Prioritizing risk mitigation measures for binary heavy metal contamination emergencies at the watershed scale using bayesian decision networks

Abstract

Water pollution accidents have the characteristics of high uncertainty, rapid evolution and are difficult to control, thus posing great threats to human health, ecological security, and social stability. During the last 10 years, China has faced the occurrence of six extraordinarily serious heavy metal contamination pollution events at the watershed scale. This has alerted governments and enterprises of the significance of emergency decision-making. To quantitatively prioritize risk mitigation strategies for heavy metal emergencies, a Bayesian Decision Network-based probabilistic model is proposed under the Drivers-Pressures-States-Impacts-Responses (DPSIR) framework. A Copula-based exposure risk model is embedded to simulate the fate of heavy metal ions for each risk reduction option, whose joint probability distributions can then be used as input parameters in the Bayesian Decision Network. This method was applied to the emergency response prioritization for acute Cr(VI)–Hg(II) contamination accidents in the Danshui River watershed. The results indicated that comprehensive measure (M5) was the best option for decreasing ecological and human health risks. As for a single risk mitigation strategy, risk source prevention (M1) was the best alternative compared to exposure pathway interruption (M2) and human/ecological receptor protection (M3–M4). This probabilistic method can not only address the uncertainties between certain risk sources and receptors in the BDN structure, but also realize the risk system optimization in a satisfactory/preferred mode under the DPSIR framework. Overall, it provides the probabilistic risk estimates for watershed-scale risk management and policy making for local risk managers and stakeholders.