Keystone indices probabilistic species sensitivity distribution in the case of the derivation of water quality criteria for copper in Tai Lake.

An alternative method for species sensitivity distribution (SSD) that considers the interaction between species and the community importance is illustrated in this article. First, a food web based on the relationship between predator and prey was constructed, and the keystone indices of species were obtained based on the food web. Then, the probability density distribution of the sensitivity for each species was derived from all of the available endpoints of each species. Finally, the species sensitivity distribution for ecosystem was constructed by sampling a specific number of values from the probability density distribution of the sensitivity for each species. Data of copper toxicity to the aquatic organisms in Tai Lake were selected to derived site-specific water quality criteria (WQC). Ninety-seven endpoints of acute toxicity for 47 species and 188 endpoints of chronic toxicity for 29 species were included, and the acute and chronic WQC developed by keystone indices probabilistic species sensitivity distribution (K-PSSD) were 4.982 μg/L and 0.965 μg/L, respectively. Results showed that the aquatic organisms of Tai Lake might be underprotected. Compared with the SSD, the K-PSSD coped with the interactions between species, the community importance, and the intraspecies and interspecies variation more effectively and was better at depicting the tendency and information of raw data. The K-PSSD was especially applicable to site-specific WQC and provided an alternative or supplement to the SSD.