Flow-adjusted trends in dissolved selenium load and concentration in the Gunnison and Colorado Rivers near Grand Junction, Colorado, water years 1986--2008

Abstract

As a result of elevated selenium concentrations, many western Colorado rivers and streams are on the U.S. Environmental Protection Agency 2010 Colorado 303(d) list, including the main stem of the Colorado River from the Gunnison River confluence to the Utah border. Selenium is a trace metal that bioaccumulates in aquatic food chains and can cause reproductive failure, deformities, and other adverse impacts in birds and fish, including several threatened and endangered fish species. Salinity in the upper Colorado River has been the focus of source-control efforts for many years. Although salinity loads and concentrations have been previously characterized at the U.S. Geological Survey (USGS) streamflow-gaging stations at the Gunnison River near Grand Junction, Colo., and at the Colorado River near the Colorado-Utah State line, trends in selenium load and concentration at these two stations have not been studied. The USGS, in cooperation with the Bureau of Reclamation and the Colorado River Water Conservation District, evaluated dissolved selenium (herein referred to as "selenium") load and concentration trends at these two sites to inform decision makers on the status and trends of selenium. This report presents results of the evaluation of trends in selenium load and concentration for two USGS streamflow-gaging stations: the Gunnison River near Grand Junction, Colo. ("Gunnison River site"), USGS site 09152500, and the Colorado River near Colorado-Utah State line ("Colorado River site"), USGS site 09163500. Flow-adjusted selenium loads were estimated for the beginning water year (WY) of the study, 1986, and the ending WY of the study, 2008. The difference between flow-adjusted selenium loads for WY 1986 and WY 2008 was selected as the method of analysis because flow adjustment removes the natural variations in load caused by changes in mean-daily streamflow, emphasizing human-caused changes in selenium load and concentration. Overall changes in human-caused effects in selenium loads and concentrations during the period of study are of primary interest to the cooperators. Selenium loads for each of the 2 water years were calculated by using normalized mean-daily streamflow, measured selenium concentration, standard linear regression techniques, and data previously collected at the two study sites. Mean-daily streamflow was normalized for each site by averaging the daily streamflow for each day of the year over the 23-year period of record. Thus, for the beginning and ending water years, estimations could be made of loads that would have occurred without the effect of year-to-year streamflow variation. The loads thus calculated are illustrative of the change in loads between water years 1986 and 2008, and are not the actual loads that occurred in those 2 water years. The estimated 50th and 85th percentile selenium concentrations associated with the selenium loads were also calculated for WY 1986 and WY 2008 at each site. Time-trends in selenium concentration at the two sites were charted by using regression techniques for partial residuals for the entire study period (WY 1986 through WY 2008). Annual selenium load for the Gunnison River site was estimated to be 23,196 pounds for WY 1986 and 16,560 pounds for WY 2008, a 28.6 percent decrease. Lower and upper 95-percent confidence levels for WY 1986 annual load were 22,360 and 24,032 pounds. Lower and upper 95-percent confidence levels for WY 2008 annual load were 15,724 and 17,396 pounds. Estimated 50th percentile daily selenium concentrations decreased from 6.41 to 4.57 micrograms/liter from WY 1986 to WY 2008, whereas estimated 85th percentile daily selenium concentrations decreased from 7.21 to 5.13 micrograms/liter from WY 1986 to WY 2008. Annual selenium load for the Colorado River site was estimated to be 56,587 pounds for WY 1986 and 34,344 pounds for WY 2008, a 39.3 percent decrease. Lower and upper 95-percent confidence levels for WY 1986 annual load were 53,785 and 59,390 pounds. Lower and upper 95-percent confidence levels for WY 2008 annual load were 31,542 and 37,147 pounds. Estimated 50th percentile daily selenium concentrations decreased from 6.44 to 3.86 micrograms/liter from WY 1986 to WY 2008, whereas estimated 85th percentile daily selenium concentrations decreased from 7.94 to 4.72 micrograms/liter from WY 1986 to WY 2008.