DISCUSSION: In June 2019, the National Centers for Environmental Prediction (NCEP) announced that the Global Forecasting System (GFS) model has been successfully updated. The NCEP is the agency in the National Weather Service (NWS) that handles predictions for aviation, hurricanes, severe storms, oceans, and even space weather. The NCEP also is in charge of several of the models that are used by meteorologists such as the GFS, the North American Model (NAM), and the High Resolution Rapid Refresh (HRRR) model.
These models are used by meteorologists in the private sector as well as by networks to aid with making forecasts. The usage is based on the various model resolutions as NAM and HRRR are regional models with lower resolutions at 12 km and 3 km respectively while the GFS has a bigger resolution as it is a global model. The resolution is a key factor as lower resolution models factor in local topography and convective activity such as thunderstorms and cumulus clouds. In addition, another component is the capability on how far the models forecast, as the HRRR only forecasts for 18 hours, the NAM forecast for 84 hours, and the GFS goes for 384 hours which would help give a bit of an indication down the road especially for the hurricane track cone. However, the models would be less accurate the further out the forecast Is, especially the GFS in it’s 384 hour time period, due to different conditions such as topography, biases in temperature and winds, and events such as snow on the ground and smoke from a fire would play a factor that would change gradually every timestep.
The new update for the GFS includes a change in the dynamic core. The dynamic core is the engine of the model which takes in equations related to atmospheric movements such as the water cycle or the solar heating. The new core that is being implemented is called the Finite Volume Cubed-Sphere dynamical core (FV3) and among the features of the GFS-FV3 is the ability to simulate the atmosphere in hydrostatic and non-hydrostatic conditions. Hydrostatic conditions indicate that there is constant mass and constant pressure throughout the surface of a fluid such as the atmosphere.
The updated GFS-FV3 model was tested for over a year to fix any technical bugs. Among the test cases of the GFS-FV3 was Hurricane Florence as well as the January 2018 blizzard in the Northeast. The results of the two test cases were that the GFS-FV3 was performing much better than the old version of the GFS such as being closer to the actual precipitation amount in the case of the blizzard as well as having tracked Florence more accurately. The GFS-FV3 also is more efficient and faster than the legacy GFS on the computers at NCEP and in general.
The legacy version of the GFS will continue to be functional until September of 2019 in order to help resolve some issues of the GFS-FV3 as well as to provide data to users such as researchers and forecasters. The GFS-FV3 is already available to the public on some weather model websites. The GFS-FV3 will become a topic of many talks at upcoming American Meteorological Society (AMS) and National Weather Association (NWA) conferences in the next few years regarding its performance.
To learn more about other interesting weather observation-based stories from around the world, be sure to click here!
© 2019 Meteorologist JP Kalb