Constellation, le dépôt institutionnel de l'Université du Québec à Chicoutimi

Résumé

Hurricane wind risk assessment has been significantly improved with the evolvement of synthesis methodologies from the single site probabilistic method to the hurricane track model (including five simulation components of genesis, translation, intensity, decay and boundary-layer wind). As first-step efforts towards advancing the data- driven hurricane track model, widely used by engineering community, to a physics-based framework for more accurate and reliable hurricane risk assessment, a new intensity model integrating important dynamics and thermodynamics inside the storms is developed. Furthermore, an extensive statistical analysis of hurricane trajectories is carried out to obtain an enhanced translation model and a height-resolving analytical wind model is utilized to acquire the vertical profiles of wind speed and direction between ground-surface and gradient levels. The other two simulation components (i.e., genesis and decay) of the hurricane track model are also revisited for the sake of completeness. Ten thousand years of full-track synthetic hurricanes are generated and compared with the HURDAT database at specific mileposts along the US East coast to validate the overall performance of the developed simulation framework (in terms of annual occurrence rate, intensity, translation speed and heading angle). Then, the New Jersey coastline is employed as a case study to compare the simulated hurricane wind speeds with ASCE 7–16 recommendations, to highlight the wind directionality effects on extreme wind speeds, and to investigate the joint distribution of hurricane wind speed and size.