摘要

Management decisions aimed at protecting aquatic resources following accidental chemical spills into rivers and coastal estuaries require estimates of toxic thresholds derived from realistic spill conditions: acute pulse exposures of short duration (h), information which often is unavailable. Most existing toxicity data (median lethal concentration or median effective concentration) come from tests performed under constant exposure concentrations and exposure durations in the 24-h to 96-h range, conditions not typical of most chemical spills. Short-exposure hazard concentration estimates were derived for selected chemicals using empirical toxicity data. Chemical-specific 5th percentile hazard concentrations (HC5) of species sensitivity distributions (SSD) from individual exposure durations (6-96h) were derived via bootstrap resampling and were plotted against their original exposure durations to estimate HC5s and 95% confidence intervals (CIs) at shorter exposures (1, 2, and 4h). This approach allowed the development of short-exposure HC5s for 12 chemicals. Model verification showed agreement between observed and estimated short-exposure HC5s (r(2) adjusted=0.95, p<0.0001), and comparison of estimated short-exposure HC5s with empirical toxicity data indicated generally conservative hazard estimates. This approach, applied to 2 real spill incidents, indicated hazard estimates above expected environmental concentrations (acrylonitrile), and suggested that environmental concentrations likely exceeded short-exposure hazard estimates (furfural). Although estimates generated through this approach were likely overprotective, these were derived from environmentally realistic exposure durations, providing risk-assessors with a tool to manage field decisions. Environ Toxicol Chem 2013;32:1918-1927.

  • 出版日期2013-8