Meteorologists are expressing growing concern about the impact of reduced or delayed weather balloon launches, warning that the missing data may be making severe weather forecasts less reliable. Weather balloons provide some of the most valuable data used in forecasting.
The National Weather Service, which is responsible for the launches, maintains that the lack of balloon data hasn’t degraded forecasts, but forecasters are frustrated by a dearth of research assessing the impact of the missing data following a handful of high-profile forecast misses in recent months.
Each weather balloon carries a radiosonde, an instrument that transmits measurements of temperature, humidity, air pressure, wind speed, and wind direction as it rises through the atmosphere. Observations from different levels of the atmosphere are critical to weather forecasting, especially for predicting whether thunderstorms will develop and how strong they might become.
The reduced balloon launches began in March 2025, when the Weather Service cut launches from twice daily to once daily at six locations because of staffing shortages. The following month, the Weather Service issued a notice saying launches could be temporarily reduced or suspended at selected sites “due to staffing limitations or operational priorities.”
“Under normal operations, the NWS conducts and/or supports radiosonde observations twice a day from 100 upper-air sites throughout the United States, Caribbean and Pacific Basin,” the agency said in the notice.
Staffing shortages have persisted since the Weather Service lost about 600 employees through Trump administration layoffs and buyouts last year, CBS News reported last week, even after President Trump granted an emergency authorization last August allowing the agency to hire 450 people over a one-year period ending this September.
While most sites are back to launching twice a day, launches at more than 25 locations are routinely being delayed from the early morning to the middle of the day, according to a Capital Weather analysis of upper-air data archived by the Weather Service’s Storm Prediction Center. The absence of morning balloon data became especially pronounced across a large portion of the western United States beginning in April of this year. Because weather systems often move from west to east, it’s possible the missing data could impact forecasts farther east.
In a statement to Capital Weather, Weather Service spokesperson Erica Cei said the agency has seen no evidence of an impact on forecasts, and that its models are not as reliant on data being collected at a specific time because they incorporate data gathered over the course of several hours.
“NOAA weather model performance has shown no evidence of degradation,” Cei said. “Our observing system is incredibly resilient with many redundancies, and we continue to expand our observational network by including more observations from the commercial sector, both at and in between traditional NWS upper-air sites.”
Although the Weather Service as of June had hired more than 300 of the 450 new employees authorized last August, according to the CBS News story, Cei told Capital Weather the agency doesn't have an estimate for when the delayed launches will be restored to their normal time.
Some meteorologists are pushing back on the Weather Service's assertion that forecasts haven't been impacted by gaps in balloon data. Politico reported last week that meteorologists and an emergency manager believed missing morning weather balloon data contributed to a poorly predicted tornado outbreak on April 13, near Kansas City, Kansas. The Weather Service issued a tornado watch only about 30 minutes before the first tornado. Such watches are issued when forecasters determine that atmospheric ingredients are in place that could lead to tornadoes.
In an April story about the same incident, NBC News reported that “many forecasting offices in the Great Plains did not launch weather balloons at 7 a.m. Monday, as they have for decades, and instead they released the balloons at noon,” too late to be incorporated into models that are timed to run shortly after 7 a.m.
Alan Gerard, a meteorologist who served in various leadership positions during a 35-year career at the National Oceanic and Atmospheric Administration, said the missing balloon data may have contributed to the poor model forecasts while acknowledging that thunderstorm and tornado forecasts are challenging even with the best data.
“I am not claiming that the lack of NWS balloon data upstream of these storms is the definitive reason for this poor forecast,” Gerard wrote in an analysis on Substack. “What I am saying, though, is that as an experienced meteorologist … it is hard for me to imagine that it does not have some negative impact on those model forecasts.”
Capital Weather meteorologist Matthew Cappucci said the gaps in weather balloon data are particularly troubling on days when there is potential for severe weather.
“We can’t go 18 hours without ascertaining how the atmosphere is layered, how much storm fuel has built up and if severe thunderstorms are going to erupt,” Cappucci wrote on X. “As an atmospheric scientist myself, I can say firsthand — the forecasts I’m able to offer you are less accurate than they would otherwise be. I’m not able to predict severe weather with the confidence I normally would. That is extremely concerning.”
Cei pointed to the growth of satellite observations and other data in recent decades as mitigating the impact of delayed balloon observations, noting that weather models ingest data from satellites, commercial aircraft, surface stations, radar, buoys, and other sensors in addition to weather balloons.
But NASA data from the past year show that radiosonde data from weather balloons were the most impactful observation type collected in the Northern Hemisphere. Their impact on a global measure of 24-hour forecast accuracy was roughly twice that of aircraft observations or any individual satellite instrument.
Yet the extent to which missing or delayed weather balloon data is affecting severe weather forecasts at a regional or local level remains difficult to quantify.
“Our view is that it may impact closer-by forecasting in the Plains, but less so in the Eastern U.S. Balloons are a very small percentage of the global dataset being used these days,” said Capital Weather meteorologist Matt Rogers, who also heads Commodity Weather Group, a private forecasting firm focused on the energy and agriculture markets. “Unfortunately, there is no way to know when and which situations would have been helped by balloon data, since we can’t go back and test the 'with data' versus 'without data' scenarios.”
That lack of quantifiable data is what frustrates Matt Lanza, a longtime meteorologist in the energy sector and managing editor of Space City Weather in Houston.
“Are satellites really doing the job well enough? We need to know. I've seen more than one instance of localized forecast busts in Houston, the Mid-Atlantic, Midwest, and elsewhere over the last 18 months, but I'm not sure if those are normal [thunderstorm forecasting] challenges or if it's directly related to the degradation of data,” Lanza told Capital Weather. “I've truthfully grown exasperated waiting for some sort of decisive answers.”
