Modelling hydrogeological parameters to assess groundwater pollution and vulnerability in Kashan aquifer: Novel calibration-validation of multivariate statistical methods and human health risk considerations.

Modeling and Investigation on pollution potential of aquifers is a matter of importance in terms of management, development and land-use allocation as well as quality monitoring, pollution prevention and groundwater protection. The purpose of this study is to calibrate and modeling the methods used for pollution potential assessment, in which the impact and apportionment of hydrogeological parameters on groundwater pollution of an aquifer located in Kashan is considered. To do so, Analytic Hierarchy Process (AHP) and fuzzy-statistical analysis methods are used for weighting, ranking and standardize the parameters based on research awards of experts and Ad-Hoc systems. This was performed in such a way that the level and importance of each class of classification parameters is considered equal to the final model, and is equivalent to the reclassified class of indices of groundwater quality and human health risk to nitrate pollution. After ranking and standardizing the parameters as well as final model by using fuzzy-statistical approach, the process of weighting the parameters is accomplished with aid of AHP and Factor Analysis-weighted Principal Component Analysis (FA-PCA) methods based on their apportion and impact on groundwater pollution in addition to their correlation with nitrate map. In addition to the correlation with the standardized nitrate concentration, techniques of the root mean square error (RMSE) and coefficient of variation (COV) are employed to validate the model. The results illustrated that parameters of net recharge, soil media, impact of vadose zone, hydraulic conductivity and aquifer media have created the highest apportion and impact on groundwater pollution. In addition, it was found that weight of these parameters for calibrated and validated model of GPPI (groundwater pollution potential index) is proved to be 8.5, 3.4, 3.3, 2.6, 2.2, for GPRI (groundwater pollution risk index) as the best model is 4.4, 3.7, 3.1, 2.9, 1.1, and it is 4.8 for the land use layer, respectively. Weighting procedure was conducted by FA-PCA approach and following considerations were used; R = 73, RMSE = 1.08 and COV = 20%. Moreover, based on these models with better calibration-validation than generic model, it was found high pollution potential in western margin, high pollution risk in the central parts to the western margin, while it was observed not to have that very high pollution potential and risk in Kashan aquifer.