US launches deep learning mission to forecast major thunderstorms six weeks ahead

Huge thunderstorm cloud with lightning strikes (representational image)Getty Images Predicting a major thunderstorm outbreak several weeks before it happens is still a major challenge for weather forecasters. A new US Department of Energy-backed project is now exploring whether combining deep learning with traditional weather models can provide earlier warnings of these powerful storm systems. Planette AI, the Pacific Northwest National Laboratory (PNNL), and the University of Wyoming are working together on the project, called DL4MCS. The effort is part of the DOE’s Genesis Mission Phase I program and focuses on forecasting mesoscale convective systems (MCS) across the continental United States. MCS events are large clusters of thunderstorms that can stretch across hundreds of miles. They can bring damaging winds, heavy rain, flooding, and other severe weather. They also account for a significant portion of warm-season rainfall in the US. The problem is that the atmospheric conditions that lead to these systems can evolve over long periods, while the storms themselves are driven by processes that operate at much smaller scales. This makes it difficult for existing forecasting systems to accurately predict MCS activity beyond about a week. Bridging weather scales DL4MCS will investigate whether deep learning can help close that gap by working alongside physics-based forecasting models. The researchers will target a forecasting window ranging from roughly seven days to six weeks, known as subseasonal forecasting. The idea is not to replace established weather models with a machine learning system. Instead, the project will test whether computational methods can extract useful patterns from existing forecasts and improve predictions of when and where MCS events are likely to develop. The team will examine atmospheric conditions at different scales, from broad climate patterns that can influence storm formation to cloud-scale processes that determine how storms evolve. “DL4MCS reflects Planette AI’s commitment to delivering more actionable environmental intelligence for high-stakes decisions,” said Dr. Hansi Singh, Founder and CEO. “By combining state-of-the-art AI with proven physical forecasting systems, this project aims to make weeks-ahead storm risk information more useful for the sectors and communities that depend on better foresight.” PNNL will bring expertise in Earth system modeling and evaluating atmospheric processes, while the University of Wyoming will focus on regional modeling and downscaling. This is important because large-scale weather models typically operate at resolutions too coarse to capture all the details needed for local storm forecasts. “Improving prediction of mesoscale convective systems requires advances across scales, from large-scale climate drivers to the cloud microphysics that shape storm behavior,” said Dr. Susannah Burrows, Atmospheric Scientist at Pacific Northwest National Laboratory. “This collaboration brings together complementary strengths in Earth system modeling, AI, and process-level model evaluation to explore a new path toward better subseasonal forecasts.” Turning forecasts into warnings The University of Wyoming’s contribution will include converting large-scale forecasts into higher-resolution information that can be more relevant at the regional level. “The University of Wyoming is excited to contribute its expertise in regional downscaling, as well as its responsible integration with AI forecasting, to this effort,” said Dr. Stefan Rahimi, University of Wyoming Derecho Professor. “The ability to translate coarse large-scale forecasts into higher-resolution, decision-relevant guidance is essential for improving real-world preparedness and resilience.” If the approach works, the benefits could extend beyond simply predicting storms more accurately. An earlier indication of elevated storm risk could give utilities more time to prepare for disruptions, help insurers assess potential exposure, and allow communities to plan for severe weather and flooding. The project is also part of a larger DOE effort to use advanced computing and machine learning to tackle scientific problems that remain difficult to solve with conventional methods. The Genesis Mission brings together government, industry, and academic researchers to develop new approaches across areas including energy, science, and national security. For DL4MCS, the immediate goal is narrower but potentially significant: determining whether better use of existing physical forecasts and deep learning can make major thunderstorm systems more predictable weeks before they strike.The project is supported by the US Department of Energy’s Genesis Mission Phase I program.Recommended ArticlesGet the latest in engineering, tech, space & science - delivered daily to your inbox.With over a decade-long career in journalism, Neetika Walter has worked with The Economic Times, ANI, and Hindustan Times, covering politics, business, technology, and the clean energy sector. Passionate about contemporary culture, books, poetry, and storytelling, she brings depth and insight to her writing. When she isn’t chasing stories, she’s likely lost in a book or enjoying the company of her dogs.