Evaluation of Experimental Models for Tropical Cyclone Forecasting in Support of the NOAA Hurricane Forecast Improvement Project (HFIP)

Wednesday, 17 December 2014
Paul A Kucera1, Barbara Brown1, Christopher Williams1 and Louisa B Nance2, (1)NCAR, Boulder, CO, United States, (2)NCAR/RAL, Louisville, CO, United States
The Tropical Cyclone Modeling Team (TCMT) in NCAR’s Joint Numerical Testbed (JNT) Program focuses on the verification of experimental forecasts of tropical cyclones (TCs). Activities of the team include the development of new verification methods and tools for TC forecasts and the design and implementation of diagnostic verification experiments to evaluate the performance of tropical cyclone forecast models. For the NOAA Hurricane Forecast Improvement Project (HFIP), the TCMT has designed and conducted verification studies involving various deterministic, ensemble, and statistical regional and global forecast models that participate in the annual HFIP real-time forecast Demonstration experiment. The HFIP Demonstration experiment is conducted during the months of August through October each year. The TCMT has applied new and established statistical approaches to provide statistically meaningful diagnostic evaluations of TC forecasts for storms observed during the Demonstration period. For this study, the TCMT has conducted evaluation of operational and experimental forecast performance for the 2012-2014 hurricane seasons in the North Atlantic and Eastern Pacific Oceans. This presentation will provide an overview of the Demonstration experiment along with a summary of results from the experimental model forecasts for the Atlantic and Eastern Pacific Ocean basins with the goal of documenting potential improvements to hurricane forecasts in comparison to operational baselines.