中文摘要：

候选参数的科学筛选是构建底栖动物完整性指数的关键步骤.在缺乏足够参照样点的情况下,生物参数与环境梯度的响应关系是重要依据.以广西钦江流域冬季枯水期22个样点的环境和底栖动物数据为例,应用主成分分析法提取研究流域的环境梯度,进而采用多重回归分析确定对环境干扰敏感且预测性强的构成参数.相关性分析表明,38个候选参数中有12个参数与其他参数间的信息冗余性很高（｜r｜〉0.80,P〈0.05）;余下的26个生物参数与环境梯度主成分Ⅰ（PCⅠ）、主成分Ⅱ（PCⅡ）和栖境质量参数的多重回归分析表明,有13个参数能和环境胁迫较好地响应（P〈0.05）.参考各指数的代表性和实用性,提出了适于构建该区域底栖动物完整性指数的8个生物参数：科级丰度、蜉蝣目丰度、鞘翅目丰度,EPT多度、优势种百分比,捕食者百分比,属级香农多样性指数和BI指数.本研究为类似情况下底栖动物完整性指数的构建提供了一个较好的解决途径.

英文摘要：

The selection of proper macroinvertebrate metrics was a key step in the successful construction of benthic-index of biotic integrity（B-IBI）.In the context of lack of sufficient reference sites,the responses of biotic metrics to the stress gradients have been proved crucially important in metrics selection.We explored this using a data set（twenty-two sampling sites）,where water environmental parameters,habitat quality index（HQI） and zoobenthos were measured and sampled using standardized methodology in the winter dry season（January,2010） within the Qinjiang Rvier basin,Guangxi.Water quality stress gradients were synthesized using principle component analysis.Twelve out of the 38 candidate metrics were excluded because of high co-linearity（｜r｜0.80,P0.05） with others.Multiple regression analysis of the rest twenty-six metrics,environmental gradients（PCⅠ and PCⅡ） and habitat quality index showed that thirteen metrics significantly correlated（P0.05） with the environmental stressors（PCs and HQI）.Eight of these metrics,family richness,Ephemeroptera richness,Coleoptera richness;Ephemeroptera,Plecoptera and Trichoptera（EPT） abundance,proportion of dominant species;proportion of predators;Shannon-Weaver diversity index for genus;and Biotic index（BI）,satisfied the selection criteria（e.g.,operational simplicity） and were thus candidates for developing a multimetric system to assess river health in our study region.We suggest that the response of metrics to environmental gradient is an ideal alternative for the construction of benthic-index of biotic integrity in the context of insufficient reference sites.