Predicting the risk of arsenic contaminated groundwater in Shanxi Province, Northern China

Shanxi Province is one of the regions in northern China where endemic arsenicosis occurs. In this study, stepwise logistic regression was applied to analyze the statistical relationships of a dataset of arsenic (As) concentrations in groundwaters with some environmental explanatory parameters. Finally, a 2D spatial model showing the potential As-affected areas in this province was created. We identified topography, gravity, hydrologic parameters and remote sensing information as explanatory variables with high potential to predict high As risk areas. The model identifies correctly the already known endemic areas of arsenism. We estimate that the area at risk exceeding 10 μg L−1 As occupies approximately 8100 km2 in 30 counties in the province.

► We develop a statistical model to predict arsenic affected areas of Shanxi Province.
► Holocene sediments, TWI, Rivdist, Gravity, remote sensing images are key predictors.
► Area of 8112 km2 and more than 30 counties are estimated at risk of arsenic hazard.
► Logistic regression model could be widely used to predict other emerging regions.