DEVELOPMENT OF GROUNDWATER MANAGEMENT DECISION SUPPORT SYSTEM MODELS USING OBJECT-ORIENTED ALGORITHMS IN A BASEMENT COMPLEX AREA

Abstract

The study of the groundwater potential and its vulnerability using object-driven multi-criteria approach was carried out in Akoko and Ose in Basement Complex area of Ondo state, Nigeria. Surface and sub-surface data which include remote sensing data, ancillary data, climate data, geological data, and geophysical data were used in the study. One hundred and twenty-three (123) vertical electrical soundings (VES) data were acquired over three (3) major underlying geologic rock types: migmatite gneiss, quartzite schist, and granite gneiss. The VES results gave the subsurface lithological layers’ resistivity and thickness which were used for assessing the area hydrogeological conditions. The acquired remote sensing data including Landsat image and digital elevation model (DEM) were processed to generate thematic maps viz: lineament density, land use, and slope percent. Ancillary (surface geology map) and selected climatic data were processed to produce lithology and rainfall maps, respectively. Based on the analyzed data, potentiality conditioning parameters (PCP) namely: lithology (Li), hydraulic conductivity (K), lineament density (Ld), transmissivity (T), and transverse resistance (TR) were derived for groundwater potentiality mapping while rainfall (R), land use (LU), bedrock topography (BT), recharge rate (Re), and slope percent (S) factors were considered as the vulnerability conditioning parameters (VCP). These parameters were objectively weighted with a python-based Entropy method of weightage determination approach. The weightage results established highest and lowest assigned values of 0.6 and 0.03 for Ld and K whereas 0.78 (highest) and 0.01 (lowest) for S and R, respectively. Hybridizing the Entropy(E)-based weightage results with PROMETHEE-II and TOPSIS algorithm gave birth to PROMETHEE-Entropy and TOPSIS-Entropy data mining algorithms. The results of the applied PROMETHEE-Entropy and TOPSIS-Entropy algorithms to the gridded rasters of PCP and VCP themes, produced the Groundwater Potentiality model (GPM) and Groundwater vulnerability model (GVM) maps of the area, respectively. Furthermore, the algorithms of the conventional PROMETHEE-II and TOPSIS were applied to produce different set of GPM and GVM maps, respectively. The validated results of the produced GPM maps employing the Receiver Operating Characteristic (ROC) curve technique produced accuracies of 86 % and 67 % based on PROMETHEE-Entropy and conventional PROMETHEE-II algorithms, respectively. The PROMETHEE-Entropy – based GPM map established about 19 % better accuracy more than the PROMETHEE-II – based GPM map of the area. Moreover, the TOPSIS-Entropy based GVM map accuracy result established about 22 % better reliability and accurate more than the conventional TOPSIS modeling result. With the adopted diverse methodologies, multiple decision-making models that are very relevant to groundwater resources exploration and management were developed. The developed models are viable constituent for producing groundwater hydrological database that can be explore for cost effective groundwater resources exploitation and management in the study area and the country at large.