Potential for Interspecies Toxicity Estimation in Soil Invertebrates.

Area of Science:

• Environmental Toxicology
• Ecotoxicology
• Soil Ecology

Background:

• Interspecies correlation estimation (ICE) models are established for aquatic organisms but not widely applied to soil invertebrates.
• Estimating toxicity for soil organisms is challenging due to limited data.
• A comprehensive database of soil invertebrate toxicity was needed.

Purpose of the Study:

• To develop and validate Interspecies Correlation Estimation (ICE) models for soil invertebrates.
• To assess the predictive accuracy of these models for different species and taxa.
• To demonstrate the proof-of-concept for applying ICE models to soil organism toxicity.

Main Methods:

• Compiled a database of 853 toxicity records for 192 chemicals across 12 soil invertebrate species.
• Developed eleven species-pair ICE models using log-linear regression.
• Assessed model uncertainty and prediction accuracy using leave-one-out cross-validation.

Main Results:

• ICE models demonstrated high prediction accuracy for toxicity within the same taxonomic order (e.g., earthworm to earthworm).
• Prediction accuracy was lower for models spanning different taxa (Arthropoda to Annelida).
• The study provides a foundational dataset and methodology for soil invertebrate ICE models.

Conclusions:

• Interspecies correlation estimation (ICE) models are feasible for predicting soil invertebrate toxicity.
• These models show promise for filling data gaps in soil ecotoxicology.
• Further development is needed to improve accuracy for across-taxa predictions.