Mapping Groundwater Spring Occurrence Potential using Logistic Regression Statistical Method (Case Study: Miredeh Mountainous Watershed, Kurdistan)

The purpose of this study is to produce a groundwater spring potential map based on statistical and probability logistic regression method. The spatial locations of the springs in mountainous Miradeh watershed, Kurdistan province, with 381 Km2 were determined. Based on field and local study and Water Resources Management Company, spatial location of 904 springs in the study area were registered and 304 springs selected randomly as test data set and the remaining 600 spring were used for training the network. Forward stepwise LR was used to analyze the data and to find the most influential factors on spring occurrence that are statistically significant. The accuracy of model evaluates and verified using statistics such as: -2LL, Cox, Snell and Nagelkerke R2. The accuracy of each model to predict was evaluated using calculating the area under the ROC curve values. ArcGIS was used to apply the proposed method and calculate statistical and Probability of spring occurrence potential. This can help planners and engineers in groundwater development plans and land-use planning.