Subseasonal hurricane prediction using Nested-Grid HiRAM

The GFDL global High-Resolution Atmosphere Model (HiRAM) has demonstrated excellent interannual variability and seasonal prediction skill of tropical cyclones at half- and quarter-degree resolution. We present a new version of HiRAM using the nonhydrostatic solver and grid-nesting capabilities of the GFDL Finite-Volume Cubed-Sphere dynamical core (FV3). Sub-seasonal 30-day predictions are performed during the Atlantic hurricane season, using a 25-km global cubed-sphere grid combined with an 8-km nest over the tropical Atlantic; 64 vertical levels to allow sufficient resolution of the boundary layer, upper troposphere, and stratosphere; and updated precipitation microphysics, convection parameterization, and land-surface model. The locally-enhanced resolution and new physics permit the formation of intense tropical cyclones with rainbands and eyewalls. The capability for sub-seasonal prediction of tropical cyclone risk (both for active and inactive periods) and of individual tropical cyclones in the North Atlantic will be presented.