technologically appropriate solutions to groundwater exploration in the country (Tarhule 2007;Badamasi et. al. 2016).
Geographic information system and remote sensing on their own have proven to be a very strong tool in groundwater exploration. Various GIS techniques have been used for groundwater potential zone delineation. They have proven to be useful in Analytical Hierarchical Process (AHP) (Badamasi et al. 2016), Multi Influence Factor Analysis (Das et al. 2018), Fuzzy Logic Analysis (Tiwari et al. 2017), and Multi-criteria decision analysis (MCDA) (Senanayake et al. 2015;Hussein et al. 2017), etc.
Geophysics for many years has provided useful remote and ground-truthing techniques for groundwater studies. There are different types of geophysical methods; gravity, magnetic, seismic, electrical, radiometry, etc. The electrical resistivity method has been one of the most popular geophysical methods in groundwater studies due to its ability to identify groundwater signatures using contrasts in electrical resistivity of the subsurface materials (Loke, 2013). The magnetic method is probably the most versatile of all the geophysical methods (Dobrin and Savit 1988) because it can be applied to both shallow and deep-seated targets. The magnetic method has emerged as a powerful tool in groundwater prospecting in recent years due to its ability to reveal lineaments (folds, faults, joints, and fractured zone) which may serve as openings for groundwater accumulation and storage (Oni et al, 2020;Abdulkareem et al. 2018;Muthamilselvan et al. 2017;Al-Gharni 2005). The advancement in magnetic data acquisition and processing has made it easy for its usage in a GIS environment (Oni et al, 2020). Derivative maps of Aeromagnetic data can now be added as a thematic layer together with other layers derived from various datasets for a more accurate groundwater potential zone delineation.
The ability to integrate various datasets for groundwater studies will most likely yield a better result in delineating groundwater potential zones within the basement complex terrain of Nigeria. Research integrating geophysical, GIS and RS datasets are uncommon and as such, the purpose of this present study is to delineate groundwater potential zones in Ilora, Oyo central using advanced geospatial techniques.

Study area description
The present area of study falls in between two local governments of Oyo central senatorial district, Oyo west and Afijio, Oyo state, Nigeria (Fig. 1). It is situated between latitudes 7°46′ N to 7°55′ N and longitudes 3°45′ E to 3°55′ E covering an area of approximately 245 km 2 .
Mean annual rainfall in the study area ranges from 1117.1 to 1693.3 mm. The rainfall pattern in this area has a characteristic bimodal distribution that mostly peaks around June or July and September and a period of low precipitation occurs around August with 4 months of the dry season (December-March). The annual temperature ranges from 24.6 and 31.5 °C (NiMet 2019). Mean monthly relative humidity reaches a minimum of 52% in February and a maximum of 83% in August (Erakhrumen 2008).
The area of study is well-drained with rivers flowing from the upland to the low land in an east-west direction. Elevation ranges between 151 m in the west and 352 m above sea level in the east. The climate and vegetation of the area favor the cultivation of cash crops such as yam, beans, maize, cassava, plantain, and rice primarily but also some tree crops including cocoa, oil palm, cashew as well as a number of other cash crops. Occupation of the inhabitants of the study area is mostly farming which is supported by a number of government agricultural farm settlements in the area.
Geologically, the present area of study falls within the basement complex of southwestern Nigeria, basement rocks that underlie the Ilora and environs include migmatite, quartzite, and schists which are Precambrian in age. This geological formation is primarily highly impervious except where deeply weathered or fractured. Groundwater flow in hard rocks is usually controlled by factors such as fractures, joints, geological contacts, shear zones, faults, vugs, and other discontinuities, their multifaceted interrelationship controls overall aquifer dynamics (Singhal and Gupta 1999;Eaton et al. 2007).

Materials and methods
This study made use of both primary and secondary data. The primary data consists of Landsat 8 OLI, ASTER Digital Elevation Model data retrieved from glovis.usgs.gov, and aeromagnetic data of the area retrieved from the Nigerian Geological Survey Agency. The secondary data include geology, soil, and topographical maps of the study area derived from the Nigerian geological survey agency and Center for World Food Studies. ArcGIS, Rockworks, ENVI, and PCI Geomatica Software were used for data processing.

Data processing
The ASTER DEM data was further processed to derive Elevation, Drainage density, and Slope maps of the area of research, Landsat 8 data was used to generate Land Use Land Cover and Lineament density maps, while the acquired aeromagnetic data was processed and used to produce the first vertical derivative map of the area. The first vertical derivative map (FVD) is a map of subsurface lineaments that underlies the area of research, these lineaments could be faults, folds, joints/fractures, lithological boundaries, or rock contacts which may have been formed during orogenic activities, they are all significant in groundwater exploration studies. The FVD map was further used to produce the fault proximity map of the area which signifies closeness or farness of an area to a fault. Acquired geology and soil maps of the area were georeferenced and digitized while the topographical map of the area was used to digitize the major roads, footpaths, and settlements present in the area of research. Figure 2 shows the flow chart of the methodology employed in this research.

Thematic map preparation
Further, eight thematic layers were developed from the acquired maps (Elevation, Drainage Density, Slope, Land use Land cover, Geology, Soil, Lineament Density, Fault proximity) using various software/tools; fault proximity map was extracted from the first vertical derivative map of aeromagnetic data of the area of study using Oasis Montaj software, surface lineaments and land use/land cover maps were prepared from the Landsat 8 OLI/TRS data using ENVI, PCI Geomatica, ArcGIS software, and Rockworks. Maximum likelihood classification was employed to produce a land use/land cover map of the research area. Elevation, drainage, and slope maps were extracted via Archydro tools in ArcGIS. To derive the thematic maps from the secondary data, hard copy maps (geology and soil) were scanned and imported into the ArcGIS software then georeferenced to World Geodetic System (WGS 84) coordinate system.

Analytical Hierarchical Process
The weighting of the thematic maps was carried out using the Analytical Hierarchical Process (AHP). Multi-criteria decision analysis using Analytical Hierarchical Process (AHP) is one of the most common and well-known GISbased methods for delineating groundwater potential zones (Arulbalaji et al. 2019). This method helps in integrating all thematic layers. 8 different thematic layers were used for this research, the thematic layers include slope, elevation, drainage density (DD), lineament density (LD), land use land cover (LULC), geology, fault proximity, and soil. The 8 thematic layers largely influence the flow and storage of water in the area. Weight assignment to these factors is based on groundwater occurrence within the basement complex terrain of Nigeria and literature review of several researchers (Yeh et al. 2016;Bayowa et al. 2014;Fashae et al. 2014;Nampak et al. 2014;Rahmati et al. 2015;Olasehinde 1999;Olorunfemi and Fasuyi 1993). A parameter with a high weight illustrates a layer with high impact and a parameter with a low weight illustrates a small impact on groundwater potential. The weights of each parameter were assigned according to Saaty's scale (1-9) of relative importance value. Further, the weights were assigned with consideration of the review of past studies and field experience. The Saaty's scale of relative importance value reveals that the value of 9 indicates extreme importance, 8 very strong, 7 very to extreme importance, 6 strong plus, 5 strong importance, 4 moderate plus, 3 moderate importance, 2 weak, and 1 equal importance (Saaty 1990). As per the classification, weights were assigned to the thematic layers based on their importance and water holding capacity.
The mapping of the groundwater potentials of the study area was done by the weighted index overlay method in Arc-GIS 10.8. Weight assignment was done by assigning the new weight values to the maps' sub-units (sub-criteria) computed from the AHP. The reclass tool in the spatial analyst tool of the ArcGIS 10.8 was used for this task. Groundwater potential zones map of the area of research was produced by overlaying all the thematic layers using the weighted index overlay tool. Figure 3 shows the flow chart of the methodology for groundwater potential zone mapping of Ilora and environs.
where; Wi = % weight for each thematic map. Xj = reclassified map.

Accuracy Assessment and validation
To validate the accuracy of the groundwater potential that would be generated in this research, static water levels and well yield data were collected across drilled wells in the area. The yield of the wells would be used to validate the groundwater potential map while an error matrix would be used for accuracy assessment. An error matrix is a square array of numbers set out in rows and columns which expresses the number of sample units (i.e., pixels, clusters of pixels, or polygons) assigned to a particular category relative to the actual category as verified by some reference data. The columns usually represent the reference data while the rows indicate the classification generated from the remotely sensed data. In other words, an error matrix is a comparison between sampled areas on the map generated from the remotely sensed data and those same areas as determined by WiXj. Fig. 2 Flow chart of method used for groundwater potential zone mapping of Ilora and environs some reference data (Cohen 1960;Congalton 1991;Congalton and Green 2008). An error matrix is not only used for accuracy assessment but also helps to enhance the accuracy of the output map (Doke et al. 2021). Accuracy was assessed using the following formula Congalton and Green (2008); where N is the total points, k is the number of classes, R is the test classes, and P is the classified class. Figure 3 shows the eight thematic maps produced for groundwater potential mapping of Ilora, each of these factors plays a crucial role in the availability of groundwater in an area. The geology of an area is very vital in the occurrence and distribution of groundwater in that area (Yeh et al. 2016). The geology controls the infiltration rate and flow of water into and out of rock formations. The published geological map of Oyo state (NGSA 2006) by the Nigerian Geological Survey Agency (NGSA) was used for delineating different geological units in the research area (Fig. 3a). The present area of study is situated within the Precambrian Basement Complex terrain of Nigeria as earlier stated, migmatite cover approximately 77% of the total area while banded gneiss and quartzite and schist occupy 16% and 5% respectively (Table 1). Based on evidence from drilled wells, quartzite and schist are usually highly fractured and tend to

Results and discussion
, play a crucial role in groundwater infiltration and as such as been assigned high weight followed by the migmatite. Soil types play a vital role in the amount of water that infiltrates subsurface formations and hence has a high influence on groundwater recharge (De Reu et al. 2013). Soil texture and hydraulic characteristics are the main factors considered for the estimation of the rate of infiltration. The soil map of the area of research is displayed in Fig. 3b. About 94% of the area of study is covered by Lixisol/Luvisols soils, while the remaining 6% is covered by Luvisol (Table 1). The details of the soil categories identified in this study can be found in the Center for World Food Studies (SOW-UV) (1997) soil map of Nigeria.
Lineaments are structurally controlled linear or curvilinear features on a map which can be identified from satellite imagery by their relatively linear alignments (Nag and Kundu 2016). Linear structures can be surface faults, folds, joints/fractures, lithological boundary, or rock contacts which may have been produced during orogenic activities or during various episodes of deformation that affected the basement complex terrain of Nigeria. They are quite significant in groundwater prospecting because their presence could lead to an increase in secondary porosity and permeability which are significant in groundwater potential zone mapping. Surface lineaments across the area of research were extracted from band 7 of Landsat 8 data automatically. The lineament density map was then prepared using line density in GIS software and is depicted in Fig. 4a. By carefully examining the values obtained, the data were reclassified into five categories: -Very low (0-0.73 km/km 2 ), Low (0.73-1.47 km/km 2 ), Moderate (1.47-2.2 km/km 2 ), High (2.2-2.94 km/km 2 ) and Very high (2.94-3.67 km/km 2 )  (Table 1). Groundwater potential decreases with increasing distance from the lineaments (Arulbalaji and Gurugnanam. 2016), therefore, high weight is assigned for high density and low weight for low-density classes.
Subsurface faults present in the area of study were extracted from the first vertical derivative image of the area of study using Oasis Montaj software. Faulting and fracturing within the basement complex terrain of Nigeria cause increased secondary porosity and permeability thereby affecting groundwater potential. Fault proximity refers to how close an area is to faults. It is revealed that the intensity of groundwater potential decreases with increasing distance from these faults. 0-199 m distance represents closeness to faults while 799-999 m represents far proximity (Fig. 4b). Therefore, closer areas were ranked higher than far areas (Table 1).
Drainage density is an important factor in groundwater recharge and accumulation. Drainage network in an area depends on the geology of the area and it provides an important index of infiltration rate. It is an inverse function of permeability which suggests it is an important parameter in the delineation of the groundwater potential zone. Drainage density is obtained by dividing the total length of all the rivers in a drainage basin by the total area of the drainage basin  (Yeh et al. 2016). High drainage density is a product of less infiltration and as a result, has little influence on the groundwater potential of the area. Low drainage density represents high infiltration and hence contributes more to the groundwater potential (Arulbalaji and Gurugnanam 2016). Drainage density within the study area (Fig. 5a) was reclassified and categorized as Very low (450.16-562.7 km/km 2 ), Low (337.62-450.16 km/km 2 ), Moderate (225.08-337.62 km/ km 2 ), High (112.54-225.08 km/ km 2 ) and Very high (0-112.54 km/km 2 ). For groundwater potential zonation, the high weight is assigned for low density, and the low weight is assigned for high density (Table 1). LULC provides important information on infiltration, soil moisture, groundwater, surface water, etc., in addition to indicating groundwater requirements. Ilora and environs exhibit a spectrum of land use categories which include: agricultural land, thick forest, rocks exposures, built-up areas, and water bodies (Fig. 5b). The LULC types in the area are delineated from Landsat 8 satellite images by using the maximum likelihood classification method. Out of the various class types, farmlands occupy the largest portion of the total landmass followed by thick forest. The LULC classes like forest and agriculture land hold a substantially high proportion of water than the built-up land and open land surfaces (Rajaveni et al. 2017). Therefore, high weight is assigned to the thick forest and farmlands while low weight is assigned to the built-up and rock exposures (Table 1).
Topography plays an important role in groundwater flow and accumulation as well as groundwater recharge, topographic factors have a direct effect on flow size and run-off  (Kia et al. 2012). Elevation refers to the height of an object above sea level. The processed ASTER DEM data of the study area shows that elevation values ranged from 151 to 352 m. The low elevation is much more significant in groundwater development as surface water flows from areas of higher elevation to areas of lower elevation. From the elevation map of the study area (Fig. 6a), low elevation values cover a total of 1.4% while high elevation covers 1% of the area. Intermediate values make up the remaining 49%. Higher values were assigned to low elevation while lower values were assigned to high elevation. The slope is an important terrain characteristic that gives an insight into the steepness of the ground surface of an area. A slope map provides valuable information on the nature of the geologic and geodynamic processes operating at the regional scale of an area (Riley 1999). Surface run-off and rate of infiltration are influenced essentially by the slope of the surface. Larger slopes produce smaller recharge because the water received from precipitation flows rapidly down a steep slope during rainfall, therefore, it does not have sufficient residence time to infiltrate and recharge the saturated zone (De Reu 2013). Figure 6b shows the slope map of the studied area. The slope values were reclassified and categorized into five classes such as flat (0-3.57), gentle (3.57-6.23), medium (6.23-9.28), steep (9.28-13.39), and very steep (13.39-33.8). High weight is assigned to flat and gentle slopes while low weight is assigned to steep and very steep slope (Table 1).

Groundwater potential zone and validation
Within the southwestern basement complex terrain of Ilora and environs, it was imperative to delineate the groundwater potential zones of the area due to massive water shortages and overreliance on the usage of rainfall and stream water for domestic and agricultural purposes. The groundwater potential zones in the area were delineated using the AHP technique, which is one of the effective techniques for groundwater potential zonation. The eight (8) thematic layers (Elevation, Slope, Soil, Geology, Lineament density, Fault proximity, Drainage Density, Land use land cover) were used to classify the study area into various groundwater potential zones. Figure 7 shows the groundwater potential zones in the area of study. The output groundwater potential zone map was further classified into five zones for easy understanding, they are: very low, low, moderate, high, and very high. From Table 2, a very high groundwater potential zone covers a total of 1.5% of the total area, high groundwater potential zone covers 35.6%, 54.8 by moderate zone, 7.8% for low, and 0.3% for areas that have a very low groundwater potential.
High groundwater potential zones in the study area are typified by low elevation, low slope angles, low drainage densities, high lineament densities, closeness to faults, forest and farmlands, and the presence of fractured and permeable soils. These appear in a large part of the area of study apart from the southeastern part of the area of research which was classified as a built-up area. In contrast, low groundwater potential zones in the research area are typified by high elevation, high slope, built-up areas, lack of permeable soils and fractured rocks, farness to faults, high drainage density, and little or no surface lineaments. Moderate groundwater potential zones dominate the area of research and the study area can be classified as a moderate to high groundwater potential area.
The groundwater potential zones delineated in this research were correlated with static water levels and well yield data from observation wells across the study area  Fig. 8). It is revealed that wells sited in the very high and high groundwater potential zones have static water level values in the range of 6 to 7 and water yielding capacity in the range of 70-120 L per minute (LPM), wells located in moderate groundwater potential zones have static water level values that ranged from 4 to 5 and water yielding capacity in the range of 30-70 LPM, while wells sited in very low and low groundwater potential zones have static water level values that ranged from 1 to 3 and water yielding capacity within 5-30 LPM. Ilora town recorded low static water levels together with low good yield and therefore fell within the low and very low areas of the groundwater potential zone map. Also, the majority of the farmland in the area of study has moderate to high static water levels and well yield values and as such displayed moderate to high groundwater potential.
From the thirty-six (36) observatory wells (Table 3), thirtythree (33) of the obtained values agree with the groundwater potential zone map of Ilora while the other three (L19, L30, L32) which fell within the high potential zone region of the map have moderate groundwater potential values.
To access the accuracy of the groundwater potential zone map produced in the study, the location of the observatory wells used in validation was correlated with the produced groundwater potential zone map generated by the AHP technique. The groundwater potential map was reclassified into three where very low and low potentials are classified as low and assigned an accurate value of one (1), moderate as moderate (2), and high and very high as high with an accurate value of three (3). The observed well data were also reclassified into three; with 1-3 static water levels and 5-30 LPM classified as one (1), 4-5 static water levels, and 30-70 LPM classified as two (2), while 6-7 static water levels and 70-120 LPM were categorized as three (3). The overall accuracy was computed using the following formula (from Jensen 1996): Overall accuracy = Number of correct Observation well location Total number of Observation well location = 32 36 = 88.9%.  Kappa (k) analysis is a multivariate method for assessing precisions and it provides a Khat statistic that represents a degree of accuracy. It can be calculated using the following formula (Usman et al. 2015): Table 4 shows the error matrix result of the correlation and the accuracy between the groundwater potential map and the observatory wells can be calculated to be 88.9% which shows a high correlation. Kappa (K) coefficient was calculated to be 0.74 which further established the level of accuracy between the groundwater potential map and the observation wells. This further confirms the relevance of the AHP technique in groundwater zonation investigations.

Conclusion
A proper evaluation of the groundwater potential of an area is very vital especially in the aspect of planning and sustainable development. Such information is priceless during the design and implementation of structures that can be used to make potable water available in abundance for domestic and agricultural usages. This study attempts to delineate the groundwater potential zones within Ilora and environs using a combination of GIS, Remote sensing, and geophysical techniques. Eight thematic layer maps such as Land Use/Land Cover, Geology, Lineament Density, Slope, Drainage Density, Elevation, fault proximity, and Soil were used in this study to map out the groundwater potential zones in the research area using AHP.
From the generated Groundwater Potential Zone map, the research area was classified into five groundwater potential zones namely; Very high, High, Moderate, Low, and Very low. Very high and high groundwater potential zones are predominantly located within the central part of the study area which is occupied by farmland and thick forest while very poor dominates the southeastern part of the area. The low and very low zones are the densely populated part of the study area. Moderate groundwater potential zone is present in all parts of the area, covering a total of 54.8% of the area of study. Very high and high groundwater potential zone covered a total of 37.1% while low and very low covered approximately 1.1%.
The delineated groundwater potential zones map of the area of study was further cross validated with observatory well data collected from the area. The data included static   water levels and well yield data. There is a good agreement between the observatory well data and the various zones on the groundwater potential zone map with an accuracy assessment value of 88.9% and Kappa coefficient value of 0.74. Therefore, the groundwater potential zone map produced in this research will assist decision-makers in proper planning and management of groundwater resources to efficiently put a lasting end to the current scarcity of water that is presently affecting Ilora and environs. It will be a huge plus for borehole/well drillers if groundwater potential maps such as the one in this study are consulted before embarking on drilling activities to save capital and valuable exploration time.