Yang Qiang author Columbia University. Lamont-Doherty Earth Observatory Jung Hun Bok author Marvinney Robert G. author Culbertson Charles W. author Zheng Yan author Columbia University. Lamont-Doherty Earth Observatory Columbia University. Lamont-Doherty Earth Observatory originator text Articles 2012 English A high percentage (31%) of groundwater samples from bedrock aquifers in the greater Augusta area, Maine was found to contain greater than 10 μg L–1 of arsenic. Elevated arsenic concentrations are associated with bedrock geology, and more frequently observed in samples with high pH, low dissolved oxygen, and low nitrate. These associations were quantitatively compared by statistical analysis. Stepwise logistic regression models using bedrock geology and/or water chemistry parameters are developed and tested with external data sets to explore the feasibility of predicting groundwater arsenic occurrence rates (the percentages of arsenic concentrations higher than 10 μg L–1) in bedrock aquifers. Despite the under-prediction of high arsenic occurrence rates, models including groundwater geochemistry parameters predict arsenic occurrence rates better than those with bedrock geology only. Such simple models with very few parameters can be applied to obtain a preliminary arsenic risk assessment in bedrock aquifers at local to intermediate scales at other localities with similar geology. Supporting material available at http://hdl.handle.net/10022/AC:P:12894. Environmental science 46 4 2080 2087 2012-02-21 http://dx.doi.org/10.1021/es203793x http://hdl.handle.net/10022/AC:P:13311 NNC NNC 2012-05-23 16:17:48 -0400 2013-02-18 15:34:05 -0500 7316 eng