How Are Typhoon Tracks Forecasted? Why Do Forecasts Vary Across Countries?
An explanation of typhoon track forecasting principles, numerical models, and forecast errors.
Principles of Track Forecasting
Modern typhoon track forecasting primarily relies on Numerical Weather Prediction (NWP) models: atmospheric data is input into supercomputers, which use fluid dynamics equations to simulate future atmospheric evolution. Major global models include:
- EC (European Centre for Medium-Range Weather Forecasts, ECMWF): Widely recognized for having the highest overall accuracy
- GFS (Global Forecast System, USA): High update frequency and freely available
- China CMA-GFS, Japan GSM, UK UKMO, etc.
Why Forecasts Vary Across Countries
- Model Differences: Different models use distinct physical parameterization schemes and resolutions.
- Initial Condition Differences: Variations in how observational data is assimilated lead to slight differences in starting points, with errors amplifying over time.
- Steering Flow Interpretation: Subtle differences in the strength of the subtropical high or the position of westerly troughs can cause tracks to diverge.
How Large Are Forecast Errors?
The average 24-hour track forecast error is approximately 60–70 km, around 110 km for 48 hours, and can exceed 300 km for 120 hours. The longer the forecast lead time, the more important it is to consider the spread among forecasts from multiple agencies—a larger spread indicates higher uncertainty.
Our typhoon detail pages provide an overlay comparison of forecast tracks from multiple agencies, including China, Japan, the United States, Hong Kong (China), and Taiwan (China), allowing you to visually assess these discrepancies.