Groundwater quality assessment of Birjand plain aquifer using kriging estimation and sequential Gaussian simulation methods

Abstract

The performance of geostatistical estimation methods of ordinary kriging and sequential Gaussian simulation have been investigated in this study to evaluate the qualitative characteristics of groundwater in Birjand plain. For that purpose, 81 water samples were collected from groundwater wells and thirteen hydro-chemical parameters of calcium (Ca), chlorine (Cl), electrical conductivity (EC), bicarbonate ( HCO3 −), magnesium (Mg), sodium percent (Na%), sodium (Na), sodium absorption ratio (SAR), sulfate ( SO4 2−), total dissolved solids (TDS) and total hardness (TH) were analyzed and interpreted. Variography of the variables was performed after the normalization and experimental variograms were plotted in GS+. The best theoretical models were then fitted to the experimental variograms. Validation of variograms was performed by two methods of cross-validation and residual analysis. Geostatistical estimation maps were then prepared for each groundwater variable. Since the smoothing effect is one of the major drawbacks of ordinary kriging estimations, the sequential Gaussian simulation method was also used to prepare the simulation maps of variables. The simulation results were more reliable than those obtained by ordinary kriging.