Search

Multiple flood-generating mechanisms can affect the communities and infrastructure systems across coastal areas. Reliable design flood estimation and risk assessment in these regions demand characterization of the interrelationships between different drivers as well as the corresponding compounding effects. In this study, we assess the compound flood risks across Canada's coasts considering eight bivariate flooding scenarios acquired from four major drivers including precipitation, streamflow, skew surge, and total water level. For each scenario, an initial dependency test based on Kendall's Tau is conducted. The individual variability of each driver, at different locations, is characterized through parametric extreme value distributions followed by developing the bivariate joint distributions using copula functions. Parameters of the copula functions and the marginal distributions are inferred based on the Bayesian approach. Further, compound flood risks and failure probabilities are analyzed considering the OR, AND, Kendall, and conditional hazard scenarios. Results suggest that most locations can be affected by compound flooding associated with at least two bivariate events. Among the studied scenarios, the concurrent occurrence of extreme precipitation and sea level is the most common across ~ 80% of the coastal locations. In addition, critical sites that can be affected by the majority of the bivariate events (six out of eight combinations) are identified at the Pacific, south of Lake Huron, Lake Erie, and Lake Ontario. Besides, the estimated failure probabilities show higher hydrologic risks associated with the bivariate analysis of flood drivers compared to the univariate estimates. [ABSTRACT FROM AUTHOR]

Data on ozone concentrations at four monitoring sites in central Canada are examined. Univariate extreme value methods do not allow for the required inference, and multivariate methods exploiting the joint dependence of the data are necessary. Families of multivariate extreme-value copulas with a flexible dependence structure and closed-form cumulative distribution functions have recently been derived. This article shows that maximum likelihood estimates of these parametric models may be obtained numerically and that their use can lead to increased efficiency in the estimation of margins. We do not, however, have any assurance that all data are well modeled by these distributions, because the space of multivariate extreme-value copulas is infinite-dimensional, and data on ozone levels are prone to outliers. This article addresses the robust methods required for proper analysis. [ABSTRACT FROM AUTHOR]