Evaluation of different geostatistical methods based on GIS for predicting and zoning salinity invasion of aquifers of West Azarbaijan

In coastal aquifers, infiltration and internal extending seawater is the most common reason for the increased salinity. Locating adjacent to the hypersaline lake and a sharp drop in groundwater in the region has exposed West Azerbaijan Province to salinity invasion. In this study, geostatistical methods with various models were used to evaluate the effects of the salinity invasion toward the groundwater aquifers in West Azerbaijan. In order to analyze the interpolation methods, the kriging, cokriging, and inverse distance weight (IDW) with weighted variables and without weighted variables were used. To assess the accuracy of the developed models of geostatistical estimator, the coefficient of determination (R2), root square error (RMSE), and mean bias error (MBE) were used. Based on the results of this study, it was found that spherical kriging model, the spherical cokrigings model with Ca auxiliary, Model 2 in IDW method and Model 2 with HCO3 as variable weighted in IDW method were selected as the best models. Among the methods used for interpolation of salinity invasion in the study area, cokriging method showed the lowest error. Comparison of the zoning maps of salinity invasion showed that IDW model was sensitive to minimum and maximum values ​​ and showed those values ​​in wider area in the study area. However, kriging and cokriging models are more accurate and perform the zoning with more accuracy.