水生生物基准推导中物种选择方法研究

运用Mann-Whitney U检验法对不同生物分类群内的物种对氨氮的毒性敏感性进行分析和统计检验,并基于毒理学、生物分类学,从抽样-推断的角度对物种选择进行讨论,并对物种选择与基准推导方法的关系进行梳理.结果发现,物种毒性数据中脊椎动物占绝大部分,尤其是辐鳍纲,与真实生态系统中物种类群结构不符;大部分不同分类群间物种对氨氮的敏感性差异显著,随着物种分类阶元级别降低,大部分分类群内物种的差异逐渐减小,个别分类群内物种的敏感性未呈现相似的规律,尤其是相对不敏感的物种;总体上,能在一定程度上说明同一分类群内的生物有相对接近的敏感性.由抽样-推断理论与方法可知,基准推导属于基于抽样设计的统计推断,因此,可以将生物分类类群作为物种选择的辅助变量,依据水生态系统中物种组成,运用分层随机抽样技术进行物种选择,以提高抽样效率和精度.

Species Selection Methods in Deriving Water Quality Criteria for Aquatic Life

The Mann-Whitney U test method was used to analyze the species sensitivity to ammonia toxicity. And based on the analysis, the relationship between species selection method and WQC deriving method was studied by using toxicology, biological taxonomy and sampling-inference theory. Results showed that vertebrate species, especially the Actinopterygii, accounted for the vast majority in the toxicity test species. And the species composition of toxicity test species was inconsistent with the species composition of the ecosystem. Sensitivity to ammonia toxicity among different taxa varied significantly for most species except some species in individual taxa, especially the less sensitive species. The variable coefficient of interspecies decreased with the reduction of biological classification level. To a certain extent, it showed that the species sensitivities in the same taxa to toxicant were more similar than those in different taxa. According to sampling-inference theory, the WQC for aquatic life deriving method belonged to the design-based inference. And taxonomic groups could be used as auxiliary variables to conduct a stratifactory sampling for species selection in WQC deriving which could improve the sampling efficiency and precision.