In December 2020, astronomers documented a burst of highly energetic light in one of the most distant galaxies ever observed. But less than a year later, the paper’s claims lay in limbo. Other scientists said it had merely been a passing satellite.
“I was a bit sad that the gamma ray burst turned out to be just an artificial satellite,” said Krzysztof Kamiński, an astronomer at the Astronomical Observatory Institute in Poland who said he matched the position, time and brightness of the discovery to an orbiting spacecraft.
Linhua Jiang, an astronomer at Peking University in Beijing who led the original finding, said his team stood by their work, adding that the probability of a satellite passing directly in front of the distant galaxy at exactly the right moment was minuscule at best.
The dispute likely will not be the last time that scientists argue over whether a passing satellite is being mistaken for an astronomical discovery.
Earth’s orbits are filling with satellites at an astounding pace. Already there are more than 9,000 satellites orbiting the planet, and more than 5,000 of them belong to Starlink, the constellation built by SpaceX to beam internet service down to Earth. They are to be joined by thousands of satellites from other companies and countries in the decades ahead.
The more of them there are, the greater the satellites’ interference with ground astronomy’s ability to answer questions about the cosmos — and humanity’s place in it.
SpaceX did not reply to requests for comment. But astronomers on the ground said they are not ready to give up the night skies to trains of freshly deployed satellites. They are combining new and old technologies with ingenuity to deal with the proliferating obstacles to their observations. They are also working with the industry to find fixes to darken satellites. And they are trying to persuade regulators to pay more attention to the mushrooming satellite industry.
The strategies are paying off — for now. But researchers’ quest to preserve the power of astronomy faces fundamental disadvantages. It can take decades to build new telescopes, while dozens of new satellites may be added to the night skies every week.
“The time scales are very mismatched,” said Meredith Rawls, a research scientist at the Vera C. Rubin Observatory, a powerful U.S.-funded telescope in Chile that is to come online in 2025. “The speed at which the satellite industry is designing and launching their hardware is just lightning fast compared to astronomy.”
To photograph the night sky, telescope operators for more than a century captured images on glass plates.
That began to change with the emergence of charge-coupled device detectors. First invented in 1969, CCDs are digital, snapping images around 100 times faster than film cameras.
In the 1980s, some of the first telescopes emerged with electronic CCD “eyes.” Today, telescopes around the world continue to rely on this Nobel Prize-winning technology. While CCDs are not the fastest camera technology now available, they are the most common. It also takes decades to build the most powerful ground observatories, and many were designed with 20th century levels of imaging techniques in mind.
That includes the Vera Rubin Observatory, named after an astronomer who played a central role in discovering dark matter. Its mission includes spotting planet-killing asteroids and studying the relationship between dark matter and dark energy.
The telescope relies on a behemoth CCD detector that is around the same size as the average car, but several thousands of pounds heavier. It is the largest astronomical digital camera ever constructed. Capturing a wide field of the sky, it is supposed to peer into the mysteries of objects 20 million times fainter than the human eye can see.
But as satellites fill the skies, astronomers who planned to rely on the Rubin telescope for scientific discovery are concerned.
“The whole point of Rubin is to open up this new window into the universe to find things that we didn’t even know to look for,” Dr. Rawls said. “And if instead we’re going to look through the equivalent of a windshield of bugs, you don’t know what you’re not going to see.”
Some telescopes that use CCD detectors study such a narrow slice of the sky that satellites may not interfere with them. But the Rubin telescope’s wide view poses unique problems. One study showed that, during certain times of night, almost every image taken from the telescope will be marred by at least one, if not many, satellites, searing a trail hundreds of pixels wide.
Dr. Rawls laid out two strategies for dealing with this threat to the telescope: dodging and correcting.
If astronomers know satellite paths in advance, the technology can anticipate and “dodge” the satellites by temporarily repointing the telescope.
“We use an algorithm to determine where the telescope points,” Dr. Rawls said. “The algorithm is brilliant, it can take into account lots of different weightings,” she added, including avoiding swarms of satellites.
Dr. Rawls said that dodging should remove about half of the streaks from Vera Rubin’s telescope, depending on how many satellites are in orbit.
For the correcting strategy, Dr. Rawls said that scientists are developing algorithms to scrub the satellites from data — a far more challenging task — but one that is less disruptive to observations.
But given that the software solutions are all imperfect and challenging, some experts have suggested that telescope builders think about changing their hardware.
Darren DePoy, an astronomer at Texas A&M University, was involved with some of the first telescopes in the 1980s to use CCDs. In 2018, he began testing and eventually using a much more ubiquitous detector: CMOS, for complementary metal oxide semiconductor, the same kind that is probably in your smartphone camera.
“Although the physics is very similar for CCD and CMOS detectors, how you get the signal out is a little different,” Dr. DePoy said. “For CMOS, you can read all the pixels simultaneously, while you have to wait to read each pixel sequentially on a CCD detector.”
As an example, Dr. DePoy said that while a modern CCD might require about 10 seconds to photograph a faint galaxy, the equivalent CMOS detector would take closer to 10 milliseconds — 1,000 times faster. By taking numerous rapid exposures, astronomers can excise the frames smeared by satellites or airplanes, then average the rest to create a pristine final image.
Dr. DePoy said that small CMOS detectors are already popular among amateur astronomers who own hobby telescopes. He finds it hard to imagine that CMOS isn’t the future. But, for now, he estimated that fewer than 10 larger telescopes use the technology.
Part of the sluggish embrace is because inertia is cheaper.
Buying and integrating large CMOS detectors is still expensive compared with using existing CCD detectors, said Richard Green, an astronomer at the University of Arizona and an interim director at the Center for the Protection of the Dark and Quiet Sky from Satellite Constellation Interference, an organization that sponsors research around the topic.
That problem was noted by Dr. Rawls when she was asked if the Rubin telescope could use CMOS technology.
“The concept of changing it now is just laughable,” she said. “Because that’s like you’re building a house and they’re about to put the windows in and someone’s like, ‘Hey, should we use a different foundation?’”
The United States government is both championing the commercialization of space and sponsoring telescopes like the Rubin Observatory. For that reason, Dr. Green said it was up to the government to deal with the effects on astronomy, perhaps by charging companies to pay for telescope upgrades.
“If the government says we’ll do that by assigning a fee to satellite operators, well that’s great,” he said. “Somebody in the government ought to help us deal with the fallout.”
The government so far has not moved to compel satellite operators to help pay for telescope upgrades. But some companies are attempting to address aspects of the problem.
SpaceX declined to comment when asked about the company’s work to lessen the effects of its satellites on science. But astronomers familiar with its efforts described some of the work.
When the SpaceX founder Elon Musk faced criticism in 2019 after the first Starlink satellites were launched, he said on Twitter that he had “sent a note” to engineers, asking them to reduce sunlight reflections from the company’s orbiters.
“SpaceX counts legions of astronomy nerds among its ranks, so the importance of protecting that scientific domain is not lost on them,” said Caleb Henry, the director of research at Quilty Space, which provides space industry analytics.
The first attempt involved a light-absorbing coating that darkened satellites. A prototype named DarkSat went up in 2020, according to Jonathan McDowell, an astronomer at the Harvard-Smithsonian Center for Astrophysics.
“The problem with that was that the equipment inside overheated,” he said. The satellite failed.
Dr. McDowell said SpaceX’s next step was to install shades over its satellites, an idea that was quickly scrapped because not only did the shades do little to darken the satellites, they blocked the laser cross-links SpaceX was developing to enable its satellites to communicate with each other.
The company’s most recent attempt involved a dielectric film coating. Contrary to expectations, this made the satellites more shiny. But instead of reflecting the sunlight down to Earth’s surface, the material bounced it back into space, muting the intensity of any streaks. SpaceX said it would share the coatings with other satellite manufacturers.
During the crucial twilight hours when many astronomical observations occur, SpaceX also began rolling its satellites to point their solar panels away from Earth. To compensate for the loss in solar power, it upsized the satellites’ solar panels, an extra expense.
“On the SpaceX side, they’ve taken real hits in order to try and accommodate us,” Dr. McDowell said.
Initial data indicates that the interventions may be working. In a study that has yet to go through peer review, astronomers reported that newer Starlink satellites appeared darker because of the reduced sunlight reflection to the surface.
This work by SpaceX occurred as it was coordinating with the National Science Foundation on a voluntary basis, said Ashley VanderLey, a senior adviser there.
Though the U.S. government has long required satellite operators to coordinate with operators of radio telescopes on sharing bandwidth, no federal rules have protected optical astronomers. But the rules that helped radio astronomers did provide a basis for optical astronomers to have discussions with companies like SpaceX and Amazon.
“That’s where our foot was in the door to start coordinating,” Dr. VanderLey said.
What had been voluntary conversations became mandatory in December 2022, Dr. VanderLey said, when the Federal Communications Commission formally required a series of measures by SpaceX. While many of the requirements focused on safe operations in orbit, the agency also said that SpaceX must coordinate with the N.S.F. to “mitigate the impact of its satellites on optical ground-based astronomy.”
Similar measures were required for Amazon’s Kuiper. A spokesman for Project Kuiper, Tim Kilbride, said it had consulted with N.S.F., in addition to consultations with the International Astronomical Union.
Then, after a request from SpaceX, the F.C.C. extended the requirements to some other satellite companies in August 2023. The F.C.C. also tightened debris mitigation requirements for SpaceX’s mega-constellation, to which the company responded by asking regulators to pass along the stricter measures to “any constellation of 25 or more satellites.”
Dr. VanderLey described the N.S.F.’s ongoing negotiations with SpaceX as productive and the only way to succeed. But as astronomers engage with satellite operators over these rules, a point may be reached when trying to reduce the impact of satellites no longer works, experts say.
Currently, the satellites amount to a nuisance — what Dr. Rawls called “a windshield of bugs” — rather than a true threat to science. But what happens when the number of satellites reaches the hundreds of thousands or more, as some forecasts predict, with other companies and China, Russia, and the European countries joining the orbital fray?
“It’s great to talk about mitigations,” Dr. McDowell said, “but there comes a point where nothing really helps, so I think you need a restriction on the number of satellites in the long run.”