Response of selenium concentrations in groundwater to seasonal canal leakage, lower Gunnison River Basin, Colorado, 2013

Abstract

Selenium is a water-quality concern in the lower Gunnison River Basin because irrigation water interacting with seleniferous soils derived from the Mancos Shale Formation has mobilized selenium and increased its concentrations in surface water. Understanding the occurrence of elevated selenium concentrations in groundwater is necessary because groundwater discharge is an important source of selenium in surface water in the basin. In 2013, the U.S. Geological Survey, in cooperation with the Bureau of Reclamation and the Colorado Water Conservation Board, began a study to understand how changes in groundwater levels attributed to canal leakage affected the concentrations and speciation of dissolved selenium in groundwater. The purpose of this report is to characterize the groundwater adjacent to an unlined leaky canal. Two locations, near the East Canal (W-N1 and W-N2) and farther from the East Canal (W-M1 and W-M2), were selected for nested monitoring well installations. The pressure exerted by changes in canal stage was more readily transferred to the deep groundwater measured in the W-N1 near the canal than the shallow groundwater at the W-N2 well. No definitive relation could be made between canal water-level elevation and water-level elevations in monitoring wells farther from the canal (W-M1 and W-M2). Water flowing through the East Canal before the irrigation season had much higher selenium concentrations (140 micrograms per liter) than water in the canal during the irrigation season (3.02 micrograms per liter). Total selenium concentrations in the monitoring wells near the canal initially increased to 51.8 micrograms per liter in W-N1 and 1.66 micrograms per liter in W-N2. The initial increase in groundwater selenium concentrations presumably resulted from the dissolution of salts in the unsaturated zone by rising groundwater levels associated with canal leakage. The subsequent decrease in total selenium concentrations resulted from a combination of dilution by canal leakage and selenium reduction processes. Total selenium concentrations in monitoring wells located farther from the canal were not directly affected by canal leakage. Selenite/total selenium mass ratios in the East Canal samples ranged from about 0.02 to 0.13, indicating that about 2 to 13 percent of the total selenium in canal samples was composed of selenite. The increase in total selenium at W-N1 from before the irrigation season to the early irrigation season was accompanied by a decrease in the percentage of selenite from about 10 to 1 percent, indicating that selenate was added to the groundwater. A nitrate pulse occurred with the selenate pulse in W-N1 at the beginning of the irrigation season but apparently dissipated to a low enough concentration during the irrigation season to allow for selenate reduction to occur, as indicated by the relatively high percentages of selenite in W-N1 during the late irrigation season. W-N2 generally contained higher percentages of selenite than W-N1. Percentages of selenite in W-M1 did not change in response to filling the canal and generally composed less than 1 percent of the total selenium in that well. The predominance of selenate in W-M1, and apparent lack of selenate reduction, cannot be explained by a lack of anoxic conditions in the groundwater because all the available dissolved-oxygen data indicate that concentrations were less than 0.5 milligrams per liter. The most likely explanation for the lack of selenate reduction in W-M1 is that the exceptionally high concentrations of nitrate in the groundwater (about 340 to 390 milligrams per liter as nitrogen) inhibited selenate reduction. These high nitrate concentrations presumably come from the Mancos Shale and its weathering products because there was no evidence for a human source of nitrate at the lower Gunnison River Basin wetland. The high concentrations of selenate in W-M1 may persist and eventually discharge to surface water unless nitrate concentrations are reduced to low enough levels to permit substantial selenate reduction to occur. Well W-M2 contained relatively low concentrations of total selenium and high percentages of selenite before and at the onset of the irrigation season. An increase in total selenium concentration associated with a drying and wetting period later in the summer was accompanied by a decrease in the percentage of selenite to near 0 percent, indicating that selenate was added to the groundwater. This pattern is consistent with the examples of increasing concentrations of total selenium in the other wells and presumably resulted from the dissolution of selenate-bearing salts in the unsaturated zone by rising water levels in W-M2.