Measurements of Starlink’s “VisorSat” show SpaceX has succeeded in making a less reflective satellite. But it’s still visible from dark-sky areas.
The first launch of Starlink satellites two years ago alarmed many amateur and professional astronomers. Lone satellites coursing through the night sky are commonplace, but in May 2019 observers witnessed an unprecedented parade of startlingly bright objects marching across the heavens.
That batch was the first of the currently more than 1,000 satellites that aerospace company SpaceX lofted into low-Earth orbit (LEO). In order to provide space-based broadband, SpaceX plans an initial “constellation” of 1,548 Starlinks, but ultimately it aims to fill out a network with as many as 42,000 spacecraft. Researchers and stargazers alike are afraid that Starlink — and the many other planned satellite constellations like it — might ruin dark skies everywhere.
To the company’s credit, SpaceX is attempting to address astronomers’ concerns. An initial attempt at dimming a Starlink (dubbed “DarkSat”) by painting parts of it black resulted in thermal issues.
A second attempt involved a sunshade, a visor-like appendage that reduces the sunlight reflected to observers on the ground. The first so-called VisorSat launched on June 4, 2020, on the seventh operational Starlink launch. Since the ninth such launch, on August 7, 2020, all Starlink satellites have been VisorSats.
Along with the hardware change for VisorSat, SpaceX also altered the relative orientation of the orbiting satellite bodies and solar arrays to further diminish their brightness. This change in software was instituted on all operational Starlinks.
To read more about the recent meeting of industry leaders and astronomers, see “Beyond Starlink: The Saga Continues.”
I have been studying observations of original Starlink and VisorSat satellites in order to compare their brightness. The observations discussed in this article are of the satellites at their operational altitude at 550 km, not of those still raising their orbits.
Some of the data are generated by visual observers who report their observations to the SeeSat-L email archive. They determine magnitudes in much the same way that amateur astronomers estimate the brightness of variable stars. That is, the satellites are compared to nearby reference stars of known magnitudes. The distribution of magnitudes recorded by visual observers is shown in Figure 1.
The other source of data is an automated observatory in Russia called Mini-MegaTORTORA (MMT). The imaging system is a 9-channel wide-angle sensor consisting of 71-mm-diameter f/1.2 lenses and 2160 x 2560 sCMOS detectors. Satellite magnitudes are posted on their online database. The magnitudes the MMT has measured are a close match to the visual magnitudes reported by human observers.
Starlink satellites are nearest when passing directly overhead at a distance of 550 km, and thus generally the brightest, while those observed toward the horizon are farther away and correspondingly fainter. To account for differences in distance, I adjusted all observed magnitudes to the apparent brightness that would be measured at 550 km. The average of these adjusted magnitudes provides a characteristic brightness for satellites passing overhead.
So far I have collected more than 1,000 magnitudes of original-design and VisorSat Starlink satellites. The average magnitude of the originals at 550 km is 4.63, so they would be visible under even moderately light-polluted skies. The average VisorSat magnitude is 5.92, so they’re only 31% as bright as the originals and significantly more difficult to see.
Another aspect of Starlink that affects their visibility is the variation around the average brightness. Even after accounting for differences in distance, there is still a dispersion in the observed magnitudes on the order of one magnitude, likely due to the complicated reflecting properties of these satellites' many surfaces. In any case, while the average magnitude for a VisorSat seen at zenith is around 6, the typical spread of individual magnitudes ranges from about 5 to 7. Therefore, an observer under dark skies will see some VisorSats, while others will pass by unnoticed.
In addition to the planned observations summarized above, satellite observers have reported unexpected flares of Starlink satellites that briefly increase their brightness — occasionally by 10 magnitudes or more. Thus, they sometimes exceed the brightness of the most brilliant planet, Venus. Most of these flares were reported early in 2020 on the SeeSat-L email archive. It is encouraging that no extremely bright flares have been reported since that time. The sunshade and adjusted orientation of the satellites are likely reasons for this improvement.
Another factor that limits the adverse impact of Starlink satellites is that, because of their low orbit, they are not all reflecting sunlight during the darkest part of the night. Furthermore, Earth shadowing makes satellites less visible in the eastern sky early at night and less visible in the west before dawn. So, theoretically anyway, observations can be scheduled by time and by sky region in order to avoid satellites.
Starlink satellites will continue to be a distraction to observers for now, but the significantly dimmer VisorSats represent a marked improvement. It remains to be seen if and how other satellite companies will take note and follow suit.