Spatial prediction of chloride concentration in Azarshahr plain aquifer-Iran,using EC, Ca   and Mg   as auxiliary co-kriging variables

For a less densely sampled Area, Lognormal ordinary cokriging (LnOCK) with auxiliary variables can sometimes improve estimates. In this study for groundwater quality assessment of Azarshahr plain aquifer- East Azerbaijan province- Iran (one of the Uromia lake subbasins) 39 samples have been gathered. Due to slight sampleaccumulation, geostatistics was utilized for accuracy rising in Chloride concentration prediction of study area. For this purpose, three stepswere designed; At first, spatial concentration of chloride has modelled by Lognormal ordinary kriging (LnOK), then three differentcovariant (EC, Ca   and Mg   ) that their quantities had more than 90% correlation to chloride, has been chosen for spatial prediction of its concentration separately and in third step EC and Ca   have used together as covariates to evaluate the spatial prediction. Outcomes have shown that Lognormal ordinary cokriging (LnOCK) using Ca   as an auxiliary covariant reveals more efficient results; With Mean error about 0.04 and RMSE about 0.26. Whereas adding more data set as excess covariant (Ca  and EC, together) reduced the model precision. Drown maps finally showed that Chloride concentration rises from the South-East to North-West in study area

کلیدواژه ها:

Cokriging ، ، Kriging ، Spatial prediction ، Azarshahr ، ، Iran ،