Application of non-parametric kernel density estimation for developing species sensitivity distributions of copper and silver

Species sensitivity distribution curves of copper and silver were constructed using non-parametric kernel density estimation model to protect Chinese freshwater aquatic life, and then their water quality criteria thresholds were derived. The results showed that the robustness and accuracy of non-parametric kernel density estimation method are superior to the traditional parameters models to derive water quality criteria for two transition metals of Group IB. After comparing different taxa of two metals, we found that HC5values of vertebrates, invertebrates, fish, crustaceans, other invertebrates and all aquatic organisms were inversely proportional to atomic number. The sensitivity of invertebrates was significantly higher than that of vertebrates at high trophic level. The proposed method enriched the methodological foundation for water quality criteria and provided an alternative approach for developing SSDs of the same group and period elements to support for protection of aquatic organisms.