Why Don't Scientists Know Precisely When and Where the Chinese Space Station Will Fall?

The Chinese space station known as Tiangong-1 is falling to Earth uncontrollably after a technical malfunction two years ago meant its owners could no longer carefully steer it to be destroyed over the ocean. It will likely finally tumble down this weekend, ending a months-long hubbub about its fate.

But while the satellite has been falling for two years, even the newest estimates of when it will fall can only narrow its landing down to a day-and-a-half long window. And scientists have essentially no idea where the thing will fall. All they've been able to do is rule out northern and southern swaths of the planet that the satellite has never crossed over.

That level of uncertainty may not sound particularly comforting, but there are a host of very good scientific reasons why it's so hard to predict Tiangong-1's final destination and arrival time. The descent itself is governed by the space station's interaction with the atmosphere that cocoons Earth, but the atmosphere is a complicated place, and Tiangong-1 is a complicated thing.

Read more: A Brief History of Things Falling Uncontrollably From Space

Take the satellite's weight, for example—it's relatively large, so it should fall relatively quickly, but the rate also depends on how that weight is arranged. "It's not just about how dense the upper atmosphere is, it's also about whether what's orbiting is a feather or a cannonball," Jonathan McDowell, an astrophysicist at the Harvard-Smithsonian Center for Astrophysics in Massachusetts who tracks launching and re-entering satellites on Twitter, told Newsweek.

The satellite's position can also change its speed, he added, since it experiences drag in the atmosphere just the same way a swimmer does in water. The bulk of Tiangong-1 is a fairly narrow cylinder, so if it falls side-on, that exposes much more of the surface to drag than if it falls round end first.

The atmosphere itself can vary as well, he said, since it is affected by the sun. When the sun is more active and spurts more charged particles toward Earth, satellites fall faster. That's part of why Tiangong-1 has taken longer to fall than early calculations suggested. "A few months in advance, we can be way off," McDowell said. "The estimates of when it's going to reenter have been slipping later and later because the solar activity has been less than expected."

As the Air Force tracks an object and releases new information about its orbital position, scientists work to predict when precisely it will actually fall given all the questions they can't answer about its journey. "Those are unknowns that we typically never actually know," said Roger Thompson, an engineer at Aerospace Corporation's Center for Orbital and Reentry Debris Studies, whose work includes running those calculations. "We don't just plug it into one model, we have about five different models that we use."

Each model uses a slightly different combination of factors to produce a different estimated re-entry time. Thompson and his colleagues run these models on about 50 or 60 items every year, he said, and as a reentry becomes imminent, the calculations sometimes occur multiple times per day. "I've done it twice already today," he told Newsweek around 2 p.m. ET on Friday.

Tiangong-1 launched in September 2011. Lintao Zhang/Getty Images

That's all in the hope that when Tiangong-1 finally does fall, someone will see it, and Thompson will be able to match all those model estimates against what really happened. "Every time they make a prediction and they see how it turns out they can adjust their models," he said. "It helps us improve our prediction capability."

The reentry time is important because that determines where the object falls. Tiangong-1, like everything else in orbit around the Earth, is speeding at about 17,000 miles per hour, circling the planet more than a dozen times a day. That means that even if scientists could narrow its reentry window to just one hour—compared to the 30- or 32-hour windows models are producing now—they would have a 17,000 mile stretch of Earth to consider.

"I'll be in the office [Saturday] and probably early Sunday morning," Thompson said, sounding more excited than resigned. "We'll be watching it the whole way."