Scroll Top

The impact of satellite operations on astronomy

Much has been written recently about the potential impact of the mega-constellations on optical astronomy, with Space-X’s Starlink constellation bearing the brunt of the criticism. And unfortunately, the situation could be rather worse than most people assume.

This is not, of course, the first time that astronomers have had to contend with a potential reflective threat in low Earth orbit. Back in the 1980’s the Eiffel Tower Corporation proposed a massive “space ring” to celebrate the impending anniversary of their tower’s construction.

This is how the L.A. Times reported the story in 1986.

The project involves launching a half-ton package from an Ariane rocket. It will inflate in space to form a string of 100 reflectors linked by narrow plastic tubes each 780 feet long. The ring will measure 15 miles in circumference and, orbiting at an altitude of 500 miles, will appear slightly larger than the moon to the naked eye. It will circle the Earth every 90 minutes and will reflect sunlight so that it can be seen at night. During each orbit it will be visible to a stationary observer on Earth for about 10 minutes.

Mercifully, perhaps, the design of “Anneau de Lumiere” proved to be too unstable to maintain its shape in orbit and, since there was so much resistance from astronomers, the project was eventually cancelled.
To their credit, Starlink have sought to address the problems that their satellites might cause. They have tried painting one of their missions black, (which may well have made it harder to see, but probably didn’t do much for the satellite’s thermal control), they have installed “sun-shields” on some missions to try to block light reflected from the solar arrays from reaching the Earth, and they have plans to modify the attitude of their satellites, again with the objective of reducing the amount of light that reaches the ground. This strategy may succeed but could also result in a non-optimal power budget, and hence a reduction in the satellite’s operational duty cycle – which ultimately will cost Starlink money.

This highlights a difference between the situation today and that in the mid 1980’s when the Anneau de Lumiere was proposed. At that time, much of the hardware that was placed in orbit was still “governmental” in origin. Now the majority of the satellites being launched are commercial and have a “business case” to close for their investors. Inevitably there will be limits on what the operators at Starlink, (and elsewhere), can achieve because of these financial pressures.

And sadly, the situation in the optical domain may be even more acute in the future. Many current major observatories have narrow fields of view, and tend to conduct operations well outside twilight conditions. For these reasons it has been suggested that, while they are “looking down the Earth’s shadow”, any satellites that occasionally happen to pass through their field of view will not be illuminated by the Sun, and so will not be seen. That is certainly true, but at least some of those satellites could still be in view of the Moon. Now clearly the (full) Moon is much dimmer than the Sun, (by 14 magnitudes, or a factor of roughly 400,000), but recall that the Starlink satellites have a visual magnitude of 5.5 once they are on station, (and are brighter still while they’re orbit-raising).

Then recall that the 10-degree field of view Rubin telescope, which is designed to detect threatening asteroids, is aiming to complete surveys of the sky down to a magnitude limit of between 24, (using individual 15-second images), and 27, (using stacked images), and it becomes apparent that there might be a serious image contamination problem.

There is some potential mitigation because most telescopes don’t have a field of view as wide as the Rubin telescope, (so satellites are statistically less likely to cause problems), and most astronomical observations are not routinely conducted at full Moon. Also, the brightness of the Moon is very dependent on phase angle. A half-moon is only 8% as bright as a full Moon, providing another 2.7 magnitudes of margin; (if you’re wondering why, consider the size of the un-illuminated regions in lunar craters as the phase angle increases). Moreover, it’s clear that satellites move across the field of view during a time exposure, so the light reflected from them will be distributed across the telescope’s detector. Nevertheless, from the above figures it is plausible to argue that any satellite which is visible to the naked eye when illuminated by the Sun could also be visible to the Rubin telescope when illuminated by the light reflected from a half Moon.

The problems for astronomy may not end there. If the IR community follow the optical community in developing wider field of view telescopes, then the thermal radiation they emit could become an issue. The internal temperature of satellites is typically maintained at around 20 C, but their external surfaces can reach over 100 C when illuminated by the Sun. In either case, the satellites are always at temperatures significantly above the -270 C temperature of the space background.

And the radio astronomers may not escape either. The Murchison Array in Australia has already been used to tune into commercial FM transmissions and act as a bistatic receiver for terrestrial transmissions reflected from the ISS. It has been calculated that the Murchison system could potentially detect objects as small as 0.5 m in size, but as the radio-astronomy community develops the Square Kilometre Array (SKA), sensitivities could increase to the point where the small debris population becomes an issue. The SKA is allegedly going to be capable of detecting lightning strikes on exoplanets, so might it also pick up radio reflections from the “chaff” we have placed in LEO? Estimates of the population of space debris in the 1-10 cm size range vary between 500,000 (NASA) and 900,000 (ESA); if the SKA has sufficient sensitivity to detect these objects, it may be difficult to find a “radio-quiet” location anywhere on the planet. In such a scenario, the SKA may prove to be a superb SSA sensor, but clearly that is not what it was designed to be.

Readers desiring more technical details will find material of potential interest at these locations:

Leave a comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.