Your weather forecast is about to get a lot worse

Vox

09-05-2025

is a correspondent at Vox writing about climate change, energy policy, and science. He is also a regular contributor to the radio program Science Friday. Prior to Vox, he was a reporter for ClimateWire at E&E News.
An aerial view of severe flooding in Frankfort, Kentucky, caused by days of heavy rainfall across the Midwest on April 7, 2025. Leandro Lozada/AFP via Getty Images
Did you check the weather forecast today?
Whether it was on your phone, the 5-day outlook in your newspaper, or your friendly TV meteorologist, that forecast was built on a massive government-run network of sensors and computers that get the weather right more often than not while rarely getting the attention they deserve. And now that system is being taken apart, piece by piece.
NOAA, the main US science agency that studies weather and climate, has already lost at least 2,000 workers since January thanks to a combination of layoffs, buyouts, and retirements. More job cuts may be looming. The White House says it wants to cut NOAA's $6 billion budget by almost 30 percent. The upshot is that with these cuts, efforts to make forecasts even more accurate will stall, while existing forecasts may get worse.
Weather forecasts are not just about whether or not you need an umbrella; they provide critical planning information for air travel, farming, shipping, and energy production. And they also save lives.
In April, massive spring floods hit the central and southern US. Ahead of the storm, the National Weather Service at NOAA warned of upward of 15 inches of rainfall. 'This is not your average flood risk,' according to a NWS bulletin from April 2. 'Generational flooding with devastating impacts is possible.'
The ensuing storms and floods killed at least 24 people, but given their extensive area, the death toll could have been much higher. In a report this week, scientists at the World Weather Attribution research group said that good storm predictions and effective emergency management were key to saving lives. Those forecasts and storm alerts were the product of decades of investment and infrastructure built up across the country.
'The US National Weather Service forecast the floods a week in advance and issued warnings throughout the event,' said Friederike Otto, a climatologist who leads World Weather Attribution, during a recent call with reporters. 'And as a result, people in the region knew when they needed to evacuate, and so the death toll was relatively small compared to similar events.'
But with the recent government cuts, the US is losing the data that informs these predictions and the scientists who produce them. The National Weather Service has been contending with understaffing for decades, and now the recent firings have made things worse as the US heads into another summer likely filled with more extreme weather.
Of the 122 NWS forecast offices across the US, 30 do not have a chief meteorologist at the moment.
Related Elon Musk is coming for our weather service
More people are living in vulnerable areas, and as global temperatures continue to rise, the destruction from extreme weather is getting worse. But even places that historically have avoided severe heat and torrential downpours are seeing dangerous weather become more common. That's why predicting the weather is more valuable than ever and why it's so alarming that the US is losing its capabilities.
How weather forecasts work
To build a weather prediction, you start with measurements of the earth, the sky, and the sea, sometimes from very far away. A thunderstorm in the southeastern US may have its seeds in the Pacific Ocean weeks in advance, for example. 'It all starts with data,' said Alan Sealls, president of the American Meteorological Society.
But how do we actually get all of that data? It can come from something as sophisticated as a geostationary satellite or as simple as a weather balloon. Twice a day, the National Weather Service launches these helium-filled orbs from 92 sites across the US. They take snapshots of temperature, humidity, and windspeed as they rise into the atmosphere. This data doesn't just inform weather models for the US but feeds into global models that help predict sunshine, rain, clouds, and snow all over the world.
'The weather balloons give us such detailed, precise data no other instrument can get,' Sealls said. 'If we don't have those in the area where that weather is likely to be hazardous or threatening, we potentially have a degraded forecast.'
Between balloons, radars, satellites, buoys, aircraft measurements, and ground instruments across the country, NOAA has built one of the most robust weather monitoring systems in the world, collecting 6.3 billion observations per day. The National Weather Service then plugs the data from these instruments into computer models to predict the next torrential downpour, cold snap, heat wave, or thunderstorm.
And almost all of this information is available to the public for free.
Related One agency that explains what the government actually does for you
However, most of us aren't collating our own, personal weather reports from raw data. We're getting them from any of the multitude of cell phone apps available — from the Weather Channel to Carrot Weather to Weather Underground — or from local experts who we trust. Meteorologists from news agencies or private companies also use the government's raw data and models to produce their own weather predictions that are focused on specific areas or draw on outside expertise and experience. That's why the forecasts on TV, in newspapers, or in weather apps can differ. But they still rely on the same foundational government data — especially when it comes to dangerous extremes.
'When we're on TV talking about hazardous weather, most of us around the country are in direct contact with the National Weather Service,' Sealls said. 'The National Weather Service, that is definitely the center point of all weather forecasting in the United States.'
Government researchers are also constantly improving weather forecasting. A modern 5-day forecast is as accurate as a 1-day forecast in 1980. A 72-hour hurricane track prediction these days is better than a 24-hour prediction from decades ago.
The National Weather Service's annual budget is just $1.3 billion — and yet its services add up to billions of dollars in economic benefits and untold numbers of lives saved. NOAA is now testing even better models for hurricanes and tropical storms that could provide up to five days of lead time.
Budget and job cuts are already hampering weather predictions
Already, some NWS sites in the US have reduced the number of weather balloon launches and some have stopped due to budget and staffing cuts.
'They've been short-staffed for a long time, but the recent spate of people retiring or being let go have led some stations now to the point where they do not have enough folks to go out and launch those balloons,' said Pamela Knox, an agricultural climatologist at the University of Georgia extension and director of the UGA weather network.
Since weather models rely so much on initial readings from real-world measurements like weather balloons, losing them can lower the quality of predictions. Losing personnel might also mean less maintenance on equipment like radars, leading to more outages. And having fewer staff scientists makes it more difficult to provide timely emergency alerts.
'We're becoming more blind because we are not having access to that data anymore,' Knox said. 'A bigger issue is when you have extreme events, because extreme events have a tendency to happen very quickly. You have to have real-time data.'
'If you have fewer people on staff,' she added, 'more things are going to fall through the cracks.'
At the same time, the climate is changing. That means that the historical patterns of weather no longer apply in much of the country, and continued warming will alter these trends further. The World Weather Attribution team estimated that the April extreme rainfall in the US was two to five times more likely due to warming, and that its intensity increased by 13 to 26 percent.
Yet at a time when the impacts of these changes have become impossible to ignore, the Trump administration is cutting climate research as well. The White House's budget proposal specifically 'terminates a variety of climate-dominated research, data, and grant programs, which are not aligned with Administration policy-ending 'Green New Deal' initiatives.'
The US did have a system for staying ahead of these long-term threats, but the Trump administration dismissed all the scientists working on the National Climate Assessment, a report required by law that assesses the current and future impacts of climate change to the country.
The result is a country facing a growing threat from natural forces but actively sabotaging its ability to stay ahead of them.