Methyl tert-butyl ether (MTBE) was detected in source water used by 8.7 percent of randomly selected community water systems (CWSs) in the United States at concentrations that ranged from 0.2 to 20 micrograms per liter (μg/L). The Random Survey conducted by the U.S. Geological Survey, in cooperation with the Metropolitan Water District of Southern California and the Oregon Health & Science University, was designed to provide an assessment of the frequency of detection, concentration, and distribution of MTBE, three other ether gasoline oxygenates, and 62 other volatile organic compounds (VOCs) in ground- and surface-water sources used for drinking-water supplies. The Random Survey was the first of two components of a national assessment of the quality of source water supplying CWSs sponsored by the American Water Works Association Research Foundation. A total of 954 CWSs were selected for VOC sampling from the population of nearly 47,000 active, self-supplied CWSs in all 50 States, Native American Lands, and Puerto Rico based on a statistical design that stratified on CWS size (population served), type of source water (ground and surface water), and geographic distribution (State).

At a reporting level of 0.2 μg/L, VOCs were detected in 27 percent of source-water samples collected from May 3, 1999 through October 23, 2000. Chloroform (in 13 percent of samples) was the most frequently detected of 42 VOCs present in the source-water samples, followed by MTBE. VOC concentrations were generally less than 10 μg/L 95 percent of the 530 detections and 63 percent were less than 1.0 μg/L. Concentrations of 1,1-dichloroethene, tetrachloroethene, trichloroethene, vinyl chloride, and total trihalomethanes (TTHMs), however, exceeded drinking-water regulations in eight samples.

Detections of most VOCs were more frequent in surface-water sources than in ground-water sources, with gasoline compounds collectively and MTBE individually detected significantly more often in surface water. Use of personal and commercial motorized watercraft on surface-water bodies that are drinking-water sources is probably the reason for the elevated detections of gasoline contaminants relative to ground water. MTBE detections demonstrated a seasonal pattern with more frequent detections in surface water in summer months, which is consistent with seasonal watercraft use.

The detection frequency of most VOCs was significantly related to urban land use and population density. Detections of any VOC, non-trihalo-methane compounds, gasoline compounds collectively, the specific gasoline compounds benzene, toluene, ethylbenzene, and xylenes (BTEX), MTBE, solvents, and refrigerants were significantly greater in areas with more than 60 percent urban land use and (or) population density greater than 1,000 people per square mile than in source waters from less urbanized or lower population-density areas. MTBE detections were five times more frequent in source waters from areas with high MTBE use than in source waters from low or no MTBE use, but, unlike other gasoline compounds, MTBE detections were not significantly related to the density of gasoline storage tanks near drinking-water sources.