Irrigation and climate effects on selenium in the Kendrick Irrigation District, WY, and a bioremediation strategy

Abstract

Selenium (Se) is an essential micronutrient for humans and animals, but can be toxic at higher concentrations. In the mid 1980s, high Se concentrations in the Kesterson National Wildlife Refuge, California, caused high mortality in fish and waterfowl, initiating the National Irrigation Water Quality Program (NIWQP). The NIWQP identified the Kendrick Reclamation Project in Wyoming as one of several areas at risk for Se-related environmental problems, due to seleniferous Cody shale. Project objectives were (1) to examine relationships of Se concentration with flow, load, and suspended solids in natural streams and irrigation drains, (2) use geochemical modeling to predict Se speciation and possible sorption to minerals, and (3) evaluate the effectiveness of a lab-scale rotating biological contactor (RBC) inoculated with indigenous organisms from the Kendrick District at reducing Se concentrations in influents containing Se and SO₄^2-.

We collected samples from natural streams and irrigation drains in and around the Kendrick District. Analysis was conducted using a completely randomized design with two factors; (1) inside vs. outside the district and (2) parent material (on vs. off Cody shale). Se concentrations were higher inside the district (64 and 17 μg L^-1 on and off the shale, respectively) than outside the district (5 and 3 μg L^-1 on and off the shale, respectively). Interaction between the two factors was significant (p=0.002), so independent-samples t-tests were used to separate the simple effects averages. Significant differences for parent material were found both inside the district (p<0.001) and outside the district (p=0.009). Se concentrations in irrigation drains were compared to natural streams. Concentrations were not significantly different between drains and streams in the irrigation off season (p=0.515), but were significantly higher in the streams during irrigation (p<0.001), indicating a dilution effect from irrigation. Streamflow was inversely correlated to Se concentration and positively correlated to Se loads.

Se⁶⁺ was the predominant form of Se present in the Kendrick surface waters, although thermodynamic data predicts Se⁴⁺ will predominate. The MINTEQA2 geochemical model was used to predict formation of solid phases that would adsorb Se⁴⁺ and explain the low concentration relative to Se⁶⁺. Although the model predicted formation of several minerals that sorb Se⁴⁺, their presence could not be confirmed by X-ray photoelectron spectroscopy.

A lab-scale rotating biological contactor (RBC) was constructed and inoculated with microorganisms collected from seleniferous surface waters of the Kendrick district to evaluate its use in remediation of Se-containing water. Four influents were used; 120 μg L^-1 Se⁴⁺, 110 μg L ^-1 Se⁴⁺ + 11,000 mg L^-1 SO₄ ^2-, 150 μg L^-1 Se⁶⁺, and 250 μg L^-1 Se⁶⁺ + 10,000 mg L ^-1 SO₄^2-. After a 7 day equilibration period, total Se was reduced between 89% for the Se⁴⁺ + SO₄ ^2- influent to 15% for the Se⁴⁺ influent. Because apparent high evaporation from the reactor tank, (indicated by increasing SO₄ ^2- concentrations in the last stage), may have concentrated the influents, the actual removal efficiency was likely much higher. Future work to optimize the reactor conditions may make the RBC a useful tool for bioremediation of Se-contaminated wastewater.