Aquatic Ecosystem Health and Management

p.449-458The toxicity of surface waters and interstitial waters from sediments were determined at six study sites in Todos os Santos Bay, Bahia, Brazil, to evaluate the possibiiity of chronic environmental impact induced by 40 years of exposure to the local petroleum industry. Samples coliected from four sites associated with the extraction, transportation and refinement of petroleum, and from two control sites, were tested at seven three-month intervals. Toxicological assays using acute mortality ofbrine shrimp (Artemia salina) nauphi and chronic abnormahties of sea urchin (Echinometra lucunter) and mangrove oyster (Crassostrea rhizophorae) larvae were ernpioyed. Friedman non-parametric analyses of variance integrated seasonai variations in species response patterns and revealed significant differences among the study sites. Ranging the among-site variations for each organism in each samphng period, on a scale from 0.00 (minimum response) to 1.00 (maximum response), permitted the caicuiation of a single mean value for each species and the ordination of the sites on a qualitative scale of relative impact. Aithough the ordinations varied with species, the reduction of three species response patterns to a common relative scaie also permitted their integration into a single multispecies ordination of the study sites. A ciuster anaiysis of the six sites and two aquatic substrates, based on their toxicity to ali three species, illustrated the similarities and differences between iocations. lnterstitial waters were more toxic, reveaiing an integrated ordination of Station 6 Station 5 (Station 3= Station2) (Station 4 = Station 1). The ordination based on surface waters was Station 6 Station 5 (Station 2 = Station 1 = Station 3) < Station 4. In combination, the three procedures served efficiently for the description and inferentiai testing of the multispecies responses and, compiemented with additionai data on species diversity and chemical contamination of the sediment, conflrmed the existence of chronic impact within the study arca