UNCERTAINTIES IN PROJECTED CHANGES OF TROPICAL CYCLONE RAINFALL HAZARD
Seminars
Semester 2
Tropical cyclones (TCs) rank among the most destructive natural hazards affecting coastal regions, with extreme rainfall consistently acting as a primary driver of damage. This study projects future TC rainfall hazards by coupling the synthetic storms downscaled from 14 CMIP6 climate models with physics-based rainfall simulations. Subsequently, we systematically evaluate and attribute the uncertainties in these projections.
We introduce three key methodological advancements: (1) For the synthetic dataset, an event-level bias correction method is developed to match multiple marginal distributions simultaneously while preserving the physical coherence of individual storms. (2) For the physics-based rainfall model, a spatially aware calibration scheme is proposed by rescaling frictional and topographic effects. (3) For the uncertainty assessment, we establish a global-emulation and local-attribution framework to link the uncertainties of projected rainfall changes with those of along-track storm features.
Integrating the bias-corrected TC dataset with the calibrated rainfall model yields reliable simulations for TC rainfall hazards. The uncertainty assessment results provide the potential to investigate the physical mechanisms driving changes in future TC rainfall.
For additional information, please contact Mr. Yu HU, yuhu08@connect.hku.hk.