Verification and Bias Analysis of the Forecast for Typhoon 'Samba' (2203) Using Numerical Models

doi:10.3969/j.issn.1007-6190.2025.02.015

Abstract：To evaluate the ability of numerical weather prediction models to forecast typhoon tracks, this study uses multi-source observed data, ECMWF model and CMA-TRAMS model forecasts, and ERA5 reanalysis data to verify the track and precipitation forecasts of Typhoon 'Samba' (2203) and analyze the causes of forecast biases. The results show that the maintenance of a blocky shape of the subtropical high pressure system contributed to the later northward movement of 'Samba'. The early forecast errors of the ECMWF model in the subtropical high pressure, South Asian high pressure, and upper-level outflow were the main reasons for the westward and southward deviation of its track. 'Aere' was not an immediate factor causing the track deviation of 'Samba', but the presence and interaction of two typhoons increased the uncertainty in typhoon track forecasting. The environmental field before 'Samba' made landfall was characterized by high temperature and high energy, and the maintenance of low-level jet stream and moisture channel on the southern side of the residual vortex provided favorable conditions for the occurrence of heavy rain in two stages. The precipitation forecast bias of the ECMWF model was caused by the track forecast error and the deviation in the intensity of the residual vortex, leading to a misrepresentation of the distribution of heavy precipitation, as well as the underestimation of wind speed and topographic convergence uplift due to errors in the low-level wind field. The CMA-TRAMS model had smaller errors in both track and precipitation forecasts compared to the ECMWF model, due to more accurate and stable forecasts of the subtropical high pressure pattern and the intensity of 'Samba'.