Space astronomy has been peddled as a way out of light pollution on the earth, but there is yet another form of glow that is accompanying the telescopes. With the spacecraft around low Earth orbit becoming increasingly reflective, it is increasingly difficult to keep the assumption of the dark sky upon which many survey designs rely: not because a single satellite is bright but because it is so easy to have a crossing when you have a long exposure.
The problem was addressed by a Nature analysis of satellite-trail contamination, which considered it as geometry and growth. The population of active satellites grows to between 560,000 and one million in one of the modelled futures; in a more aggressive case, to the same number. Every extra shell and aircraft boosts the probability of a spacecraft floating through a detector during a period when an observatory is accumulating dim light of galaxies, halos of matter, or other tiny bodies of the Solar System.
Simulations in the study were of four space telescopes in or near the same orbital neighborhood as the constellations: Hubble, NASA SPHEREx, the proposed ARRAKIHS of ESA and Xuntian of China. In the scenario of over 560,000 satellites, over 96 percent of SPHEREx, ARRAKIHS and Xuntian exposures contained at least one trail, and Hubble with its severely limited field of view still had to deal with the fact that it was seeing a significant portion of images. This identical modeling framework also predicted the number of trails that might be observed per exposure, whether in a single-digit crossing in a wide-field survey or dozens or more when long exposures and low altitudes are used together.
These lines are not just eye shadows. Pixels can be saturated with a trail, which covers subtle features in the image under additional brightness, and produces residual artifacts despite automated masking. In surveys with statistical power based on a stack of thousands of images, tiny per-exposure losses cause patchiness in depth over the sky, and the systematic depth biases and systematic measurement biases of surveys that rely on background subtraction.
The information that defines the degree of impact of a telescope is unexpectedly mission-dependent. There are field of view considerations, and there are also pointing considerations like avoiding Earth-limbs and having limited zenith angle; there are often extremely restrictive limits that can be used to avoid encounters by avoiding parts of the sky where satellites are most visible. The altitude of the orbit has two contradictory effects: satellites with lower altitudes can be shaded longer by the Earth and travel faster, but a low-orbit observatory can be located in the busiest layer. Exposure time is also important; a 600-second imaging sequence just provides more time to cross than to come to the point, and spectroscopy may be more difficult still since the signals are not so easy to see and to eliminate.
Mitigation is not a switch, therefore. On the engineering front there are unanimous suggestions to include dimmer spacecraft, reduced number of high amplitude flares on the time of attitude changes and improved characterization of surfaces prior to launch. Within the operations arena, astronomers are still gradually requiring accurate and updated ephemerides in order to facilitate avoidance scheduling and purer masking. The gap lies in the fact that popular orbit formats are not tailored to the fineness of some of the corrections at space-telescope scales, a mismatch which the discussion of accuracy requirements of the Nature study point out as a problem of determining the precise path of a trail showing up in a detector.
Certain observatories already consider satellite contamination as a normal calibration issue. Hubble In Hubble data, better trail-finding algorithms have been developed based on mathematical image transforms; as Dave Stark of STScI wrote, We created a new tool to find satellite trails, which is more sensitive than our previous software. However, the tools that have been successful in the situation where a small percentage of exposures is compromised are different when trails are the rule instead of the exception.
The long-term engineering implication is clear: any future low-orbit observatory must have orbital crowding as a first-order boundary condition. When the number of satellites rises into the hundreds of thousands, reflected light becomes a structural aspect of the near-Earth environment – an aspect that cannot be processed away without sacrificing time, sky coverage or scientific accuracy.
