The Forecast for the Weather Service Is Bad
If you have tips about the remaking of American climate science, environmental policy, or disaster response, you can contact Zoë on Signal at @zoeschlanger.99
At 4 p.m. ET yesterday, Andrew Hazelton got a form email telling him his work as a hurricane modeler at the federal government would be officially over at 5 p.m. that day. In his five months as a federal employee, his job was to help improve the models that serve as the basis for the National Hurricane Center’s forecasts. Now, he told me, “on my particular team there won’t be hurricane expertise.” He had been hired specifically for his storm experience, which he had built over nearly nine years working for the federal government.
Only, those were spent in contract positions. Before Hazelton joined NOAA’s Environmental Modeling Center, which keeps federal weather models running, he was on a team that developed NOAA’s next-generation hurricane-modeling system, which successfully predicted the rapid intensification of Hurricanes Milton and Helene last year. He also worked for a time on “Hurricane Hunters” missions that fly directly into storms to collect data. But because he’d joined the agency as a federal employee only in October, Hazelton was one of hundreds of people at the National Oceanic and Atmospheric Administration whose jobs were terminated yesterday in a purge of so-called probationary employees, who had been in their positions less than one year.
As I reported last summer, Project 2025—the compendium of policy proposals published by the Heritage Foundation prior to the 2024 elections, several authors of which are now serving under President Donald Trump—stated that an incoming administration should all but dissolve the NOAA. Among its other duties, NOAA employs thousands of people to help accurately predict the weather, through the National Weather Service; privatizing federal weather data has long been a project of some conservative lawmakers, too. Trump’s pick for commerce secretary, Howard Lutnick, said in his confirmation hearing that he would keep NOAA intact, but this first cut of NOAA staff is similar to those recently made to other federal agencies. NOAA was also one of the agencies that DOGE employees marched into earlier this month, and it is now bracing for further staff and budget cuts.
It costs the public roughly $4 per year per person to keep the National Weather Service functioning. In return, the NWS provides its own raw weather data to anyone who wants the information, and publishes its own public-facing weather reports; virtually every private forecast relies on freely available NOAA data as the basis for their weather reports. Dan Satterfield, a veteran television meteorologist, told me on Bluesky that many times he’d notice signs in the data that a storm could turn into a tornado, and that “NWS would have the warning out before I could get on air.”
The National Weather Service is already operating with the smallest workforce it has in years, a current NOAA employee, who requested anonymity out of fear for her job, told me. It’s doing highly technical work on a shoestring budget, and each and every cut will be felt. Already, a National Weather Service office in Alaska announced it would cease sending out weather balloons, which collect weather data, because of a lack of staffing. The chief meteorologist for CBS Detroit said on X that he gets most of his raw data for television weather reports from those balloon launches.
The average local National Weather Service forecasting office has about 20 employees. They work in shifts, ensuring that there are eyes on satellite data, radar, surface measurements, and other weather information 24 hours a day. After analyzing those data, they write forecasts that are shared directly with the public, and when a weather event develops, they coordinate with local emergency managers and county and state officials to develop warnings and make choices about measures like evacuations. Similarly, at NWS’s aviation-weather offices, meteorologists provide weather reports to pilots and Federal Aviation Administration air traffic controllers. Every fishing and sailing vessel off the coast similarly relies on daily NWS reports.
The loss of probationary employees is a particularly strong blow to the National Weather Service, the NOAA employee told me: Because they often have years of experience, but are still early enough in their careers, they tend to bring energy and ambition to the workforce. They were the plan for the agency’s future. Without them, “we have no pipeline of future employees,” she said.
Employees at NOAA were scrambling to find out exactly who among their colleagues were cut last night: Managers were not told in advance whom they were losing, and some employees had to tell their supervisors that they’d been cut, the employee said. But IT employees at local forecast offices were among those who lost their jobs, and “right now the IT infrastructure in the weather service is very fragile,” the employee said. It often suffers major data outages during vital moments, such as during a flooding disaster in the Midwest last summer. Without IT teams, outages could be more frequent and last longer. More Americans will go without forecasts, likely when they need them most.
Those forecasts also depend on work like Hazelton was doing to improve modeling for hurricanes and other fast-moving weather phenomena. Models use real-time data—including information collected by Hurricane Hunter missions, weather balloons, and satellite data—to set their initial conditions, then simulate how weather phenomena will behave. Hazelton worries what will happen to America’s ability to accurately predict hurricane behavior without him and his colleagues; he’s heard of a dozen other people who were cut from the Environmental Modeling Center last night.
Models need constant tending to use large streams of new data that come in, and hurricanes need continuous study; any slippage could lead to worse catastrophes. Hazelton estimated that the system he’d helped develop had made hurricane forecasts 10 to 15 percent better, more able to predict the rapid increases in storm intensity that are becoming more common as the climate warms. He was the most proud of that work. It almost certainly saved lives. But, he told me, “there still are some that sneak up on us sometimes. I don't want that to become more common now as a result of all this.”
The risks of disasters, and so the need for accurate forecasts, are only growing. Last year was the hottest ever for the contiguous United States, and the hurricane season was among the costliest on record. The country’s security is intimately linked to our ability to accurately predict the weather, particularly as the climate warms and extreme weather grows only more extreme.
Of course, NOAA itself has provided some of the key data that confirm that the climate is warming; it houses one of the most significant repositories of climate data on Earth, detailing shifting atmospheric conditions, the health of coastal fisheries, ice-core and tree-ring data, and countless other data sets. Project 2025 targeted the agency for exactly this reason: NOAA's research is “the source of much of NOAA’s climate alarmism," the policy book states. But climate and weather are closely related. Reducing the country’s understanding of and responsiveness to either will actively shove the public more squarely into harm’s way. The ability to prepare for turbulence that one cannot prevent is the only defense against it. For now, the forecast looks bad.