Starlink led bright satellite passes, but Guowang had a higher rate

In a preprint submitted to arXiv on October 7, 2026, researchers reported that 14,997 of 15,591 satellite passages recorded at twilight exceeded an altitude-adjusted brightness recommendation during observations at Ontario’s Elginfield Observatory from October 7 to November 11, 2024. The paper reports the share as 96%. Its authors conducted the survey; the International Astronomical Union’s Centre for the Protection of the Dark and Quiet Sky from Satellite Constellation Interference (IAU CPS) supplied the scientific benchmark. The study compares Starlink’s share of observed bright passages with the rate for each constellation’s possible passes, which produces different rankings.

What the survey measured

Six cameras recorded 24,812 satellite tracks from 5,118 unique satellites over 36 observing nights. The team had 230 clear observing hours out of 407 possible hours at the single Ontario site.

The headline 96% refers to recorded twilight passages above the study’s brightness threshold. It is not a percentage of all satellites in orbit. Most tracks in the survey appeared within two hours after sunset or the final two hours before sunrise, when the ground was dark but satellites could still be illuminated by the Sun.

Why Starlink and Guowang/Hulianwang rank differently

Starlink accounted for roughly two-thirds of the observed passages above the threshold, the largest absolute share in the survey. But its rate was about 25% when researchers compared above-threshold passages with theoretically possible observations. Guowang/Hulianwang’s rate on that same basis was more than 50%; OneWeb’s was about 6%.

The possible-pass rate uses a different denominator from Starlink’s two-thirds share. Researchers estimated how many observations were possible when cataloged satellites would enter the cameras’ fields of view, accounting for factors such as clouds, camera downtime, atmospheric attenuation and Earth’s shadow. That rate compares the likelihood of an above-threshold passage per observing opportunity; the two-thirds figure counts Starlink’s share of all above-threshold passages recorded. A large absolute contribution and a higher per-pass rate answer different questions.

Every Guowang/Hulianwang passage actually recorded in the survey exceeded the threshold. That observed-pass result is distinct from the more-than-50% rate calculated against possible passes.

What the IAU CPS recommendation means

The benchmark concerns apparent visual brightness, expressed in V magnitude. A larger magnitude means a fainter object. The recommendation adjusts for orbital altitude: at altitudes above 550 kilometers, it calls for a fainter apparent magnitude because a satellite’s slower apparent motion can make its trail brighter per unit area in an image. The IAU CPS recommendation is a scientific benchmark, not a law.

How far the results reach

The measurements describe one site and one campaign in late 2024, rather than a worldwide census. Although the cameras had a nominal sensitivity of about +9.0 V magnitude, the survey’s effective completeness limit was +6.15. The authors say some reported rates for populations mostly fainter than that limit are lower bounds. Fixed camera orientations also affect which satellite paths the survey could observe, and most satellites still raising their orbits after launch were not represented because public orbital data was generally unavailable.

What the Rubin Observatory estimate says

The researchers modeled about 0.34 satellite trails per 30-second twilight exposure at the Vera C. Rubin Observatory, using measured satellite flux and assumptions about the telescope and exposure. That is a projection for those observing conditions, not a measured rate from Rubin images.