FGOALS-f3 simulated tropical cyclones：Characteristics and biases analysis

The High Resolution Model Intercomparison Project (HighResMIP) is a new component of the Sixth Coupled Model Intercomparison Project (CMIP6). Within this framework, this study employs the China Meteorological Administration (CMA) Best Track Dataset (2002—2014), the Integrated Multi-satellitE Retrievals for Global Precipitation Measurement (GPM IMERG) precipitation product, and European Centre for Medium-range Weather Forecasts reanalysis data (ERA5) to compare the performance of high- and low-resolution versions of the FGOALS-f3 climate model in simulating Tropical Cyclone (TC) tracks, intensity, precipitation, and further investigates the sources of simulation biases. The results show that the model's ability to simulate TC frequency and intensity is improved with increasing resolution. Specifically, the correlation coefficient between simulated TC track frequencies and observations increases from 0.85 to 0.90 using the high-resolution version; moreover, the high-resolution model is able to capture super typhoons with wind speed exceeding 53 m/s, whereas the low-resolution version fails to reproduce strong typhoons with wind speed greater than 41.5 m/s. In terms of precipitation, the high-resolution version demonstrates improvements in both the spatial distribution and relative contribution of TC rainfall, more realistically capturing precipitation over the ocean, and increasing the TC rainfall contribution rate by 18.02% compared with the low-resolution model. These improvements are mainly attributed to the high-resolution model's better representation of regional vorticity distribution and vertical wind shear, which are key large-scale environmental factors. However, the model significantly overestimates island precipitation, particularly in the high-resolution version, where the maximum overestimation reaches 121%. This bias is likely due to the model's overestimation of orographic uplift effects and moisture convergence over islands. Future simulations should focus on reducing the overestimation of precipitation associated with island topography.