Quantifying the precision of ecological risk: Conventional assessment factor method vs. species sensitivity distribution method

Abstract

In ecological risk assessment, the Predicted No Effect Concentration (PNEC) of a substance is generally derived by one of two methods: either by applying an Assessment Factor (AF) or by using a Species Sensitivity Distribution (SSD). With the AF method, which is the conventional way, the PNEC is determined by dividing the lowest No Observed Effect Concentration (NOEC) by an AF of a certain fixed magnitude. With the SSD method, which is becoming increasingly used in the European Union and the United States, an HC5 value (Hazardous Concentration for 5% of species) is estimated from the NOEC and then divided by an AF to derive the PNEC. This study aimed to explore the most appropriate AF and the most effective application of each method. The performances of the SSD and AF methods were compared on the assumption that the better method is that in which more PNECs are lower than HC5. We concluded that the performance of these methods depends on sample size and variation in species sensitivity. As the sample size increases (i.e., if there are more toxicity data), the performance of each method increases. The performance of the AF method is better when variation in species sensitivity is small (i.e., all species tend to have a similar NOEC), but it declines as variation in sensitivity rises, implying that persisting with either of the methods may misrepresent the ecological risk. Our results suggest that the variation in sensitivity is an important factor affecting the ecological risk and more effort should be paid to understanding why the variation varies depending on chemical substances.