I’m trying to learn more about how satellite trails affect measurement-focused observing, especially photometry.
A lot of the discussion I’ve seen online comes from general imaging or astrophotography, where people often talk about stacking, rejection, or removing trails in post-processing. But I’m more interested in the practical side for people who care about the measurements themselves.
For those of you doing photometry, I’d really like to know:
When do satellite trails become a real problem in your workflow?
Is the main issue loss of usable data, extra rejection/QA work, reduced precision, or something else?
Does it depend a lot on target type, field size, cadence, or observing setup?
Are there tools or habits you already use that make the problem mostly manageable?
I’m asking because papers often say satellite contamination is a serious issue for astronomy, but I’m trying to understand what that means in actual observing practice for photometry users.
Thanks very much — any real workflow examples would be very helpful.
In my experience, unless the satellite trail is unusually bright, they affect the photometry only when the trail comes close to the variable or the comparison/check stars. That is, when the trail would be included in the measurement aperture or the background apertures of the stars.
I have been using the free ASTAP software which has a very helpful streak detection function to automatically inspect each of my images. It works very well as it detects even the faintest satellite trails. When it detects a trail, I go back and visually inspect those images and determine if the trail might distort the photometry.
Of course if you see something that looks amiss in the photometry itself, you should always go back and double check the images themselves and your measurement process.
You must take a good look at each image before you start doing the photometry. If the satellite trail does not cross over your target or comp star(s) they are not a problem. Of course if the setellite is the ISS or a large space telescope there might be an affect frome a near miss, but I’ve never had that happen to me. So,no, IMO satellite trails are not a significant problem for our photometry.
A few years ago a member of my local astronomy club showed the light curve of a time series of a short period eclipsing binary. One point of the curve was an obvious, solitary outlier, brighter than points on either side. Inspection of the image showed a satellite trail passing through the image of the variable.
Starlink satellites move about 4 degrees per minute. If we assume 40K satellites (coming soon!), that is 160K degrees covered by the satellites. If we assume each streak contaminates a path 15arcsec wide (typical photometer apertures), then the satellites contaminate 40K square arcmin, or about 10 square degrees in the sky. Since there are about 40K square degrees in the sky, this means that, in a typical 1-minute exposure, the chances of the satellite passing close to a single target is about 1/4000. A typical night is ~10 hours long, or about 600 minutes. This means on a typical night, less than one target will be contaminated by a satellite, once the approved constellations reach their expected size. It gets worse if you are doing ensemble photometry, as then you have to reject contaminated comp stars too. It gets worse if you have multiple targets in the same field of view.
Satellite trails will become the major source of data rejection within a few years. Telescopes like the Vera Rubin Observatory are even more affected, as the satellites saturate and cause instrumental problems, and their field of view is many square degrees.
There are many papers about the impact of Starlink and other constellations. I’m doing a VERY back-of-the-envelope calculation here, and undoubtedly have made some mistakes. In my own photometry, I see multiple satellites per night even now. When there are 40K satellites up there, you will see a satellite on every image (just not over your target).
At USNO, I found that transcontinental jet traffic was a bigger problem (but that was before Starlink!). Not only do you get contrails that act like clouds, but the jet exhaust disturbed the seeing, and the navigation lights acted like a very bright, unfocussed stars. A jet trail is much wider than a satellite trail.
My guess is that we’ll have to invest more time in observing techniques to mitigate the problem. I fear astroimaging will suffer first.
40K satellites is a lower-bound for the total number of LEO satellites going up. Amazon, more Musksats, the Chinese, U.K./France, BlueOrigin and I understand the Russians too are getting into the mega-satellite-constellation game. It should be interesting to see how things play out.