658 Development and Performance of the GFDL Global Prediction System—SHiELD

Tuesday, 14 January 2020
Hall B (Boston Convention and Exhibition Center)
Linjiong Zhou, NOAA/GFDL, Princeton Univ., Princeton, NJ; and S. J. Lin, L. Harris, K. Gao, B. Xiang, M. A. Bender, and J. H. Chen

The Geophysical Fluid Dynamics Laboratory (GFDL) has developed the fvGFS model, which couples the Finite-Volume Cubed-Sphere Dynamical Core (FV3) to an improved Global Forecast System (GFS) physical parameterization suite. This model served as a prototype for the new FV3-based GFSv15 which became operational by the National Weather Service on 12 June 2019. The model has been recently updated and renamed the System for High-resolution modeling for Earth-to-Local Domain (SHiELD). In the 2019 version of SHiELD, the dynamical core and physical parameterizations have been substantially improved. The physics upgrades include further development of GFDL cloud microphysics and cloud-radiation interaction. The options of using Yonsei University (YSU) planetary boundary layer, mixed-layer ocean model, and ocean surface roughness are also included. The 2019 SHiELD is currently used as a near real-time forecast system at GFDL that provides 10-day weather forecast every 6 hours. Beyond 10-day weather forecasts, the 2019 SHiELD is capable of convective-scale to sub-seasonal-to-seasonal predictions. The model is regarded as the GFDL’s Unified Forecast System (UFS) for weather-to-seasonal predictions and can leverage contributions from and to other UFS and FV3-based models. In this talk, the details of model development for the 2019 SHiELD as well as its prediction of convective-scale storms, global precipitation, hurricane track and intensity, and Madden-Julian Oscillation (MJO) will be presented.
- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner