基于物种敏感度分布法建立中国土壤中锑的环境基准

基于Sb（锑）的植物及动物毒理学数据缺乏以及保护生态受体的土壤Sb的环境基准尚未建立的现状，通过收集和筛选文献中Sb的毒理学数据并补充开展不同土壤类型的跳虫和植物的毒理学试验，建立了Sb的生物毒性预测模型，并以此为依据对收集及试验毒理学数据进行归一化处理，以消除土壤性质的影响.此外，进一步利用SSD（species sensitivity distribution，物种敏感度分布法）推导我国4种典型情景土壤中Sb的HC₅（hazardous concentration，能够保护95%物种的生态安全阈值），最终建立基于土壤性质参数的环境基准计算模型.结果表明:①不同土壤中Sb对跳虫的毒性差异较大，跳虫毒性阈值EC₁₀（effect concentration，10%抑制效应浓度）与土壤pH呈负相关，与w（SOC）（SOC为土壤有机碳）呈正相关，即随着土壤pH增加或w（SOC）降低，Sb对跳虫的生物有效性随之增加，进而导致EC₁₀降低.②通过毒性阈值与土壤性质之间的多元回归分析可知，土壤pH和SOC可较好地预测Sb的生物毒性，植物和无脊椎动物的R2（决定系数）分别为0.778和0.867.③利用SSD得到11个物种在4种典型情景土壤中的HC₅分别为55.12、28.28、28.08及14.55 mg/kg，推导出PNEC_total（predicted no effect concentration，预测无效应浓度）分别为28.96、15.54、15.44及8.68 mg/kg，计算模型为PNEC_total=-5.811pH+0.587SOC]+55.480+C_b（SOC]为土壤有机碳含量，C_b土壤Sb背景浓度）.鉴于此，建议以中性土壤中Sb的环境基准值作为我国农用地土壤Sb污染风险筛选值制订的参考依据，即农用地土壤w（Sb）限值定为15 mg/kg. 

Deriving Soils Environmental Criteria of Antimony in China by Species Sensitivity Distributions

For lack of toxicology data of plants and animals on antimony (Sb) and the soil environmental criteria for conservation of ecological receptors have not yet been established in China, through collecting and screening the toxicological data in literature and carrying out supplementary toxicological tests on Folsomia candida and plants under different soil conditions, this study create a biological toxicity prediction model of Sb, and on the basis of this model normalized the collected and experimental toxicological data in order to eliminate the influence of soil properties. Then, the hazardous concentration (HC₅) of soil Sb in four typical scenarios by species sensitivity distribution (SSD) was derived and the environmental criteria model based on the soil properties indexes was established. As a result:(1) The toxicity differences of Sb on F. candida were significant in the different soils. There was a negative correlation between soil pH and toxicity threshold of F. candida (EC₁₀). On a contrary, a positive correlation was demonstrated between soil organic carbon and EC₁₀ in soils. Along with the increased soil pH or decreased soil organic carbon, bioavailability of F. candida by Sb in soils was improved. (2) The multiple regression analysis between toxicity thresholds (EC₁₀) and soil properties showed that the two-factor predictive models based on soil pH and SOC could predict EC₁₀ on plants and invertebrates with (determination coefficients) R² of 0.778 and 0.867, respectively. (3) Through the SSD, HC₅ of 11 species under four soil scenarios were 55.12, 28.28, 28.08 and 14.55 mg/kg, respectively; and the PNEC_total were 28.96, 15.54, 15.44 and 8.68 mg/kg, respectively. The calculation model was PNEC_total=-5.811pH+0.587SOC]+55.480+C_b (SOC] is the soil organic carbon content, C_b is background concentration of Sb in soil). When deriving the risk screening values for Sb of agricultural land, the environmental quality criteria of Sb in the neutral soils was recommended as a reference, i.e. the criteria of w(Sb) was 15 mg/kg in agricultural land.