Simulation of the population-level effects of 4-nonylphenol on wild Japanese medaka (Oryzias latipes)

Here we report a structured matrix model for wild Japanese medaka (Oryzias latipes) and the use of the model to predict the population-level effects of 4-nonylphenol (4-NP) using a stochastic simulation approach. In the model, the natural fecundity and mortality rates of wild medaka were considered inhibited by 4-NP, based on Weibull dose-response models estimated from a full life-cycle toxicity test. The matrix model was simulated according to three scenarios: (1) stochastic daily growth under optimal conditions, (2) stochastic annual growth under conditions of seasonal variation and (3) population dynamics over 3 years undergoing density-regulation. Accordingly, a finite population growth rate, λ, was applied as an endpoint in the first and second scenarios and a quasi-extinction risk in the third scenario. The median and 95% confidence interval (CI) of 4-NP concentrations corresponding to λ = 1 (Cλ=1) in the first and second scenarios were 27.5 μg/L (CI: [20.2, 33.0] μg/L) and 17.0 μg/L (CI: [16.0, 17.9] μg/L), respectively. The quasi-extinction risks mounted quickly when the exposure concentration approached Cλ=1. The influence of uncertainties was analyzed, and the Cλ=1 was found to be robust against uncertainties in both the mean and the variation of vital rates, whereas the quasi-extinction risk was moderately sensitive to the mean mortality rate. The ignorance of whether the survival of adult madaka should be considered inhibited by 4-NP or not was found to remarkably affect the population-level effects. This study demonstrates an application of Cλ=1 in a stochastic sense as a population-level ecological risk assessment (PLERA) endpoint in chemical risk management.