▎ACHIEVEMENTS

Tropical cyclones (typhoons) play a crucial role in supplying water resources to East Asia during the summer months. However, the intense rainfall they deliver sometimes poses significant threats to Taiwan and the surrounding regions. Improving the accuracy of tropical cyclone forecasts is essential not only for water resource management and supply chain efficiency but also for disaster preparedness, helping to mitigate potential risks and losses. Understanding tropical cyclone genesis and the underlying sources of track predictability remain fundamental challenges for atmospheric scientists.

Unlike the deterministic weather forecasts (3-7 days), which are mainly influenced by initial atmospheric conditions, or seasonal forecasts, which are heavily affected by boundary conditions, such as ocean-atmosphere interactions, subseasonal forecasts (2-5 weeks) fall into a region of traditionally low signal-to-noise ratios, often referred to as the “Predictability Desert.” In a recent study, researchers at NTU employed an entropy-minimization approach from information theory to identify areas where tropical cyclone activity exhibits the longest predictability lead time. By applying this method, they demonstrated that forecasts of tropical cyclone tracks and genesis locations can be extended up to three weeks in advance. Their significant findings were recently published in npj Climate and Atmospheric Science.

This study was a collaborative effort between Professor Chun-Chieh Wu’s Typhoon Dynamics Research Lab and Assistant Professor Kai-Chih Tseng’s Chaos and Predictability Laboratory; the principal author was Research Assistant Chi Lok Loi. By analyzing the ECMWF S2S Reforecast dataset and applying statistical information methods, the researchers found that tropical cyclone predictability at three-week lead times is primarily driven by tropical intra-seasonal oscillations, which influence the effective beta gradient and relative vorticity stretching in the South China Sea. These large-scale atmospheric changes determine tropical cyclone genesis frequency and track density.

Additionally, they found that at shorter timescales (approximately 12 days), predictability signals primarily originate from mixed Rossby-gravity waves, which play a key role in the sequential formation of tropical cyclones in the major oceanic regions east of Taiwan and the Philippines. Moreover, experimental hindcasts incorporating enhanced initial conditions with improved signal-to-noise ratios were found to successfully extend the forecast range for Northwest Pacific summer tropical cyclones to 20 days.

Operational and Scientific Implications of the Study:

• Operational Impact: Offers improved early-warning capabilities for weather centers, water resource management, and disaster preparedness.
• Scientific Contribution: Identifies the physical connections between tropical wave activity and typhoon behavior, offering new insights into cross-scale interactions between typhoon dynamics and climate variability.

Citation
Loi, C. L., K.-C. Tseng, and C.-C. Wu*, 2025: Predictability of Tropical Cyclone Track Density in S2S Reforecast. npj Clim. Atmos. Sci. 8, 24 (2025). https://doi.org/10.1038/s41612-025-00909-0