Fusion of GIS, remote sensing, geophysics and Dempster Shafer theory of evidence for mapping groundwater prospectivity: A case study of the central parts of Lagos State, Nigeria

Abstract

Water utilization for different human activities is universally crucial, but it is not readily available for consumption in some areas, such as the central parts of Lagos State, Nigeria. Moreover, there are many groundwater controlling factors (GWCFs) spanning geological, geophysical, and hydrological factors that contribute to this scenario, coupled with uncertainties that necessitate their careful selection using appropriate modelling techniques. In this study, the Dempster–Shafer Theory of Evidential Belief Function (DST-EBF) model (a coupling approach) was deployed to produce a groundwater prospectivity zonation (GWPZ) map for the study areas. The choice of the selected locations was informed by the continued dearth of water supplies, necessitating the need to discover new locations or re-appraise the existing ones for groundwater resource development. Furthermore, remote sensing, geological, field geophysical, and hydrological datasets that constituted GWCFs were integrated into a Geographic Information System (GIS) environment. Next, the computed values for the evidence of the mass functions (i.e., belief, disbelief, uncertainty, and plausibility) were combined using the Dempster–Shafer combination rule and then interpolated using the Inverse Distance Weighted (IDW) method. Subsequently, the GWPZ map was generated and classified into five zones, ranging from very low to high prospectivity zones. The GWPZ map was validated using cross-validation to estimate statistical errors, the receiver operating characteristic (ROC) curve, and the use of inverted resistivity models from the 2D electrical resistivity imaging surveys. All the metrics used for the validation provided good account for the classified GWPZ map. The north-eastern and the south-central parts are the most promising regions for groundwater, which could be harnessed for sustainable development through borehole drilling. Thus, the integration of the DST-EBF model and GIS for effective groundwater resources and uncertainty mapping was quite successful and impressively reliable.