Toxicodynamic modeling of zebrafish larvae to metals using stochastic death and individual tolerance models: comparisons of model assumptions, parameter sensitivity and predictive performance

Yongfei Gao, Jianfeng Feng, Lin Zhu

Process-based toxicodynamic (TD) models are playing an increasing role in predicting chemical toxicity to aquatic organism. Stochastic death (SD) and individual tolerance distribution (IT) are two often used assumptions in TD models which could lead to different consequences for risk assessment of chemicals. Here, using the toxicity data of single (Cu, Zn, Cd, and Pb) and their binary metal mixtures on survival of zebrafish larvae, we assessed the parameter sensitivity and evaluated the predictive performance of SD and IT models. The sensitivity analysis indicated the parameters related to toxicodynamics such as k k and threshold, had a great influence on the SD model's output and α had a great influence on the IT model's output. The predicted survival probability was highly sensitive to the assumptions of SD or IT models, and the SD model explained toxicity of single metal and binary metal mixtures better than IT model. Our results suggested that SD model is more suitable in assessing the metal toxicity to zebrafish larvae. Moreover, different combinations of laboratory metal-specific and species-specific experiments with SD and IT models need further study for better understanding and predicting toxic effects for different metals and organisms.