Analysis shows wind speed and significant wave height asymmetry in cyclones, indicating new modeling approaches using XGBoost.
Key Points
As wind speeds exceed 20 m/s, wind radii correlations increase, while significant wave height radii show a decreasing trend.
Using over 600 dual-polarized Sentinel-1 images, the study explores wind and wave asymmetries across 300 tropical cyclones.
Machine learning technique XGBoost leads to improved predictions of significant wave height, reducing prediction errors compared to traditional models.
The new model is particularly effective under high-wind conditions, enhancing storm warning and mitigation strategies.