Probabilistic hazard assessment of contaminated sediment in rivers

We propose a probabilistic framework rooted in multivariate and copula theory to assess heavy metal hazard associated with contaminated sediment in freshwater rivers that provide crucial ecosystem services such as municipal water source, eco-tourism, and agricultural irrigation. Exploiting the dependence structure between suspended sediment concentration (SSC) and different heavy metals, we estimate the hazard probability associated with each heavy metal at different SSC levels. We derive these relationships for warm (spring-summer) and cold (fall -winter) seasons, as well as stormflow condition, to unpack their nonlinear associations under different environmental conditions. To demonstrate its efficacy, we apply our proposed generic framework to Fountain Creek, CO, and show heavy metal concentration in warm season and under stormflow condition bears a higher hazard likelihood compared to the cold season. Under both warm season and stormflow conditions, probability of exceeding maximum allowable threshold for all studied heavy metals (Cu, Zn, and Pb, in recoverable form) at a standard hardness of 100 mg/1 CaCo3 and at a high level of SSC (95th percentile) is consistently more than 80% in our study site. Moreover, a longitudinal study along the Fountain Creek demonstrates that urban and agricultural land use considerably increase likelihoods of violating water quality standards compared to natural land cover. The novelty of this study lies in introducing a probabilistic hazard assessment framework that enables robust risk assessment with important policy implications about the likelihood of different heavy metals violating water quality standards under various SSC levels. (C) 2019 Elsevier B.V. All rights reserved.