The large-scale use of the gasoline oxygenate methyl tert-butyl ether (MTBE), and its high solubility, low soil adsorption, and low biodegradability, has resulted in its detection in ground water and surface water in many places throughout the United States. Studies by numerous researchers, as well as many State and local environmental agencies, have discovered high levels of MTBE in soils and ground water at leaking underground gasoline-storage-tank sites and frequent occurrence of low to intermediate levels of MTBE in reservoirs used for both public water supply and recreational boating.

In response to these findings, the American Water Works Association Research Foundation sponsored an investigation of MTBE and other volatile organic compounds (VOCs) in the Nation's sources of drinking water. The goal of the investigation was to provide additional information on the frequency of occurrence, concentration, and temporal variability of MTBE and other VOCs in source water used by community water systems (CWSs). The investigation was completed in two stages: (1) reviews of available literature and (2) the collection of new data. Two surveys were associated with the collection of new data. The first, termed the Random Survey, employed a statistically stratified design for sampling source water from 954 randomly selected CWSs. The second, which is the focus of this report, is termed the Focused Survey, which included samples collected from 134 CWS source waters, including ground water, reservoirs, lakes, rivers, and streams, that were suspected or known to contain MTBE. The general intent of the Focused Survey was to compare results with the Random Survey and provide an improved understanding of the occurrence, concentration, temporal variability, and anthropogenic factors associated with frequently detected VOCs.

Each sample collected was analyzed for 66 VOCs, including MTBE and three other ether gasoline oxygenates (hereafter termed gasoline oxygenates). As part of the Focused Survey, 451 source-water samples and 744 field quality-control (QC) samples were collected from 78 ground-water, 39 reservoir and (or) lake, and 17 river and (or) stream source waters at fixed intervals for a period of 1 year.

Using a common assessment level of 0.2 μg/L (micrograms per liter) (2.0 μg/L for methyl ethyl ketone), 37 of the 66 VOCs analyzed were detected in both surveys. However, VOCs, especially MTBE and other gasoline oxygenates, were detected more frequently in the Focused Survey than in the Random Survey. MTBE was detected in 55.5 percent of the CWSs sampled in the Focused Survey and in 8.7 percent of those sampled in the Random Survey. Little difference in occurrence, however, was observed for trihalomethanes (THMs), which were detected in 16.4 and 14.8 percent of Focused Survey and Random Survey CWSs, respectively. This may indicate a pervasive occurrence of THMs in several source-water types, regardless of CWS size or geographic location.

Using data at or above the method detection limit to assess temporal variability and anthropogenic factors associated with frequent detection of select VOCs, concentrations in the Focused Survey in ground-water, reservoir, and river source waters were typically less than 1 μg/L. Also, at a 95-percent confidence interval, no statistically significant differences were observed in comparing concentrations in the first and second ground-water sample. A weak seasonal pattern was observed in samples collected from reservoirs and lakes where gasoline oxygenates and other gasoline compounds were detected more frequently during spring and summer, presumedly a result of increased use of motorized watercraft during these seasons. In contrast, seasonal patterns were not observed in samples collected from rivers and streams. The lack of seasonal differences in river and stream source waters sampled may indicate a common and continuous source of contamination.

The most frequently detected VOC (MTBE) and the two most frequently occurring subgroups of VOCs (gasoline oxygenates and THMs) detected in CWS source waters were further characterized to evaluate some anthropogenic factors that may better explain their frequent occurrence. Gasoline oxygenates were detected in 73.9 percent of all CWSs sampled. The concentration of gasoline oxygenates was slightly correlated with watercraft use on reservoirs inside MTBE high-use areas (r2=0.3783) but not outside of these areas (r2=0.0242). In general, the concentration of gasoline oxygenates increased as watercraft use increased. THMs were detected in 47.8 percent of the CWSs supplied by surface water. The frequent occurrence of THMs in reservoir source waters was determined to be an artifact of disinfection and the recycling of chlorinated water to these reservoirs. All CWSs with frequent occurrence of THMs served by a reservoir indicated that chlorine was added to waters for various reasons and that the chlorinated water was then released back to,
or upstream of, the reservoir or lake that was sampled.