Review of 15 years of research on ecotoxicology and remediation of land contaminated by agricultural drainage sediment rich in selenium.

The consequences of elevated Se accumulation at the Kesterson Reservoir National Wildlife Refuge in the Central Valley of California created adverse effects on wildlife and led to extensive research on the behavior of Se in both the wetland and upland ecosystems. Selenium concentrations in water entering the Kesterson Reservoir averaged 300 microg L(-1). In pond waters 20-30% of the Se was selenate, while only 2% was selenite in the drainage water entering the reservoir. Submerged rooted aquatic plants fed on by water birds were found to contain 18-390 mg Se kg(-1) dry weight. Mosquitofish collected from the San Luis Drain contained 332 mg Se kg(-1), and those collected from the ponds ranged from 339 to 380 mg kg(-1). Livers of water birds had Se concentrations ranging from 19.9 to 127 mg kg(-1). The high concentrations of Se accumulation in the food chain of the wetland strongly suggest that Se bioaccumulation was the cause of death and deformity of embryos of the waterfowl nesting at the wetland habitat. In June 1986, the Kesterson Reservoir was closed to drain-water inputs, and the wetland was transformed to an upland grassland. New remedial plans were proposed. These new plans involved soil, water, and vegetation management to dissipate Se by bioaccumulation and volatilization through soil microorganisms and plants. The investigations of the potential transfer of Se from farm land into the crop and vegetables in the Central Valley indicated that plant tissue Se concentrations generally fall in a nonseleniferous category, except that the highest Se concentration of cotton was at a threshold where toxicity in animals could occur at a relatively low frequency. At the Kesterson upland grassland habitat, average total Se concentrations ranged from 500 to 8000 microg kg(-1) and water-extractable Se ranged from 10 to 700 microg kg(-1) in the top 15 cm of soil and varied greatly, by a factor greater than 100, among soil samples. Uptake of Se by the plants was profoundly affected by the soil available Se concentration, soil moisture, pH, soil salinity, soil sulfate concentration, soil reoxidation condition, kind of plant species, and soil-management practices. The rate of soil Se dissipation at the Kesterson grassland system was from 1% (low methylation rate) to 5% (high methylation rate) Se inventory per year and it will take from 46 to 230 years to bring the soil Se down to a normal level, 4 mg Se kg(-1) soil. However, the Kesterson upland grassland habitat had Se bioaccumulation values less than 10% of those of the previous wetland. The potential food-chain contamination at the existing Kesterson grassland is much less problematic. No negative impact on wildlife has been reported for the upland habitat. Plants may contribute to the Se reoxidation process and be able to reduce the movement of Se in the soil. At the Kesterson grassland, the distribution of soil Se is extremely uneven; high levels of soil Se concentrated only in isolated spots. Therefore, leaching of soil Se is not at an area level. It is unlikely that problems of transport of Se from the Kesterson soil to the adjacent uncontaminated environment by leaching can occur.