Impacts of Large Satellite Constellations on Astronomy

[su27k]

You do realize Elon Musk already said Starlink can rotate its solar panel in one axis?

https://twitter.com/elonmusk/status/1132906066423889920

See this is the problem with the anti-Starlink crowd, you don't even have the basic knowledge to comprehend the problem, yet you come here as if you know everything and demand SpaceX to stop immediately.

You do realize momentum wheels / torque rods need be much beefier for a maneuver you propose doing within a few minutes, also significantly reducing the spacecraft's power margin, than for a bbq roll over a whole semi-orbit, right?

1. Why does it have to be done in a few minutes?
2. How much beefier? Let's see some numbers.
3. How much reduction of power margin? Where are these reduction comes from?

But these are besides the point, the point is you didn't even know Starlink has this capability, you just assumed it doesn't, just like you assumed the impact to astronomy is significant, even though AAS itself doesn't know the impact and is asking the observatories to provide the details.

[eeergo]

1. Why does it have to be done in a few minutes?
2. How much beefier? Let's see some numbers.
3. How much reduction of power margin? Where are these reduction comes from?

1. Because otherwise you're unnecessarily reducing power margins when you're not a hindrance to ground-based night-sky visibility: you need to feather the array as quickly as possible in the few minutes where the sats are illuminated but the ground is not, for this idea to have any sense. As mentioned earlier, this is anyway just a partial solution (s/c body).

2. Any solid information about Starlink's design or performance are proprietary to SpaceX, so no idea.

3. Any solid information about Starlink's design or performance are proprietary to SpaceX, so no idea. The reduction comes from having as little as 0% power generation (fully side-on towards the sun vector) at a time where, nominally, the sat could be ~100% power positive, twice during a single orbit - while having to provide extra juice to the power-hungry actuators/wheels/torque rods and, obviously, keep the payloads working (add an extra of difficulty when the intersatellite laser links are added, because of pointing accuracy requirements).

But these are besides the point, the point is you didn't even know Starlink has this capability, you just assumed it doesn't, just like you assumed the impact to astronomy is significant, even though AAS itself doesn't know the impact and is asking the observatories to provide the details.

Aha. Me missing a two-liner tweet from May, when the sats were not even the final iteration (v0.9), is EXACTLY equal to estimating why tens of thousands of new bright satellites will have a non-negligible impact in most astronomical observations from the ground. Sure! Not a hyperbole at all!

[su27k]

1. Why does it have to be done in a few minutes?
2. How much beefier? Let's see some numbers.
3. How much reduction of power margin? Where are these reduction comes from?

1. Because otherwise you're unnecessarily reducing power margins when you're not a hindrance to ground-based night-sky visibility: you need to feather the array as quickly as possible in the few minutes where the sats are illuminated but the ground is not, for this idea to have any sense. As mentioned earlier, this is anyway just a partial solution (s/c body).

2. Any solid information about Starlink's design or performance are proprietary to SpaceX, so no idea.

3. Any solid information about Starlink's design or performance are proprietary to SpaceX, so no idea. The reduction comes from having as little as 0% power generation (fully side-on towards the sun vector) at a time where, nominally, the sat could be ~100% power positive, twice during a single orbit - while having to provide extra juice to the power-hungry actuators/wheels/torque rods and, obviously, keep the payloads working (add an extra of difficulty when the intersatellite laser links are added, because of pointing accuracy requirements).

In other words you have no way of knowing if this idea works or not, yet somehow you imply it can't work?

And yes, rotating the solar panel is a partial solution, I never said it is the entire solution. The full solution would also include coating the bottom of the body, which SpaceX is already doing.

But these are besides the point, the point is you didn't even know Starlink has this capability, you just assumed it doesn't, just like you assumed the impact to astronomy is significant, even though AAS itself doesn't know the impact and is asking the observatories to provide the details.

Aha. Me missing a two-liner tweet from May, when the sats were not even the final iteration (v0.9), is EXACTLY equal to estimating why tens of thousands of new bright satellites will have a non-negligible impact in most astronomical observations from the ground. Sure! Not a hyperbole at all!

It's pretty clear doing the impact analysis is not easy, otherwise AAS wouldn't need to send the survey, this much is obvious. Yet you insist that your 10 minutes estimate based on nothing except handwaving is correct, even though you admitted you don't have much experience with optical astronomy. I think that's enough evidence to show you don't know what you're talking about. This latest display of lack of knowledge about basic design of Starlink is just the cherry on top.

[eeergo]

In other words you have no way of knowing if this idea works or not, yet somehow you imply it can't work?

And yes, rotating the solar panel is a partial solution, I never said it is the entire solution. The full solution would also include coating the bottom of the body, which SpaceX is already doing.

Any idea would work with a large enough change of the spacecraft. That may mean that the Starlink system has to be strongly modified because of these concerns, that's what we're discussing. And, by the way, this was brought about by discussion of the solar panel being the major problem, which you denied before. Just "coatings" will not solve specular reflections, again.

It's pretty clear doing the impact analysis is not easy, otherwise AAS wouldn't need to send the survey, this much is obvious. Yet you insist that your 10 minutes estimate based on nothing except handwaving is correct, even though you admitted you don't have much experience with optical astronomy. I think that's enough evidence to show you don't know what you're talking about. This latest display of lack of knowledge about basic design of Starlink is just the cherry on top.

Doing a detailed impact analysis is not easy, nobody claimed that, and I explicitly have repeated it several times already in the last few posts.

I have provided much more than the 10-minute (I didn't actually specify a number, but whatever) estimate, which by the way is obvious to anyone with a passing knowledge of orbital mechanics, no need to be a professional optical astronomer - but you obviously are grasping at straws now.

I have no interest on further fundamentalist, aggressive rants from you that ignore 95% of the information/calculations I post, so as previewed in my response to your latest "any critic to SpaceX is propaganda from anti-SpaceX conspiracists", expect no further answer. Peace be with you!

[su27k]

In other words you have no way of knowing if this idea works or not, yet somehow you imply it can't work?

And yes, rotating the solar panel is a partial solution, I never said it is the entire solution. The full solution would also include coating the bottom of the body, which SpaceX is already doing.

Any idea would work with a large enough change of the spacecraft. That may mean that the Starlink system has to be strongly modified because of these concerns, that's what we're discussing. And, by the way, this was brought about by discussion of the solar panel being the major problem, which you denied before. Just "coatings" will not solve specular reflections, again.

Yes, it may mean Starlink has to be modified, or it may not, you have no data to judge either way. Yet you insist solar panel is a major problem, despite the fact that nobody (not SpaceX, not astronomers) mentioned anything about solar panels.

It's pretty clear doing the impact analysis is not easy, otherwise AAS wouldn't need to send the survey, this much is obvious. Yet you insist that your 10 minutes estimate based on nothing except handwaving is correct, even though you admitted you don't have much experience with optical astronomy. I think that's enough evidence to show you don't know what you're talking about. This latest display of lack of knowledge about basic design of Starlink is just the cherry on top.

Doing a detailed impact analysis is not easy, nobody claimed that, and I explicitly have repeated it several times already in the last few posts.

I have provided much more than the 10-minute (I didn't actually specify a number, but whatever) estimate, which by the way is obvious to anyone with a passing knowledge of orbital mechanics, no need to be a professional optical astronomer - but you obviously are grasping at straws now.

No need to be a professional optical astronomer to estimate the impact to astronomy? Then why is AAS asking professional optical astronomers to do the estimate, why don't they just pick a random guy from the street to do it? You act as if you care about astronomers, yet you claim anybody can do their job, no professional knowledge needed, if that's not hyperbole I don't know what is.

If you actually read the AAS survey, it's pretty clear AAS is asking for an estimate, not detailed analysis, they even used the word "estimate" twice. There's no way you could fit detailed analysis to a textbox online anyway.

[ZChris13]

Can we please not drown out potentially interesting news in pointless back and forths?

[Dizzy_RHESSI]

Again, the key is the integration time for a single exposure being orders of magnitude different for a satellite than a CR. That was the question you formulated, and I showed mathematically what qualitative reason suggested. If you performed a single raw integration lasting tens of seconds or minutes without subtracting noise, you'd end up overwhelmed by CRs, thermal noise and atmospheric fluorescence, among other things, as correctly pointed out by yourself a few posts ago...

VISTA stands for VISIBLE and Infrared Survey Telescope for Astronomy - so the numbers are indeed valid for what we were talking about. In any case, IR will also suffer from constellations, even more than in visible perhaps. Most other large professional visible telescopes have similar FOVs and pixel size (order-of-magnitude), see for example the Canarian GTC: gtc.iac.es/instruments/instrumentation.php

Optical astronomers regularly do single exposures lasting ten minutes or more. There are two reasons for this. One is read-out noise, which is added per read of the detector. If you split an observation into more sub-exposures you get more read noise as a result. But the big reason is overheads, it takes time to read out the detector. The slower the read out the lower the read noise.

https://www.eso.org/sci/facilities/paranal/cfp/cfp103/overheads.html

If you look at this list for some of the visible instruments on the VLT the overheads for read-out, they can be up to a minute. During this time the telescope isn't taking new data. If you try to do 0.1 second exposures you will waste 99% of your limited telescope time.

Dark current, thermal noise (which doesn't affect optical) and the sky all scale as exposure time, so you don't gain anything from taking two exposures instead of one. Cosmic rays do increase as a function of exposure time but on the ground they aren't that severe, you only need a couple sub-exposures to remove CRs from a stack. Also VISTA, despite it's name, never had a visible instrument. One was planned but it was never built.

The discrepancy with optical work is not the pixel scale, but the integration times used. Infrared astronomy uses very short exposures, most optical work does not. In infrared astronomy the sky level is very high and you have non-destructive read outs, so the observational strategy is very different. Most optical programs don't require sky offsets. Infrared astronomy is not likely to be as affected as optical astronomy, because of the shorter exposures, smaller fields of view and the fact that sunlight is less intense in the infrared. Optical astronomy is still bigger than infrared astronomy, at least on the ground.

http://doi.eso.org/10.18727/docs/1

If one looks at the VLT, there are roughly as many infrared instruments as visible (more if you count the interferometer). All the high publication rate instruments (UVES, MUSE, VIMOS, X-Shooter, FORS2) are visible instruments, although X-Shooter also covers the near infrared. The visible instruments dominate the publication rate.

[Star One]

The astronomer royal in the UK Martin Rees has now weighed in on the topic of satellite constellations like Starlink. He does attempt to be even handed in his comments.

https://amp.theguardian.com/commentisfree/2019/dec/29/be-wary-of-elon-musk-despoiling-the-vault-of-heaven

[envy887]

The astronomer royal in the UK Martin Rees has now weighed in on the topic of satellite constellations like Starlink. He does attempt to be even handed in his comments.

https://amp.theguardian.com/commentisfree/2019/dec/29/be-wary-of-elon-musk-despoiling-the-vault-of-heaven

From the article:

This entails launching up to 40,000 spacecraft into orbit ... There would be roughly one in every square degree over the sky (the area on the sky covered by a small coin held at arm’s length).

You would think that the Astronomer Royal would know that even though there are 41253 square degrees in the sky, and (potentially) 40000 satellites, that does NOT mean that there would be roughly 1 visible satellite per square degree of visible sky. While you can see about half the sky from any given location, you can only seen a much, much smaller fraction of all the satellites, because the vast majority are near or below the horizon.

This is such a gross and basic error that it's hard to put any weight to the rest of his article.

[niwax]

The astronomer royal in the UK Martin Rees has now weighed in on the topic of satellite constellations like Starlink. He does attempt to be even handed in his comments.

https://amp.theguardian.com/commentisfree/2019/dec/29/be-wary-of-elon-musk-despoiling-the-vault-of-heaven

From the article:

This entails launching up to 40,000 spacecraft into orbit ... There would be roughly one in every square degree over the sky (the area on the sky covered by a small coin held at arm’s length).

You would think that the Astronomer Royal would know that even though there are 41253 square degrees in the sky, and (potentially) 40000 satellites, that does NOT mean that there would be roughly 1 visible satellite per square degree of visible sky. While you can see about half the sky from any given location, you can only seen a much, much smaller fraction of all the satellites, because the vast majority are near or below the horizon.

This is such a gross and basic error that it's hard to put any weight to the rest of his article.

More than that, 1° from the ground != 1° in orbit. 1° difference in a 400km orbit is 17° for an observer on the surface looking straight up.

The article also mentions satellites popping into pictures randomly and unexpectedly. How will we ever know when a satellite is likely to pass?

[pochimax]

The astronomer royal in the UK Martin Rees has now weighed in on the topic of satellite constellations like Starlink. He does attempt to be even handed in his comments.

https://amp.theguardian.com/commentisfree/2019/dec/29/be-wary-of-elon-musk-despoiling-the-vault-of-heaven

From the article:

This entails launching up to 40,000 spacecraft into orbit ... There would be roughly one in every square degree over the sky (the area on the sky covered by a small coin held at arm’s length).

You would think that the Astronomer Royal would know that even though there are 41253 square degrees in the sky, and (potentially) 40000 satellites, that does NOT mean that there would be roughly 1 visible satellite per square degree of visible sky. While you can see about half the sky from any given location, you can only seen a much, much smaller fraction of all the satellites, because the vast majority are near or below the horizon.

This is such a gross and basic error that it's hard to put any weight to the rest of his article.

¿¿¿because the vast majority are near or below the horizon.

How could this even be phisically  possible???

Do you know what are you talking about?? And you are questioning Mr. Rees??!!

[envy887]

The astronomer royal in the UK Martin Rees has now weighed in on the topic of satellite constellations like Starlink. He does attempt to be even handed in his comments.

https://amp.theguardian.com/commentisfree/2019/dec/29/be-wary-of-elon-musk-despoiling-the-vault-of-heaven

From the article:

This entails launching up to 40,000 spacecraft into orbit ... There would be roughly one in every square degree over the sky (the area on the sky covered by a small coin held at arm’s length).

You would think that the Astronomer Royal would know that even though there are 41253 square degrees in the sky, and (potentially) 40000 satellites, that does NOT mean that there would be roughly 1 visible satellite per square degree of visible sky. While you can see about half the sky from any given location, you can only seen a much, much smaller fraction of all the satellites, because the vast majority are near or below the horizon.

This is such a gross and basic error that it's hard to put any weight to the rest of his article.

¿¿¿because the vast majority are near or below the horizon.

How could this even be phisically  possible???

Do you know what are you talking about?? And you are questioning Mr. Rees??!!

An observer on the surface of Earth can only see ~3.8% of the total area of a 550 km altitude shell. Any satellite at that altitude more than ~2700 km away is below the horizon. Unless the satellites are not evenly distributed but are congregating over one particular area for some reason, the remaining 96+% of the satellites won't be visible.

Put another way, if 40,000 Starlinks cover 70% of the globe in a 550 km shell, focusing on low and mid latitudes, that's still 10,000 km of shell area per satellite. Satellites will be 100 km from each other, which for an observer 550 km away puts them 10 degrees apart in the sky. You will need a truly massive coin to cover two objects 10 degrees apart in the sky. They will be closer together near the horizon, of course, but counting all the ones above 30 degrees elevation they will still be 50 times more sparse than "one per square degree".

Or to put it another way, at the areal densities Rees is suggesting, some 10,000 satellites would be visible above 30 degrees elevation for any observer!

[Stan-1967]

¿¿¿because the vast majority are near or below the horizon.

How could this even be phisically  possible???

Do you know what are you talking about?? And you are questioning Mr. Rees??!!

It seems pretty clear Mr. Rees is now a "Flat Earther".  He can add this alarmism to his long resume of concern trolling the LHC, & other kooky unprovable cosmological speculations on the universe.

[Star One]

¿¿¿because the vast majority are near or below the horizon.

How could this even be phisically  possible???

Do you know what are you talking about?? And you are questioning Mr. Rees??!!

It seems pretty clear Mr. Rees is now a "Flat Earther".  He can add this alarmism to his long resume of concern trolling the LHC, & other kooky unprovable cosmological speculations on the universe.

So don’t actually address what he says just character assassinate the man. If anyone is engaging in trolling here it isn’t Mr Rees.

[Dizzy_RHESSI]

It seems pretty clear Mr. Rees is now a "Flat Earther".  He can add this alarmism to his long resume of concern trolling the LHC, & other kooky unprovable cosmological speculations on the universe.

This is some strange concern trolling here:

"There is no risk [in LHC collisions, and] the LSAG report is excellent."

- Prof. Lord Martin Rees, UK Astronomer Royal and former President of the Royal Society of London. 

A quote that CERN feature on their LHC safety page.

And with respect to his current book which got some media attention, even in the mainstream media articles you can see he clearly isn't pointing to the LHC:

"Prof Rees said the third risk from particle accelerators, such as the Large Hadron Collider at Cern, was from a ‘catastrophe that engulfs space itself.’... Prof Rees pointed out that particles of much higher energies than are created in accelerators already collide frequently in the galaxy, without ripping space apart."

As for his "other kooky unprovable cosmological speculations", I think that is a deeply ignorant statement. Rees predicted the apparent superluminal motion in the jets of radio galaxies, which was observed a few years later with very long baseline interferometery. He was also one of the first to propose massive black holes as the central engines of quasars and other active galactic nuclei, which I'd consider proven correct by now. Rees coauthored some of the most important papers in structure formation. He was the first to consider the polarisation of the cosmic microwave background, which was detected decades later by DASI.

You can say you disagree with someone's assessment without attempting to dismiss their entire career.

[pochimax]

The astronomer royal in the UK Martin Rees has now weighed in on the topic of satellite constellations like Starlink. He does attempt to be even handed in his comments.

https://amp.theguardian.com/commentisfree/2019/dec/29/be-wary-of-elon-musk-despoiling-the-vault-of-heaven

From the article:

This entails launching up to 40,000 spacecraft into orbit ... There would be roughly one in every square degree over the sky (the area on the sky covered by a small coin held at arm’s length).

You would think that the Astronomer Royal would know that even though there are 41253 square degrees in the sky, and (potentially) 40000 satellites, that does NOT mean that there would be roughly 1 visible satellite per square degree of visible sky. While you can see about half the sky from any given location, you can only seen a much, much smaller fraction of all the satellites, because the vast majority are near or below the horizon.

This is such a gross and basic error that it's hard to put any weight to the rest of his article.

¿¿¿because the vast majority are near or below the horizon.

How could this even be phisically  possible???

Do you know what are you talking about?? And you are questioning Mr. Rees??!!

An observer on the surface of Earth can only see ~3.8% of the total area of a 550 km altitude shell. Any satellite at that altitude more than ~2700 km away is below the horizon. Unless the satellites are not evenly distributed but are congregating over one particular area for some reason, the remaining 96+% of the satellites won't be visible.

Put another way, if 40,000 Starlinks cover 70% of the globe in a 550 km shell, focusing on low and mid latitudes, that's still 10,000 km of shell area per satellite. Satellites will be 100 km from each other, which for an observer 550 km away puts them 10 degrees apart in the sky. You will need a truly massive coin to cover two objects 10 degrees apart in the sky. They will be closer together near the horizon, of course, but counting all the ones above 30 degrees elevation they will still be 50 times more sparse than "one per square degree".

Or to put it another way, at the areal densities Rees is suggesting, some 10,000 satellites would be visible above 30 degrees elevation for any observer!

If you can see a 3.8% shell this is like  1.600 square degrees. So, obviously, you will see an all sky with 1.600 satellites or one satellite per square degree as Mr. Rees said.

It is not necessary to confuse anybody here with wrong calculations that don't take into account that this are not fixed (geostationary) satellites but low orbital, moving ones, crossing the sky.

Edit. It doesn' t matter satellites are moving. Simply, your calculations are nonsense. At any point on the surface you will always see 50% of the sky sphere, not only 3.8%. No matter how low or high satellites are.

[pochimax]

And again, is a wrong policy to think only about current astronomy capabilities. It is very important to think how this constellations will affect near and medium future capabilities, specifically those all-sky or wide-field.

Astronomy is not frozen. It develops, everyday.

[edzieba]

Edit. It doesn' t matter satellites are moving. Simply, your calculations are nonsense. At any point on the surface you will always see 50% of the sky sphere, not only 3.8%. No matter how low or high satellites are.

There's a big difference in visibility between objects at apparent infinity (stars) and objects a mere 500km above the surface you are observing from. You cannot see half the starlink satellites form a point on the surface of the Earth, but only a very small fraction at any one time. And these are spread mostly around the horizon: overhead the coverage is very sparse. The sparse overhead coverage is the reason LEO constellations require so many satellites in the first place!

[niwax]

The astronomer royal in the UK Martin Rees has now weighed in on the topic of satellite constellations like Starlink. He does attempt to be even handed in his comments.

https://amp.theguardian.com/commentisfree/2019/dec/29/be-wary-of-elon-musk-despoiling-the-vault-of-heaven

From the article:

This entails launching up to 40,000 spacecraft into orbit ... There would be roughly one in every square degree over the sky (the area on the sky covered by a small coin held at arm’s length).

You would think that the Astronomer Royal would know that even though there are 41253 square degrees in the sky, and (potentially) 40000 satellites, that does NOT mean that there would be roughly 1 visible satellite per square degree of visible sky. While you can see about half the sky from any given location, you can only seen a much, much smaller fraction of all the satellites, because the vast majority are near or below the horizon.

This is such a gross and basic error that it's hard to put any weight to the rest of his article.

¿¿¿because the vast majority are near or below the horizon.

How could this even be phisically  possible???

Do you know what are you talking about?? And you are questioning Mr. Rees??!!

An observer on the surface of Earth can only see ~3.8% of the total area of a 550 km altitude shell. Any satellite at that altitude more than ~2700 km away is below the horizon. Unless the satellites are not evenly distributed but are congregating over one particular area for some reason, the remaining 96+% of the satellites won't be visible.

Put another way, if 40,000 Starlinks cover 70% of the globe in a 550 km shell, focusing on low and mid latitudes, that's still 10,000 km of shell area per satellite. Satellites will be 100 km from each other, which for an observer 550 km away puts them 10 degrees apart in the sky. You will need a truly massive coin to cover two objects 10 degrees apart in the sky. They will be closer together near the horizon, of course, but counting all the ones above 30 degrees elevation they will still be 50 times more sparse than "one per square degree".

Or to put it another way, at the areal densities Rees is suggesting, some 10,000 satellites would be visible above 30 degrees elevation for any observer!

If you can see a 3.8% shell this is like  1.600 square degrees. So, obviously, you will see an all sky with 1.600 satellites or one satellite per square degree as Mr. Rees said.

It is not necessary to confuse anybody here with wrong calculations that don't take into account that this are not fixed (geostationary) satellites but low orbital, moving ones, crossing the sky.

Edit. It doesn' t matter satellites are moving. Simply, your calculations are nonsense. At any point on the surface you will always see 50% of the sky sphere, not only 3.8%. No matter how low or high satellites are.

This is so wrong, it an be disproven with MS Paint. As I said earlier, unless you take into account orbit height, angle measurements from the center and a viewpoint on the surface are completely incomparable. Looking straight up, 1° spacing between satellites at 400km turns into 19° for an astronomer at 0km. This is highly nonlinear of course, but as a rule of thumb, the solid angle fov of a telescope is less than 1% of the corresponding solid angle measured from the center of the orbit. So a 1°x1° telescope would only be impeded on average by satellites spaced closer than 0.1°x0.1°, or 4 million randomly distributed.

[alexterrell]

I suppose that an effective communications array will provide a view of about two satellites (a minimum of one, but usually more), not too near (perhaps 20 degrees above) to the horizon, that could interfere with astronomy.  GPS and Galileo - being higher - have several satellites in view each, as a user needs to "see" four and they also need to operate where the user has poor view of the sky. It is this design criteria - and the number of competing satellite arrays - that will determine the number of satellites visible.

I suppose the difference is that these are going to streak across the sky, providing a line across any long exposure shots.

The effect will be small compared to micrometeorite entry, and I assume software is there to remove these "streaks".

Long term, astronomy needs to move more and more off-planet, which means astronomers really should be encouraging SLS and other big, low cost launchers.