New engine is driving NOAA’s flagship weather forecast model

Engines combine strengths of climate and weather models

As NOAA launches a major upgrade of its flagship weather forecast model today, an important part is the Global Forecast System’s new dynamical core. The story of how scientists developed the dynamical core or engine for the model is a view into how scientific invention works.

The dynamical core, called the Finite-Volume Cubed-Sphere, or FV3, is a key model component that computes wind and air pressure for successful numerical weather prediction. It is expected to bring a new level of accuracy and efficiency to the Global Forecast System’s representation of atmospheric processes from the jet stream, to thunderstorms, to hurricanes, to winter blizzards. In the future it has a potential to improve simulations of clouds and precipitation at resolutions not yet seen in an operational weather forecast model.

The result is expected to be more accurate and timely weather forecasting, and over time, longer lead times for weather forecasts. This will save lives and protect property and strengthen our economy that depends on accurate weather prediction.

“This week’s launch of the upgraded Global Forecast System with FV3 is a great example of why we need continued investment in long-term sustained research,” said Craig McLean, NOAA assistant administrator for Oceanic and Atmospheric Research. “It took many years of dedicated scientists’ commitment to convert an idea into an operational model’s dynamical core.”

Dynamical core team

A NOAA team led by Shian-Jiann Lin, center back row, developed the Finite-Volume Cubed-Sphere dynamical core and adapted it for NOAA's Global Forecast system. Credit: NOAA

Like many great scientific advancements, the FV3 began as research to solve another problem.

More than two decades ago, a young scientist named Shian-Jiann Lin joined NASA’s Goddard Space Flight Laboratory to work on models to predict how chemical substances in the air, including those that destroy the Earth’s protective ozone layer, travel through the atmosphere.

Modeling ozone destroyers

Working with other scientists, Lin developed a model to represent how flowing air carries these substances. The new model divided the atmosphere into cells or boxes and used computer code based on the laws of physics to simulate how air and chemical substances move through each cell and around the globe.

The model paid close attention to conserving energy, mass and momentum in the atmosphere in each box. This precision resulted in dramatic improvements in the accuracy and realism of the atmospheric chemistry.

Richard Rood, now a professor of atmospheric science at the University of Michigan, who had hired Lin to work with him at NASA, said Lin has a unique ability to visualize the dynamic flows of the atmosphere, called fluid dynamics. “Rather than write down the basic mathematical equations and start to derive relationships to represent the atmosphere, S.J. jumps ahead and visualizes the algorithm that can physically solve the problem.”

NASA began using this dynamical core, called the Finite-Volume dynamical core, as a key component of its weather and climate models. Soon others wanted in on the action. The National Center for Atmospheric Research and Harvard University adopted FV for its climate and atmospheric chemistry models in the early 2000s.

Highpowered engine for climate models

In 2003, Lin was recruited by Ants Leetma, director of NOAA’s Geophysical Fluid Dynamics Laboratory (GFDL), to join GFDL in Princeton, New Jersey.

“I love the place,” Lin said of GFDL recently, when describing how easy it was for him to make the switch from NASA to NOAA. It was a coming home for Lin who had completed his doctorate in atmospheric science at Princeton University where he’d worked with scientists at NOAA GFDL.

Very quickly Lin worked to advance the FV dynamical core to improve the world class climate models created by and used at GFDL. In 2007, a GFDL atmosphere and ocean climate model with the FV core outperformed all other climate models during a worldwide comparison by the Intergovernmental Panel on Climate Change in preparation for its fourth climate assessment.

Over the next few years, scientists at GFDL continued to improve climate models aided by high-powered super computers. As FV became FV3, models went from simulating climate on a grid of 60 to 120 square miles to more precise areas of less than two square miles.

The cubes of the FV3

The FV3 dynamical core divides the atmosphere into small cubes arranged on a grid and computes the changes in winds and pressure within each cube as part of a model's forecast. Models using the FV3 have the capability to zoom in on storm systems to improve predictions. Here, we see a satellite image of Hurricane Sandy with a stretched version of FV3's cubed sphere grid, zoomed in on the hurricane. Credit: NOAA

Then along came Hurricane Sandy in 2012, the worst hurricane the U.S. had faced since Hurricane Katrina in 2005. Prompted by the disastrous hurricane, Congress passed the Disaster Relief Appropriations Act which included funding for NOAA to dramatically improve U.S. hurricane forecasting and develop a forecast system that would be second to none. Lin recalls getting a call around this time from Tim Schneider at NOAA’s Earth System Research Laboratory, who asked him to join the NOAA effort to improve weather prediction of hurricanes and other high intensity events.

Combining strengths of climate and weather models

Lin and his colleagues began adapting the FV dynamical core for use in NOAA’s next generation Global Forecast System. It would be the first time in nearly 40 years that the GFS would have a new dynamical core.

To drive innovation and find the best dynamical core for its new global forecast model, NOAA collaborated with partners across academia and the weather community. After a thorough review of different dynamical cores, NOAA chose the GFDL dynamical core, citing its superior accuracy, skill with global and regional forecasts, and computational efficiency.

Over the last three years, scientists from NOAA’s National Weather Service and NOAA GFDL have worked together to transition the FV3 to become the vital engine of the Global Forecast System. During extensive testing of the FV3-based GFS and other upgraded model components, it has outperformed the older GFS in predicting tropical cyclone track and intensity.

With a strong history of powering climate models that predict long-range weather patterns on scales of seasons to decades, the FV3 is now in the operational trenches helping NOAA forecast day-to-day weather to protect American lives and communities.  

“This is an exciting time for weather forecasting,” said Brian Gross, Ph.D., director of NOAA’s Environmental Modeling Center within the National Weather Service. “The GFS with the FV3 dynamical core brings together the superior dynamics of global climate modeling with the day-to-day reliability and speed of operational numerical weather prediction.”

For more information, please contact Monica Allen, NOAA Research Communications, at 301-734-1123 or by email at monica.allen@noaa.gov