Overview of groundwater quality in the Piceance Basin, western Colorado, 1946--2009

Groundwater-quality data from public and private sources for the period 1946 to 2009 were compiled and put into a common data repository for the Piceance Basin. The data repository is available on the web at http://rmgsc.cr.usgs.gov/cwqdr/Piceance/index.shtml. A subset of groundwater-quality data from the repository was compiled, reviewed, and checked for quality assurance for this report. The resulting dataset consists of the most recently collected sample from 1,545 wells, 1,007 (65 percent) of which were domestic wells. From those samples, the following constituents were selected for presentation in this report: dissolved oxygen, dissolved solids, pH, major ions (chloride, sulfate, fluoride), trace elements (arsenic, barium, iron, manganese, selenium), nitrate, benzene, toluene, ethylbenzene, xylene, methane, and the stable isotopic compositions of water and methane.

Some portion of recharge to most of the wells for which data were available was derived from precipitation (most likely snowmelt), as indicated by δ2H [H2O] and δ18O[H2O] values that plot along the Global Meteoric Water Line and near the values for snow samples collected in the study area. Ninety-three percent of the samples were oxic, on the basis of concentrations of dissolved oxygen that were greater than or equal to 0.5 milligrams per liter.

Concentration data were compared with primary and secondary drinking-water standards established by the U.S. Environmental Protection Agency. Constituents that exceeded the primary standards were arsenic (13 percent), selenium (9.2 percent), fluoride (8.4 percent), barium (4.1 percent), nitrate (1.6 percent), and benzene (0.6 percent). Concentrations of toluene, xylenes, and ethylbenzene did not exceed standards in any samples. Constituents that exceeded the secondary standard were dissolved solids (72 percent), sulfate (37 percent), manganese (21 percent), iron (16 percent), and chloride (10 percent). Drinking-water standards have not been established for methane, which was detected in 24 percent of samples. Methane concentrations were greater than or equal to 1 milligram per liter in 8.5 percent of samples. Methane isotopic data for samples collected primarily from domestic wells in Garfield County indicate that methane in samples with relative high methane concentrations were derived from both biogenic and thermogenic sources. Many of the constituents that exceeded standards, such as arsenic, fluoride, iron, and manganese, were derived from rock and sediment in aquifers. Elevated nitrate concentrations were most likely derived from human sources such as fertilizer and human or animal waste.

Information about the geologic unit or aquifer in which a well was completed generally was not provided by data sources. However, limited data indicate that Quaternary deposits in Garfield and Mesa Counties, the Wasatch Formation in Garfield County, and the Green River Formation in Rio Blanco County had some of the highest median concentrations of selected constituents. Variations in concentration with depth could not be evaluated because of the general lack of well-depth and water-level data.

Concentrations of several important constituents, such as arsenic, manganese, methane, and nitrate, were related to concentrations of dissolved oxygen. Concentrations of arsenic, manganese, and methane were significantly higher in groundwater with low dissolved-oxygen concentrations than in groundwater with high dissolved-oxygen concentrations. In contrast, concentrations of nitrate were significantly higher in groundwater with high dissolved-oxygen concentrations than in groundwater with low dissolved-oxygen concentrations. These results indicate that measurements of dissolved oxygen may be a useful indicator of groundwater vulnerability to some human-derived contaminants and enrichment from some natural constituents.

Assessing such a large and diverse dataset as the one available through the repository poses unique challenges for reporting on groundwater quality in the study area. The repository contains data from several studies that differed widely in purpose and scope. In addition to this variability in available data, gaps exist spatially, temporally, and analytically in the repository. For example, groundwater-quality data in the repository were not evenly distributed throughout the study area. Several key water-quality constituents or indicators, such as dissolved oxygen, were underrepresented in the repository. Ancillary information, such as well depth, depth to water, and the geologic unit or aquifer in which a well was completed, was missing for more than 50 percent of samples.

Future monitoring could avoid several limitations of the repository by making relatively minor changes to sample- collection and data-reporting protocols. Field measurements for dissolved oxygen could be added to sampling protocols, for example. Information on well construction and the geologic unit or aquifer in which a well was completed should be part of the water-quality dataset. Such changes would increase the comparability of data from different monitoring programs and also add value to each program individually and to that of the regional dataset as a whole. Other changes to monitoring programs could require greater resources, such as sampling for a basic set of constituents that is relevant to major water-quality issues in the regional study area. Creation of such a dataset for the regional study area would help to provide the kinds of information needed to characterize background conditions and the spatial and temporal variability in constituent concentrations associated with those conditions. Without such information, it is difficult to identify departures from background that might be associated with human activities.