Satellite Megaconstellations Are Flooding Night Skies With Light, and Ground-Based Astronomy Is Paying the Price

SpaceX's Starlink network had roughly 6,000 active satellites in low Earth orbit by early 2024. OneWeb added several hundred more. Amazon's Project Kuiper has regulatory approval for 3,236. When all currently licensed megaconstellations are fully deployed, the total number of active satellites in low Earth orbit could exceed 65,000. To put that against history: the entire 60-year record of human spaceflight, from Sputnik in 1957 to about 2019, produced roughly 9,000 satellite launches total. We are about to multiply that legacy count sevenfold in a decade.

The problem for astronomy is geometric. Satellites in low Earth orbit, typically between 340 and 1,200 kilometres up, catch sunlight long after the ground has gone dark. During twilight and the hours around it, they are bright enough to saturate the sensors on professional telescopes. A single Starlink train shortly after launch, before the satellites disperse into their operational orbits, can put 60 or more streaks across a single long-exposure image.

The Vera C. Rubin Observatory in Chile, which will conduct the Legacy Survey of Space and Time starting in 2025, expects to image the entire visible sky every few nights. Its wide-field camera is precisely the instrument most vulnerable to satellite trails: wide field means more sky per frame, which means more satellites per frame. A 2020 analysis by astronomers at the National Optical-Infrared Astronomy Research Laboratory estimated that at full Starlink deployment, up to 30 percent of Rubin's twilight exposures could be compromised.

Compromised does not mean ruined in every case. Algorithms can flag and mask satellite trails. But masking removes real data, a trail running across a galaxy or a near-Earth asteroid's predicted path deletes the science underneath it. An asteroid that was in the masked zone is an asteroid that was not tracked. Planetary defence is not an abstract concern: the European Space Agency's Hera mission and NASA's DART impact in 2022 both depend on the kind of ground-based follow-up observation that becomes harder when the sky is busier.

After the first Starlink launches in 2019 drew immediate protests from the astronomy community, SpaceX introduced a visor design called VisorSat on later batches, intended to block sunlight from reflecting off the satellite's antenna panel. Independent measurements by the satellite brightness tracking project, run by astronomers including those at the American Astronomical Society's working group on satellite constellations (SATCON), found that VisorSat reduced Starlink brightness by about a factor of five, meaningful, but not enough to drop the satellites below the threshold where they stop affecting sensitive instruments. The later Generation 2 Starlink satellites are larger, which creates a new brightness problem even if the visor approach is retained.

OneWeb, operating at a higher altitude of around 1,200 kilometres, is actually dimmer during operations but stays sunlit for longer each night because of its altitude. The trade-off is not obviously better for observatories.

India has a direct professional interest in this. The Indian Astronomical Observatory at Hanle in Ladakh, operated by the Indian Institute of Astrophysics, sits at 4,500 metres, one of the highest-altitude observatory sites in the world. Its location was chosen specifically for the dark, dry skies above the Himalayan plateau. The 2-metre Himalayan Chandra Telescope there has been used for everything from exoplanet transit observations to monitoring active galactic nuclei. Satellite streaks affect Hanle exactly as they affect observatories in Chile or Hawaii: the physics of sunlit metal in low orbit does not care about borders.

ISRO's own scientific ambitions also intersect here. AstroSat, India's first dedicated astronomy satellite launched in 2015, observes from space and sidesteps the ground-based pollution problem. But AstroSat operates in specific wavelength bands and cannot replace wide-field optical surveys done from the ground. The argument that space telescopes solve the megaconstellation problem misreads what each instrument class actually does. Ground-based telescopes and space telescopes are not substitutes, they are complements, each covering territory the other cannot.

The International Astronomical Union issued a statement in 2021 calling for satellite operators to keep their spacecraft below naked-eye visibility and to work with the astronomy community on mitigation. The statement has no enforcement mechanism. The ITU, which governs radio spectrum and orbital slots, has no mandate to regulate optical brightness. There is currently no international treaty that gives astronomy a legal claim on dark skies.

Mitigation software is improving. The Rubin Observatory has developed a pipeline specifically to detect and mask satellite trails in real time. But software mitigation has a ceiling: if enough satellites are bright enough often enough, the masked fraction of data climbs past the point where statistical science is still valid. The sky is a shared resource that was never formally designated as one.

The constellation operators have a genuine product, broadband internet for underserved regions, including remote parts of India. That is not a trivial benefit. The tension is real, and it will not resolve by declaring one side obviously right. What is clear is that the decisions being made now, at the pace of rocket launches rather than treaty negotiations, are setting the conditions under which ground-based astronomy will operate for the next several decades. The observatories being built today will still be running when those 65,000 satellites are overhead. The astronomers designing those instruments already know they are designing around a sky that has changed.