Preliminary assessment of MPAS-A ensemble tropical cyclone forecasting skill in the East China Sea: A case study over Zhejiang and adjacent offshore waters

This study provides a preliminary evaluation of the ensemble forecasting capabilities of the Model for Prediction Across Scales-Atmosphere (MPAS-A) for tropical cyclones (TCs) affecting Zhejiang Province and adjacent offshore waters in the East China Sea. Five recent high-impact TCs were retrospectively simulated using a global 60-3 km variable-resolution mesh and an eight-member ensemble that samples key physics parameterizations, including cumulus, microphysics, boundary layer, and surface layer schemes. Ensemble-mean and spread characteristics of track and intensity forecasts were assessed against best-track data from the China Meteorological Administration (CMA), while near-surface wind predictions were evaluated using tower-based observations during Typhoon Muifa (2022). Track forecasts exhibited promising skill, with median errors under 50 km at 24 h and 100 km at 48 h, while the ensemble mean was typically more accurate than the median and comparable to operational forecasts. Intensity forecasts showed larger spread and systematic biases, particularly in maximum wind speed, highlighting the influence of boundary-layer physics on ensemble variance. Case studies illustrate how differing physics choices drive divergence in storm translation, steering flow, and inner-core structure. Comparison with tower observations confirmed that the ensemble has the potential to bracket uncertainty in near-surface wind forecasts, although storm position errors remain a key limiting factor. Despite high computational cost, MPAS-A ensembles demonstrate strong potential for probabilistic TC forecasting and offshore wind risk management in this region. ● MPAS-A ensemble simulations capture TC track and intensity evolution in the East China Sea, especially within the first 48 hours. ● Ensemble-mean track errors are maintained below 50 km at 24 h and 100 km at 48 h, comparable to operational forecasting systems. ● Case studies reveal that physics-driven variations in cumulus and boundary-layer schemes produce divergence in steering flow and inner-core structure. ● Tower observations during Typhoon Muifa (2022) confirm that near-surface wind uncertainty can be effectively bracketed by the ensemble framework.