Modeling and study of the mechanism of mobilization of arsenic contamination in the groundwater of Nepal in South Asia

A study on arsenic (As) contamination in the groundwater of the Gaur Municipality, Rautahat district, Nepal was undertaken. We have designed a model, by backward stepwise regression analysis, to predict As concentration (R ² = 67 %) when information on some covariates such as nitrate, bicarbonate, and total iron (Fe) is available. The model was applied to Nepal groundwater and verified for Bangladesh groundwater. This multivariate model may be used to predict As concentration in given anions and cations. The analysis of variance test showed that the model designed is statistically significant (p < 0.001). Among all the covariates, Fe content showed the strongest positive correlation with As. The multivariate model for predicting the As from the existing three explanatory variables (NO₃, HCO₃, and Fe) for Nepal groundwater is found as: \( {\hat{\text{A}}\text{s}} = 19.349 - 0.221\;{\text{NO}}_{3} - 0.041\; {\text{HCO}}_{3} + 3.274.50 \; {\text{Fe with }}R^{2} = 67\;\% , \) where As in μg/L, Nitrate bicarbonate, and Fe in mg/L. The release of As from the soil was found according to the Fe reduction hypothesis. The As concentration varied from 1 to 62 μg/L in the groundwater sample (n = 50) of both the private and public shallow tube-wells. Among analyzed sample units, one of the tube-well’s As concentration exceeded 50 μg/L, 18 were between 10 and 50 μg/L, and the rest (31) were below 10 μg/L.