The 1990 Clean Air Act Amendments require fuel oxygenates to be added to gasoline used in some metropolitan areas to reduce atmospheric concentrations of carbon monoxide or ozone. Methyl tert-butyl ether (MTBE) is the most commonly used fuel oxygenate and is a relatively new gasoline additive. Nevertheless, out of 60 volatile organic chemicals analyzed, MTBE was the second most frequently detected chemical in samples of shallow ambient groundwater from urban areas that were collected during 1993−1994 as part of the U.S. Geological Survey's National Water-Quality Assessment program. Samples were collected from five drinking water wells, 12 springs, and 193 monitoring wells in urban areas. No MTBE was detected in drinking water wells. At a reporting level of 0.2 μg/L, MTBE was detected most frequently in shallow groundwater from urban areas (27% of 210 wells and springs sampled in eight areas) as compared to shallow groundwater from agricultural areas (1.3% of 549 wells sampled in 21 areas) or deeper groundwater from major aquifers (1.0% of 412 wells sampled in nine areas). Only 3% of the shallow wells sampled in urban areas had concentrations of MTBE that exceed 20 μg/L, which is the estimated lower limit of the United States Environmental Protection Agency draft drinking water health advisory. Because MTBE is persistent and mobile in groundwater, it can move from shallow to deeper aquifers with time. In shallow urban groundwater, MTBE generally was not found with benzene, toluene, ethylbenzene, or xylene (BTEX) compounds, which commonly are associated with gasoline spills. This disassociation causes uncertainty as to the source of MTBE. Possible sources of MTBE in groundwater include point sources, such as leaking storage tanks, and non-point sources, such as recharge of precipitation and stormwater runoff.

The effects of digital elevation model (DEM) map scale and data resolution on watershed model predictions of hydrologic characteristics were determined for TOPMODEL, a topography‐based watershed model. The effects of topography on watershed hydrology are represented in TOPMODEL as the distribution of ln (a/tan B), where ln is the Napierian logarithm, a is the upslope area per unit contour length, and tan B is the gravitational gradient. The minimum, maximum, mean, variance, and skew values of the ln (a/tan B) distribution were computed from 1:24,000‐scale (24K) DEMs at 30‐ and 90‐m resolutions and from 1:250,000‐scale (250K) DEMs at 90‐m resolution for 71 areas in Pennsylvania, New York, and New Jersey. An analysis of TOPMODEL showed that model predictions of the depth to the water table, the ratio of overland flow to total flow, peak flow, and variance and skew of predicted streamflow were affected by both the DEM map scale and data resolution. Further TOPMODEL analyses showed that the effects of DEM map scale and data resolution on model predictions were due to the sensitivity of the predictions to the mean of the ln (a/tan B) distribution, which was affected by both DEM map scale and data resolution. DEM map scale affected the mean of the ln (a/tan B) distribution through its influence on the mean of the ln (a) distribution, which characterizes land‐surface shape, and the mean of ln (1/tan B) distribution, which characterizes land‐surface slope. DEM resolution, in contrast, affected the mean of the ln (a/tan B) distribution primarily by its influence on the mean of the ln (a) distribution.

As part of the National Water-Quality Assessment Program of the U.S. Geological Survey, an assessment of 60 volatile organic compounds (VOCs) in untreated, ambient groundwater of the conterminous United States was conducted based on samples collected from 2948 wells between 1985 and 1995. The samples represent urban and rural areas and drinking-water and nondrinking-water wells. A reporting level of 0.2 μg/L was used with the exception of 1,2-dibromo-3-chloropropane, which had a reporting level of 1.0 μg/L. Because ambient groundwater was targeted, areas of known point-source contamination were excluded from this assessment. VOC concentrations generally were low; 56% of the concentrations were less than 1 μg/L. In urban areas, 47% of the sampled wells had at least one VOC, and 29% had two or more VOCs; furthermore, U.S. Environmental Protection Agency drinking-water criteria were exceeded in 6.4% of all sampled wells and in 2.5% of the sampled drinking-water wells. In rural areas, 14% of the sampled wells had at least one VOC; furthermore, drinking-water criteria were exceeded in 1.5% of all sampled wells and in 1.3% of the sampled drinking-water wells. Solvent compounds and the fuel oxygenate methyl tert-butyl ether were among the most frequently detected VOCs in urban and rural areas. It was determined that the probability of finding VOCs in untreated groundwater can be estimated on the basis of a logistic regression model by using population density as an explanatory variable. Although there are limitations to this national scale model, it fit the data from 2354 wells used for model development and adequately estimated the VOC presence in samples from 589 wells used for model validation. Model estimates indicate that 7% (6−9% on the basis of one standard error) of the ambient groundwater resources of the United States probably contain at least one VOC at a reporting level of 0.2 μg/L. Groundwater is used in these areas by 42 million people (35−50 million based on one standard error); however, human exposure to VOCs from this ambient groundwater is uncertain because the quality of the finished drinking water is generally unknown.

No abstract

No abstract