In mid-September 2025, an experimental forecasting system developed by the U.S. National Science Foundation National Center for Atmospheric Research (NSF NCAR) demonstrated remarkable accuracy in predicting the development and intensity of Super Typhoon Ragasa. This powerful storm rapidly intensified into a Category 5 event, reaching wind speeds of 165 mph, and was the most potent storm of 2025. The NSF NCAR's innovative approach utilizes the Model for Prediction Across Scales (MPAS), a sophisticated computer modeling system that simulates Earth's atmosphere with unprecedented detail.
This high-resolution global forecasting capability, operating at a 3.75-kilometer resolution, can capture thunderstorms worldwide and reveal how distant weather systems influence tropical storm evolution. "Essentially, this brings the weather into high-definition all over the globe," stated NSF NCAR scientist Falko Judt, who led the effort. "We believe this can significantly improve forecasting of extreme events like hurricanes and flash floods on a global scale." The experimental forecasts were generated in real-time throughout September, a period that typically marks the peak of the Atlantic hurricane season.
The MPAS model, powered by the advanced Derecho supercomputer, proved its efficacy not only with Ragasa but also by capturing the rapid intensification of Hurricane Gabrielle over the Atlantic. The primary focus of this initiative was on tropical cyclones in the Atlantic, eastern Pacific, and western Pacific, with an additional aim to assess performance in predicting extreme rainfall events. This cutting-edge modeling effort builds upon similar work by NSF NCAR. While a previous initiative focused on mid-latitude extreme weather with 3-kilometer forecasts, the current tropical cyclone focus involves forecasts extending up to 120 hours at a slightly coarser 3.75-kilometer resolution for computational efficiency.
Weather models represent the atmosphere using a grid system, and higher resolution, meaning closer grid points, generally leads to more accurate predictions. Judt noted that simulating the entire globe at high resolution could not only capture storms wherever they occur but also aid in anticipating tropical cyclones before they even form, potentially extending accurate forecasts 7 to 10 days into the future. This was evident with Super Typhoon Ragasa, as Judt observed, "MPAS forecasted this system to be a super typhoon even before the storm had formed," and at an earlier stage and with better intensity prediction than many operational models. The development of such high-resolution, high-quality data from models like MPAS also holds significant promise for training future generations of artificial intelligence weather models. The successful prediction and tracking of Super Typhoon Ragasa underscore the transformative potential of advanced modeling techniques in enhancing the accuracy and timeliness of weather forecasts, particularly for the most extreme events. Super Typhoon Ragasa itself was noted as the strongest storm of 2025, reaching sustained winds of 165 mph, equivalent to a Category 5 hurricane, and highlighting the increasing intensity of tropical cyclones in a changing climate.