Hurricane season brings with it a certain kind of dread—not the storm itself, but the hours leading up to it, when you watch the forecast cone change on your screen and wonder how accurate any of it is. Most people believe that the best data is somewhere. They believe the system is operational. That assumption needs to be seriously examined right now.
The National Oceanic and Atmospheric Administration has spent decades developing what many meteorologists believe to be the world’s most advanced weather forecasting infrastructure, including ocean buoys sending live data from the Atlantic, satellites measuring ocean temperatures, and weather balloons rising twice a day from hundreds of stations. Researchers describe the current cuts to that infrastructure as “cutting into bone” rather than “trimming the fat.”
The proposed budget for the fiscal year 2026 by the Trump administration would cut NOAA’s overall funding by approximately 25%, or more than $1.5 billion. A 74% reduction would be made to the Office of Oceanic and Atmospheric Research, which serves as the center of the agency’s climate science efforts. It was stated clearly in one budget memo that was leaked: “At this funding level, OAR is eliminated as a line office.” It might not endure a 57-year collaboration between NOAA and Princeton University that created what are widely regarded as the most sophisticated climate modeling systems in the world.
Apart from the politics, it’s difficult to ignore what’s going on at the operational level. Some National Weather Service offices have already had to spend the night without on-site forecasters due to staffing reductions. The number of balloon launches, those unglamorous but crucial twice-daily missions that gather vertical atmospheric data, has decreased. These losses are not symbolic. Weather models are only as good as the data they are fed, and when there are fewer observations, the models have less to work with.
It’s hard to disagree with how NOAA’s former acting chief scientist, Craig McLean, put it: weather accumulated over time becomes climate. The weather forecast will eventually deteriorate if the climate research is cut. The two systems are not distinct from one another. They depend on one another, and cutting off that connection results in gradual, cumulative errors that nobody will notice until a storm catches someone off guard rather than an immediate failure.
NOAA’s recent investment in AI-powered forecasting makes this especially challenging. In an effort to increase speed and accuracy, the agency released a set of artificial intelligence models late last year that were trained on centuries’ worth of historical weather data. Now, the timing seems ironic, perhaps even painful. In the words of Monica Medina, a senior NOAA official during the Obama administration, “you need AI to process more data faster, but what you’re actually doing right now is collecting less data.” While the raw materials that power the technology are getting smaller, the technology itself is growing.
Scientists in this field believe that the damage won’t be apparent right away, and that’s exactly what makes it risky. Errors in weather forecasting are not immediately apparent. They show up when a hurricane intensifies more quickly than the models predicted or when a tornado follows a path no one anticipated. By then, no one wants to discuss how the failure was caused by a budget cut made two years prior.
The timing of these cuts could hardly be worse, with NOAA expected to release its 2026 Atlantic hurricane season outlook soon and a possible super El Niño already identified as a driver of elevated storm activity. Whether Congress will significantly oppose the proposed cuts is still up in the air. However, the scientists who actually construct the models appear to agree that the direction is incorrect and that it will take longer than anyone would like to reverse it.
