Electronic Thesis and Dissertation Repository

Thesis Format

Integrated Article


Doctor of Philosophy


Civil and Environmental Engineering


Mohammad Reza Najafi


The communities settling in the Canadian coastal regions are threatened by multiple flood-generating mechanisms including riverine, pluvial, and sea level forces. 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, as our first step, we assess the compound flood risks across Canada’s coasts considering eight bivariate flooding scenarios acquired from four flooding drivers including total water level, streamflow, precipitation and the skew surge at 41 sites located at three main regions of the Pacific, the Great Lakes (GL) and the Atlantic. For each scenario, an initial dependence test based on Kendall’s Tau is conducted. Their joint probability is constructed using copulas. Further, compound flood risks and the 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 types of bivariate events.

In the second step, we characterize the dependence structure between the three drivers of total water level, streamflow and precipitation based on the C-vine copula statistical approach and create their multivariate joint distribution for different locations. This is followed by calculating the OR, AND, and Kendall compound flooding joint return periods (JRPs) and their corresponding failure probabilities (FPs) and comparing them with the univariate and independent JRP values. Further, the CHR index is applied to quantify possible under- or overestimations of the flooding risks when individual drivers are assessed, independently. The results show that multivariate JRPs are less than those of univariate and independent multivariate hazard estimates.

In our third objective, we try to explore the univariate and multivariate trends of four flooding drivers at all sites. The univariate Mann Kendall trend test and its extension to the multivariate case namely the Covariance Inversion Test, Covariance Sum Test, and Covariance Eigenvalue Tests are applied to see the univariate (change in the intensity and frequency) and joint nonstationary behavior of the three drivers, respectively. The results show increased risks of individual and compound flooding over the Atlantic coast, and various trends in the Pacific and the GL regions.

Finally, we assess the compound flooding hazard under a nonstationary framework for all the locations. To this end, the time-varying behavior of the three drivers of step 2 and also the interdependencies between them are captured using linear and polynomial models. This process leads to producing a time-dependent joint occurrence/probability of the drivers. Then, the temporal variations of the compound flood hazard are assessed concerning the OR, and AND hazard scenarios and the CHR index. The results highlight the decline and increase in the AND JRPs and CHR values over time at 23 locations, especially in the Atlantic region.

Summary for Lay Audience

Approximately half of the global population lives within 200km of coastlines. The communities and infrastructure systems in the coastal environments are at risk of flooding caused by one or multiple mechanisms. Understanding the compounding effects of the drivers of flooding and quantifying the corresponding uncertainties are critical for flood risk analysis and the development of effective resilience strategies. To address this objective, we investigate compound flood events considering terrestrial (both precipitation, and streamflow which reflects the effects of snow/ice melt in addition to rainfall) and coastal mechanisms across Canada's Atlantic, Pacific and Great Lakes' coasts, with distinct hydroclimate characteristics, based on a state-of-the-art statistical approach. The proposed design flood estimation method addresses the limitations in traditional approaches that neglect the interdependencies between two or multiple drivers of flooding. Further, the proposed approach identifies areas that are at high risk of compound flooding and identifies the main contributing factors. We also investigate whether the frequency and intensity of the co-occurrence of these flooding sources have altered from 1960 to 2015 or not. The results suggest that the risk of flooding can increase up to 50% if flood mechanisms are analyzed holistically and the interrelationships are accounted for, compared to estimates from the traditional approach. Precipitation and sea levels are the major factors that contribute to compound flooding, in particular at the Atlantic coast. The obtained findings of this research help the coastal managers with designing more sustainable and long-lasting coastal protective infrastructures concerning all three flooding sources, and the interconnection between them.