Comparison of performance of WIOA, FR, SI, and MIF for groundwater recharge identification: an integrated framework

Identifying areas with high potential for groundwater recharge has become the most important factor for groundwater resource management. The potential for groundwater recharge depends on the different geo-environmental characteristics of the surface, sub-surface of ground, and meteorological parameters. In this study, it is tried to identify areas with the potential for groundwater recharge of the Madian River basin. Using the weighted index overlay analysis (WIOA), multi-influence factor (MIF), frequency ratio (FR), and statistical index (SI), performance comparison of models is conducted. The parameters used in this study include soil, geomorphology, drainage density, groundwater level, elevation map, slope, geology, and land use. The relationship between these parameters was determined based on the effect of each parameter, and accordingly, each layer was scored and weighted; then, the MIF model was implemented by overlapping layers. WIOA model was carried out based on the same theory, with the exception that each class is scored in the range of 1 to 9. The models of SI and FR were carried out based on the number of wells with discharges above 10 cubic meters per hour in the region, so that, they were randomly divided to 70% as training wells ( n = 106) and 30% as validation wells ( n = 45), the necessary analyses were performed on class and layers based on the theory of models, and finally, the map of groundwater recharge potential of two models of SI and FR were produced. ROC curve method was used to evaluate the performance of the models and the AUC results showed SI model with 99.6% AUC has the maximum efficiency compared to other used models. Models of FR, WIOA, and MIF with AUC of 99%, 94.6%, and 89.8%, respectively, are next in rank in terms of performance.