60 Verifying Hindcast Simulations of Recent Major Hurricanes in Variable-Resolution CAM

Tuesday, 17 April 2018
Champions DEFGH (Sawgrass Marriott)
Alyssa M Stansfield, SUNY Stony Brook, Stony Brook, NY; and K. A. Reed, C. M. Zarzycki, and M. F. Wehner

With fifteen named tropical systems and six major hurricanes, the 2017 Atlantic hurricane season has been exceptionally active. Hurricanes Harvey, Irma, and Maria were particularly costly, due to extensive damages through different combinations of strong winds, storm surge, and extreme rainfall. By exploring the Community Atmospheric Model (CAM)’s ability to simulate the observed intensity, size, and rainfall totals of these three storms, further use of this model to investigate the impact of climate change on the storms will be justified. CAM's Spectral Element dynamical core is set up in a variable-resolution configuration (CAM-VR) with a base grid spacing of 100 km and a refined region over the North Atlantic ocean basin with a grid spacing of 14 or 28 km. CAM-VR will be initialized with atmospheric analyses from NOAA’s Global Forecast System (GFS) 4-7 days before the time of interest (e.g., landfall or intensification) for each individual storm. A suite of ensemble members with slightly different perturbations in their initialization will be completed to account for model uncertainty in storm characteristics. Output from these simulations such as accumulated rainfall, storm size (both radius of maximum wind and overall size), and intensity changes will be compared to real-time observational datasets, both ground based and satellite when appropriate. Results with the 28-km CAM-VR will be compared against the 14-km CAM-VR configuration to explore potential improvements in the representation of relevant storm characteristics from increasing the horizontal resolution.
- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner