Probabilistic spill occurrence simulations and quantitative water quality risk analysis for chemical spill management.

摘要

Thousands of inland chemical spills occur as a result of accidents or natural disasters each year in the world and threaten human health and the environment. More than 700 recorded inland chemical spills involving more than 1,000 types of chemical occur every year in Southern Ontario, resulting in multiple environmental impacts. Eleven regional municipalities involving 77 municipalities had experienced chemical spills in the period of 1988-2007. The majority of these chemical spills occurred at industrial plants, while pipe/hose leaks accounted for the highest proportion of total chemical spills, resulting in the largest portion of chemical spills causing surface water impacts.;Benzene spills into the St. Clair River Areas of Concern are used as a case study to demonstrate the models. The probabilistic occurrences of various NAICS codes are found to be 1.2 to 5.1 over a 10-year period. The violation-causing NAICS-based spill occurrences and the associated risks of drinking water quality impairments at the Ontario's intakes are found to be less than 1.4 and 37%, respectively. No drinking water quality is found to be impaired at the Michigan intakes. Uncertainty analysis indicates that simulated spill characteristics can be described by lognormal distributions and the NAICS-based risks of violation at the Ontario's intakes are Weibull distributed. A hypothetical case, benzene spills in the Mimico Creek watershed is used to investigate the possibility of spill characteristic transfer from one area to another area.;A comprehensive spill management planning framework is proposed to facilitate the development of municipal spill prevention, control, and emergency response plans. In order to develop a spill management framework, simulation models termed MMCS (MATLAB-based Monto Carlo Simmulation) and EMMCS (Extended MMCS) that characterizes temporal and spatial randomness and quantifies statistical uncertainty have also been developed. The MMCS model simulates the probabilistic quantifiable occurrences of inland chemical spills by time, magnitude, and location based on North America Industry Classification System (NAICS) codes, while the EMMCS model quantifies the risk of drinking water quality violation due to inland chemical spills. The models can also quantify aleatory and epistemic uncertainties through integrated bootstrap resampling technique.