Every six hours this week, in a bit of meteorological choreography, dozens of National Weather Service (NWS) offices have been releasing weather balloons in unison. It is an effort to help answer a big question mark that has been hanging ominously over Florida all week as Hurricane Irma creeps closer: What part of the state—if any—will bear the brunt of a hurricane that is one of the strongest ever measured in the Atlantic Ocean, a storm that has already left a trail of destruction across the Caribbean?

High-resolution images from satellites and the dramatic “hurricane hunter” flights into the category 5 storm get all the glory. But whereas they are important methods of hurricane forecasting, it is the old-fashioned weather balloon—a stalwart of the meteorological world since the 1930s—that can provide some of the most detailed, reliable data on the interlocking puzzle of large-scale weather patterns that ultimately determine where storms like Irma go. After they are released as simply as a child loses a grip on a party balloon, the instruments they carry measure the state of the atmosphere from the ground up into the stratosphere, providing an instantaneous, three-dimensional cross-section of the atmosphere over the entire eastern U.S. That data is then fed into weather models to, it is hoped, give forecasters a clearer picture of Irma’s future. “This one will be devastating to someone,” says Mike Griffin, a meteorologist with the Springfield, Mo., NWS office.

Forecasters have known for days that Irma is going to turn northward toward the mainland U.S., but they did not know when exactly that would happen—and in this case the “when” makes all the difference for where the storm finally hits, what kind of damage it might do and the ability of officials to order evacuations as early as possible. To figure that out, forecasters have to look at the same weather patterns bringing sunny skies to Missouri or rain to Virginia because they also provide the pushes and pulls that move hurricanes around.

So far, Irma is being shoved westward by a high-pressure area over the western Atlantic. But it is expected to take that northward turn courtesy of the pull of a low-pressure trough developing over the eastern U.S. If the trough is strong and digs deep into the southeastern U.S., Irma would turn sooner and might rake the coasts of Florida, Georgia and South Carolina, bringing a devastating storm surge and winds to a densely populated coastline. But if the trough is weak and stays farther north, Irma could turn later in its journey and move up through the center of Florida or along its west coast, minimizing the storm surge risk (because its strongest winds would be blowing offshore).

Although state-of-the-art satellites do continuously monitor that bigger weather picture, they measure things such as wind and water vapor indirectly, based on the wavelengths of light they detect, so they are not as accurate as direct measurements, says Sim Aberson of the National Oceanic and Atmospheric Administration’s Hurricane Research Division. That is where the high-quality direct measurements provided by weather balloons come in.

The technology is nothing fancy. Each five-foot-diameter balloon carries a small, expendable instrument package that is encased in Styrofoam and measures humidity, temperature and wind as it rises from the ground up to 100,000 feet. It radios the data back, providing a detailed cross-sectional snapshot of the state of the atmosphere over most of its depth. This information gets plugged into weather models to provide the initial conditions for forecasts. The measurements also help provide ground truth for what satellites are seeing.

On a normal day weather balloons are lofted simultaneously every 12 hours at NWS offices across the country, a schedule that is usually enough to see what the atmosphere is doing, Aberson says. But with an uncertain forecast for a potentially historic storm, the National Hurricane Center on Sunday asked several NWS offices to conduct extra balloon releases to get more high-quality data into the models, in the hopes of improving the forecast. “The thinking here is that the more detailed information, with relatively high confidence in accuracy, is probably going to benefit the forecast,” says Greg Carbin, chief of the Forecast Operations Branch of the NWS Weather Prediction Center. This would give forecasters a better sense of how strong and deep the trough over the eastern U.S. will be.

The extra releases began in the central U.S., then expanded eastward to include 52 NWS offices, according to NOAA, the NWS’s parent organization. Such a large and coordinated effort involving extra launches is fairly rare, usually reserved for hurricanes that could do major damage and come with a particularly uncertain forecast, says Griffin, whose office is participating in the extra releases. One of the last times such an effort occurred was in the lead-up to Hurricane Sandy in 2012, when forecasters were trying to gauge the progression of a cold front that would affect Sandy’s track, says Daryl Williams, a forecaster with the Norman, Okla., NWS office, which is also taking part in the Irma effort. Extra releases are also sometimes used on a regional scale, when forecasters expect an outbreak of severe weather and tornadoes, for example.

Forecasters say the NWS offices will conduct the extra launches as long as the National Hurricane Center asks for them—likely at least through Friday, when they hope to have a better handle on who is in the storm’s path.