The number of proposed giant constellations of satellites, mostly in low Earth orbit and mostly devoted to providing broadband internet service anyplace on Earth, just keeps growing, with no international regulations yet in place to limit these new megaconstellations. Within a few years, the number of such satellites in orbit — presently more than 5,000 — could increase by at least an order of magnitude, possibly even to hundreds of thousands, threatening the scientific output of planned and existing optical and radio observatories.
The International Astronomical Union (IAU) held a conference in October in La Palma, in Spain’s Canary Islands, devoted to estimating the extent of the problem and examining various technical and regulatory measures to attempt to mitigate the impact. Already, images taken by ground-based telescopes, radio telescope arrays, and even the Hubble Space Telescope have been affected by streaks of light or streams of radio noise produced by the rapidly growing fleet of StarLink satellites launched by Elon Musk’s SpaceX and those of other companies offering similar services.
How satellites affect astronomy
There are three main issues potentially posed by this dramatic increase in the number of satellites in orbit, explains Richard Green, the interim director of the IAU’s Center for the Protection of the Dark and Quiet Sky from Satellite Constellation Interference. These are the optical interference, the streaks and flashes of reflected light that can especially interfere with wide-field survey observations; the background sky glow and other issues associated with debris in space and the increasing risk of collisions as low Earth orbits become more crowded; and the various kinds of radio interference, both intentional and inadvertent, associated with these satellites.
As far as the optical effects, the impact depends on the telescopes involved. For some of the newest generation of giant telescopes, it may not matter much, says Robert Kirshner, director of the Thirty Meter Telescope (TMT) project, slated for installation on Hawaii’s Maunakea (if a suitable agreement is reached with the local population). “We have done a study, looking at what if there were 80,000 satellites,” he says, to see what percentage of TMT observations might be impacted. It depends a bit on which of several instruments are being used, he says, but “the answer is, our field of view is so small that it’s about a tenth of a percent.” What’s more, he says, most of their detectors will be read out 10 times a second, so it would be easy to throw out the few tenths of a second that happened to be affected.
But for wide-field telescopes like the 8.4-meter Simonyi Survey Telescope at theVera Rubin Observatory, currently nearing completion in Chile, he says, “it’s going to be a real headache.” For them, “every once in a while, the thing you most want — the galaxy with a supernova that’s lensed by a cluster, something that’s really important and interesting — is going to have a streak across it.”
Green says that software can to some extent remove streaks from an image like those that will be produced by observatories such as Rubin. But it’s not only the streaks themselves, because they can also produce artifacts in the pixels adjacent to the streak: “If the streak itself is too bright, then you enter into a regime that’s no longer easily calibrated, and then you lose 10 times the area, and it gets to be significant,” he says. Studies show that in order to avoid seriously interfering with such observations, reflections from satellites should be kept to no brighter than 7th magnitude as seen from the ground. “Coincidentally, that is below the limit where most normal human eyes can see a star,” he points out, “so it would also make them not readily visible and not change the appearance of the night sky.”
SpaceX itself has been working hard to reduce the brightness of its satellites, since its license from the U.S. Federal Communications Commission requires it to coordinate with the National Science Foundation to find acceptable ways of mitigating their effects on astronomy. Attempts to reduce the Starlink satellites’ reflectance by painting them a dark black backfired, causing the satellites to absorb too much solar radiation and overheat. A system using thin sunshades to block sunlight from reflecting off the satellites has been more successful, though not quite sufficient to meet the goal of keeping their brightness below the magnitude 7 limit. “They’ve taken the problem very seriously, and they’ve devoted real engineering effort to try to address it,” Green says. But the next generation of Starlink satellites will be twice as big, potentially exacerbating the issue.
In September, the FCC issued similar requirements for coordinating with the NSF to other companies planning large satellite constellations: OneWeb, Project Kuiper, Iceye, and Planet Labs. In addition to reducing their satellites’ brightness, the companies are supposed to avoid beaming any radio signals in the vicinity of radio observatories, and to provide information on planned orbits and operations to observatories that might be affected.
As for the potential for collisions and the resulting debris, that’s something that could pose a traffic hazard for spacecraft, and an overall brightening of the night sky that could hamper many ground-based observations. These are effects “no one wants,” Green says, so all involved parties have an interest in finding solutions. “We are aligned with everyone else when we say what’s needed is strict traffic management, and careful disposal of satellites at the end of their mission,” Green says.
Patrick Seitzer, professor emeritus of astronomy at the University of Michigan and an expert on orbital debris, says “orbital debris is a huge problem. Right now, I think you could go to a meeting on that topic every week.” At present, the FCC does require any company launching a constellation to have an orbital debris mitigation plan in place before launch, he says, and the European Space Agency has similar rules, but there are still no formal international regulations.
The effects of these satellites on radio astronomy are a bit more complex. Major radio telescope installations are placed in designated radio-quiet zones, ranging from parts of West Virginia to Western Australia and South Africa. “However, there is no legal or other protection from radio sources flying overhead,” Green says, “so right now it requires voluntary agreements with the operators that they won’t downlink a powerful radio signal right onto a radio telescope, because it would be so many billions of times brighter than the sources they’re looking at. And the dynamic range of the receivers just can’t tolerate that kind of illumination.” The companies so far have been very cooperative, he says, but there is no requirement at the moment.
But satellites also emit a lot of radio-frequency radiation besides what is intentionally beamed out. For one thing, there is always some leakage of frequencies outside the desired band, and there is also always some directional spread far outside the targeted zone — for example, in side lobes. “A radio telescope can see a satellite anywhere above the horizon. So there’s this growing cacophony of radio noise at frequencies that used to be accessible to radio astronomers for observations,” Green says.
In addition, every electrical component on a spacecraft has the potential to produce radio-frequency emissions. Electric motors used to aim solar panels, and even wiring harnesses connecting components, can produce low-frequency radiation. “Starlink satellites have actually been detected” at those low frequencies by a radio array in a Western Australia radio-quiet zone, he says — and those low frequencies are of interest to astronomers because they represent objects at the highest redshifts, or the greatest distances from Earth (and the earliest times in our universe).
Off to a good start, but concerns linger
“We’ve gotten off on a good foot,” Green says of the cooperation from satellite companies so far, “but on the other hand, this is a world problem.” China is planning a large constellation that of course won’t be subject to any FCC rules, and a registry in Rwanda has proposed a 300,000-satellite constellation. “We’re in the days now where there are a few thousand. If there are tens of thousands, we’ll be having a different problem.”
Seitzer says “it is the cumulative effect, the aggregate effect of all these constellations that could pose a significant issue. And we also have very large structures being flown in low Earth orbit, which are very, very bright.”
At present, Green says, with many companies and nations making their own decisions about satellite rules and numbers, “there’s little to no structure to take aggregate effects into account. How many satellites can you pack into a given orbit realistically?”