Estimating the Likelihod of MTBE Occurrence in Drinking Water Supplied by Ground-Water Sources in the Northeast and Mid-Atlantic Regions of the United States.

摘要

A multivariate logistic regression model was developed to help explain the presence or absence of methyl tert-butyl ether (MTBE) at concentrations equal to or exceeding 0.5 micrograms per liter in ground water from wells in parts of the Northeast and Mid-Atlantic regions of the Nation that are used primarily for drinking water. The model was developed from a calibration data set consisting of information on MTBE in water samples collected from 1,042 wells. MTBE was detected at concentrations equal to or exceeding 0.5 micrograms per liter in 104 samples, but at concentrations equal to or exceeding 5.0 micrograms per liter in only 14 samples. Thus, the model was developed to predict the occurrence of MTBE at greater than or equal to 0.5 micrograms per liter. A number of factors that describe the conditions in the vicinity of the well were related to the frequency of detection of MTBE but three factors, or variables, most effectively explain MTBE occurrence in a multivariate logistic regression model: MTBE use in gasoline in the study area, the density of above and underground storage tanks, and a soil erodability factor.