Predicting the thresholds of metals with limited toxicity data with invertebrates in standard soils using quantitative ion character-activity relationships (QICAR)

Abstract

Terrestrial invertebrates are often used as indicator organisms in ecological risk assessments. However, determining the risk of metals to invertebrates is laborious and time-consuming due to the lengthy testing and ethical approval procedures. In this study, a review of the literature was conducted to provide toxicity data for two standard soils (OECD and LUFA 2.2). An attempt was made to establish models for predicting the toxicity of elements to invertebrates using quantitative ion character-activity relationships (QICARs). In OECD soil, the element toxicity of four groups (Enchytraeus albidus mortality and reproduction, Folsomia candida and Eisenia fetida reproduction) showed significant correlations with atomic number, atomic mass and atomic ionization potential (0.852 ≤ R2 ≤ 0.989, P < 0.05). For LUFA 2.2 soil, polarization force parameters and boiling point were most significant parameters for toxicity values of F. candida and Enchytraeus crypticus, respectively (0.866 ≤ R2 ≤ 0.962, P < 0.05). Finally, QICAR models were established, and LC50 or EC50 of elements were predicted. Then, models were verified using standard and natural soils, and showed that errors between observed and predicted logLC50/EC50 were mostly < 0.5 orders of magnitude. Thus, the developed QICAR models have potential for predicting the toxicity of elements for soils.