The effects of selenium on agricultural crops are summarized in Table 11.01.
Plant species from the genera Astragalus, Xylorhiza, Oonopsis, and Stanleyea are capable of growing in seleniferous soils and accumulating high concentration of Se (Rosenfeld and Beath 1964). These plants, classified as Se accumulators, are considered to have an unpleasant odour associated with high concentrations of Se-containing organic compounds, and can be toxic to animals grazing on them.
Typical agricultural crops have a much lower tolerance to Se. The symptoms of Se toxicity include snow-white chlorosis of leaves and pink rot tissue in grasses, and general stunting, yellow chlorosis, and pink leaf veins in other plants. Apparently, the toxic effects of Se to plants are caused largely by interference with sulphur (S) metabolism. The small physical and chemical differences between Se and S result in significant biological differences within a plant. The similarity of Se-amino acids to their S-analogues of cysteine and methionine can disrupt the normal biochemical reactions and enzyme function within the cell. Selenium accumulator plants, however, have a mechanism that prevents Se incorporation into protein, thus avoiding Se-induced phytotoxicity (Mikkelsen et al. 1989).
Toxicity of naturally occurring Se to agricultural crops in the field has not been documented in the literature. Published values of tissue selenium concentrations associated with 10% yield reduction in the laboratory range from 19-94 mg Se/kg (selenate) and 12-30 mg Se/kg (selenite) for alfalfa, and 2-3 mg Se/kg (selenate) for rice, mustard, pea, and wheat plants (Mikkelsen et al. 1989). Tolerance to selenium depends upon plant and selenium species. In a review of published values, Mikkelsen et al. (1989) noted that the lowest concentrations of selenium in plant tissue, causing 10% reduction in yield, were 3 mg Se/kg (dry weight - mustard, pea, and wheat) for selenate and 10-15 mg Se/kg (dry weight - wheat) for selenite. The authors did not report the Se concentrations in the soils.
In a growth chamber study, Soltanpour and Workman (1980) reported that soils treated with selenite-Se and yielding 2 mg Se/kg dry weight (ammonium bicarbonate-DTPA extractable) were highly toxic to alfalfa. Jump and Sabey (1989) reported that the ammonium bicarbonate-DTPA extractable selenium in soil might range from 1% to 17% of the total selenium, depending on the soil type. Singh and Singh (1978) studied selenium toxicity to nonaccumulator plants in a greenhouse. They reported wheat-grain production of 100%, 72%, 17%, and 11%, respectively, in soils treated with 0, 2.5, 5, and 10 mg Se/kg as selenite. The grain contained 0.6, 10.2, 24.5, and 36.8 mg Se/kg dry weight, respectively.
Broyer et al. (1966) found that 0.025 mg Se/L in a nutrient solution decreased the yield of alfalfa. Mild toxicity was displayed in bush beans at 8 mg Se/L, as Na2SeO3, in solution culture (Wallace et al. 1980). In sand cultures, Davis et al. (1978) found that the critical levels of selenite-Se causing toxicity to young barley plants were 5 mg/L (in solution) and 30 mg/kg dry-weight (in tissue). In the literature, a selenate-Se concentration as low as 0.1 mg/L, in solution cultures void of sulphate, was reported to produce detectable injury to wheat plants, but this was probably due to sulphur deficiency, since signs of injury did not occur in the presence of sulphur (Adriano 1986). More recently, Mikkelsen et al. (1988) reported that the growth of alfalfa, in sand cultures, decreased when exposed to a solution containing 0.25 mg/L selenate-Se.
11.2 Uptake by Plants
All plants are capable of taking up Se from soils. While Se is an important element in animal nutrition, the evidence that Se is essential to plants is not convincing. However, there are plant genera, called `indicator, or `accumulator' plants, where Se may be required for normal growth (Adriano 1986).
The Se-accumulating plants have been divided arbitrarily into three groups, depending upon their ability to accumulate selenium: (1) `Primary Se-Accumulator, or Indicator, Plants' are believed to require Se for growth and may accumulate large amounts (often several thousand mg/kg), but levels <100 mg Se/kg are common. These plants are good indicators of seleniferous soils and include species from the genera Astragalus, Machaeranthera, Happlopappus, and Stanleya; (2) `Secondary Se Absorbers' do not require Se for growth and rarely contain more than a few hundred mg Se/kg tissue, although lesser amounts are common. They may be found in species of Atriplex, Aster, Gutierrezia, Catelleja, Grindelia, Machaeranthera, and Mentzelia; (3) `Grasses, Small Grains, and Other Shallow-Rooted Plants' usually accumulate toxic amounts of Se ranging from 5 to 12 mg/kg, but seldom more than 50 mg/kg. The Se accumulation is dependent on factors other than genera, species, and availability of soil Se. Some plants such as white clover (Trifolium repens L.), buffalograss (Buchloe dactyloides), and gamma grass (Bouteloua spp.) take up very little Se, even when growing on seleniferous soil. Other plants, such as those belonging to the Crucifera family (certain mustards, cabbage, broccoli, and cauliflower), may concentrate relatively high levels of Se if grown on high-Se soils (James et al. 1989).
The mobility and availability of selenium in soils depend upon several factors, including plant characteristics (e.g., species, age), soil characteristics (e.g., pH, texture and mineralogy, and organic matter content), presence of competitive ions such as sulphate and phosphate, and Se species (selenite, selenate, etc.) and concentration in soils (Adriano 1986). As a result, the concentration of Se in any one crop may vary widely. In general, however, the root and bulb crops contain the highest concentration of Se (mean = 0.407 mg Se/kg dry-weight) followed by field crops (mean = 0.279 mg Se/kg dry-weight), leafy vegetables (mean = 0.110 mg Se/kg dry weight), seed vegetables (mean = 0.066 mg Se/kg dry weight), vegetable fruits (mean = 0.054 mg Se/kg dry weight), and tree fruits (mean = 0.407 mg Se/kg dry weight) (Mikkelsen et al. 1989).
Soil pH influences plant Se uptake by altering the oxidation state of Se in soils and the capacity of clays and ferric oxide to adsorb Se. In general, plants accumulate more Se at higher pH (Mikkelsen et al. 1989). Hamdy and Gissel-Nielsen (1977) showed that clays and ferric oxide quickly remove Se4+ from soil solution, thus making Se less available to plants. They also found that 1:1 clays (e.g., kaolinite) had a greater Se+4 fixation capacity than 2:1 clays. Among the 2:1 clays minerals, vermiculite had a greater Se+4 adsorption capacity than montmorillonite, but Fe2O3 adsorbed more Se+4 than all the minerals tested. Reports concerning the effect of soil organic matter on the uptake of Se by plants are mixed.
Westerman and Robbins (1974) reported that the addition of gypsum to soils reduced significantly the Se concentration in alfalfa. Since sulphur (in gypsum) also increased the yield of alfalfa, this reduction was attributed to a dilution effect; i.e., although the tissue-Se concentration decreased, the total uptake of Se by the crop increased. Reduction in the Se uptake by red clover and perennial ryegrass, from soils naturally high in selenite-Se, was also reported in response to ammonium sulphate treatment (Williams and Thornton 1972). Mikkelsen et al. (1988) reported a negative correlation between tissue Se concentration and sulphate level in irrigation water for alfalfa grown in sand cultures. However, the interaction between Se and sulphur was not observed in soils low in sulphur and Se (Spencer 1982) or soils producing vegetation with inadequate Se for animal nutrition (Gupta and Winter 1975). From these data, Mikkelsen et al. (1989) concluded that the presence of SO42- in soil generally reduces the concentration of plant Se either through antagonism or may simply reflect a dilution of plant Se due to increased growth.
In a field experiment, Cary and Allaway (1973) applied 0, 2.2, and 4.5 kg selenite-Se/ha to five silt loam soils cropped with corn, oat, or various forage crops for four years. They reported that a single application of 2.2 kg Se/ha was sufficient to raise plant Se concentrations to levels sufficient to protect animals from Se deficiency for at least 4 years on perennial pasture and two to three years on corn. Application of 4.5 kg Se/ha further raised tissue Se concentrations, but did not produce vegetation with Se levels potentially harmful to animals. Gupta et al. (1982) reported that annual application of small amounts of selenite-Se were more effective than a single, large application for annual crops. They also reported that the residual effect of selenite-Se application to soil may be lost for barley following a single cropping season, whereas, for perennial pasture, a single application was found to have a residual effect of increasing plant Se concentrations for four to five years.
In a more recent field trials, Banuelos et al. (1992) reported that the first clipping (after 60 days) of the plant species, grown in soils containing 0.7 mg Se/kg dry weight, accumulated selenium to the levels of 1.3 mg Se/kg dry matter in tall fescue (whole plant), 4.7 mg Se/kg dry matter in alfalfa (whole plant), 3.1 mg Se/kg dry matter in birdsfoot trefoil (whole plant), 34.1 mg Se/kg dry matter in wild mustard (whole plant), and 51 mg Se/kg dry matter in canola (leaves). In a review of the literature, these investigators also noted that the desirable Se level (or safe level for consumption by animals) in cereals and forages is between 0.05 and 2 mg Se/kg dry matter. Obviously, a concentration of 0.7 mg/kg selenium in soil may be excessive for some crops (e.g., wild mustard and canola), producing feed that is not safe for animals. Gupta and Winter (1981) reported that a selenium salt rate of 1 to 2 kg/ha would enhance plant Se concentration sufficiently to meet the dietary requirements of cattle and not cause Se toxicity. Based on a soil density of 1300 kg/m3 and a leaching depth of 0.15 m, these rates corresponded to soil concentrations of 0.5 to 1 mg Se/kg.
Gissel-Nielsen and Bisbjerg (1970) tested the uptake of elemental Se, selenite-Se, and selenate-Se from soils by red clover, alfalfa, mustard, and sugar beets. Except for mustard, elemental Se uptake was very little. Selenite-Se uptake was roughly proportional to its soil concentration, regardless of the solubility of the salt added. However, the uptake of Se by plants from soils treated with selenate-Se salts was 20 to 50 times more than for the selenite-Se salts treatment. During the two-year field experiment, they found that mustard recovered 0.01% of elemental Se, 4% of K2SeO3, and 30% of the K2SeO4 applied to the soil. Carlson et al. (1991) have confirmed the greater availability to plants of selenate than selenite, which probably reflects the tendency of selenite to become adsorbed on soil clays and hydrous oxides.
The atmosphere plays an important role in the biochemical cycle of Se, which may also affect concentrations in vegetation. In a study of plant and environment interactions, Haygarth et al. (1995) reported that both atmospheric and soil Se was accumulated in ryegrass. The percent contribution of Se in the pasture leaves from the soil was 47% (pH 6.0) and 70% (pH 7.0), with by inference, the remainder coming from the atmosphere.
11.3 Summary of Existing Guidelines
The CCREM (1987) recommended that the concentration of total selenium in irrigation waters should not exceed 0.02 mg Se/L for continuous use on all soils, and 0.05 mg Se/L for intermittent use on all soils. USEPA (1973) recommended a maximum concentration of 0.02 mg Se/L in irrigation water for continuous on all soils. Ontario (OMOE 1984) also recommended 0.02 mg Se/L and Manitoba (Williamson 1988) recommended 0.05 mg Se/L for field crops and 0.02 mg Se/L for greenhouse crops.
11.4 Recommended Water Quality Guidelines
It is recommended that the concentration of total selenium in irrigation water supplies should not exceed 0.01 mg/L.
Two procedures were used to derive the guideline for irrigation water supplies:
1. The lowest observed effect concentration (LOEC) in soil causing 28% reduction in yield was reported to be 2.5 mg Se/kg dry-weight (Singh and Singh 1978). Using the CCME protocol (1993), the no observed effect concentration (NOEC) was estimated to be (2.5 ÷ 4.5) or 0.56 mg Se/kg. Based on the literature LOEC, the estimated NOEC, and the uncertainty factor of 10, the acceptable soil concentration (ASC) was computed as follows:
ASC = (2.5 mg Se/kg x 0.56 mg Se/kg)0.5 ÷ 10 = 0.12 mg Se/kg
Assuming a soil bulk density of 1300 kg/m3 and a leaching depth of 0.15 m, the allowable contaminant mass (ACM) was calculated as follows:
ACM = ASC (mg Se/kg) x 1 300 kg/m3 x (100 m x 100 m x 0.15 m)/ha
ACM = 0.12 x 1 300 x 100 x 100 x 0.15 = 234 000 mg Se/ha (or 0.234 mg/ha)
The allowable contaminant mass thus calculated is about three times lower than 0.7 mg Se/kg soil treatment that raised the selenium level in wild mustard and canola above the acceptable level for animal consumption (Banuelos et al. 1992). Also, the computed ACM appears to be in line with the Cary and Allway (1973) recommendation (if expressed on an annual basis) that soils treated at the rate of 2.2 kg Se/ha will raise Se levels in plants sufficient to protect animals from selenium deficiency for at least four years on perennial pastures.
Assuming an annual irrigation rate (IR) of 1.2 x 107 L/ha, the species maximum acceptable toxicant concentration (SMATC), or the recommended water quality guideline was calculated as follows:
SMATC = (ACM ÷ IR) = 234 000 mg/ha ÷1.2 x 107 l/ha = 0.0195 mg/L
SMATC 0.02 mg/L
2. Mikkelsen et al. (1988) reported that the lowest observed effect concentration (LOEC) of 0.25 mg Se/L caused a 19.5 to 25% reduction in the yield of alfalfa grown in a greenhouse sand culture. The no observed effect concentration (NOEC) was estimated to be 0.25 ÷ 4.5 = 0.056 mg Se/L, in accordance with the procedure suggested in the CCME (1993) protocol. Assuming an uncertainty factor (UF) of 10, the species maximum acceptable concentration (SMATC) was calculated as follows:
SMATC = (LOEC x NOEC)0.5 ÷ UF = (0.25 x 0.056)0.5 ÷ 10 0.010 mg Se/L.
It is recommended that the lowest (0.01 mg Se/L) of the two values computed above be accepted as the water quality guideline.
The selenium guideline may be influenced by factors such as availability of selenium, form of selenium, and sulphate content of irrigation water. For instance, Albasel et al. (1989) proposed a guideline of 0.1 mg Se/L in irrigation waters for the west side of the San Joaquin Valley in California. The specific conditions that justified this guideline were as follows: (i) Se in the drainage water of the valley was largely in the selenate form, which is not appreciably adsorbed by soils and is thus readily leached from the root zone; (ii) the waters containing high concentrations of Se also had substantial quantities of sulphate that dramatically reduced selenate-Se adsorption by plants, and (iii) the high salinity of the valley water demanded a high leaching fraction of irrigation water, which caused a fraction of the Se applied in irrigation water to leach through the root zone. These factors were not considered in this document because their quantitative relationships with Se toxicity/uptake could not be established from the published literature.