Assessing predictive uncertainty in comparative toxicity potentials of triazoles

Area of Science:

• Environmental Chemistry
• Ecotoxicology
• Computational Chemistry

Background:

• Comparative toxicity potentials (CTPs) assess ecotoxicological impacts per emission unit.
• CTPs integrate environmental fate, exposure, and hazardous concentrations.
• Predictive models are crucial when empirical data are scarce.

Purpose of the Study:

• To evaluate how predictive uncertainty in substance properties affects triazole CTPs.
• To quantify the influence of various uncertainty sources on CTP predictions.

Main Methods:

• Quantitative structure-activity relationships (QSARs) predicted physicochemical and toxic properties.
• Uncertainty quantification used underlying QSAR data and probability distributions.
• Monte Carlo simulations propagated uncertainties in degradation half-lives and hazardous concentrations.

Main Results:

• CTP 90% confidence intervals often spanned four orders of magnitude.
• Uncertainty in soil sorption and degradation rates, plus limited species sampling, were primary drivers of CTP uncertainty.
• Uncertainty in species-specific toxicity predictions had minimal impact.

Conclusions:

• Enhancing experimental data for soil sorption and degradation is key to improving triazole CTP reliability.
• Developing toxicity QSARs for a broader range of species will increase prediction accuracy.
• Addressing key uncertainty sources is vital for robust ecotoxicological risk assessment.