“Will the next great cosmic discovery be erased by a broadband satellite?” A new NASA led study warns that the accelerating deployment of constellations of satellites in low Earth orbit or LEO will soon overwhelm the imaging capabilities of major space telescopes. The projections are stark: by the time the currently filed fleets are fully operational 560,000 satellites by the mid‑2030s up to 96% of images from observatories like NASA’s SPHEREx, ESA’s ARRAKIHS, and China’s Xuntian will be marred by bright streaks called satellite trails. Even the Hubble Space Telescope with a narrow field of view could see one in three exposures compromised.
The mechanism of contamination has its roots in orbital geometry and optics. LEO satellites, typically operating at altitudes in the range of 340 800 km, reflect sunlight, Earthshine, and even thermal infrared radiation toward telescope detectors. A satellite crossing the field of view of a telescope during an exposure will have its reflected light recorded as a linear artifact, usually much brighter than the faint astrophysical targets under study. Simulations using realistic orbital parameters and survey plans predict that SPHEREx will average 5.6 trails per image, ARRAKIHS about 69, and Xuntian roughly 92. These can reach surface brightness levels of µ18-23 mag arcsec⁻² that is, orders of magnitude above the detection threshold and render the affected pixels scientifically useless.
Efforts to reduce optical brightness have had very limited success. The “DarkSat” coating and “VisorSat” sunshade applied by SpaceX reduced apparent magnitudes from about 4.6 to 5.9 – 6.0, which dims them for the naked eye but leaves them glaringly bright to sensitive instruments. Newer direct‑to‑cell satellites, featuring 125 m² solar arrays, are expected to reach visual magnitudes near 0-1, rivaling the brightest stars. High amplitude flares can be produced by changes in orientation. Once a satellite loses attitude control at end‑of‑life, unpredictable tumbling can increase reflectivity in ways that make effective correction algorithms impossible.
These temporal patterns of contamination are particularly harmful to some science programs. Since most of the trails are visible during pre‑sunrise and post‑sunset twilight, the imaging must be stopped by the observatories at that time. This is a key loss for asteroid surveys, which use twilight for the detection of potentially hazardous near‑Earth objects. In some instances, ground based telescopes can stop the exposure for long enough to avoid the known passes but space telescopes have tighter operational constraints and centimeter‑level orbital accuracy needed to predict and dodge trails is beyond the current public tracking capabilities.
From an engineering point of view, a number of mitigation strategies have been suggested: limits to the satellite reflectivity, a reduction in flares induced by orientation, support for a global monitoring network for light contamination, and publication of BRDF data on spacecraft surfaces. Other measures still include capping constellation altitudes below those of vulnerable telescopes to exploit Earth’s shadow, keeping an open archive of active and defunct spacecraft orbits, and improving the precision of orbital elements for avoidance planning.
But there are trade-offs with reducing orbits: increased atmospheric drag cuts satellite lifetimes and increases launch cadence and re-entry rates, injecting aluminum oxide nanoparticles into the stratosphere materials that could deplete ozone and change global temperatures by up to 1.5 °C. Image-processing techniques can mask or interpolate over trails, but as lead author Alejandro S. Borlaff cautions, “the information under those pixels is forever lost.” For transient phenomena such as supernovae or fast-moving asteroids there is no second chance.
In wide-field surveys, multiple trails per frame can destroy large fractions of the data. At projected constellation densities, simulations show ARRAKIHS could lose over 22% of its field of view to trails in a single exposure quadrupling the pixel loss rate from cosmic rays in Hubble imagery. The problem does not stop at visible wavelengths. Satellites emit thermal infrared radiation from their electronics and radiators, contaminating mid‑IR observations even when in Earth’s shadow. They also leak low‑frequency radio emissions, threatening radio astronomy.
Multi‑band interferences like these imply that even telescopes designed for non‑optical science are at risk. Rapidly increasing LEO traffic also presents serious collision risk, too. Denser orbital shells raise the chances of debris‑generating impacts that further threaten telescopes either through micrometeoroid‑like strikes or by forcing costly avoidance maneuvers. Collision risk modeling underlines the need for coordinated traffic management, but the international regulatory frameworks are falling ever further behind the pace of deployment.
What’s taking shape is a new regime in which the industrialization of near‑Earth space is directly competing with humanity’s ability to study the universe. As Borlaff puts it, “Space is for everyone, and it will affect everybody.” Unless rigorous engineering controls, orbital zoning, and industry‑to‑science data sharing can be implemented, the next generation of space telescopes may be forced to spend most of its observing time not looking at the cosmos, but at the glare of our own machines.
