In 2012, Hurricane Sandy came barreling down on the East Coast. It slammed into the New York region and sat there for several days, dumping torrential rain that caused over a hundred deaths, flooded entire communities, and wrecked local infrastructure.
The destruction almost certainly would have been higher without the detailed, precise forecasts of how the storm would proceed along its track, which scientists were able to feed to emergency management personnel well before the storm made landfall.
Meteorological science has gotten better and better over past decades, squeezing ever more information out of the data gathered at Earth’s surface, through the atmosphere, and from instruments mounted on satellites spinning overhead. The result is increasingly sophisticated, far-reaching, and accurate forecasts.
But the precision we’ve become accustomed to from those forecasts may be under threat, scientists are warning. Our ability to predict with confidence what’s coming down the road weather wise could be set back by 40 years, and a key forecasting tool could be seriously degraded.
Telecommunications technologies like 5G internet need space on the electromagnetic spectrum, the range of all types of electromagnetic radiation that includes microwaves, infrared and ultraviolet light, gamma and X-rays. Today that space is at a premium. And much of the information that feeds into sophisticated weather models comes from parts of the spectrum that are right next door to areas telecommunications companies want to use for the new technologies.
“It’s like an apartment building of sorts,” explains Jordan Gerth, an atmospheric scientist at the University of Wisconsin, Madison. “There’s some general expectation that everybody keeps relatively quiet. In the spectrum land, we have our meteorological application, our science applications, and those that require a very quiet environment and adjacent environment. But the telecom signals are typically very loud, and are also susceptible to leaking outside their space.
“It’s like trying to run a daycare for small children who want to take a nap, but one that’s adjacent to a sports bar. There may be a wall between them, but you’re still going to get noise bleeding through.”
Over the past month delegates from countries and trade groups have gathered at the World Radiocommunications Conference to decide on international rules about how strictly to protect the “bands” of the electromagnetic spectrum crucial to weather forecasting—in other words, how much noise from the sports bar they’ll allow to be heard in the nap room.
In the end they came to a decision that some scientists—including Jim Bridenstine, the administrator of NASA—say may degrade the forecasts in a dangerous way, perhaps irreparably.
What’s at stake
One of the crucial bands, says William Blackwell, an atmospheric scientist and engineer at MIT, is around 23.8 GHz. Water vapor absorbs in this microwave band, leaving behind a faint signal that can be read by satellite-mounted instruments that look at the microwave part of the spectrum. The problem now is that telecommunications companies are interested in using parts of the spectrum right next to that water vapor signal.
The electromagnetic spectrum is like water in a river: There’s only so much of it. Some of the water is necessary to keep the habitat healthy, just like some of the spectrum is necessary for making weather forecasts. But most of the rest of the spectrum has already been allocated for all different kinds of wireless communication—GPS, radio navigation, satellite controls, telecommunications, and more. So demands on the remaining clear bits are growing.
“The reason we’re in this tug of war, it’s because of all these cell phones, just like the one that I’m holding,” says Tom Ackerman, an atmospheric scientist at the University of Washington.
In the past, the communications uses were kept far away from the bands used for weather and climate science work.
“But we’re running out of spectrum real estate,” says Ackerman. “Before, we could coexist nicely, but now the sandbox is full.”
Earlier this year, the U.S. Federal Communications Commission auctioned off part of the microwave spectrum right next to the 23.8 GHz water vapor band. Companies, eager for access to the new space, bid over $2 billion.
Prior to the auction, though, Jim Bridenstine, the administrator of NASA, warned that interference—“leaking” of the big 5G signal into the faint water vapor signal in the 23.8GHz band—could degrade forecast quality to levels not seen since before the microwave sounder era, in the mid-1970s.
At around the same time, NOAA’s acting deputy administrator, Neil Jacobs, told a congressional committee that telecommunications activity in the nearby parts of the spectrum could degrade forecasting accuracy by 30 percent, and could cause the lead time on forecasts of hurricanes to decrease by 2 to 3 days, he said.
They and other scientists asked for strict limits on how “loud” the next-door emissions could be—asking for something like the World Meteorological Association had suggested, a limit of -42 decibel watts (more negative numbers mean stronger limits). Instead, the FCC decided to use a limit of -20 decibel watts.
At the World Radiocommunications Conference this month, the decisions fell in between. The interference, decision-makers landed on, could be -33 decibel watts until 2027, at which point the limits would strengthen slightly to -39 decibel watts.
That’s better than what the FCC proposed, says Gerth, but still far from ideal. “This problem isn’t one that’s going to go away,” he says.
The leading trade group for the U.S. wireless industry, the Cellular Telecommunications and Internet Association (CTIA), disagrees. Their executive vice president, Brad Gillen, wrote in a blog post that the NOAA and NASA analyses were based on the wrong microwave sounder instrument, and if more modern ones are considered, the problem goes away. But NOAA and NASA and the Navy disagree.
The internal NOAA and NASA studies analyzing this particular issue are not yet public, so non-government weather scientists are still unable to vet the claims directly.
The satellite era changed weather forecasting
A century ago, the best weather forecasts in the world were mostly well-informed guesswork. Cloud patterns and the feel of wind could provide hints about what the atmosphere might get up to in the next few hours, but looking beyond that was impossible. Today scientists can look more than a week into the future and make a solid prediction of what to expect: rain, snow, sunshine, hurricanes.
By the 1970s, scientists had built the bones of the weather forecasting system we know today. They had developed computer models that described the complicated physics that controls the way air flows around the atmosphere. The more they honed in on the details of the physics, the better their predictions got.
But they found that the atmosphere was a fickle beast to understand. To predict what would happen with the weather in the future, they needed to know exactly what the weather conditions were right now, they found: The physics only worked if they had really good understandings of exactly where things started.
The game, scientists and engineers around the world recognized, was to get the absolute best possible data about the current state of the atmosphere. That would get them the best possible predictions of the future, they thought.
More and more effort went into developing instruments that could very precisely map the entire three-dimensional expanse of the atmosphere, things like exactly how warm the air was from the surface up to the stratosphere, and how much water vapor was streaked through the lower sections of the atmosphere versus higher up, over Chicago and Jakarta and in the middle of the ocean.
One of the critical developments was figuring out exactly how much and where the water vapor was. To do that, scientists relied on the fact that water vapor absorbs electromagnetic radiation at several different frequencies.
As the satellite observations improved, so did the precision and accuracy of weather forecasts. Today’s five-day forecasts are as accurate as a one-day forecast in the early 1980s.
More recently, observations from the many instruments called microwave sounders attached to different satellites orbiting the planet have become even more valuable to forecasters. In the last decade, research from the European Center for Medium-Range Weather Forecasting (ECMWF, considered the premier weather forecasting agency in the world) shows that microwave frequency data play a critical role in short-term weather forecasting, providing about 20 percent of the information critical to the forecast models.
Microwave sounders mounted on satellites like NOAA's Joint Polar Satellite System and the European Meteorological Operational (MetOp) satellites sense the amount of microwave radiation coming out of the atmosphere each time they pass overhead. They can tell how much water vapor is present by looking at the emissions in a suite of different bands—one of which is at 23.8 GHz. The water vapor signal in that band, though, is small, like a streamlet. The microwave sounders are now very good at measuring that faint signal.
Even if the part of the spectrum used for weather forecasting is protected thoroughly—and MIT’s Blackwell is not convinced it will be—there are many more bands crucial to weather forecasting that are under threat of similar encroachment.
“The same thing is happening in other parts of the spectrum,” he says. “The 5G spectrum is cozying up to those bands, to this sacred spectral territory. And that’s going to be a problem.”