NOAA Will Develop and Implement a New Global Forecast Model in Three Years

Tom Moore
Published: July 28,2016

With a lofty goal of building the world's best global forecast model, the National Oceanic and Atmospheric Administration (NOAA) is set to develop and implement a vastly improved version over the next three years.
There are a number of computer global forecast models that have been developed around the world for use by meteorologists. The current computer model used by NOAA is the (GFS) Global Forecast system.
Over the past few years, some critics of the GFS model pointed out that it lagged a bit behind the European model (ECMWF) in overall forecast accuracy.

Sample of current GFS model forecast (500 mb)

Creating The World's Best Global Forecasting Model

NOAA has made developing the world's best global weather model a top priority, and on July 27, 2016, they announced that a new state-of-the art global forecasting model will replace the current GFS model.
The new model will also be called the GFS, and during its developmental stage, it will be run in a test mode before it becomes fully operational in about three years. The model will be run on powerful supercomputers.
NOAA has identified some goals for the new GFS model:

• A unified system to improve forecast accuracy beyond eight to 10 days
• Better model forecasts of hurricane tracks and intensity
• Extension of weather forecasting through 14 days, and for extreme events, three to four weeks in advance

There was actually a competition of several models, one of which would be selected as the major "driver" of the new GFS model. These models were tested over a number of months, and a NOAA selection committee picked the Finite Volume on a Cubed Sphere (FV3) as the model that would be used.
This model was developed by NOAA's Geophysical Fluid Dynamics Laboratory (GFDL) in Princeton, New Jersey.
NOAA indicated that the FV3 core brings a new level of accuracy to the model's representation of atmospheric processes, such as air motions. This makes possible simulations of clouds and storms with a much higher resolution than what is used now.
As a result, the new GFS model will be able to provide localized forecasts for several weather events while generating a global forecast every six hours.
NOAA'S National Weather Service director Dr. Louis W. Uccellini issued the following statement: "Using our new dynamic core, which drives the model and the newest modeling techniques, we are poised to develop and run a more accurate and reliable global model that is used as a basis for all weather forecasts in the U.S."
He went on to say that they are "collaborating with the best model developers in the U.S. and around the world to ensure that the GFS has the most recent advances in the weather prediction modeling so we can accelerate improvements in the model as they are developed."
It's expected to take three years to develop the new GFS model.
There are still some critics of the decision to use the FV3 model as the "driver" for the new GFS model. There was another model (MPAS), developed by the National Center for Atmospheric Research in Boulder, Colorado, that survived until the final cut.
Dr. Cliff Mass, professor of Atmospheric Science at the University of Washington, believes that the MPAS model should have been the choice for NOAA.
"The Europeans are working on an updated version of their model that is similar to the MPAS," Mass said.
His concern is that the U.S. will end up further behind in weather forecasting.

Information obtained from weather satellites is one source of data used by computer models to produce forecasts
(NOAA)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

How Computer Forecast Models Work

A computer weather forecast model ingests a large amount of atmospheric data and uses principles of math and physics to assimilate the data and predict future weather.
Some of the atmospheric elements that are used by the computer model to make forecasts include temperature, pressure, wind, humidity, cloud cover and precipitation from around the globe and high up into the atmosphere.
The data for these projections comes from sources such as weather satellites, weather balloons and surface and oceanic observations.
There are many computer forecasts models that are used around the world.
The accuracy of computer model forecasts, especially over time, vary due to factors such as computing power, physics and data initially ingested into the model and resolution (grid size in degrees of latitude and longitude).
The higher the resolution, the more detailed and precise the forecast can be. For example, NOAA indicated that the FV3's capability to zoom in on critical weather events can be a great aid in forecasting hurricanes or significant storms, even more than two weeks in advance.
It will be very interesting to see if NOAA can reach its forecasting goals over the next decade with their new GFS model.
MORE: Wyoming Computers