Comparison of the performance of different interpolation methods in order to prepare groundwater level map

Groundwater is one of key water resources in arid and semi-arid countries. One of the most important requirements to undertake any planning and management in groundwater resources is obtaining accurate groundwater level map. The choice of interpolation method to prepare the groundwater map of a particular study area is not an easy task. In this study, spatial interpolation methods including Inverse Distance Weighting (IDW), Kriging, universal kriging, and Spline were applied to estimate the groundwater level at unknown points in an aquifer located in southern part of Tehran province, Iran. To choose the best method, data is divided into two groups: testing and training data. The result was validated with cross validation using different univariate statistical analysis for training and testing data sets to pick the best method. The evaluation of the results indicates that for the aquifer under study, the kriging with Spherical function for semivariogram model with the coefficient of determination 0.93 is the best method for surface interpolation of groundwater followed by universal kriging, kriging with other semivariogram model, IDW1, and Spline.