Risk of five polycyclic aromatic hydrocarbons in a terrestrial environment: influence of data variability

The risk of five different pyrogenic polycyclic aromatic hydrocarbons (PAHs) toward two soil-dwelling organisms (i.e., springtail [Folsomia fimetaria] and earthworm [Eisenia veneta]) has been investigated with respect to lethality and reproduction at two soil depths in a typical Danish soil. Predicted environmental concentrations (PECs) are calculated with a model describing diffusion, bulk flow, and microbial degradation. Predicted no-effect concentrations (PNECs) are derived from laboratory experiments performed with nominal soil concentrations in the range from 0 to 300 microg PAH/g dry weight. Risk is estimated through a stochastic approach as well as with the conventional point estimate. The point estimate predicts a potential risk for pyrene, log PEC/PNEC = -0.01, with respect to springtail reproduction at 5 cm soil depth. In all other scenarios, the point-estimate log-ratios are significantly lower than 0. For the stochastic approach risk is defined when the probability for risk (i.e., the probability for log PEC/PNEC > 0), is larger than 5%. The results show that risk is present only for springtail and in the following five scenarios: For anthracene, the probability for risk with respect to lethality is 12% at 5 cm soil depth, and 17 and 5% with respect to reproduction at 5 and 50 cm soil depth, respectively; for pyrene the probability for risk with respect to reproduction is 49 and 14% at 5 and 50 cm, respectively. The results show that risk cannot be defined unambiguously with the two approaches. The probabilistic approach is less restrictive, and even small probabilities may be used as early-warning indications that risk may be posed under unfavorable circumstances.