Quantitative Ecological Risk Assessment of Industrial Accidents: The Case of Oil Ship Transportation in the Coastal Tropical Area of Northeastern Brazil

ABSTRACT

Accidents such as toxic spills can cause massive damage to local ecosystems and hamper the sustainable development of hazardous industries. Models that only consider regularly occurring pollution are unable to truly quantify ecological risks (ecorisks) from these industries. This work presents a methodology capable of quantifying ecorisks related to rare and extreme events such as industrial accidents. We developed a procedure that integrates information from different studies that contributes to characterize ecorisks from industrial accidents: (1) reliability analysis, (2) fate and transport modeling, (3) individual-level toxicological assessment, and (4) population modeling. The methodology is exemplified by an application to oil ship transportation to supply Brazil's Suape industrial complex. A fish population was strategically chosen to represent the ecosystem's health of Suape beach. For the critical accidental scenarios, their frequencies of occurrence were estimated and the space–time evolution of oil simulated. The ecorisks were quantified in terms of time and population probability of fish extinction, categorized and compared against a no-accident scenario. The total ecorisks from all scenarios were presented as a FN curve, where N is the average number of deaths in the population and F the cumulative frequency of accidents with potential to cause N or more deaths.

Key Words:

• probabilistic ecological risk assessment
• oil spill
• Suape Port
• frequency estimates
• stochastic population modeling
• oil plume dispersion.

ACKNOWLEDGMENTS

An assessment of this magnitude relies on the cooperative work of many people. At LEPT, we collectively thank the dozens of observers who collected the fish population data. At CEERMA, thanks go to the oceanographers who collected oceanographic and meteorological data and supported us in fate and transport modeling, especially colleagues M. A. Silva and R. A. Silva. We acknowledge the financial support provided by the CEERMA, the Brazilian Council for Scientific and Technological Development (CNPq), and the Brazilian Council for the Improvement of Higher Education (CAPES).