Photometric Characterization and Trajectory Accuracy of Starlink Satellites: Implications for Ground-Based Astronomical Surveys

TL;DR

This study quantitatively characterizes Starlink satellites' brightness and trajectory accuracy, assessing their impact on ground-based astronomy and evaluating light pollution mitigation strategies.

Contribution

It provides the first extensive quantitative analysis of Starlink satellite brightness and TLE accuracy, including effects of mitigation technologies like DarkSat and VisorSat.

Findings

Starlink satellites have an average magnitude of 5.5 in GAIA G band.

DarkSat is approximately 7.6 times fainter than V1.0 Starlinks.

TLEs are accurate within 0.12 degrees in RA and -0.08 degrees in Dec.

Abstract

Starlink is a low-Earth orbit (LEO) satellite constellation operated by Space Exploration Technologies Corp. (SpaceX) which aims to provide global satellite internet access. Thus far, most photometric observations of Starlink satellites have primarily been from citizen scientists' visual observations without using quantitative detectors. This paper aims to characterize Starlink satellites and investigate the impact of mega constellations on ground-based astronomy, considering both the observed magnitude and two-line element (TLE) residuals. We collected 353 observations of 61 different Starlink satellites over a 16-month period and we found an average GAIA G magnitude of 5.5 +/- 0.13 with a standard deviation of 1.12. The average magnitude of V1.0 (pre-VisorSat) Starlinks was 5.1 +/- 0.13 with a standard deviation of 1.13. SpaceX briefly used a low-albedo coating on a Starlink satellite called DarkSat to test light pollution mitigation technologies. The brightness of DarkSat was found to be 7.3 +/- 0.13 with a standard deviation of 0.78, or 7.6 times fainter than V1.0 Starlinks. This concept was later abandoned due to thermal control issues and sun visors were used in future models called VisorSats. The brightness of VisorSats was found to be 6.0 +/- 0.13 with a standard deviation of 0.79, or 2.3 times fainter than V1.0 Starlinks. Over the span of the observations, we found that TLEs were accurate to within an average of 0.12 degrees in right ascension and -0.08 degrees in declination. The error is predominantly along-track, corresponding to a 0.3 second time error between the observed and TLE trajectories. Our observations show that a time difference of 0.3 +/- 0.28 seconds is viable for a proposed 10 second shutter closure time to avoid Starlinks in images.