Improving ERA5 maritime gust forecasts in East China using objective synoptic weather classification and random forest models

The ERA5 gust reanalysis data is crucial for offshore wind resources assessment, but its weather-dependent biases cannot be adequately captured by existing correction models. To address the issue, this study used the objective weather classification method to classify synoptic weather patterns (SWPs) and evaluate ERA5 gust in eastern coastal China from 2013-2022. On this basis, a Weather Type-based Random Forest (WTRF) model was developed by integrating random forest regression with SWP labels. The model's performance was evaluated using an independent test set in 2023. The results showed that ERA5 demonstrated notable variability across nine weather types in eastern China, with RMSE differences reaching up to 0.9 m/s, a correlation coefficient (R) varying by 0.16, and the critical success index (CSI) for gust force six differing by more than 0.2. The WTRF model effectively captured the differentiated mechanisms by which weather type influences gust formation, achieving a 12.2% reduction in correction RMSE compared to the non-classified approach. For extreme wind events, the WTRF model reduced both the MAE and MBE by 1−4 m/s and improved the CSI by 0.05−0.23. Overall, the WTRF model provided more accurate information for offshore wind resource assessment and enhanced forecasting capabilities for extreme gust events.