Contamination of spectroscopic observations by satellite constellations

Abstract

The number of satellites on low orbit has dramatically increased over the past years, raising concerns among the astronomical community about their impact on observations. Spectroscopic observations represent a large fraction of professional data, and spectrographs lack spatial information that can reveal the presence of a satellite. We simulated how often satellites contaminate spectrograph observations by using realistic constellations with over 400,000 objects. We also measured how a spectrum is affected by using real data from different scientific targets and a scaled solar analogue as the satellite, and using standard tools to measure astrophysical parameters and compare them with the clean spectrum. The fraction of affected spectra varies dramatically with the elevation of the sun, with a maximum of 10% at twilight and a nightly average of about 2%. Because of the fast motion of the satellites and the limiting magnitude of the spectrographs, high-resolution instruments are essentially blind to most satellites. For lower resolution spectrographs, the effect on the measured astrophysical parameters depends strongly on the signal-to-noise of the exposure, longer exposures on brighter targets being the least affected at <=1%. Satellites that are brighter and/or higher than the constellation satellites, while less numerous, can also contaminate spectra. While the fraction of affected spectra is likely to remain low, some of these contaminated spectra will be difficult to identify, as it is already the case with existing satellites and asteroids. The best mitigation is to ensure that their brightness is fainter than V=7, that their absolute magnitude V1000km is also fainter than 7, and, whenever possible, to shoot multiple exposures.