The Weather Company Advances Global Weather Forecasting with First Operational Implementation of JEDI System

Next-generation Weather Prediction Technology Will Lead to a Better Understanding of Forecast Uncertainty, Promising Up to 15% Improvement in Forecast Accuracy

ATLANTA, March 20, 2025 /PRNewswire/ -- The Weather Company today announced it is now the first operational user of the Joint Effort for Data Assimilation Integration (JEDI), a groundbreaking weather forecasting framework developed by the Joint Center for Satellite Data Assimilation (JCSDA) and hosted by the University Corporation for Atmospheric Research (UCAR). This milestone implementation underscores The Weather Company's commitment to pushing the boundaries of weather science and technology, especially with probabilistic forecasting to help improve decision making based on weather.

In an era where extreme weather events are increasingly impacting lives and economies, The Weather Company's implementation of JEDI within its global high-resolution atmospheric forecasting system (GRAF) marks a significant step forward in weather prediction capabilities. Internal testing by The Weather Company indicates forecast accuracy improvements of up to 15% – a significant advancement in meteorological science – are possible as its team continues to unlock the full potential of JEDI with advanced weather observations.

Proprietary to The Weather Company, the existing GRAF model updates 6x more frequently and up to 3x the resolution compared to other leading global models.¹ GRAF drives short-term forecasting for the day ahead, which is critical in planning for people and businesses.

In collaboration with JCSDA – a research center of excellence based on a multi-agency partnership among NOAA, NASA, and the U.S. Navy and Air Force, with key collaborators including the U.K. Met Office and the U.S. National Research Foundation National Center for Atmospheric Research (NSF NCAR) – this first commercial use of JEDI within GRAF demonstrates The Weather Company's leadership in adopting cutting-edge technology to deliver meaningful weather insights to millions of consumers and businesses worldwide.

"As the first institution in the world to operationalize JEDI for weather forecasting, The Weather Company is not just implementing new technology – we're revolutionizing how weather forecasting will be done in the future," said Dr. Peter Neilley, SVP of science and forecasting operations for The Weather Company. "We continue our commitment to innovate in weather forecasting technology and collaborate with the scientific community to solve global problems such as the increasingly disruptive impacts of weather."

JEDI provides the highest quality analysis available of the current state of the atmosphere, which is crucial in making future weather predictions. An open-community software seven years in the making, JEDI integrates vast amounts of eclectic and asynchronous observations – including radar, satellites, weather balloons, aircraft, weather stations and more – and turns them into a consistent format to be fed into forecast models like GRAF with unprecedented efficiency and accuracy.

"Weather doesn't operate in silos, and neither should we," said Dr. Thomas Auligné, director of the JCSDA. "This successful first implementation of JEDI demonstrates how cross-sector collaboration can help bridge the gap between government research and operational forecasting, ultimately driving breakthrough innovations that benefit society as a whole."

As the first operational user of JEDI, The Weather Company will create benefits within its GRAF forecasts and insights shared with consumers and customers. For example:

• Increased forecast accuracy: Improved data assimilation will enable material improvement in forecast accuracy and day-ahead skill for years to come, as the team continues to advance and upgrade the system.
• New cutting-edge weather observations: JEDI will enable The Weather Company to make better predictions by incorporating more types of weather observations into its GRAF system, such as specialized types of satellite observations, five-second updating data from aircraft, and mobile device pressure readings.
• Enhanced probabilistic forecasting: JEDI enables ensemble modeling, or multiple forecast runs, which delivers more accurate predictions of weather's uncertainty. Seeing a range of potential outcomes can be crucial for critical decision making based on weather. An example of probabilistic forecasting can be seen in the newly released snowfall feature in The Weather Channel Premium app.

Behind the Technology: A New Era in Weather Forecasting
While recent attention has focused on single weather models, a high-performing prediction system requires much more than a single model. A true modeling system like GRAF has three basic components: (1) a data assimilation system that brings observational data into models; (2) a model core, which for GRAF is based on MPAS by NSF NCAR and tailored by The Weather Company team; and (3) a post-processing system that takes the model forecasts and turns them into useful products.

While media attention often highlights AI-based weather model cores, there is a growing body of evidence that the most important element in an accurate modeling system is the data assimilation. It serves as the foundation for all forecasting by providing more accurate initial conditions for weather predictions. Even newer AI models – such as the first-of-its-kind kilometer-scale model The Weather Company is developing with NVIDIA – will still benefit from physics-based models and technologies like GRAF and JEDI. And combining GRAF forecasts with artificial intelligence will also improve short-term precipitation forecasts, such as rain alerts down to the scale of a thunderstorm, to help people and businesses plan for what's ahead.

As the world's most accurate weather forecaster,² The Weather Company's comprehensive approach combines advanced data assimilation through JEDI, the MPAS-based GRAF modeling system, and AI-powered blending of approximately 100 forecast models – including both physics-based and AI-based models. Expert meteorologist oversight then helps deliver unparalleled accuracy for location-specific forecasts, particularly crucial for high-impact local weather events affecting communities and businesses worldwide.

This advancement will further enhance The Weather Company's ability to provide the most accurate, actionable weather insights across its consumer and enterprise solutions. For more information about The Weather Company's efforts and its impact on weather forecasting, visit weathercompany.com.

About The Weather Company
The Weather Company helps people and businesses around the world make more informed decisions and take action in the face of weather. With its deep industry expertise and highly accurate, high-volume weather data combined with advanced technology and AI, The Weather Company provides insights and solutions that harness the power of weather in a scalable, privacy-forward way. The world's most accurate forecaster, the company serves hundreds of enterprise customers across media, advertising, aviation and more, and is trusted by hundreds of millions of monthly active users via digital properties from The Weather Channel (weather.com) and Weather Underground (wunderground.com). For more, visit weathercompany.com.

About JCSDA
The Joint Center for Satellite Data Assimilation (JCSDA, jcsda.org) is a collaborative research center hosted by the University Corporation for Atmospheric Research (UCAR). It brings together key U.S. operational agencies — including NOAA, NASA, the U.S. Air Force, and the U.S. Navy —— and collaborators — including the Met Office and NSF-NCAR —to advance the use of satellite data in weather, ocean, and environmental forecasting. At the heart of JCSDA's efforts is the Joint Effort for Data assimilation Integration (JEDI), an open-source software initiative that accelerates the transition of research into the next generation of operational forecasting systems. By fostering interagency cooperation, harnessing state-of-the-art computational infrastructure, and engaging a broad scientific and user community, JCSDA ensures that satellite observations are fully leveraged to enhance environmental prediction capabilities..

About UCAR
The University Corporation for Atmospheric Research is a nonprofit consortium of more than 130 North American colleges and universities focused on research and training in Earth system science. It manages the U.S. National Science Foundation National Center for Atmospheric Research on behalf of NSF.

¹GRAF model updates its global forecasts every hour, unlike other conventional global models that typically update every six hours. Current global weather models forecast down to 9-13 kilometers (5.6-8.0 miles), while GRAF forecasts down to 4 kilometers (2.4 miles).

² ForecastWatch, Global and Regional Weather Forecast Accuracy Overview, 2017-2022, https://forecastwatch.com/AccuracyOverview2017-2022, commissioned by The Weather Company.

SOURCE The Weather Company