Assessing parametric rainfall models in reproducing tropical cyclone rainfall characteristics

Abstract

Parametric models have the potential to be used to assess the rainfall hazard induced by Tropical Cyclones (TCs), by taking advantage of the physical nature of these types of models as well as their high computational efficiency. This study details the development of a physics-based parametric TC rainfall model. Together with existing three rainfall models (one statistical and two parametric), extensive examinations of this model are conducted by observing the spatial structure of TC rainfall fields and the parametric analysis of TC rainfall processes. The correlation between rainfall and vertical winds, the contribution of radial and vertical moisture fluxes to the rainfall, and the sensitivity of the model to several important parameters are investigated using two virtual TCs. The spatial rainfall distributions of two historical TCs, Typhoon Lekima (2019) and Typhoon IN-FA (2021), are then simulated and compared with meteorological observations to examine the model capability to reproduce real TC rainfall scenarios. Contributions of TC dynamic process, inhomogeneous topography, vertical wind shear, and changes of TC vorticity to the total TC rainfall are also analyzed to study rainfall mechanisms. Furthermore, rainfalls of 118 historical TCs landed in the China coastal region from 1990 to 2018 are simulated and compared with the observation records to reveal the model ability to generate TC rainfall statistics. Inadequacy of parametric rainfall models in generating extreme rainfall events and spatial rainfall distribution are pointed out, and future work to improve the performance of TC rainfall models are discussed.