Projection of Design Wind Speeds in the Southeastern Coastal Regions of China Under Climate Change Using a Physics-Driven Tropical Cyclone Model

Description

Tropical cyclones (TCs), also known as hurricanes or typhoons, are destructive natural hazards that can have significant impacts on coastal regions and even inland areas. Current research indicates that climate change significantly influences TC behaviors, as it leads to changes in occurrence frequency, increased intensity, shifts in track motion, and rapid intensification. Preparing for potential changes in TC behaviors is vital to mitigate risks and protect vulnerable populations. While existing data-driven TC models have been applied to predict the behavior of these storms, they may not be entirely sufficient to address the inherent complexities and non-stationarity introduced by climate change. Moreover, linear wind field models based on symmetric pressure models cannot accurately represent realistic TC events. These models do not consider the nonlinearity of the wind field and the nonuniformity of surface conditions after landfalls. Therefore, extreme wind estimations and track predictions provided by these models often deviate significantly from observational data. The aim of the proposed research is to enhance the performance of physics-driven TC models to provide more reliable projections of design wind speeds across the southeastern coastal regions of China under potential climate change. The research will involve four specific tasks: (1) the reconstruction of 3D TC wind fields using physics-informed neural networks; (2) the investigation of TC vortex structures for improved track prediction; (3) the modeling of TC intensity evolution considering topographic effects; and (4) large regional and city-scale mappings of design wind speeds under climate change. In task (1), comprehensive numerical simulations of TCs will be conducted using the Weather Research and Forecasting model. These simulations will be validated against global positioning system dropsonde measurements and routine observations from over 50 weather stations operated by the Hong Kong Observatory. The 3D wind field will be reconstructed using the observed data and the momentum equations of the air. Task (2) involves identifying asymmetries in the 3D wind field when TCs make landfalls. Azimuth-dependent pressure models will be established, and the relationship between track motion and winds at different pressure levels will be explored to determine the critical pressure levels that govern TC motion. Regarding track predictions, typical vortex structures will be tested and compared with the asymmetric wind field model in the vorticity equation. In task (3), two intensity models will be proposed. The first is a hybrid model that simulates intensity over the ocean using two coupled ordinary differential equations and over land using a terrain-following numerical model. The second is a revised denoised diffusion model, which provides probabilistic spatiotemporal predictions of intensity evolutions. In task (4), environmental conditions will be predicted using global climate models, and variations in TC behavior under the influence of climate change will be examined. Subsequently, regional (e.g., southeastern coastal regions of China) and city-scale (e.g., Hong Kong) mappings of design wind speeds under climate variability will be completed to provide references for future wind engineering applications.