video: A satellite-view video showcasing the movement and structural evolution of the typhoon. The black track represents the ground truth, while tracks in various colors depict forecasts from different refinement configurations. The colorful dashed lines highlight regions refined to convection-permitting resolution, contrasted with coarser resolution areas in the rest of the globe.
Credit: ©Science China Press
Accurately predicting typhoon tracks is critical for mitigating disaster risks, yet recent advancements have plateaued triggering the discussion about whether the limit of typhoon track predictability has been reached. Researchers from the University of Science and Technology of China and collaborators have made a significant breakthrough by employing a global convection-permitting model at a 3-km resolution. Their study, primarily focused on Typhoon In-fa (2021), achieved unprecedented accuracy, reducing track errors to below 100 km over a 120-hour forecast period. Notably, the model successfully predicted Typhoon In-fa’s sudden track changes and dual landfall locations, outperforming current operational forecasts.
The team also developed an innovative variable mesh refinement strategy, which balances computational demands with accuracy by targeting key weather systems influencing typhoon motion. This method reduced computational costs by over 90% compared to a global convection-permitting model, while maintaining comparable precision.
These results mark a new era in typhoon prediction, demonstrating that convection-permitting resolution, combined with smart mesh refinement strategies, can enhance disaster preparedness and response with least computational burden.
This groundbreaking research will be published in Science Bulletin. The team plans to further validate their method across different basins and improve model physics to broaden its application globally.
See the article:
Pronounced advance on typhoon track forecast with global convection-permitting model
Journal
Science Bulletin