Impacts of Large Satellite Constellations on Astronomy

[eeergo]

Astronomers Appeal
https://astronomersappeal.wordpress.com/
[...]

That's a poorly-written, terribly-researched, alarmist piece of propaganda. And I've made my position clear on megaconstellations in this very thread before - but these kinds of worthless pompous prose do nothing except murk the waters and radicalize positions. Count me in the "these people are aggressive luddites" field in this occasion.

In other news, DarkSat hasn't failed to provide the intended mitigation (i.e. what was obvious before is patent now: a smooth flat surface will show similar brightness when specularly reflecting light no matter its color). Back to the drawing board I guess, if they ever even reached that point:

[laszlo]

Astronomers Appeal
https://astronomersappeal.wordpress.com/
[...]

That's a poorly-written, terribly-researched, alarmist piece of propaganda... these kinds of worthless pompous prose ...

The content is one thing, but you might want to give them a break on the literary style. The English version of the document reads like a translation from the original Italian. What you call poorly-written and pompous seems to be simply a foreigner attempting to communicate in a non-native language.

[envy887]

DarkSat hasn't failed to provide the intended mitigation (i.e. what was obvious before is patent now: a smooth flat surface will show similar brightness when specularly reflecting light no matter its color). Back to the drawing board I guess, if they ever even reached that point

I think you meant to say "has failed", but that's not confirmed yet.

According to SpaceX most of the brightness during orbit-raising is off the solar array which is not dark-coated and is in a "low-drag" configuration.

So any noticeable decrease in brightness would not be expected to happen until the satellite is in the operational orbit and configuration. After 3.5 weeks, it is at 420 km, almost exactly half-way to the operational orbit.

And I doubt most of the reflection is specular, as that would result in flaring when a large surface was angled correctly to the observer, but much lower visibility from any other angle. The relatively constant brightness across a large angular range indicates diffuse reflection (or possibly a large number of curved surfaces, but the photos we've seen of Starlinks up close shot them to be mostly faceted, not curved).

[eeergo]

Astronomers Appeal
https://astronomersappeal.wordpress.com/
[...]

That's a poorly-written, terribly-researched, alarmist piece of propaganda... these kinds of worthless pompous prose ...

The content is one thing, but you might want to give them a break on the literary style. The English version of the document reads like a translation from the original Italian. What you call poorly-written and pompous seems to be simply a foreigner attempting to communicate in a non-native language.

I live and work (in astrophysics) in Italy, and use both Italian and English daily for work, while not being a native speaker in either - so I make lots of mistakes and sometimes express myself poorly, yet I'd never dream of stooping to publish something like this.

If you're trying to publish your work, even in the arXiv, it's good practice and a sign of basic respect towards the community that's reading you to make sure your manuscript doesn't look like something you'd utter in a bar after a couple of drinks, certainly with respect to content (and this paper is a pamphlet using unreferenced secondary material), but also regarding style - if necessary asking a colleague for at least a cursory proof read. The arXiv should not be treated as a notice board. These people surely have published in English before, if they work in INAF. It's pompous to use the Declaration of Human Rights in CAPITAL LETTERS to advance half-cooked, poorly-researched arguments.

It's relevant to note the lead author is actually a technologist specialized in data storage and computing, outspoken on the dangers of cellphone antennas and vaccines on Facebook among other channels. The second author has a grand total of 4 collaborative peer-reviewed papers instead, as fifth or later non-alphabetical author.

[eeergo]

DarkSat hasn't failed to provide the intended mitigation (i.e. what was obvious before is patent now: a smooth flat surface will show similar brightness when specularly reflecting light no matter its color). Back to the drawing board I guess, if they ever even reached that point

According to SpaceX most of the brightness during orbit-raising is off the solar array which is not dark-coated and is in a "low-drag" configuration.

So any noticeable decrease in brightness would not be expected to happen until the satellite is in the operational orbit and configuration. After 3.5 weeks, it is at 420 km, almost exactly half-way to the operational orbit.

And I doubt most of the reflection is specular, as that would result in flaring when a large surface was angled correctly to the observer, but much lower visibility from any other angle. The relatively constant brightness across a large angular range indicates diffuse reflection (or possibly a large number of curved surfaces, but the photos we've seen of Starlinks up close shot them to be mostly faceted, not curved).

One of the main problems with Starlinks is their unexpectedly high reflectivity when in their initial orbits (which as we're seeing is not a transient state that can be measured in hours or a few days), compounding with the fact that the constellation will require frequent such positioning of large numbers of units in (de)comissioning. If the attempt at a solution attacks a sub-leading term to the problem, is it a solution? OTOH I agree the solar array is the most important source of light pollution so far, as I've stated before.

Maybe specular is not the precise term in this case - what I mean is a flat surface covered in smooth material, even if it has some slight wrinkles. Check out for instance this building's facade at face-on illumination - it's painted completely black.

[envy887]

DarkSat hasn't failed to provide the intended mitigation (i.e. what was obvious before is patent now: a smooth flat surface will show similar brightness when specularly reflecting light no matter its color). Back to the drawing board I guess, if they ever even reached that point

According to SpaceX most of the brightness during orbit-raising is off the solar array which is not dark-coated and is in a "low-drag" configuration.

So any noticeable decrease in brightness would not be expected to happen until the satellite is in the operational orbit and configuration. After 3.5 weeks, it is at 420 km, almost exactly half-way to the operational orbit.

And I doubt most of the reflection is specular, as that would result in flaring when a large surface was angled correctly to the observer, but much lower visibility from any other angle. The relatively constant brightness across a large angular range indicates diffuse reflection (or possibly a large number of curved surfaces, but the photos we've seen of Starlinks up close shot them to be mostly faceted, not curved).

One of the main problems with Starlinks is their unexpectedly high reflectivity when in their initial orbits (which as we're seeing is not a transient state that can be measured in hours or a few days), compounding with the fact that the constellation will require frequent such positioning of large numbers of units in (de)comissioning. If the attempt at a solution attacks a sub-leading term to the problem, is it a solution? OTOH I agree the solar array is the most important source of light pollution so far, as I've stated before.

Maybe specular is not the precise term in this case - what I mean is a flat surface covered in smooth material, even if it has some slight wrinkles. Check out for instance this building's facade at face-on illumination - it's painted completely black.

That's probably still much darker than a building clad in polished bright aluminum under the same lighting. Starlinks don't need to disappear completely, losing a couple magnitudes would help a lot. For one thing, they would then not be naked-eye visible except under excellent conditions. And the saturation and ghosting issues would be at least partly mitigated.

Since the coating is specifically intended to reduce visibility, I'm sure SpaceX has at least considered the coating's reflectance across a range of incidence angles.

I'm not sure that the deorbit phase would have the same issue. The satellites should certainly not be in the "low drag" configuration for deorbit. And for deorbit, the time is a function of thrust + drag, while on ascent the time is a function of thrust - drag. Depending on the ratio of thrust to drag at low altitudes, this time difference could be quite significant.

Only about 2% of the constellation would be orbit-raising assuming a steady state of operations, 6 week obit raising time, and 6 year operational life. That's an average of 88 at any time for the 4425 satellite LEO constellation. 88 new very bright objects are unlikely to have any significant impact on observations, because they would be clustered spatially and temporally, easily avoided, and rarely pass through any given FOV.

On the other hand, the VLEO constellation would probably be immediately be inserted to the ~330 km operational orbits, and spend minimal time raising. It's not clear if they would use a low drag orientation, or what their brightness would be.

[eeergo]

That's probably still much darker than a building clad in polished bright aluminum under the same lighting. Starlinks don't need to disappear completely, losing a couple magnitudes would help a lot. For one thing, they would then not be naked-eye visible except under excellent conditions. And the saturation and ghosting issues would be at least partly mitigated.

Since the coating is specifically intended to reduce visibility, I'm sure SpaceX has at least considered the coating's reflectance across a range of incidence angles.

I'm not sure that the deorbit phase would have the same issue. The satellites should certainly not be in the "low drag" configuration for deorbit. And for deorbit, the time is a function of thrust + drag, while on ascent the time is a function of thrust - drag. Depending on the ratio of thrust to drag at low altitudes, this time difference could be quite significant.

Only about 2% of the constellation would be orbit-raising assuming a steady state of operations, 6 week obit raising time, and 6 year operational life. That's an average of 88 at any time for the 4425 satellite LEO constellation. 88 new very bright objects are unlikely to have any significant impact on observations, because they would be clustered spatially and temporally, easily avoided, and rarely pass through any given FOV.

On the other hand, the VLEO constellation would probably be immediately be inserted to the ~330 km operational orbits, and spend minimal time raising. It's not clear if they would use a low drag orientation, or what their brightness would be.

Of course a black panel is less bright than a mirror, why state the obvious? Point was the color actually doesn't have that much influence on the brightness once the reflection is strong - just the material's roughness. In the (illustrative, not apples-to-apples accurate) picture I posted you can actually see the albedo difference there is between the balcony edges, painted matte white, and the most reflecting area of the black paneling: little to none.

I don't know to what depth SpaceX has studied the coating but so far there is zero evidence of it having any effect. They also should have checked how bright their array would be and didn't - it surprised even their manufacturers, when its behavior should be quite straightforward. On the other hand, if the problem was coming from just the huge array's fixed position being somehow very visible precisely at twilight, the satellites should exhibit strong flaring as you mentioned before, which hasn't been reported and leads to believe there's more to their reflectivity.

It's not too worthwhile to talk about a 4000-unit constellation, or the (quite unproven) 2% of it in lower orbits: while representing a huge change in the orbital environment, nobody is claiming such numbers would seriously impede astronomical observations, apart from needing a more or less complex overhaul of the streak avoidance/removal bag of tricks. The satellites "in decomissioning", by the way, should not just include those actively deorbiting - also those that fail along the maneuver, or on their way up. You cannot plan for 100% nominal operations, and even 1% of tens of thousands is a large number.

[edzieba]

Of course a black panel is less bright than a mirror, why state the obvious?

Because the existing satellites are mirrors. Not being a mirror should therefore a significant and beneficial change.

Given that - as per the conference slides posted previously - the surprise from Starlink was not the number of satellites but their apparent brightness compared to objects of a similar altitude, "not being a mirror" may well be sufficient to bring the problem from a significant observation issue back down to the same issue as existing orbiting objects.

I don't know to what depth SpaceX has studied the coating but so far there is zero evidence of it having any effect.

Do you have a link to observations of Starlink 1130?

[thirtyone]

In other news, DarkSat hasn't failed to provide the intended mitigation (i.e. what was obvious before is patent now: a smooth flat surface will show similar brightness when specularly reflecting light no matter its color). Back to the drawing board I guess, if they ever even reached that point:

So, before we all continue making conclusions here - I'd like to point out, once again, that even the professional astronomers who specialize in tracking man-made space objects for the community, have emphasized that it will not be possible to determine how effective DarkSat will be until it's in its final operational orbit later in February IIRC. Many of these details are in the AAS video earlier. They are aware of the unavoidable satellite brightness during orbit raising and know that the coating has no effect in the no-drag orientation.

I myself would much prefer reading an article instead of watching 1+ hour of talk, but I haven't yet found a transcript, and there are really many useful details directly from professional astronomers (not SpaceX, and definitely not whoever wrote that paper random PDF). Recommend for anyone actively following the topic, because it's really the only trustworthy info so far IMO that comes from professional astronomers with some actual background.

[envy887]

That's probably still much darker than a building clad in polished bright aluminum under the same lighting. Starlinks don't need to disappear completely, losing a couple magnitudes would help a lot. For one thing, they would then not be naked-eye visible except under excellent conditions. And the saturation and ghosting issues would be at least partly mitigated.

Since the coating is specifically intended to reduce visibility, I'm sure SpaceX has at least considered the coating's reflectance across a range of incidence angles.

I'm not sure that the deorbit phase would have the same issue. The satellites should certainly not be in the "low drag" configuration for deorbit. And for deorbit, the time is a function of thrust + drag, while on ascent the time is a function of thrust - drag. Depending on the ratio of thrust to drag at low altitudes, this time difference could be quite significant.

Only about 2% of the constellation would be orbit-raising assuming a steady state of operations, 6 week obit raising time, and 6 year operational life. That's an average of 88 at any time for the 4425 satellite LEO constellation. 88 new very bright objects are unlikely to have any significant impact on observations, because they would be clustered spatially and temporally, easily avoided, and rarely pass through any given FOV.

On the other hand, the VLEO constellation would probably be immediately be inserted to the ~330 km operational orbits, and spend minimal time raising. It's not clear if they would use a low drag orientation, or what their brightness would be.

Of course a black panel is less bright than a mirror, why state the obvious? Point was the color actually doesn't have that much influence on the brightness once the reflection is strong - just the material's roughness. In the (illustrative, not apples-to-apples accurate) picture I posted you can actually see the albedo difference there is between the balcony edges, painted matte white, and the most reflecting area of the black paneling: little to none.

Based on photos of Starlinks, the dark-coated surfaces on Darksat were previously bright machined aluminum, which is highly reflective and exhibits a combination of specular and diffuse reflection. Covering machined aluminum with a dark coating will lower the average reflectance considerably, even if it does not change the ratio of diffuse and specular reflection.

Specular reflection by a polished dark surface can certainly be similar to diffuse reflection from a matte light surface, but the observer needs to be in the right location to see the specular reflection. Your building picture probably shows a specular reflection from the black surfaces, and maybe saturation of the CCD on the diffuse white surfaces. It's pretty easy to take a photo of a polished black surface and a matte white surface together under lighting at an appropriate angle to show the specular reflection and get this result. But at every other angle the black surface will appear much darker.

Astronomers would probably prefer occasional flares from polished surfaces when they happened to be at the wrong angle, over continuous but dimmer diffuse reflection. At least as long as the average duration between flares was much longer than dwell time of the satellite in the FOV, which even for survey telescopes is only on the order of 5 seconds.

I don't know to what depth SpaceX has studied the coating but so far there is zero evidence of it having any effect.

Some people observed it before it reached operational configuration and noticed that it wasn't darker. You find this to be evidence of... what, exactly? SpaceX said that it probably wouldn't be darker until it reached operational configuration.

[freddo411]

In other news, DarkSat hasn't failed to provide the intended mitigation (i.e. what was obvious before is patent now: a smooth flat surface will show similar brightness when specularly reflecting light no matter its color). Back to the drawing board I guess, if they ever even reached that point:

It's kind of telling to view this video.    We see a tiny point of light slowly cross a pretty large area of the sky.   Then a bit later another tiny point of light, and so on.

Not at all the like the still photos of huge swaths of light bars blocking the whole picture.

[envy887]

It's kind of telling to view this video.    We see a tiny point of light slowly cross a pretty large area of the sky.   Then a bit later another tiny point of light, and so on.

Not at all the like the still photos of huge swaths of light bars blocking the whole picture.

That only tells us that the video has a short exposure per frame.

[FutureSpaceTourist]

https://www.bbc.co.uk/news/science-environment-51334423

Mega-constellation firms meet European astronomers
By Jonathan Amos
BBC Science Correspondent
31 January 2020

Leading satellite mega-constellation companies SpaceX and OneWeb have met with astronomers in Europe to discuss the impact their operations could have on observations of the Universe.

There's concern that the size and brightness of the firms' planned fleets could interfere with the work of professional telescopes.

The parties discussed the issues in a private meeting at the Royal Astronomical Society in London, UK.

The talks were described "as positive".

Present for OneWeb was Dr Timothy Maclay, the start-up's director of mission systems engineering; and for SpaceX, the participant was Patricia Cooper, the California company's vice president of satellite government affairs.

Article goes on to talk about two research studies being conducted:

One, from the University of Southampton, has investigated the reflectivity of SpaceX's Starlink satellites and what's driving their brightness in the sky.

The second, from the European Southern Observatory organisation, has attempted to model how much observing time might be lost by the world's major telescope facilities if the mega-constellations' interference is as bad as some fear it could become.

[Rondaz]

Because OneWeb's satellites will be about twice as high as Starlink satellites... Steckel said they should not be visible to the naked eye. In addition, he said, OneWeb has been in dialogue with both the Royal Astronomical Society in London and the American Astronomical Society.

https://twitter.com/OneWeb/status/1224374990352658434

[thirtyone]

Because OneWeb's satellites will be about twice as high as Starlink satellites... Steckel said they should not be visible to the naked eye. In addition, he said, OneWeb has been in dialogue with both the Royal Astronomical Society in London and the American Astronomical Society.

https://twitter.com/OneWeb/status/1224374990352658434

I don't think some of the astronomical conventions have been mentioned in the OneWeb thread yet, but the last I heard OneWeb will actually be more problematic for quite a few of the serious astronomical surveys. They are apparently not visible to the naked eye, but bright enough to saturate at least one major project in the works (sounds almost like it's just at the sweet spot to cause actual problems for astronomers while being completely invisible to the general public). It's not obvious, but the higher altitude actually causes the satellites to be visible (to astronomical imaging equipment) throughout the night, instead of right and dawn and dusk, which causes quite a few more problems over the current starlink altitude:



There's a slide deck posted somewhere in the megconstellations/astronomy general thread as well.

[meekGee]

Because OneWeb's satellites will be about twice as high as Starlink satellites... Steckel said they should not be visible to the naked eye. In addition, he said, OneWeb has been in dialogue with both the Royal Astronomical Society in London and the American Astronomical Society.

https://twitter.com/OneWeb/status/1224374990352658434

I don't think some of the astronomical conventions have been mentioned in the OneWeb thread yet, but the last I heard OneWeb will actually be more problematic for quite a few of the serious astronomical surveys. They are apparently not visible to the naked eye, but bright enough to saturate at least one major project in the works (sounds almost like it's just at the sweet spot to cause actual problems for astronomers while being completely invisible to the general public). It's not obvious, but the higher altitude actually causes the satellites to be visible (to astronomical imaging equipment) throughout the night, instead of right and dawn and dusk, which causes quite a few more problems over the current starlink altitude:



There's a slide deck posted somewhere in the megconstellations/astronomy general thread as well.

No, that's wrong (or misleading).

If the satellites are higher, they'll be lit for longer times when then ground below them is in the dark, and what's worse said lit satellites will occur at higher apparent elevations - encroaching into the sky zones and times where it actually matters.

[abaddon]

Starlink vs the Astronomers: https://spacenews.com/starlink-vs-the-astronomers/

[Lars-J]

I watched this video, and was happy to see a very level headed discussion on the topic from these astronomers. (quite different from some of the loud people on twitter)

Well worth watching, no matter if you are in favor of or against constellations like Starlink.

[ChrisWilson68]

I watched this video, and was happy to see a very level headed discussion on the topic from these astronomers. (quite different from some of the loud people on twitter)

Well worth watching, no matter if you are in favor of or against constellations like Starlink.

Yes, the video was quite informative.  A couple of the questioners were on the hysterical side, but all the panelists were quite reasonable.

One of the things I found interesting that was brought up several times was that astronomers were fully aware that Starlink was coming.  They've had issues with satellites and other human-made space objects impacting optical astronomy for years.  They were aware of Starlink and thought that they'd be able to deal with it, but it turned out the Starlink satellites were much brighter than they expected.  And that's not just the temporary string-of-pearls orbit-raising phase.  Even after they reach their operational orbit, they are just barely visible in perfect conditions to the human eye, and several times brighter than astronomers expected.  They are also much brighter than SpaceX expected.

I'd be curious to learn why they are so much brighter than expected.  The astronomers say they are brighter than they expected with their experience with active and inactive satellites, rocket bodies, and other debris.  Why?

[thirtyone]

I'd be curious to learn why they are so much brighter than expected.  The astronomers say they are brighter than they expected with their experience with active and inactive satellites, rocket bodies, and other debris.  Why?

I've been really curious as well. My working theory was that they initially made estimates based on typical brightness for a satellite of that mass - but because Starlink has an abnormally high surface area (probably a bunch of novel design methodologies to make it flat packed and all), it has far more reflective surface area than is typical for a satellite of that mass. If that were the case, though, you'd expect SpaceX could have had some idea during design, but I'm guessing it just wasn't something on their minds at the time.