The value of adding new toxicity data to species sensitivity distributions in ecological risk assessment

A species sensitivity distribution (SSD) is a statistical distribution of the toxicity values of chemicals and is increasingly used in ecological risk assessment (ERA). When an SSD is used with limited toxicity data, additional testing is sometimes requested in the expectation that more data will reduce assessment uncertainty. Such an expectation is often valid, but the extent of the reduction depends on the number and variability of the available data. If a newly obtained toxicity value does not reduce uncertainty, acquiring it is of limited value and may be even wasteful. However, there has been no studies quantify how a new toxicity value changes assessment uncertainty. Here, we quantitatively and analytically investigated how indicators used in ERAs, such as the concentration that protects 95% of species (HC5) and the size of the required uncertainty factor, change when a new toxicity value is obtained, assuming that an SSD follows a log-normal distribution. We found that when a new toxicity value was added, HC5 tended to shift upward more frequently than downward. However, the magnitude of increases in HC5 was generally small, whereas the magnitude of decreases could be very large. For example, when the sample size is 10 and the standard deviation of SSD is 1, the probability of an increase is 66%, but the magnitude of this increase is at most about 1.3‑fold. In contrast, there is a 0.5% probability that HC5 becomes one‑tenth of its original value and a 0.03% probability that it becomes one‑hundredth of its original value. Our findings provide one of the first rigorous theoretical bases for careful and informed planning of additional toxicity testing in species sensitivity distribution–based ecological risk assessment.