One of the most significant impacts of tropical cyclones is the copious amount of rainfall they often produce. Drowning from inland flooding is the leading cause of death from landfalling tropical cyclones in the United States in the past 30 years. Despite this fact, little work to date has been done to assess the skill of numerical models at predicting rainfall from tropical cyclones.

This seminar will present a comparison of rainfall forecasts for U.S. landfalling tropical cyclones during the period 1998-2003 from the GFDL, GFS and Eta models. Results will be presented for approximately 28 storms that made landfall during that period. Results will be compared and contrasted for all 3 models and will be compared to forecasts from the benchmark statistical- climatological rainfall model (R-CLIPER) for all of the same cases.

New methods for validating tropical cyclone rainfall forecasts that have been developed as part of this project will also be described.