Impacts of Large Satellite Constellations on Astronomy

[high road]

Sure there is. Put those telescopes in space, where there is no pesky atmosphere, planes, communication etc messing with the observations. It's not like satellites are the most important hindrance by far. With launch costs dropping, launchable sizes about to increase, and cubesat technology becoming ever more potent, space telescopes become ever more feasible.

Regardless of launch costs, free-flying telescopes are incredibly expensive, with much more expensive instruments, a huge lead time, and a limited lifetime with few (if any) opportunities for repair or upgrade.

Ground based instruments can be repaired, maintained and upgraded relatively easily, which has enormous advantages.

Simplistic "just launch space telescopes" stuff is not helpful.

--- Tony

Free flying telescopes are incredibly expensive and have huge lead times now because they were designed for expensive launches, and little or no maintenance possibilities. That is changing right now. Constellations of cheaper sats (of which some are allowed to fail without compromising the mission) can be used to create much larger telescopes. Cheaper launches allow on-orbit maintenance to be cost-effective. Bigger launch vehicles mean telescopes can be bigger before they need origami designs. And maybe on orbit assembly using spiderfab trusses and individual segments built on Earth becomes a thing too. Interpreting this as a simplistic 'just launch space telescopes' is not helpful indeed.

[pochimax]

Free flying telescopes are incredibly expensive and have huge lead times now because they were designed for expensive launches, and little or no maintenance possibilities. That is changing right now. Constellations of cheaper sats (of which some are allowed to fail without compromising the mission) can be used to create much larger telescopes. Cheaper launches allow on-orbit maintenance to be cost-effective. Bigger launch vehicles mean telescopes can be bigger before they need origami designs. And maybe on orbit assembly using spiderfab trusses and individual segments built on Earth becomes a thing too. Interpreting this as a simplistic 'just launch space telescopes' is not helpful indeed.

things are expensive because they are difficult.
https://www.airbus.com/newsroom/press-releases/en/2019/11/airbusbuilt-telescope-for-esas-euclid-mission-takes-shape.html

The Euclid Payload Module contains the largest telescope with such optical performance ever designed and integrated by Airbus.

After mounting the truss on the base plate, both made of silicon carbide, Airbus is now completing the final phase of integration: the optical alignment. The telescope will then be sent to Thales Alenia Space, where it will go through further testing and be integrated with the platform.

The whole payload integration necessitates extreme clean room discipline with a perfect cleanliness, temperature stability and lack of vibration. For example, during the mechanical assembly, precision of operations is calculated in µm, where for the optical alignment, the scale of precision is in nm (1 million nm in a millimetre).

The Euclid Payload Module has three primary features, the main telescope, a 1.2m diameter mirror and a three-mirror Korsch telescope. The main telescope is made entirely of  lightweight silicon carbide (SiC) which gives it excellent thermal stability and enables operation at 130 K (-140°C), making it invisible to near infrared wavelengths.

You can' t do space maintenance with micron precission.

[high road]

Free flying telescopes are incredibly expensive and have huge lead times now because they were designed for expensive launches, and little or no maintenance possibilities. That is changing right now. Constellations of cheaper sats (of which some are allowed to fail without compromising the mission) can be used to create much larger telescopes. Cheaper launches allow on-orbit maintenance to be cost-effective. Bigger launch vehicles mean telescopes can be bigger before they need origami designs. And maybe on orbit assembly using spiderfab trusses and individual segments built on Earth becomes a thing too. Interpreting this as a simplistic 'just launch space telescopes' is not helpful indeed.

things are expensive because they are difficult.
https://www.airbus.com/newsroom/press-releases/en/2019/11/airbusbuilt-telescope-for-esas-euclid-mission-takes-shape.html

The Euclid Payload Module contains the largest telescope with such optical performance ever designed and integrated by Airbus.

After mounting the truss on the base plate, both made of silicon carbide, Airbus is now completing the final phase of integration: the optical alignment. The telescope will then be sent to Thales Alenia Space, where it will go through further testing and be integrated with the platform.

The whole payload integration necessitates extreme clean room discipline with a perfect cleanliness, temperature stability and lack of vibration. For example, during the mechanical assembly, precision of operations is calculated in µm, where for the optical alignment, the scale of precision is in nm (1 million nm in a millimetre).

The Euclid Payload Module has three primary features, the main telescope, a 1.2m diameter mirror and a three-mirror Korsch telescope. The main telescope is made entirely of  lightweight silicon carbide (SiC) which gives it excellent thermal stability and enables operation at 130 K (-140°C), making it invisible to near infrared wavelengths.

You can' t do space maintenance with micron precission.

So all telescopes on earth go through clean room maintenance with maintenance precision after their initial construction? By which I mean, the telescopes that are so much cheaper than their in-space counterparts that were now doomed according to the post I responded to.

Hard to compare systems with very specific capabilities. There will always be a need/use for expensive telescopes, both on Earth and in space, regardless of our capability to build cheaper variants for other uses.

[envy887]

- First and foremost, I think the blame lies not only in the companies aiming to deploy large constellations, but also in the competent regulatory bodies. I can appreciate such matters represent a complicated issue to tackle, but it's also true there was little to no discussion as to how to orderly mitigate impacts before dozens of actual satellites started being orbited.

This discussion did take place in front of the FCC.  Perhaps not to your liking, but it did take place.

Could you point me to its conclusions, or at least the considerations they took into account? I'm not aware of any such discussion and it isn't being quoted as a roadmap, but we might be missing something. On the other hand, I'm guessing the FCC will deal with issues narrowly focused on impacts to communications, not the larger visibility and scientific impacts, much less orbital dangers.

The FCC does evaluate orbital dangers and requires that satellite operators meet international guidelines for deorbit/graveyard orbits and debris mitigation, as well as an evaluation of hazards posed by reentering space objects. All this data is in the various constellation operators FCC applications alongside the relevant RF data.

[envy887]

You're greatly overstating the general visual impact of even the largest constellation. These satellites are somewhere between 3rd to 6th magnitude in operation. The vast majority of the general public will never even notice they are there. The vast majority of people don't even notice ISS when it flies over at a (relatively) blinding -4 magnitude, or if they do they think it's an airplane.

3th to 6th magnitude in astronomy are very very very bright objects. the appearance of an object so bright in your FOV ruins your observation

My point was that something like 99.99% of the people in the world are not astronomers and do not care in the least about a 3rd to 6th magnitude object in the sky. They don't even care about objects 1,000 times brighter. Comparing the visual impact of such an object to a wind turbine or a high-rise building is absurd.

The impact of such objects in any general environmental assessment would that they have "no to negligible impact". Any assessment which would find such objects to have significant impact would be one focusing strictly on optical astronomy.

I could turn that round and say 99.99% of people in the world are not the kind of space advocates you get on here and don’t care about their interests. I bet a high percentage don’t even know or care about Starlink or whether it doesn’t or does exist. Or will ever use it in spite of claims. As the biggest users of this are actually likely to be by far the money markets of the world and the military who both prioritise the high response speeds.

So... you want to posit that more people care about getting good astronomical data than care about getting good internet connectivity?

[RedLineTrain]

- First and foremost, I think the blame lies not only in the companies aiming to deploy large constellations, but also in the competent regulatory bodies. I can appreciate such matters represent a complicated issue to tackle, but it's also true there was little to no discussion as to how to orderly mitigate impacts before dozens of actual satellites started being orbited.

This discussion did take place in front of the FCC.  Perhaps not to your liking, but it did take place.

Could you point me to its conclusions, or at least the considerations they took into account? I'm not aware of any such discussion and it isn't being quoted as a roadmap, but we might be missing something. On the other hand, I'm guessing the FCC will deal with issues narrowly focused on impacts to communications, not the larger visibility and scientific impacts, much less orbital dangers.

Attached is the order on the original ~4,000 satellite Starlink constellation.  Depending on your level of detail desired on the back-and-forth, you can pick through the footnotes and read those documents online.  The orbital debris discussion is in paragraphs 11 through 17.  The radio astronomy discussion is in paragraph 18.

There is nothing in the order regarding optical/IR telescope interference because nobody in the optical/IR astronomy community wrote to the FCC about it (the radio astronomers did write).  While I don't know of anything in the law giving the FCC power over optical/IR, the FCC is a political animal and would have reviewed such correspondence seriously.  Anything that SpaceX wants in the future from the FCC will take such considerations into account, which is in part why SpaceX is so sympathetic to astronomers about this.

In any event, if you think the FCC isn't doing a good enough job on this as the first regulatory body approving this, keep in mind that Starlink has to get landing rights in each country.  The idea that SpaceX or others who follow them are or could ever be running amok with their constellations is ludicrous.

[su27k]

No offense to astronomers, but I believe it's crazy to suggest our society cares more about astronomy than telecommunication. Here's some numbers:

1. National Science Foundation (NSF) Division of Astronomical Sciences (AST) annual budget: ~$250M, source: https://www.nsf.gov/mps/ast/astpresentations/AAS231TownHall.pdf

2. US satellite services industry revenue per year: ~$50B, source: https://www.sia.org/wp-content/uploads/2017/07/SIA-SSIR-2017.pdf

The difference is two orders of magnitude, so any impact on astronomy, even as high as 20%, would be miniscule comparing to the potential revenue. Now I'm not suggesting satellite industry just do whatever they want, if NGSO does become successful and has a sizable impact on astronomy, I would support some kind of compensation scheme. But this would only be needed if the impact becomes real, and at the same time only when it is real will the NGSO companies have the money to pay for it, so the problem kind of generates its own solution.

To put it another way, there're basically two future scenarios:

a. NGSO fails to reach its potential, in which case the constellation will either be deorbited, or remains at a small scale, which will have minimal impact to astronomy. In this scenario, astronomers have nothing to worry about. Note both SpaceX and Amazon's satellites are in a low enough orbit (less than 600km) such that without propulsion, they'll naturally decay in 5 years or so. So even if the NGSO companies went bankrupt and was unable to actively deorbit their satellites, it would not cause any long term issue.

b. NGSO becomes very successful, we have 10,000+ satellites and tens of billions of revenue, and there're modest impact on astronomy. Note the exact amount of impact is still being researched, and probably will depend on the number of satellites in orbit, but the highest number I've seen is 20% for LSST, the lowest is less than 1% from ESO, so the average is between the two. In this scenario, NGSO companies fork over some revenue to compensate for the impact on astronomy would be warranted.

[Asteroza]

You can' t do space maintenance with micron precision.

There are a number of industry people who might say otherwise, with the caveat that they assume active adaptive optics with active metrology of the telescope. There was a proposal about assembling a diffraction limited telescope at ISS apparently, that assumed assembly and metrology on-orbit (not clear if that included continuous active metrology during operation or not).

If you were doing something like OCCAMS, that would appear to be all active by default and by design. An all active design heavily favors maintainability, since recalibration is a given, and active parts will break eventually so the maintenance is a given.

[eeergo]

Attached is the order on the original ~4,000 satellite Starlink constellation.  Depending on your level of detail desired on the back-and-forth, you can pick through the footnotes and read those documents online.  The orbital debris discussion is in paragraphs 11 through 17.  The radio astronomy discussion is in paragraph 18.

There is nothing in the order regarding optical/IR telescope interference because nobody in the optical/IR astronomy community wrote to the FCC about it (the radio astronomers did write).  While I don't know of anything in the law giving the FCC power over optical/IR, the FCC is a political animal and would have reviewed such correspondence seriously.  Anything that SpaceX wants in the future from the FCC will take such considerations into account, which is in part why SpaceX is so sympathetic to astronomers about this.

In any event, if you think the FCC isn't doing a good enough job on this as the first regulatory body approving this, keep in mind that Starlink has to get landing rights in each country.  The idea that SpaceX or others who follow them are or could ever be running amok with their constellations is ludicrous.

Thank you for that, I was unaware of the existence of this document.

However

This is a very restricted discussion (pertaining mostly to RF interference and immediate orbital debris risks), approved for a scaled-down constellation proposal before any satellite was orbited and while details were still sparse and under development. For a system with such a potential wide impact, this process needs to be enlarged, updated and systematically discussed using the principle of caution as needed.

In that sense, already for the current document there are many caveats to the allowable limits of Starlink:

[...] the unprecedented number of satellites proposed by SpaceX and the other NGSO FSS systems in this processing round will necessitate a further assessment of the appropriate reliability standards of these spacecraft, as well as the reliability of these systems’ methods for deorbiting the spacecraft. Pending further study, it would be premature to grant SpaceX’s application based on its current orbital debris mitigation plan.
[...]
no consensus exists on what the two reliability numbers should be, and a design and fabrication reliability on the order of 0.999 or better per spacecraft may be prudent to mitigate the risk of malfunction in a 4,000+ spacecraft constellation (my comment: not achieved so far, although obviously at very early stage to tell) [...]
Although not a condition to this authorization, SpaceX should be aware of these facts and contact the National Science Foundation Spectrum Management Unit to assist with coordination and information on radio astronomy sites (so not done)

[Star One]

You're greatly overstating the general visual impact of even the largest constellation. These satellites are somewhere between 3rd to 6th magnitude in operation. The vast majority of the general public will never even notice they are there. The vast majority of people don't even notice ISS when it flies over at a (relatively) blinding -4 magnitude, or if they do they think it's an airplane.

3th to 6th magnitude in astronomy are very very very bright objects. the appearance of an object so bright in your FOV ruins your observation

My point was that something like 99.99% of the people in the world are not astronomers and do not care in the least about a 3rd to 6th magnitude object in the sky. They don't even care about objects 1,000 times brighter. Comparing the visual impact of such an object to a wind turbine or a high-rise building is absurd.

The impact of such objects in any general environmental assessment would that they have "no to negligible impact". Any assessment which would find such objects to have significant impact would be one focusing strictly on optical astronomy.

I could turn that round and say 99.99% of people in the world are not the kind of space advocates you get on here and don’t care about their interests. I bet a high percentage don’t even know or care about Starlink or whether it doesn’t or does exist. Or will ever use it in spite of claims. As the biggest users of this are actually likely to be by far the money markets of the world and the military who both prioritise the high response speeds.

So... you want to posit that more people care about getting good astronomical data than care about getting good internet connectivity?

I didn’t say. Try answering what I actually asked.

I am quite surprised the FCC approved this without at least conducting a wide ranging and extensive consultation with all relevant stakeholders.

[ChrisWilson68]

for private profit reasons and moreover setting jurisprudence for other such actors - not a handful of unique, non-profit scientific instruments.

In a properly functioning market economy, private profit comes from bettering people's lives.

Starlink is designed to provide better, cheaper communications to millions of people.  That's where the profit comes from -- from helping millions of people.  If they don't help millions of people, they don't make money.

Scientific instruments also provide a societal good, but implying that they are superior simply because they are not run by a for-profit organization is not justified.

[RedLineTrain]

Thank you for that, I was unaware of the existence of this document.

However

This is a very restricted discussion (pertaining mostly to RF interference and immediate orbital debris risks), approved for a scaled-down constellation proposal before any satellite was orbited and while details were still sparse and under development. For a system with such a potential wide impact, this process needs to be enlarged, updated and systematically discussed using the principle of caution as needed.

In that sense, already for the current document there are many caveats to the allowable limits of Starlink:

[...] the unprecedented number of satellites proposed by SpaceX and the other NGSO FSS systems in this processing round will necessitate a further assessment of the appropriate reliability standards of these spacecraft, as well as the reliability of these systems’ methods for deorbiting the spacecraft. Pending further study, it would be premature to grant SpaceX’s application based on its current orbital debris mitigation plan.
[...]
no consensus exists on what the two reliability numbers should be, and a design and fabrication reliability on the order of 0.999 or better per spacecraft may be prudent to mitigate the risk of malfunction in a 4,000+ spacecraft constellation (my comment: not achieved so far, although obviously at very early stage to tell) [...]
Although not a condition to this authorization, SpaceX should be aware of these facts and contact the National Science Foundation Spectrum Management Unit to assist with coordination and information on radio astronomy sites (so not done)

Regarding the radio astronomers, on June 19, subsequent to that Order, SpaceX reached a Coordination Agreement with the National Science Foundation.

Regarding debris, that Order was for only the first ~4,000 satellites.  For the next 8,000 v-band satellites, additional discussion on orbital debris is available in that Order (attached, see paragraphs 11 through 16).  And then for the next 30,000 satellites, you can be sure that this will be discussed again.

There have been and will be numerous bites at this apple in the US and in each other country where Starlink seeks landing rights, so your desires regarding additional regulation ring hollow.  This is already a very heavily regulated industry.

[Steven Pietrobon]

If an asteroid that has our number on it is missed because of all the satellites above interfered with viewing when it was best visible (low down at the horizon during dawn or dusk), then it could be a very big deal. A civilisation ending big deal!

[su27k]

If an asteroid that has our number on it is missed because of all the satellites above interfered with viewing when it was best visible (low down at the horizon during dawn or dusk), then it could be a very big deal. A civilisation ending big deal!

NASA is already working on building a space telescope for asteroid defense: https://forum.nasaspaceflight.com/index.php?topic=49105.0

[ChrisWilson68]

New article on the impact of Starlink and its ilk on astronomy especially the LSST. Contains a lot of quotes I’ve not seen before.

https://www.theguardian.com/science/2019/nov/22/not-cool-telescope-faces-interference-from-space-bound-satellites

There's nothing new in this article.  It's a rehashing of things we've discussed in detail on this thread.  And it's all very one-sided and misleading.  There's nothing in the article that mentions that most of the night sky for most of the night is not vulnerable to the sun's light reflected off LEO satellites.

That’s incorrect as I said in my OP the quotes are new. For you to call it one sided and misleading is disingenuous.

The new quotes have nothing new in them.  They are just quotes of different people rehashing things that have already been discussed here.

Especially as your own bias in this topic is clear for all to see.

What's clear for all to see is that I don't personally attack people.  I have confidence in the strength of my arguments to stand on their own.

[ChrisWilson68]

If an asteroid that has our number on it is missed because of all the satellites above interfered with viewing when it was best visible (low down at the horizon during dawn or dusk), then it could be a very big deal. A civilisation ending big deal!

It could also happen that Starlink provides enough economic benefit to let SpaceX finish development of Starship and Starship's ability to launch large amounts of mass to orbit on short notice lets us deflect an asteroid that we otherwise wouldn't have been able to deflect.

[Comga]

AURA Statement on the Starlink Constellation of Satellites

. A very conservative upper limit on the number of LSST pixels affected by Starlink satellites is about 0.01%, and quite likely smaller. Therefore, for LSST, even a constellation of about 10,000 Starlink satellites would be a nuisance rather than a real problem.

The bold part is the only quantitative estimate of the negative impact of Starlink to astronomy I can find so far, looks like at least in this case impact is basically non-existent.

The bold part is a quantitative estimate of the negative impact of Starlink to the LSST, not to astronomy.  By virtue of very long exposure times, satellites crossing the LSST's image field in a given exposure can be filtered out.

This is incorrect.
LSST has very short exposures.
It takes two 15s exposures per pointing.
These exposures then do not go into a pile of images, they are then put into what amounts to a database of objects after a very long pipeline to properly categorise image artifacts.

The CCDs in LSST are 56000 pixels across, with a 1/20 radian field of view.
This means that (if overhead), the field of view is 25km. and during one exposure the satellite will move 140km. A starlink satellite will 'always' only affect one frame of two, so will never be picked up as a false object.

The resolution of LSST is (at 500km) 0.5m, though obviously it has a 9m diameter diameter of a point.
Due to a starlink sat being 10m across, it's probably reasonable to assume it wipes out a 20m stripe (40 pixels wide* 16000 pixels/s), if it does not bloom.
The pixel depth before blooming is 175k electrons, or multiplied by the 64K pixels, 10 billion photoelectrons/s from a 60m^2 dish.
Or 150M photoelectrons/s/m^2.
This is 6*10^-11W/m^2.
Or, equivalent to a magnitude 10 star.

That is - if a source of an equivalent brightness is streaked over the LSST, it will not saturate it and have any science impact beyond the tiny fraction of the sky that is its track.
6*10^-11W/m^2 * 500km^2 *6(hemispherical)*60(area of LSST) = 12W that can be radiated by the starlink satelite without causing blooming.

The Starlink sat is reflecting ~10kW of light, so it is probably safe to say that without active measures on starlinks part, it will cause blooming.

I have annoyingly been unable to find the effect of blooming for various magnitude sources over 10.

https://www.youtube.com/watch?v=W2Ght1xYfk4?t=1043
17:10 details this process, i strongly recommend this talk, and the ones in the 'Mario' series if you are interested in how modern astronomy with piles of data is going to be done. 

The only time a Starlink satellite intercepts 10kW is above the region around the terminator. At that point, the sunlight falls on the dark solar panels, and any specular reflection above dusk goes away from the Earth. IIRC. However, at those times the 8(?) sq meter lower surface can be illuminated at shallow angles, so much less solar power.

Can Starlink satellites “flare”?

That image is the Starlink tracks across the LSST image field has the tracks dotted. Why is that?

These tracks are all deterministic. If astronomical imaging systems are sufficiently disrupted by satellite constellations they could be engineered to not record for the brief intervals the satellites track across the individual detector chips or, for some detector architectures, strips or areas of those arrays. That would result in some small degradation of sensitivity, but not massive loss.

As to casual visual observations, when I point out the ISS or Iridium flare to strangers no one has ever said that they have noticed them before. People won’t see them, never mind be bothered by them.

That’s not to say there is no problem. It just seems manageable.

PS. Anyone who talks about moving ANYTHING from the ground to orbit probably hasn’t built spaceflight hardware. I have.

[pochimax]

So all telescopes on earth go through clean room maintenance with maintenance precision after their initial construction? By which I mean, the telescopes that are so much cheaper than their in-space counterparts that were now doomed according to the post I responded to.

Yes, they are. Check it on the net.

Here a picture of a complex instrument on one of ESO's VLT. It is a very complex instrument. I disagree with your comment intention of " much cheaper than their in-space counterparts". This instruments aren' t cheap. They are expensive. The problem is that space counterpart will be incredible expensive (compared to ground instruments)

¿could you imagine an astronaut trying to fix something on this instrument?

[pochimax]

These tracks are all deterministic. If astronomical imaging systems are sufficiently disrupted by satellite constellations they could be engineered to not record for the brief intervals the satellites track across the individual detector chips or, for some detector architectures, strips or areas of those arrays. That would result in some small degradation of sensitivity, but not massive loss.

That’s not to say there is no problem. It just seems manageable.

It was manageable until now. But, it probably won' t be in the future, with all this constellations.

I don' t think the majority of observatories have enough money to anticipate the satellites nor to manage and use this contaminated data. Probably they will lost the observation.

[niwax]

These tracks are all deterministic. If astronomical imaging systems are sufficiently disrupted by satellite constellations they could be engineered to not record for the brief intervals the satellites track across the individual detector chips or, for some detector architectures, strips or areas of those arrays. That would result in some small degradation of sensitivity, but not massive loss.

That’s not to say there is no problem. It just seems manageable.

It was manageable until now. But, it probably won' t be in the future, with all this constellations.

I don' t think the majority of observatories have enough money to anticipate the satellites nor to manage and use this contaminated data. Probably they will lost the observation.

Even a 10x increase in satellites is barely noticeable and certainly nowhere near unmanageable (or expensive in any way). Even a very wide 5° shot only covers 34.9km on the 400km orbital shell. That's 0.0002% of the sky at that height.

Note that telescope fov angles are measured from the surface while distance between satellites and orbital planes is measured from the center of Earth. A 5° observation angle is equivalent to less than 0.29° from the center at a 400km orbit. So you can fit a whole bunch of observations even between tightly spaced orbital planes. A constellation of a few thousand satellites randomly distributed across the surface still gives minimum distances orders of magnitude above that. And in reality, telescopes are closer to the equator while satellites are most widely spaces, improving even on that.