Tracing Uncertainty Contributors in the Multi‐Hazard Risk Analysis for Compound Extremes

In this study, an iterative factorial multimodel Bayesian copula (IFMBC) framework was developed for revealing uncertainties in risk inferences of compound extremes under the consideration diverse model structures and parameter sets. Particularly, analysis (IFA) method would be advanced IFMBC to track dominant contributors imprecise predictions multi-hazard risks. The applied assessment floods at two estuarine systems (i.e., Washington Philadelphia) US. results indicate that most likely events, predefined return periods, exhibit noticeable uncertainties. Those also present multiple hotspots which may attributed different impacts from factors. By applying IFA method, suggest structure ranked as one top 2 impact factors failure probabilities (FPs) scenarios AND, Kendall, with its contributions higher than 30% FP Kendall (more 40% Washington) more 25% (larger Philadelphia). comparison, not pose a visible effect on predictive uncertainty OR, contribution possibly less 5% long-term service time periods. However, marginal distributions have effects OR other FPs. distribution extreme variable high skewness kurtosis values tends significant OR. Also, overall parametric both dependence models cannot neglected all three FPs their probably larger 20% short period. Compared traditional multilevel analysis, can provide reliable characterization analyses, since seems significantly overestimate