Selenium mobilization during a flood experiment in a contaminated wetland: Stewart Lake Waterfowl Management Area, Utah

Abstract

Constructed and natural wetlands can accumulate elevated levels of Se; however, few data are available on cost-effective methods for remobilization and removal of Se from these areas. A field experiment was conducted to assess the effectiveness of flooding on the removal of Se from dry surface sediments. The 83-m2 flood-experiment plot contained 10 monitoring wells, a water-quality minimonitor (continuous measurement of pH, specific conductance, water temperature, and dissolved O2), a down-hole Br electrode, and 2 pressure transducers. Flooding was initiated on August 27, 2002, and a Br tracer was added to water delivered through a pipeline to the flood plot during the first 1.2 h. Standing water depth in the flood plot was maintained at 0.3 m through September 1, 2002. The Br tracer data indicate a dual porosity system that includes fracture (mud cracks) and matrix flow components. Mean vertical water velocities for the matrix flow component were estimated to range from 0.002 to 0.012 m/h. Dissolved (less than 0.45 ??m) Se increased from pre-flood concentrations of less than 10 ??g/L to greater than 800 ??g/L during flooding in samples from deep (2.0 m below land surface) ground water. Selenium concentrations exceeded 5500 ??g/L in samples from shallow (0.8 m below land surface) ground water. Ratios of Se to Br in water samples indicate that Se moved conservatively during the experiment and was derived from leaching of near-surface sediments. Cumulative Se flux to the deep ground water during the experiment ranged from 9.0 to 170 mg/m2. Pre- and post-flood surface soil sampling indicated a mean Se flux of 720 mg/m2 through the top 15 cm of soil. Ground-water samples collected 8 months after termination of the flood experiment contained Se concentrations of less than 20 ??g/L. The minimonitor data indicate a rapid return to chemically reducing conditions in the deep ground water, limiting the mobility of the Se dissolved in the water pulse introduced during the flood experiment. Ratios of Se to Br in deep ground-water samples collected 8 months after the experiment confirmed the removal of Se from the aqueous phase. Based on the median Se flux rate estimated during the experiment of 0.65 mg/h/m2 (n = 52), 7 flooding cycles would be required to meet the 4 ??g/g remediation goal in surface soils from the SLWMA wetland.