Houghton Lectures

Tropical cyclones are a major hazard and an energetic element of the climate system. There are thus both societal and scientific drivers to understand and predict the year-to-year variations of tropical cyclone activity, in order to support decisions and to test the hypotheses underlying the prediction methodologies. Predictions of tropical cyclone
activity aggregated over an entire basin (e.g., the North Atlantic) from months prior to the tropical cyclone season are generated regularly, and even show some measure of skill. However, for many downstream applications information at a more regional scale is necessary. In addition, regional prediction targets allow for more refined assessment of prediction methodologies. This lecture will introduce a new experimental high-resolution seasonal prediction system that enables the skillful prediction of regional tropical cyclone activity across much of the Northern Hemisphere months and seasons in advance. Key limitations and steps to improve these promising results will also be discussed.