Study on the Intensity and Track Prediction of Typhoon Yagi (2024) Based on Ensemble Kalman Filter Assimilation of Satellite and Radar Data

Typhoons are among the most important natural disasters affecting coastal regions. Their track and intensity are crucial for disaster prevention and loss reduction. However, typhoon prediction still faces challenges, especially in simulating rapid intensification. This study investigates the impact of assimilating satellite and radar data using the Ensemble Kalman Filter (EnKF) on the prediction of Typhoon Yagi (2024). Based on the Weather Research and Forecasting (WRF) model and the EnKF method, a series of sensitivity experiments with different assimilated observational data types were conducted, including conventional observations, TCVitals data, Himawari-9 satellite radiance and brightness temperature data, and radar data. The study shows that, in the analysis fields, satellite brightness temperature assimilation can significantly improve the model representation of typhoon clouds and convective structures. Compared with the experiments without satellite brightness temperature assimilation, the assimilation of Himawari-9 brightness temperature data produces more concentrated strong convection in low-brightness-temperature regions, clearer rotational structures and spiral rainbands, and better agreement with observations. During the initial rapid intensification stage (RI1), satellite brightness temperature assimilation markedly improves the representation of typhoon cloud and convective structures. During the RI2 and RI3 stages, data assimilation can effectively enhance the dynamical inner-core structure of the typhoon, as reflected by lower minimum sea-level pressure, tighter isobars, and a more compact low-level circulation. Overall, the satellite brightness temperature assimilation experiments tend to produce a stronger and more compact typhoon structure at multiple times, while the conventional observation assimilation experiments also show some advantages at certain stages, indicating that the effectiveness of different assimilation schemes has clear stage-dependent characteristics. In terms of forecast fields, all experiments generally reproduced the typhoon's northwestward movement and its approach to the South China coast, but there was a consistent northward bias, and the track forecasts were quite sensitive to initialization time. Quantitative statistical results indicate that during the RI1 stage, the satellite brightness temperature assimilation experiments generally exhibited smaller track errors at most initialization times, outperforming other experiments. During the RI2 stage, the radar data assimilation experiments produced better track forecast results than the other experiments. In terms of intensity forecasting, all experiments were able to capture the overall trend of the typhoon's second rapid intensification phase to some extent. However, they generally underestimated the typhoon's intensity and failed to effectively capture the third rapid intensification process. As the initialization time was delayed, the ability of all experiments to simulate the rapid intensification process weakened further. In summary, data assimilation can significantly improve the structure of typhoon analysis fields and enhance track and intensity forecast skill. Satellite brightness temperature assimilation shows clear advantages in improving typhoon cloud and convective structures and enhancing the inner-core structure. However, the accurate representation of rapid intensification still needs further improvement.

Grantor dc:publisher

Helsingin yliopisto

Year dc:date.issued

2026

Author dc:creator

• Hu, Jie

• weather forecasting
• tropical cyclones
• storms
• simulation
• remote sensing
• Ensemble Kalman Filter
• Typhoon Yagi
• Data Assimilation
• WRF Model
• Rapid Intensification

Statement dc:rights

• In Copyright 1.0

Language dc:language.iso

eng

Handle dc:identifier.uri

http://hdl.handle.net/10138/634318

OAI identifier oai:identifier

oai:helda.helsinki.fi:10138/634318