Impacts of Large Satellite Constellations on Astronomy

[eeergo]

What a tricky balance. Watch the stars from Earth and reach for the stars from above?

That's what we all space enthusiasts should be yearning for IMO. Clearly some have more urgent (and mundane) priorities where the very same science that allowed to formulate those is not a basic part or even important - but rather an annoyance.

[Frogstar_Robot]

It'd be laughable if it wasn't so sad these points of view are flooding any debate.

Science can be used for good or evil. But those are subjective terms. In this case, what the masses might think is a good thing, some scientists think is bad.

Personally, I am a fan of science. But a lot of people don't care much what scientists think, and there seems to be an increasing distrust of scientists, leading to popular and damaging anti-science beliefs.

It always puzzles me why scientists naively take a high handed approach, rather than pop down to the psychology dept, and ask, "what is the best way to present our arguments and get people on board?". Like it or not, they might need some good PR and marketing to make their case.

Scientists may not get their way. Denigrating the wider population is not going to help their case. Saying "we are scientists and you must do what we say" is not going to work. So I'm just saying, welcome to the real world.

[Dizzy_RHESSI]

You still provided no numbers, references, or calculations. I am not going to take the word of someone who has trouble with the definition of someone who has trouble with the definition of a simple word like "constant." (see below.) Your answer boils down to "because you said so."

You didn't ask for numbers, references or calculations. You said "what parameters and what instrument". If you really want a calculation then you should have asked.

The output from the FORS2 ETC is attached. FORS2 has a pixel scale of 0.25'' with the low res collimator[1], compared to LSST's 0.2''[2]. That doesn't really matter however as the trails are extended. My calculation uses the R_SPECIAL+76 filter[1], which is approximately the LSST r band but slightly broader[2]. The sites do have similar backgrounds[3][4], LSST even base their sky model on data from Paranal. The exact backgrounds won't really matter anyway because I match in SNR. I used the MIT red-optimized CCD, a Moon FLI of 0.50 and an airmass of 1.50 with an exposure time of 15 seconds to match LSST[2]. I assumed an extended source and chose a magnitude which reached a SNR of about 100 per pixel, this was 18.0 R_AB per arcsec^2. I then accounted for the difference in read noise between FORS2 (3.150 e-/pixel [1]) and LSST ( 9. e-/pixel [3]).

All this is to get counts per pixel of objects with an SNR of 100 per pixel in a system much like LSST, to compare the size of the noise terms. So per exposure the noise components are sqrt(11398) from noise in the object counts, sqrt(1755) from the background and 9.0 from the read noise (which is not square rooted). As you can clearly see the noise in the object counts dominates. Anyone with any experience of imaging could have told you this, as I correctly did.

[1]https://www.eso.org/sci/facilities/paranal/instruments/fors/doc/VLT-MAN-ESO-13100-1543_P06.pdf
[2]https://www.lsst.org/scientists/keynumbers
[3]https://smtn-002.lsst.io/
[4]https://www.eso.org/gen-fac/pubs/astclim/paranal/skybackground/

I'm just going to ignore all the bickering about the meaning about words. I have made myself clear and I'm tired of repeating myself. Building an argument based on ignoring my entire first post is not in good faith.

What context is there to explain? Jg tried to claim that satellite trails would be a problem for LSST detecting objects at a fraction of a photon per pixel level, which obviously needs stacking to detect at all. I already explained how stacking makes throwing out pixels from a single frame have marginal impact in the case of any reasonable number of frames.

As I said before, lots of the interesting science will be done in individual exposures.

I am me, you were addressing me, there is no "you" which includes both me and the scientists working LSST. I have done no goalpost moving, you are the one who keeps conflating different things and trying to rewrite old posts.

https://en.wikipedia.org/wiki/Generic_you

I have described why, because adding signal always increases the noise. You eventually seemed to agreed with this point (" Things are neither linear nor constant. ") and hence, the noise has to increase with the presence of a trail. But instead of staying on point you've diverted the discussion onto semantics.

The semantics is because you are straight up lying about things that were previously said by both me and you. When you deny stating something that you said a word for word definition of, I am going to pull out a dictionary. (and please don't try to say it wasn't word for word, "always" falls into the category "a period of time")
I gave the mathematical example where the noise started at 100 and "10" more noise actually meant 0.1 more noise. A change like this cannot be meaningfully called an increase or realistically claimed to have any impact on the science data. You have provided no counterargument to this.

"A change like this cannot be meaningfully called an increase" So now 100.1 is equal to 100.0. This is a deeply silly argument. As I originally said the noise always increases, even if it is marginal. I'm glad you now agree.

[eeergo]

It'd be laughable if it wasn't so sad these points of view are flooding any debate.

Science can be used for good or evil. But those are subjective terms. In this case, what the masses might think is a good thing, some scientists think is bad.

Personally, I am a fan of science. But a lot of people don't care much what scientists think, and there seems to be an increasing distrust of scientists, leading to popular and damaging anti-science beliefs.

It always puzzles me why scientists naively take a high handed approach, rather than pop down to the psychology dept, and ask, "what is the best way to present our arguments and get people on board?". Like it or not, they might need some good PR and marketing to make their case.

Scientists may not get their way. Denigrating the wider population is not going to help their case. Saying "we are scientists and you must do what we say" is not going to work. So I'm just saying, welcome to the real world.

I'm not interested in debating philosophical epistemology in this *technical* forum, much less with such tone-deaf arguments as "I <3 science BUT... scientists should get off their high horse" while the numbers and technical arguments are summarily dismissed.

Suffice to say the outreach PR and marketing, the leniency and spirit of cooperation towards corporations, and the uncalled-for weight on their overburdened shoulders imposed by those who are acting the offense, are very much present in the case of astronomers.

From megaconstellation operators (not just SpaceX, which at least is in the limelight and reportedly doing something about it), the public is just getting spin, wild untenable promises amid corporate wars, dismissive attitudes about legitimate proven concerns, derogatory comments such as yours where science has suddenly become an elite's unimportant game, and half-hearted assurances with no binding obligations or regulations that EverythingWillBeFine(TM) when it's been proven time and again not to be the case (not just in this specific instance of astronomy, but that's another topic).

Funny you are the spokesperson of "the masses". Those "masses" I've spoken casually to, who have basic formal education or just no particular interest in space or science, are horrified when the idea is presented to them of thousands of objects owned by private corporations, beaming stuff day and night on them, everywhere they go, for their private profit, with the only benefit for these "masses" being increased (paywalled) Internet connectivity.

Bowing, thanking and paying heftily to hand universal patrimony for free and without in-depth criticism to very few and notoriously ethics-challenged individuals is not "the real world", it has another name.

[Frogstar_Robot]

Funny you are the spokesperson of "the masses". Those "masses" I've spoken casually to, who have basic formal education or just no particular interest in space or science, are horrified when the idea is presented to them of thousands of objects owned by private corporations, beaming stuff day and night on them, everywhere they go, for their private profit, with the only benefit for these "masses" being increased (paywalled) Internet connectivity.

Previously you said, "these points of view are flooding any debate", but now you are saying a majority are in support? If so, there is nothing to get exercised about. [sarc]There will be mass demonstrations in the streets and the politicians will be forced to ban satellite constellations. [/sarc] Unfortunately there is a bit of a queue at the moment with many other issues people wish to protest about.

Most people don't care about climate change, even when it's impacts directly affect them. The reports of impacts on astronomy of large constellations make barely a ripple in the media, and the lack of comments under those articles show people are not interested.

The vast majority of people don't care about astronomy, but they do care about cheap internet access. As you said yourself, you are fighting a losing battle. I am sorry, but your irrational ranting about evil billionaires and private corporations etc is not going to have any positive effect.

[rsdavis9]

Unfortunately I gave up on backyard observing when the battle was won by streetlights everywhere. I live in a suburbia. I like to be close to a big city(bos 1hr) and close to the mountains and ocean. In my neighborhood there are 4 street lights in the corners and many young people that like to leave their outside lights on all night for what I can assume to be a level of added security. I live a mile from downtown which again is lit all night long. I have a 4 inch reflector and a 12 inch dobsonian reflector. I used to know all the bright stars for all the constellations. Satellites are fun to watch and don't bother me at all.

Maybe Elon will offer free launches for large telescopes in orbit. He could also provide a common mass produced bus for the telescope to be mounted to. It could provide: ion thrusters, data communication, etc. Fine pointing might be better left to the telescope part.

[eeergo]

Previously you said, "these points of view are flooding any debate", but now you are saying a majority are in support? If so, there is nothing to get exercised about.

Willfully uncontextualized: obviously I meant debate HERE or similar websites.

[sarc]There will be mass demonstrations in the streets and the politicians will be forced to ban satellite constellations. [/sarc] Unfortunately there is a bit of a queue at the moment with many other issues people wish to protest about.

Mass demonstrations aren't a necessary condition for a topic to be relevant, even critical, and sometimes they're precisely the contrary - but you already know that and are trolling.

Most people don't care about climate change, even when its impacts directly affect them. The reports of impacts on astronomy of large constellations make barely a ripple in the media, and the lack of comments under those articles show people are not interested.

You're making a wonderful point, the only one why I bother to answer: climate change is clearly an existential crisis for many, as are many other economic measures taken in the last few years that most people squarely impacted by barely know exist, let alone debate about. Yet the lack of "comments under those articles" (as a laughable proxy for public interest/concern) should make it an irrelevant topic? Or what's your point (rhetorical question, no need to waste more bytes)?

The vast majority of people don't care about astronomy, but they do care about cheap internet access. As you said yourself, you are fighting a losing battle. I am sorry, but your irrational ranting about evil billionaires and private corporations etc is not going to have any positive effect.

Starlink isn't about cheap Internet access, it's about connectivity in underserviced areas by its creators' own admission. Of course, it's really mainly about something else much less philanthropic. Still, most people DO care about environmental issues, corporate overreach and power grabbing, and -unfoundedly- about "radiation being beamed on us from space" - much more than about cheap entertainment. It may be a losing battle, but one worth fighting for most of us.

Anyway, enough of this worthless back-and-forth for me; since you're choosing to ignore and derise all hard data and authoritative arguments just posted, instead favoring ranting with trolling banners, you're hardly in a position to claim irrationality. If you wish to continue deriding science as an elitist optional game you're probably in the wrong forum - here every topic is directly anchored to it.

[Welsh Dragon]

It'd be laughable if it wasn't so sad these points of view are flooding any debate.

Science can be used for good or evil. But those are subjective terms. In this case, what the masses might think is a good thing, some scientists think is bad.

Personally, I am a fan of science. But a lot of people don't care much what scientists think, and there seems to be an increasing distrust of scientists, leading to popular and damaging anti-science beliefs.

It always puzzles me why scientists naively take a high handed approach, rather than pop down to the psychology dept, and ask, "what is the best way to present our arguments and get people on board?". Like it or not, they might need some good PR and marketing to make their case.

Scientists may not get their way. Denigrating the wider population is not going to help their case. Saying "we are scientists and you must do what we say" is not going to work. So I'm just saying, welcome to the real world.

If you think scientists don't use PR and marketing, you don't know what you're talking about. I am one, I should know. Fact remains that a very significant part of the population (in an country) couldn't care less, and no amount of PR is going to change that. This is equally true for politics or the economy, or whatever other important item you want to discuss. They're more interested in football (either kind) or celebrity gossip, and it is entirely their right to be. But don't lay all the blame on the scientists.

[meberbs]

You still provided no numbers, references, or calculations. I am not going to take the word of someone who has trouble with the definition of someone who has trouble with the definition of a simple word like "constant." (see below.) Your answer boils down to "because you said so."

You didn't ask for numbers, references or calculations. You said "what parameters and what instrument". If you really want a calculation then you should have asked.

And in the very next sentence I specifically requested evidence. What you provided in this post is evidence, what you previously provided in response was just more assertions.

From the numbers provided here, it is enough information to approximate where thresholds of satellite brightness may be to eliminate the impact. Much of the video seemed to be using mag 6 as a reference case. At mag 7 (like recent measurement of visorsat), the contribution to total signal from satellite brightness would drop an order of magnitude. Rounding to make the numbers convenient, and ignoring the read noise contribution for simplicity, a total signal of 3000 produces noise of 55, while a total signal of 1700 produces noise of 41, this can no longer be said to be dominated by the satellite signal in any sense. It is not quite to the point of the satellite signal contribution being negligible, but it is close. This indicates making the satellite signal have negligible effect in all cases is reasonable. This ties in to recommendation 3 from the report that suggests detailed simulations of SNR to determine this threshold in a more rigorous fashion.

I'm just going to ignore all the bickering about the meaning about words. I have made myself clear and I'm tired of repeating myself. Building an argument based on ignoring my entire first post is not in good faith.

Making your argument by ignoring what you previously said is what you are doing here though, and you refusal to even acknowledge what I wrote speaks to the fact that you have no counterargument.

What context is there to explain? Jg tried to claim that satellite trails would be a problem for LSST detecting objects at a fraction of a photon per pixel level, which obviously needs stacking to detect at all. I already explained how stacking makes throwing out pixels from a single frame have marginal impact in the case of any reasonable number of frames.

As I said before, lots of the interesting science will be done in individual exposures.

You asked for context on clarification of a point about science that requires stacking not being affected, and your response is just to ignore that context and refuse to address whether jg's claim that such science would be significantly affected is correct or wrong? I take this to mean that you have no argument that this science would be affected, and are just refusing to say so.

I am me, you were addressing me, there is no "you" which includes both me and the scientists working LSST. I have done no goalpost moving, you are the one who keeps conflating different things and trying to rewrite old posts.

https://en.wikipedia.org/wiki/Generic_you

As that says in that link, that usage is for a generic, unspecified person, but we know who presented the option of removing trails, it is not me, and it is not a random hypothetical generic person. (There since has been printed that set of recommendations, which references removing trails as well.)

But I thought you were dropping arguments about the meaning of words. I guess you are keeping it up here to avoid addressing what I said after mentioning that it was not my idea: "Currently no one has the information required to know exactly how effective it will be, but it at least leaves open the possibility of making the impact on the final science data negligible, even if the satellites are still visible in the frame "

"A change like this cannot be meaningfully called an increase" So now 100.1 is equal to 100.0. This is a deeply silly argument. As I originally said the noise always increases, even if it is marginal. I'm glad you now agree.

This is in the context of scientific measurements, where there are error bars on everything, including estimates of noise. That much difference in a hypothetical noise value would be in the margin of error in practice. Even if extremely careful measurements of the noise were made to make this effect measurable though, it seems absurd to claim that this would make any significant impact on the actual science result.

[Frogstar_Robot]

Anyway, enough of this worthless back-and-forth for me; since you're choosing to ignore and derise all hard data and authoritative arguments just posted, instead favoring ranting with trolling banners, you're hardly in a position to claim irrationality. If you wish to continue deriding science as an elitist optional game you're probably in the wrong forum - here every topic is directly anchored to it.

"you're choosing to ignore and derise all hard data and authoritative arguments" - that's exactly the type of high-handedness I referred to, which you dismissed. The point is the public doesn't care about hard data and authoritative arguments. You choose to shoot the messenger.

Either way, anyone who doesn't agree with you must be a troll  It does seem you are not interested in debate or discussion, just asserting your point of view. Your comments are littered with ad hom attacks, not facts. It's kinda like you are shouting into the wind, cathartic for you perhaps, but not the way to win hearts and minds.

As spokesman for the astronomy community, you have helped make up my mind that astronomers' objections are overblown, and seemingly based on irrational sentiment. So if it comes to a vote, I'll be dropping my ball in the mega-constellation column. Astronomers will have to deal with the issue, just like everyone else has to deal with negative impacts of technology on their daily lives. Astronomers are really smart people, they will figure something out in cooperation with the satellite operators.

But it doesn't matter what I think, these constellations are going ahead anyway. Lawmakers are not showing any interest in the problems faced by astronomers. The opposite in fact, there is already a proposal to buy Starlink terminals by one US state, and the UK government investing substantial amounts in Oneweb.

[edzieba]

Can we quit it with the bandwidth-wasting ad-moninems and straw-men already?
"Constellations are perfect and need no changes": not happening
"Constellation are unfixable and should not be launched": not happening

As of today, we have one constellation in full deployment swing with active mitigations going into place, and one zombie partial constellation that is already in the process of retooling for a new mission (if it can ever get its ducks in a row to start launching again in the first place) with unknown mitigations planned.

[Frogstar_Robot]

If you think scientists don't use PR and marketing, you don't know what you're talking about. I am one, I should know.

Oh my bad. I have just never noticed it.

Fact remains that a very significant part of the population (in an country) couldn't care less, and no amount of PR is going to change that. This is equally true for politics or the economy, or whatever other important item you want to discuss. They're more interested in football (either kind) or celebrity gossip, and it is entirely their right to be. But don't lay all the blame on the scientists.

Sadly you are right about all that. It seems to me that scientists are smart people, and have access to all the psychological know-how to persuade public opinion. They should really be able to work out what the public's hot buttons are, and  create a killer PR campaign. Presenting the facts, and assuming that people will agree with their authority, is a naive approach and doesn't work in the face of propaganda campaigns from multinational companies. Generally people trust scientists more than politicians, but that is getting seriously undermined by groups like the anti-vaxxers.

For example, in this case, I think the impact to highlight would be the ability to detect "killer asteroids". Yes, I know that is "NEOs" is the scientific term, but you have to adopt the language. It's something that the public could possibly be concerned about.

[Frogstar_Robot]

Can we quit it with the bandwidth-wasting ad-moninems and straw-men already?
"Constellations are perfect and need no changes": not happening
"Constellation are unfixable and should not be launched": not happening

As of today, we have one constellation in full deployment swing with active mitigations going into place, and one zombie partial constellation that is already in the process of retooling for a new mission (if it can ever get its ducks in a row to start launching again in the first place) with unknown mitigations planned.

I agree, I am not sure why it descended so quickly into a "either for or against" bunfight. I think everyone here is both interested in science and space, at least I am.

Clearly, these constellations are going to be a reality, and the impacts must be dealt with.  I am somewhat interested in reading about what mitigations are possible, but we can leave out the other stuff.

[Dizzy_RHESSI]

From the numbers provided here, it is enough information to approximate where thresholds of satellite brightness may be to eliminate the impact. Much of the video seemed to be using mag 6 as a reference case. At mag 7 (like recent measurement of visorsat), the contribution to total signal from satellite brightness would drop an order of magnitude. Rounding to make the numbers convenient, and ignoring the read noise contribution for simplicity, a total signal of 3000 produces noise of 55...

No they wouldn't. One magnitude is a factor of ~2.5 in flux, not an order of magnitude. So even if you were correct, the source noise would still dominate. For some reason you divided the counts by 4 which is neither an astronomical magnitude nor an order of magnitude, 4537 does not round to 3000.

And secondly your premise is totally wrong, Tony Tyson said very clearly that the SNR~100 is after getting below the non-linear mag 7 limit. From his slides in the EAS conference (attached):

"However, even if that works, the satellite trails will clearly be in the data at S/N~100 – complicating data analysis and limiting discoveries"

I'm not going to argue whether 100.1 is greater than 100. because it's now irreverent. You now agree the source noise is not negligible, which is what I said from the beginning. And I'm not going to respond to your constant attempts to put words in my mouth.

[su27k]

It's a little bit disappointing that their first listed mitigation is basically "don't do it".  That seems unrealistic and not particularly helpful.

They clarified this during the press briefing: Finding 2 is just a list of options to mitigate the impact, not all options are feasible, obviously "don't do it" is not feasible due to industry objection.

[meberbs]

No they wouldn't. One magnitude is a factor of ~2.5 in flux, not an order of magnitude. So even if you were correct, the source noise would still dominate. For some reason you divided the counts by 4 which is neither an astronomical magnitude nor an order of magnitude, 4537 does not round to 3000.

I said total signal, and 3000 also corresponds to the total counts for mag 7. That was shown in the graph in the video for most channels. (One channel being down near 1000.) Notably, some channels start levelling off at that point in the graph, which is expected for the case of the trail being comparable to background signals.

You are right that I confused astronomical magnitude with base 10 log scale, but at the same time, this makes it seem that your comparison and estimate of the signal was not particularly accurate. Using my numbers gives results for mag 7 comparable to the data in the video.

And secondly your premise is totally wrong, Tony Tyson said very clearly that the SNR~100 is after getting below the non-linear mag 7 limit. From his slides in the EAS conference (attached):

"However, even if that works, the satellite trails will clearly be in the data at S/N~100 – complicating data analysis and limiting discoveries"

This gets to one point that was rather confusing in the video, he often talks about order of magnitude better than v 0.9 so he is often talking about mag 6, while occasionally mentioning mag 7 as a goal. When I watch the video, I see him clearly saying and showing a slide that says the SNR of 100 is for when it is a magnitude dimmer than v 0.9 and it is clear in the graph that mag 6 is where the artifacts start being removable, so that is what that bullet point should be referring to.

I'm not going to argue whether 100.1 is greater than 100. because it's now irreverent. You now agree the source noise is not negligible, which is what I said from the beginning. And I'm not going to respond to your constant attempts to put words in my mouth.

Except it is relevant, because the data we have is not good enough to know the final values, what we know about visorsat is that it is no brighter than mag 7. The final real values may very well actually be negligible contributors to the noise. If they aren't it is possible that future visits could be made darker until it is negligible.

You claim that I am putting words in your mouth, but I have not done so, I pointed out exactly what you have said and what your words mean. Meanwhile you take my example numbers that show the noise contribution being small, and potentially could become negligible, and you turn that into a general agreement that it is not negligible, which is not supported by what I have said.

Edit: add screenshot from video that is notably different than what you posted in a rather important way.

Edit 2: There are a bunch of "approximate" and rounding happening in the video, so much was what is in the bullet points simply doesn't add up with the shown data unless you account for that. You aren't going to get exact results with 2-3 significant figures starting with order of magnitude or 1 significant figure numbers.

[Dizzy_RHESSI]

I said total signal, and 3000 also corresponds to the total counts for mag 7. That was shown in the graph in the video for most channels. (One channel being down near 1000.) Notably, some channels start levelling off at that point in the graph, which is expected for the case of the trail being comparable to background signals.

You are right that I confused astronomical magnitude with base 10 log scale, but at the same time, this makes it seem that your comparison and estimate of the signal was not particularly accurate. Using my numbers gives results for mag 7 comparable to the data in the video.

Their estimate is not the same as my calculation. You can't just add together the background noise and the source counts from two completely different assumptions. The point of using the ETC is that the background and source are treated consistently, same pixel scale, same throughput, same filter. If you add together different calculations you will get nonsense. My calculation was SNR matched, matching by counts makes no sense. Also note the levelling off primarily happens in the y band, which is less sensitive and has the highest background. The bands don't level out at the same magnitude because the backgrounds, count rates and throughputs are different, my calculation was for the r band. The r band is barely flattening at 7th.

My calculation was more conservative that the plot shown in the video. I assumed an airmass of 1.5, whereas they assumed zenith. I assumed 0.5 Moon FLI while they assumed dark sky (which they don't actually define). These factors will mean that the background I calculated was higher than they will estimate and the throughput was lower. Because I matched in SNR, this will mean the counts needed to reach SNR of 100 will be higher.

Once can work out background count rate from the information provided on the page below. These use the same assumptions as the video.
https://smtn-002.lsst.io

m_sky = -2.5 * log10(counts/t_exp) + m_zeropoint

Where m_sky is the sky surface brightness in mags/arcsec^2 and the last term is the magnitude zeropoint which corresponds to a flux which generates one electron per second. Using the data in the reference 8873 counts/arcsec^2 or 355 counts per 0.2'' pixel. Note this is still an order of magnitude less than the source counts from the plot, so the source noise dominates and hence the trail will significantly increase the noise even at 7th mag. The overall result is unchanged, the ratio of source to sky is very similar to the ETC. And this shows under their assumptions that it will take about another order of magnitude improvement to become comparable to the sky noise (~10th mag), which still as an effect on the total noise.

This gets to one point that was rather confusing in the video, he often talks about order of magnitude better than v 0.9 so he is often talking about mag 6, while occasionally mentioning mag 7 as a goal. When I watch the video, I see him clearly saying and showing a slide that says the SNR of 100 is for when it is a magnitude dimmer than v 0.9 and it is clear in the graph that mag 6 is where the artifacts start being removable, so that is what that bullet point should be referring to.

 v0.9 was ~5th mag not 6th. Read the slide you posted again. It says a factor of 10 better than v0.9, i.e. an order of magnitude, not an astronomical magnitude. It doesn't really matter what magnitude he was referring to for the purpose of comparing the noise.

[meberbs]

Their estimate is not the same as my calculation. You can't just add together the background noise and the source counts from two completely different assumptions. The point of using the ETC is that the background and source are treated consistently, same pixel scale, same throughput, same filter. If you add together different calculations you will get nonsense.

I was assuming that you made calculations that were actually relevant to the system being discussed. Since you did not do so it is incorrect to compare to the data in the presentation. This only shows that you cannot just translate parameters from another system to get results, what you need (and apparently neither of us have) is the parameters for the actual system.

Also note the levelling off primarily happens in the y band, which is less sensitive and has the highest background. The bands don't level out at the same magnitude because the backgrounds, count rates and throughputs are different, my calculation was for the r band. The r band is barely flattening at 7th.

It is clear that multiple bands are starting to level at that point. What this actually tells us is that doing calculations in a single band will not tell the whole story.

My calculation was more conservative that the plot shown in the video. I assumed an airmass of 1.5, whereas they assumed zenith. I assumed 0.5 Moon FLI while they assumed dark sky (which they don't actually define). These factors will mean that the background I calculated was higher than they will estimate and the throughput was lower. Because I matched in SNR, this will mean the counts needed to reach SNR of 100 will be higher.

And that you are no longer comparing to the magnitude of Starlink, but something brighter.

To summarize all of the different details you are adding on, it is that this is not a simple calculation that you can reliably cover all details in a back of the envelope calculation. If it was, the astronomer's recommendations would not have including going off to do detailed SNR calculations to determine what threshold brightness would result in a significant reduction in impacts. (In other words, when does it get to the point that either other noise sources dominate, or at least when would techniques like trail removal reduce the total resulting noise to something that is acceptable for the science goals.) If it were that easy that we could solve that question right here with the data we have, it would not have been a recommendation, they would have already done it.

This gets to one point that was rather confusing in the video, he often talks about order of magnitude better than v 0.9 so he is often talking about mag 6, while occasionally mentioning mag 7 as a goal. When I watch the video, I see him clearly saying and showing a slide that says the SNR of 100 is for when it is a magnitude dimmer than v 0.9 and it is clear in the graph that mag 6 is where the artifacts start being removable, so that is what that bullet point should be referring to.

 v0.9 was ~5th mag not 6th. Read the slide you posted again. It says a factor of 10 better than v0.9, i.e. an order of magnitude, not an astronomical magnitude. It doesn't really matter what magnitude he was referring to for the purpose of comparing the noise.

You clearly misunderstood what I said, because I never claimed that v0.9 was mag 6, but that mag 6 is the point where all bands reach the threshold labelled as able to remove artifacts. The point when this threshold is passed is what is referred to repeatedly as a factor of 10 below v0.9. This is where some of the information presented is not entirely consistent, most likely due to rounding. (Nothing against the presentation, the point of the presentation was not to provide data at the level needed for the calculations you are attempting.)

[Dizzy_RHESSI]

I was assuming that you made calculations that were actually relevant to the system being discussed. Since you did not do so it is incorrect to compare to the data in the presentation. This only shows that you cannot just translate parameters from another system to get results, what you need (and apparently neither of us have) is the parameters for the actual system.

It is relevant. The conclusion was correct, something you have entirely ignored. That doesn't mean you can combine different calculations, that is your mistake not mine. The r band zeropoints are basically the same between FORS2 and LSST, confirming my original claim. I literally just did the calculation with LSST's parameters.

LSST does not even have adaptive optics. No it is not the same as doing adaptive optics, deformable mirrors in the pupil or conjugated planes cannot magically remove objects from an image. Don't say it's trivial without an actual idea of how that would work. AO is not easy, it took decades to develop and it's still expensive and limited. It is not currently applicable to survey telescopes, and isn't for the foreseeable future.

To summarize all of the different details you are adding on, it is that this is not a simple calculation that you can reliably cover all details in a back of the envelope calculation. If it was, the astronomer's recommendations would not have including going off to do detailed SNR calculations to determine what threshold brightness would result in a significant reduction in impacts. (In other words, when does it get to the point that either other noise sources dominate, or at least when would techniques like trail removal reduce the total resulting noise to something that is acceptable for the science goals.) If it were that easy that we could solve that question right here with the data we have, it would not have been a recommendation, they would have already done it.

No. They want astronomers to simulate the effect of the constellation on scientific programs. There is no button on any ETC that says "add 4000 LEOsats", nor can you do it with pen and paper. You can only do that from simulations. ETCs exist to ensure you ask for the right about of time do you your science, they do not account for every aspect of being an observer. Estimating the total impact is nothing like figuring out which noise terms dominate in a single image, which is a simple calculation that have already shown you. Converting magnitudes into counts (and vice-versa) is trivial even at an undergraduate level. Your handwaving does not change this.

http://slittlefair.staff.shef.ac.uk/teaching/phy217/lectures/principles/L04/index.html#calibration

[meberbs]

It is relevant. The conclusion was correct, something you have entirely ignored. That doesn't mean you can combine different calculations, that is your mistake not mine. The r band zeropoints are basically the same between FORS2 and LSST, confirming my original claim. I literally just did the calculation with LSST's parameters.

They are not the same, or the resulting counts would be the same. The counts you ended up with were significantly different, per your own statement.

LSST does not even have adaptive optics. No it is not the same as doing adaptive optics, deformable mirrors in the pupil or conjugated planes cannot magically remove objects from an image. Don't say it's trivial without an actual idea of how that would work. AO is not easy, it took decades to develop and it's still expensive and limited. It is not currently applicable to survey telescopes, and isn't for the foreseeable future.

What in the world are you talking about? I said nothing about adaptive optics at all. This seems to be you changing the subject again, rather than address what was actually under discussion.

To summarize all of the different details you are adding on, it is that this is not a simple calculation that you can reliably cover all details in a back of the envelope calculation. If it was, the astronomer's recommendations would not have including going off to do detailed SNR calculations to determine what threshold brightness would result in a significant reduction in impacts. (In other words, when does it get to the point that either other noise sources dominate, or at least when would techniques like trail removal reduce the total resulting noise to something that is acceptable for the science goals.) If it were that easy that we could solve that question right here with the data we have, it would not have been a recommendation, they would have already done it.

No. They want astronomers to simulate the effect of the constellation on scientific programs. There is no button on any ETC that says "add 4000 LEOsats", nor can you do it with pen and paper. You can only do that from simulations. ETCs exist to ensure you ask for the right about of time do you your science, they do not account for every aspect of being an observer. Estimating the total impact is nothing like figuring out which noise terms dominate in a single image, which is a simple calculation that have already shown you. Converting magnitudes into counts (and vice-versa) is trivial even at an undergraduate level. Your handwaving does not change this.

And another complete non-sequiter. Read the recommendation #1 and #3 for observatories that were listed in the report attached to this post:
https://forum.nasaspaceflight.com/index.php?topic=48302.msg2124381#msg2124381

Nothing in those 2 bullets requires modelling 4000 satellites. It is talking about software processing to remove known trails from an image and detailed SNR calculations to determine how much residual error would be present in the images after the trails are removed. Bullet 3 explicitly requests figuring out the signal to noise that would remain present in a single image after the trail is removed, which is an operation that would be done on a single frame. Actually doing this full SNR calculation correctly is non-trivial, sure there may be undergrad level back of the envelope versions, but it is rarely that simple.

At some combination of high knowledge and/or low brightness, the residual effect goes to zero. Figuring out exactly where this happens is difficult, and also what threshold for the effect being negligible to the scientific results is also non-trivial. One of the only things that can be stated with certainty is that this point should be at some brightness where the satellite can still be detected in a single image, since if the satellite was too dim to be detected, then meeting the "no effect" level is trivial.