Identification of hotspots and cold-spots of groundwater potential using spatial statistics

The Guna-Tana landscape is located in Ethiopia. This landscape is seriously facing water scarcity problems, that’s why we studied this landscape and provided the hotspots of groundwater potential areas in this region. In this study the hotspots and cold-spots of groundwater potential at, 99, 95, and 90 % confidence levels has been deciphered. Using Gi-Bin values, four classes has been identified viz., 2–3 (highly favorable), 0–1 (fairly favorable), −2 to −1 (poorly favorable) and −3 (very poorly favorable). The hotspots was subjected to ordinary least squared (OLS) regression to understand the impact of chosen parameters (viz., geology, land-use, soil, rainfall, slope, and distance to rivers) towards groundwater potential. The absence of redundancy among the selected parameters was indicated by the VIF values of the parameters, which were determined to be less than 7.5. It was discovered that the Robust Probability (Robust_Pr) was statistically significant (p < 0.01). The OLS model appears to have captured the variability of exploratory variables, as evidenced by the decreased values of Akaike's Information Criterion (AICc). The Adjusted R-squared value of 0.9119 indicates that exploratory variables has successfully explained 91.19 % of the variance of the model. Figure a) Hotspots of groundwater potential at 99 %, 95 %, and 90 % confidence, b) Groundwater potential with respect to Gi-Bin values, and c) Distribution of standardized residuals • Analyzed hotspots and cold-spots of groundwater potential. • Gi-Bin values were classified based on natural breaks. • Explained 91.19 % of the variance in groundwater potential using the model. • GWR was applied to assess coefficient and residual distribution.