Surprise tornadoes in Kansas prompt concerns about changes at the National Weather Service

A tornado outbreak near Kansas City, Kansas, on Monday night came as a surprise.

At least three injuries were reported after at least five tornadoes developed in areas southwest of the city. Several homes were damaged, trees were downed, and RVs were overturned.

But in its Monday afternoon outlook, the National Weather Service’s Storm Prediction Center, which forecasts severe thunderstorms and tornadoes, did not anticipate a tornado threat for the Kansas City area. The disconnect has prompted concerns among some outside meteorologists that ongoing changes to staffing and weather balloon releases at the NWS might be leaving forecasters in the dark about threats.

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 — a change that several meteorologists think was made because of staffing issues.

“We are missing data at the normal times,” said Chris Vagasky, a meteorologist and research manager at the Wisconsin Environmental Mesonet, a statewide network of weather monitoring stations. He added that the staggered balloon launches Monday left a “big area over the southern Plains in the central United States without that weather balloon data, which might have caused the models to not forecast the day’s activity as well as it could have.”

The strongest tornado in the Kansas City area Monday was rated EF2, according to the enhanced Fujita scale, which rates tornadoes by wind speed and destruction. That tornado’s wind speeds reached about 125 mph, according to preliminary damage reports.

Forecast models did not show the threat of tornadoes until the risk was on the doorstep. The NWS issued watches at about 6:35 p.m., about half an hour before the first tornado. Local forecasting offices in Topeka, Kansas, and Kansas City, Missouri, sent out several warnings later. (A tornado watch means conditions are ripe for tornadoes to form; a tornado warning means a twister has been spotted on radar or is expected to form.)

Notice ahead of a possible tornado is critical to give people time to take shelter. Some local governments also sound tornado sirens to alert residents when forecasters issue tornado warnings. Ideally, a watch is issued several hours before an event, and then a warning goes out when a tornado is imminent. Research suggests tornado warnings typically go out about 15 minutes in advance.

Erica Grow Cei, a spokesperson for the weather service, said the changing cadence of weather balloons has not affected forecasts.

“NOAA’s weather model performance has not been impacted by any changes to the schedule of our twice-daily radiosonde (weather balloon) launches,” she said in a statement. “NOAA’s Environmental Modeling Center regularly evaluates the performance of the Agency’s weather models and publishes its findings on the EMC’s website. NOAA’s weather model performance shows no evidence of degradation.”

Weather balloons, which are filled with hydrogen or helium, carry small instruments called radiosondes into the upper atmosphere to measure temperature, humidity and wind speed and transmit data back to the ground. Because storms generally track from west to east in the U.S., the balloons often provide key information about what conditions might be expected downwind in the hours and days after they fly.

For decades, the NWS released weather balloons at a clockworklike cadence at more than 100 sites across the country, as well as over the Pacific and the Caribbean. But the Trump administration cut the agency deeply last year through buyouts and by firing probationary workers, leaving some forecasting offices short-staffed and scrambling to fill key roles. As a result, the NWS halted balloon launches at some locations and is delaying them elsewhere.

Alan Gerard, a meteorologist who retired last year as the director of analysis and understanding at the National Oceanic and Atmospheric Administration’s National Severe Storms Laboratory, said he thinks the balloon release schedule was changed because of continuing staffing issues.

“They’ve been deferring them to the day shift — midday — for when they have more staff,” he said.

Gerard said the NWS should study the issue: “It’s like we’re conducting a real-time experiment without any way to evaluate what the impacts of it are.”

He added that some forecasting models are timed to run in the morning — just after 7 a.m. local time. But Monday, balloon data from the Central Plains did not come in until after those morning models had run, so it was not factored in.

“You’re missing data from the Four Corners region, from the northern Plains, from part of the central Plains, and that’s where the storms were developing,” Vagasky said.

The unexpected tornadoes in Kansas followed a similar, much-scrutinized event in Michigan last month, when a tornado watch was not issued for southern Michigan and four people died. NWS offices did issue warnings when tornadoes were imminent, however.

Democratic Sens. Gary Peters and Elissa Slotkin of Michigan sent a letter asking NWS Director Ken Graham why a watch was not issued, whether short staffing played a role and how the NWS planned to improve tornado alerts. A spokesperson said the NWS had written a response directly to the senators but did not share it with NBC News.

Although the NWS has taken steps to bolster its ranks since the staffing shortages began, outside meteorologists said effects linger.

“It was something that was not going to be solved overnight, and I still don’t think it is solved, given what we’re seeing with the balloons,” Vagasky said. “I’m concerned with any severe weather day … and the real heart of tornado season is late April and into May.”

Forecasters at the NWS office in Kansas City said they viewed the atmosphere Monday as “volatile” and were aware that storms could grow severe if they formed. But the meteorologists thought storms had only a low chance of developing.

“There was a pretty strong possibility there would be no showers or thunderstorms at all — you’d have a primed environment where nothing ever develops or nothing happens,” said Brad Temeyer, a meteorologist in the Kansas City office. “It was a low-probability event of it occurring, but given that it did occur, it had high impact.”

Temeyer said the agency surveyed the tornado damage Tuesday.

“We didn’t have any significant injuries or fatalities, and in going out and doing the storm surveys today, everybody said they were prepared for potential tornadoes and took shelter once those warnings were in effect,” he said.

Brian LaMarre, a weather consultant who was the meteorologist-in-charge at the National Weather Service’s weather forecasting station in Tampa Bay, Florida, until last April, said he reviewed Monday’s forecasting models before and after the later-than-usual release of weather balloons.

In the morning, none of the models identified a thunderstorm risk for the area southwest of Kansas City, he said. However, once the models incorporated the balloon data, LaMarre said, they began to hint at clusters of thunderstorms. Further study would be needed to understand whether the shift was because of the added balloon data.

“It has to be investigated in terms of researching why did it not happen. Why did it not capture it? What was missed?” he said.

More severe weather could be coming for the Midwest. The Storm Prediction Center expects a risk of severe thunderstorms into the weekend.

For the Kansas City area, Temeyer said, “we want people to remain vigilant.”

This article was originally published on NBCNews.com