GIS-Based Bivariate Statistical Model Prediction of Groundwater Potential Mapping for Sustainable Developments in Suez Governorate, Egypt

Abstract

One of the primary challenges for sustainable development in semi-arid regions like Egypt, is the scarcity of freshwater, making it critical to assess groundwater potential. The purpose of the current study is to predict spatially potential groundwater zones in Suez Governorate (SG), Egypt using (relative frequency prediction rate) integration and Shannon entropy (SE) bivariate statistical models. Sixteen causal factors affecting groundwater instances were assessed in terms of geo-environmental. The results obtained from the current study revealed that these two models can be effectively working for spatial prediction modeling. Furthermore, the RF-PR model results have shown that most paramount factors in groundwater instances in study region were observed in soil units, depth to water table, LU/LC and drainage density whereas SE model reflects LU/LC, lithology, Distance to stream, soil units, and depth to water table respectively. Following by validation analysis of AUCs for both relative frequency-prediction rate and Shannon's models are 0.749 and 0.745, correspondingly, representing that RF-PR outperforms the Shannon's. Finally, groundwater potential zones prediction maps (GPZPm) obtained from both models were categorized into five classes. Current research results are useful for multi-criteria decision makers such as water resources authorities and decision architects to broadly assess the groundwater investigation for future planning

Keywords

• relative frequency
• predictor rate
• shannon entropy
• multi-criteria decision makers
• sdgs