Enhancing Tropical Cyclone Prediction Through AI Weather Model Ensemble Forecasts and Multi‐Source Data Integration
5 affiliations
- 1School of Computer Science and Technology Tongji University Shanghai China
- 2Laboratory of Cloud‐Precipitation Physics and Severe Storms Institute of Atmospheric Physics Chinese Academy of Sciences Beijing China
- 3Artificial Intelligence Innovation and Incubation Institute Fudan University Shanghai China
- 4Shanghai Academy of Artificial Intelligence for Science Shanghai China
- 5School of Mechanical Engineering Tongji University Shanghai China
Abstract
Tropical cyclone (TC) forecasting is essential for disaster preparedness, yet accurate intensity prediction remains challenging, particularly during rapid intensification. This study proposes TianXing‐TC, a lightweight probabilistic post‐processing framework that leverages ensemble forecasts from the AI weather prediction model TianXing and integrates multi‐source TC‐related information. The framework first generates perturbed TianXing ensemble forecasts and extracts TC‐centered environmental trajectories along the predicted storm tracks. A TC Correction Network then fuses these environmental trajectories with geostationary satellite cloud‐pattern imagery and storm‐state statistical factors to predict maximum sustained wind, minimum sea‐level pressure, longitude, and latitude up to 96 hr ahead. Compared with the raw TianXing ensemble, TianXing‐TC substantially reduces systematic intensity underestimation and improves track forecasts, particularly at longer lead times. Homogeneous comparisons with official forecasts show that TianXing‐TC consistently reduces intensity forecast errors and achieves competitive track forecast skill, with the largest long‐lead track improvements in the western North Pacific basin. Stratified verification further indicates useful short‐range skill during rapid‐intensification events, although rapid intensity changes remain underestimated. Integrated‐gradient analyses and case studies reveal physically plausible attribution patterns related to environmental flow, convective organization, and storm‐state persistence. These results suggest that AI weather model generated ensemble trajectories, when combined with satellite and statistical information, can provide an efficient complementary approach for TC forecast post‐processing.
Related
- Multivariate phase-space interpretation of ensemble forecasts: A comparison of physical and AI models
- Ensemble reliability and the signal‐to‐noise paradox in ECMWF subseasonal forecasts
- Rigorous uncertainty quantification of probabilistic AI weather forecasts with conformal prediction
- Ensemble Forecast Updates without Model Re-integration Based on Ultra-rapid Data Assimilation: Idealized Experiments with a Heavy Rainfall Case
- Skillful Seasonal Tropical Cyclone Prediction with AI Weather--Climate Models Forced by Dynamically Predicted SSTs
BibTeX
@article{yuan2026enhancing,
title = {{Enhancing Tropical Cyclone Prediction Through AI Weather Model Ensemble Forecasts and Multi‐Source Data Integration}},
author = {Shijin Yuan and Guansong Wang and Bin Mu and Feifan Zhou and Hao Li and Yake Zhang},
journal = {Earth and Space Science},
year = {2026},
month = oct,
doi = {10.1029/2026ea005040},
url = {https://agupubs.onlinelibrary.wiley.com/doi/10.1029/2026EA005040},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - Enhancing Tropical Cyclone Prediction Through AI Weather Model Ensemble Forecasts and Multi‐Source Data Integration
AU - Shijin Yuan
AU - Guansong Wang
AU - Bin Mu
AU - Feifan Zhou
AU - Hao Li
AU - Yake Zhang
PY - 2026
DA - 2026/10/10
JO - Earth and Space Science
DO - 10.1029/2026ea005040
UR - https://agupubs.onlinelibrary.wiley.com/doi/10.1029/2026EA005040
AB - Tropical cyclone (TC) forecasting is essential for disaster preparedness, yet accurate intensity prediction remains challenging, particularly during rapid intensification. This study proposes TianXing‐TC, a lightweight probabilistic post‐processing framework that leverages ensemble forecasts from the AI weather prediction model TianXing and integrates multi‐source TC‐related information. The framework first generates perturbed TianXing ensemble forecasts and extracts TC‐centered environmental trajectories along the predicted storm tracks. A TC Correction Network then fuses these environmental trajectories with geostationary satellite cloud‐pattern imagery and storm‐state statistical factors to predict maximum sustained wind, minimum sea‐level pressure, longitude, and latitude up to 96 hr ahead. Compared with the raw TianXing ensemble, TianXing‐TC substantially reduces systematic intensity underestimation and improves track forecasts, particularly at longer lead times. Homogeneous comparisons with official forecasts show that TianXing‐TC consistently reduces intensity forecast errors and achieves competitive track forecast skill, with the largest long‐lead track improvements in the western North Pacific basin. Stratified verification further indicates useful short‐range skill during rapid‐intensification events, although rapid intensity changes remain underestimated. Integrated‐gradient analyses and case studies reveal physically plausible attribution patterns related to environmental flow, convective organization, and storm‐state persistence. These results suggest that AI weather model generated ensemble trajectories, when combined with satellite and statistical information, can provide an efficient complementary approach for TC forecast post‐processing.
SN - 2333-5084
SN - 2333-5084
ER -