Impacts of Large Satellite Constellations on Astronomy

[edzieba]

https://twitter.com/nasaspaceflight/status/1650234052572585984

An interview with astrophysicist Dr. Jonathan McDowell (@planet4589) from the Harvard-Smithsonian Center for Astrophysics on the concerns about megaconstellations photobombing observations from ground-based telescopes.

By Ian Atkinson (@IanPineapple)

https://www.nasaspaceflight.com/2023/04/megaconstellations-effects-on-astronomy/

Second image caption:

A Starlink v1.5 satellite (left) compared to a Starlink v1.0 satellite (right). Note the comparably larger size of the solar panel, which causes significant reflections when not angled away from Earth. (Credit: SpaceX)

The 1.0 satellite in the image only looks like it has a smaller array because the source is an image that had the array cropped out from the start - note the distinctive stairstep aliasing in the shadow on the bus, and how the corner of one of the array panels is clipped off due to the partially rotated image.

[FutureSpaceTourist]

https://twitter.com/planet4589/status/1651312934763593730

Oh brilliant. Think you've discovered a binary star using spectroscopy? Guess again, the second component is just reflected light from a Starlink sat.
Another new way that constellations can cause bogus astronomy discoveries if you're not very careful.

twitter.com/johnbarentine/status/1651310932537245698

MNRAS preprint showing contamination of LAMOST medium-resolution object spectra with sunlight reflected from Starlink satellites.

https://www.lamost.org/publications/upload/2023/Mikhail%20Kovalev-MNRAS(2023).pdf

[edzieba]

That's certainly an interesting failure mode! They determined that operational Starlink and Oneweb satellites would have been far too dim for the spectrograph to even detect, so this would have had to have be from a brighter source. The HST and ISS are given as examples of sufficiently bright sources, but they also estimated a recently launched Starlink train could also have been bright enough, and may have been in the right place at the right time. From the preprint:

While the satellites on operational orbit will not be a major
concern for LAMOST, the compact trains of very low satellites can
affect the observations. The probability of such a train crossing a
telescope field of view is low, but considering that constellations
will need to be regularly replenished, new satellites will need to
be continuously launched. Considering 100 000 satellites with a
life-time of 5 years, this would result in about one launch per day
(each with 60 satellites). If the satellites stay one month in low
orbit, this would result in about 60 trains in orbit, at various stage
of dispersion. It is therefore important that the satellite operators
also keep the brightness of the satellites to the absolute minimum
possible during their stay on transit orbit. The changes of satellite
attitude implemented by Starlink illustrate the improvements than
can be made.

[Twark_Main]

Considering 100 000 satellites with a
life-time of 5 years, this would result in about one launch per day
(each with 60 satellites).

Considering Starlink 2.0, the 100k estimate is increasingly implausible.

[FutureSpaceTourist]

https://twitter.com/jeff_foust/status/1665799882354098176

At the NSF town hall during #AAS242, Debra Fischer says the NSF has signed a coordination agreement with OneWeb to mitigate satellite interference on astronomy, similar to an agreement with SpaceX announced in January. More details about the OneWeb agreement "soon."

[deltaV]

Has anyone made a quantitative estimate of how much reflections from large constellations will cost astronomy in lost productivity and in mitigations? From the little I know it seems astronomers should be able to work around the satellite reflections with measures such as taking many short exposures and discarding the bad pixels in each frame, not using telescopes near dawn and dusk when they may see sun-lit satellites, or using more space-based telescopes. These mitigations could be expensive (especially if they have to use space telescopes) but probably not expensive enough to justify making internet significantly more expensive for millions of people.

[Twark_Main]

Has anyone made a quantitative estimate of how much reflections from large constellations will cost astronomy in lost productivity and in mitigations?

Hard to say.

What's the cost (in money) of a scientific breakthrough you didn't make, you furthermore one you don't know you didn't make? 

These mitigations could be expensive (especially if they have to use space telescopes) but probably not expensive enough to justify making internet significantly more expensive for millions of people.

Reasonable brightness mitigations do not make satellite internet "significantly more expensive", as Starlink's trailblazing work has demonstrated.

[Robotbeat]

Has anyone made a quantitative estimate of how much reflections from large constellations will cost astronomy in lost productivity and in mitigations?

Hard to say.

What's the cost (in money) of a scientific breakthrough you didn't make, you furthermore one you don't know you didn't make? 

These mitigations could be expensive (especially if they have to use space telescopes) but probably not expensive enough to justify making internet significantly more expensive for millions of people.

Reasonable brightness mitigations do not make satellite internet "significantly more expensive", as Starlink's trailblazing work has demonstrated.

I think deltaV is referring to preventing Starlink expansion entirely as the “making internet more expensive” option.

Starlink and it’s ecosystem (including Starship) is most certainly ALSO going to lead to scientific breakthroughs. The goal for everyone is to make sure the net benefit to astronomy and science and humanity are positive.

[edzieba]

Preprint: Starlink Generation 2 Mini Satellites: Photometric Characterization

Starlink Generation 2 Mini satellites are fainter than Gen 1 spacecraft despite their larger size. The mean of apparent magnitudes for satellites in brightness mitigation mode is 7.06 +/- 0.10. When these magnitudes are adjusted to a uniform distance of 1,000 km that mean is 7.87 +/- 0.09. The brightness mitigation mode reduces distance-adjusted satellite luminosity by a factor of 12 relative to spacecraft that are not mitigated.

7 was the target magnitude recommended at SATCON 1 (Recommendation 5).

[FutureSpaceTourist]

Crosspost. SpaceX video (which first tweet below quotes) gives very clear view of the dielectric mirrors:

https://twitter.com/starlink/status/1703132933102649527

Developed in-house, the dielectric mirrors on the surface of the satellites and extremely dark black paint for angled surfaces or those not conducive to mirror adhesion help absorb and redirect light away from the ground

We firmly believe in the importance of protecting the night sky for all to enjoy, which is why the Starlink team has been working with leading astronomers around the world to reduce satellite brightness

https://twitter.com/starlink/status/1703135000206745949

The Starlink team continues to share best practices for brightness mitigation, and has made the dielectric mirror film available at cost to other satellite operators → https://api.starlink.com/public-files/BrightnessMitigationBestPracticesSatelliteOperators.pdf

[Twark_Main]

Has anyone made a quantitative estimate of how much reflections from large constellations will cost astronomy in lost productivity and in mitigations?

Hard to say.

What's the cost (in money) of a scientific breakthrough you didn't make, you furthermore one you don't know you didn't make? 

These mitigations could be expensive (especially if they have to use space telescopes) but probably not expensive enough to justify making internet significantly more expensive for millions of people.

Reasonable brightness mitigations do not make satellite internet "significantly more expensive", as Starlink's trailblazing work has demonstrated.

I think deltaV is referring to preventing Starlink expansion entirely as the “making internet more expensive” option.

If that's what they meant, then deltaV was constructing a false dichotomy.

[FutureSpaceTourist]

https://www.nature.com/articles/s41586-023-06672-7

Published: 02 October 2023
The high optical brightness of the BlueWalker 3 satellite

Authors

Sangeetha Nandakumar, Siegfried Eggl, Jeremy Tregloan-Reed, Christian Adam, Jasmine Anderson-Baldwin, Michele T. Bannister, Adam Battle, Zouhair Benkhaldoun, Tanner Campbell, J. P. Colque, Guillermo Damke, Ilse Plauchu Frayn, Mourad Ghachoui, Pedro F. Guillen, Aziz Ettahar Kaeouach, Harrison R. Krantz, Marco Langbroek, Nicholas Rattenbury, Vishnu Reddy, Ryan Ridden-Harper, Brad Young, Eduardo Unda-Sanzana, Alan M. Watson, Constance E. Walker, John C. Barentine, Piero Benvenuti, Federico Di Vruno, Mike W. Peel, Meredith L. Rawls, Cees Bassa, Catalina Flores-Quintana, Pablo García, Sam Kim, Penélope Longa-Peña, Edgar Ortiz, Ángel Otarola, María Romero-Colmenares, Pedro Sanhueza, Giorgio Siringo & Mario Soto

Nature (2023)

We are providing an unedited version of this manuscript to give early access to its findings. Before final publication, the manuscript will undergo further editing. Please note there may be errors present which affect the content, and all legal disclaimers apply.

Abstract

Large constellations of bright artificial satellites in low Earth orbit (LEO) pose significant challenges to ground-based astronomy1. Current orbiting constellation satellites have brightnesses between apparent magnitudes 4 and 6, while in the near-infrared Ks band they can reach magnitude 22. Satellite operators, astronomers, and other users of the night sky are working on brightness mitigation strategies3,4. Radio emissions induce further potential risk to ground-based radio telescopes that also need to be evaluated. Here we report the outcome of an international optical observation campaign of a prototype constellation satellite, AST Space Mobile’s “BlueWalker 3”. BlueWalker 3 features a 64.3 m2 phased-array antenna as well as a Launch Vehicle Adapter (LVA)5. The peak brightness of the satellite reached an apparent magnitude of 0.4. This made the new satellite one of the brightest objects in the night sky. Additionally, the LVA reached an apparent V-band magnitude of 5.5 four times brighter than the current International Astronomical Union recommendation of magnitude 73,6; it jettisoned on 2022-11-10 (UT), and its orbital ephemeris was not publicly released until four days later. The expected build-out of constellations with hundreds of thousands of new, bright objects1 will make active satellite tracking and avoidance strategies a necessity for ground-based telescopes.

[deadman1204]

Has anyone made a quantitative estimate of how much reflections from large constellations will cost astronomy in lost productivity and in mitigations? From the little I know it seems astronomers should be able to work around the satellite reflections with measures such as taking many short exposures and discarding the bad pixels in each frame, not using telescopes near dawn and dusk when they may see sun-lit satellites, or using more space-based telescopes. These mitigations could be expensive (especially if they have to use space telescopes) but probably not expensive enough to justify making internet significantly more expensive for millions of people.

As of today, its more than 10% of all observations are heavily impacted or thrown away. That number will only increase. I don't mean theres a streak in the corner. That 10% is of quite significant impact.

However, it goes way beyond this. Astronomers have spent alot of time and money trying to find alternate ways to observe to mitigate starlink. Huge projects to design new ways to ignore all the starlinks blazing across the observations. Some hardware is better at this than other. ALOT of time and money has been spent over dealing with starlink. However its probably near impossible to quantify. So many countries with soo many unrelated teams all building new algorithms to salvage the data.

This also only looks at visible light observations. Radio is a bigger problem. Starlink and other constellations will blast the ground with soo many colors of radio light. All the mitigations spaceX is trying to do for starlink only extends to visible light. The only way to not blind radio telescopes is to turn off over certain regions.

[edzieba]

Has anyone made a quantitative estimate of how much reflections from large constellations will cost astronomy in lost productivity and in mitigations? From the little I know it seems astronomers should be able to work around the satellite reflections with measures such as taking many short exposures and discarding the bad pixels in each frame, not using telescopes near dawn and dusk when they may see sun-lit satellites, or using more space-based telescopes. These mitigations could be expensive (especially if they have to use space telescopes) but probably not expensive enough to justify making internet significantly more expensive for millions of people.

As of today, its more than 10% of all observations are heavily impacted or thrown away. That number will only increase. I don't mean theres a streak in the corner. That 10% is of quite significant impact.

That depends on the proportion of observations that would otherwise make scientific contributions vs. those that are for other purposes or are rejected for other issues (e.g. aircraft overflights, streaks from other local objects, etc). For Starlink in particular (does not apply to Oneweb and other higher orbits) the illumination peak is during twilight, where astronomical observations are more opportunistic. If 10% of observations are impacted but those 10% impacted would have had a 90% rejection rate without satellite streaks due to background light, that's a big difference to impacting 10% of observations at midnight which would ahve a much higher baseline acceptance rate.

This also only looks at visible light observations. Radio is a bigger problem. Starlink and other constellations will blast the ground with soo many colors of radio light. All the mitigations spaceX is trying to do for starlink only extends to visible light. The only way to not blind radio telescopes is to turn off over certain regions.

The RF issue has been known since before the first Starlink bird launched. SAR birds are a far bigger threat than downlink RF: downlink RF is much lower power and almost always beamformed to target users (why waste precious power budget), whereas RADAR is desigend for uniform coverage and has sufficient power not just to blind receivers but to physically damage them (burning out frontend amplifiers). The solution for both is the same though: turning off emitters over astronomy sites. This is the solution currently in use.

[deltaV]

Astronomers have spent alot of time and money trying to find alternate ways to observe to mitigate starlink.

If all ~10,000 professional astronomers (https://astronomy.stackexchange.com/questions/6328/how-many-astronomers-are-there-in-the-world-today) have or will spend $100k worth of time and other resources each that's $1B. That's significant but much smaller than the value that the constellations will provide. This suggests that the solution should be to let the constellations glow but impose a fee/tax on satellite nighttime brightness that's used to fund astronomy.

[deadman1204]

Has anyone made a quantitative estimate of how much reflections from large constellations will cost astronomy in lost productivity and in mitigations? From the little I know it seems astronomers should be able to work around the satellite reflections with measures such as taking many short exposures and discarding the bad pixels in each frame, not using telescopes near dawn and dusk when they may see sun-lit satellites, or using more space-based telescopes. These mitigations could be expensive (especially if they have to use space telescopes) but probably not expensive enough to justify making internet significantly more expensive for millions of people.

As of today, its more than 10% of all observations are heavily impacted or thrown away. That number will only increase. I don't mean theres a streak in the corner. That 10% is of quite significant impact.

That depends on the proportion of observations that would otherwise make scientific contributions vs. those that are for other purposes or are rejected for other issues (e.g. aircraft overflights, streaks from other local objects, etc). For Starlink in particular (does not apply to Oneweb and other higher orbits) the illumination peak is during twilight, where astronomical observations are more opportunistic. If 10% of observations are impacted but those 10% impacted would have had a 90% rejection rate without satellite streaks due to background light, that's a big difference to impacting 10% of observations at midnight which would ahve a much higher baseline acceptance rate.

Huh? How do I unpack this. All the observations all the big telescopes make ARE for science. They are so over subscribed its not like they are just randomly pointing for fun half the time.

Your "twilight" is multiple hours after sunset and before dawn. Musk fans hand wave it away, but it actually adds up to close to half the night. 2-3 hours after sunset, 2-3 hours before dawn? Thats 4-6 hours. Only half the night is being generous.
ALL observations are the best time to make them, because differnet things are above the horizon at different times. Many things may never get that high above the horizon. So you observe when its at the best position to. Midnight isn't a magic time. Its just the middle of the dark part of the day.  That lost data I mentioned is a greater than 10% loss of ALL Science observations. Yes, ALL the observations are scientific and matter. Your saying that observations more than 50% of the night (2-3 hours on each end) don't matter. Now we are talking about a 70% decrease in science.


Note, I'm saying starlink only here, because 99% of all impacts are starlink. Airline paths don't go over the major observatories, and most all satellites are starlink. However, we haven't discussed IR either, cause starlink blazes in those colors too. Yet not only is nothing being done about that, making them darker to visible light actually makes the IR problem vastly worse.

The RF issue has been known since before the first Starlink bird launched. SAR birds are a far bigger threat than downlink RF: downlink RF is much lower power and almost always beamformed to target users (why waste precious power budget), whereas RADAR is desigend for uniform coverage and has sufficient power not just to blind receivers but to physically damage them (burning out frontend amplifiers). The solution for both is the same though: turning off emitters over astronomy sites. This is the solution currently in use.

There are what, a few dozen SAR satellites up there? Well over a thousand starlinks? With many thousand more coming? Math seems to be hard


I would ask instead why you are so keen to dismiss the concerns of the entire professional astronomy community. Its their job to know this stuff. Why not take what they say at face value? I would posit that if starlink was owned by bezos, no one would be ignoring and second guessing every single scientist on the planet involved in astronomy.

[edzieba]

Has anyone made a quantitative estimate of how much reflections from large constellations will cost astronomy in lost productivity and in mitigations? From the little I know it seems astronomers should be able to work around the satellite reflections with measures such as taking many short exposures and discarding the bad pixels in each frame, not using telescopes near dawn and dusk when they may see sun-lit satellites, or using more space-based telescopes. These mitigations could be expensive (especially if they have to use space telescopes) but probably not expensive enough to justify making internet significantly more expensive for millions of people.

As of today, its more than 10% of all observations are heavily impacted or thrown away. That number will only increase. I don't mean theres a streak in the corner. That 10% is of quite significant impact.

That depends on the proportion of observations that would otherwise make scientific contributions vs. those that are for other purposes or are rejected for other issues (e.g. aircraft overflights, streaks from other local objects, etc). For Starlink in particular (does not apply to Oneweb and other higher orbits) the illumination peak is during twilight, where astronomical observations are more opportunistic. If 10% of observations are impacted but those 10% impacted would have had a 90% rejection rate without satellite streaks due to background light, that's a big difference to impacting 10% of observations at midnight which would have a much higher baseline acceptance rate.

Huh? How do I unpack this. All the observations all the big telescopes make ARE for science. They are so over subscribed its not like they are just randomly pointing for fun half the time.

All* observations are intended for science. Not all will produce useful data. Clouds unfortunately still exist, sometimes atmospheric distortion is unacceptable, non-Starlink orbiting bodies can intrude in the frame, background illumination levels can be too high to observe your target, etc. Every* frame has a purpose, but not every one can satisfy that purpose.

However, we haven't discussed IR either, cause starlink blazes in those colors too. Yet not only is nothing being done about that, making them darker to visible light actually makes the IR problem vastly worse.

The dichroic coatings developed for Starlink are also effective in nIR and mid-IR, not just the visual spectrum. Even in thermal IR, emission is dependant on insolation just as with visual-range emission, so they glow in the same twilight conditions as in the visual range (where sky temperature makes IR observation even more difficult than normal). Observations of and calculations of Starlink impacts are also including IR effects, such as the ZTF studies.
 

I would ask instead why you are so keen to dismiss the concerns of the entire professional astronomy community. Its their job to know this stuff. Why not take what they say at face value? I would posit that if starlink was owned by bezos, no one would be ignoring and second guessing every single scientist on the planet involved in astronomy.

Rather than 'dismiss the concerns of the entire professional astronomy community', I instead listen to the IAU and similar bodies and the outcomes of the SATCON workshops and similar as to the impacts of constellations and their the effectiveness of mitigations. If you're even bothering to care about system ownership, then you're in the realm of crappy tabloid rags rather than actual publications. Finding other sources is recommended.


*Most. There will be some observations for engineering purposes too.

[deadman1204]

Has anyone made a quantitative estimate of how much reflections from large constellations will cost astronomy in lost productivity and in mitigations? From the little I know it seems astronomers should be able to work around the satellite reflections with measures such as taking many short exposures and discarding the bad pixels in each frame, not using telescopes near dawn and dusk when they may see sun-lit satellites, or using more space-based telescopes. These mitigations could be expensive (especially if they have to use space telescopes) but probably not expensive enough to justify making internet significantly more expensive for millions of people.

As of today, its more than 10% of all observations are heavily impacted or thrown away. That number will only increase. I don't mean theres a streak in the corner. That 10% is of quite significant impact.

That depends on the proportion of observations that would otherwise make scientific contributions vs. those that are for other purposes or are rejected for other issues (e.g. aircraft overflights, streaks from other local objects, etc). For Starlink in particular (does not apply to Oneweb and other higher orbits) the illumination peak is during twilight, where astronomical observations are more opportunistic. If 10% of observations are impacted but those 10% impacted would have had a 90% rejection rate without satellite streaks due to background light, that's a big difference to impacting 10% of observations at midnight which would have a much higher baseline acceptance rate.

Huh? How do I unpack this. All the observations all the big telescopes make ARE for science. They are so over subscribed its not like they are just randomly pointing for fun half the time.

All* observations are intended for science. Not all will produce useful data. Clouds unfortunately still exist, sometimes atmospheric distortion is unacceptable, non-Starlink orbiting bodies can intrude in the frame, background illumination levels can be too high to observe your target, etc. Every* frame has a purpose, but not every one can satisfy that purpose.

However, we haven't discussed IR either, cause starlink blazes in those colors too. Yet not only is nothing being done about that, making them darker to visible light actually makes the IR problem vastly worse.

The dichroic coatings developed for Starlink are also effective in nIR and mid-IR, not just the visual spectrum. Even in thermal IR, emission is dependant on insolation just as with visual-range emission, so they glow in the same twilight conditions as in the visual range (where sky temperature makes IR observation even more difficult than normal). Observations of and calculations of Starlink impacts are also including IR effects, such as the ZTF studies.
 

I would ask instead why you are so keen to dismiss the concerns of the entire professional astronomy community. Its their job to know this stuff. Why not take what they say at face value? I would posit that if starlink was owned by bezos, no one would be ignoring and second guessing every single scientist on the planet involved in astronomy.

Rather than 'dismiss the concerns of the entire professional astronomy community', I instead listen to the IAU and similar bodies and the outcomes of the SATCON workshops and similar as to the impacts of constellations and their the effectiveness of mitigations. If you're even bothering to care about system ownership, then you're in the realm of crappy tabloid rags rather than actual publications. Finding other sources is recommended.


*Most. There will be some observations for engineering purposes too.

Wow. Who decides what is "useful data". You? 

[Asteroza]

If the claim that terrestrial astronomy is irrevocably broken with the advent of megaconstellations is going to be pursued, then a commitment to a permanent tax by FCC equivalents on spacecraft with substantial terrestrial impact (read any megaconstellation, SAR operators, large area spacecraft) to fund national space based astronomy projects, and permanently abandon any terrestrial project not in active use or under construction has to be made (under the assumption that they will be retired in exchange for replacement space based assets).

You can't say terrestrial astronomy is screwed and still try to ask for terrestrial astronomy handouts. Give up and go orbital.

There might be a case for VLBI astronomy on the ground due to physical limitations, but all the big mirror/big dish telescopes are a lost cause. Big RF dish radiotelescopes are the basic existence proof for space structure manufacturing (truss and mesh stringing work for a spiderfab/archinaut). JWST basically proves that segmented mass production mirrors with AO for space telescopes are a thing. Cranking out sub-2m mirror segments for mass produced space telescopes operating as part of an international timeshare constellation that universities sign up for is a logical thing.

[meekGee]

Has anyone made a quantitative estimate of how much reflections from large constellations will cost astronomy in lost productivity and in mitigations? From the little I know it seems astronomers should be able to work around the satellite reflections with measures such as taking many short exposures and discarding the bad pixels in each frame, not using telescopes near dawn and dusk when they may see sun-lit satellites, or using more space-based telescopes. These mitigations could be expensive (especially if they have to use space telescopes) but probably not expensive enough to justify making internet significantly more expensive for millions of people.

As of today, its more than 10% of all observations are heavily impacted or thrown away. That number will only increase. I don't mean theres a streak in the corner. That 10% is of quite significant impact.

That depends on the proportion of observations that would otherwise make scientific contributions vs. those that are for other purposes or are rejected for other issues (e.g. aircraft overflights, streaks from other local objects, etc). For Starlink in particular (does not apply to Oneweb and other higher orbits) the illumination peak is during twilight, where astronomical observations are more opportunistic. If 10% of observations are impacted but those 10% impacted would have had a 90% rejection rate without satellite streaks due to background light, that's a big difference to impacting 10% of observations at midnight which would ahve a much higher baseline acceptance rate.

Huh? How do I unpack this. All the observations all the big telescopes make ARE for science. They are so over subscribed its not like they are just randomly pointing for fun half the time.

Your "twilight" is multiple hours after sunset and before dawn. Musk fans hand wave it away, but it actually adds up to close to half the night. 2-3 hours after sunset, 2-3 hours before dawn? Thats 4-6 hours. Only half the night is being generous.
ALL observations are the best time to make them, because differnet things are above the horizon at different times. Many things may never get that high above the horizon. So you observe when its at the best position to. Midnight isn't a magic time. Its just the middle of the dark part of the day.  That lost data I mentioned is a greater than 10% loss of ALL Science observations. Yes, ALL the observations are scientific and matter. Your saying that observations more than 50% of the night (2-3 hours on each end) don't matter. Now we are talking about a 70% decrease in science.


Note, I'm saying starlink only here, because 99% of all impacts are starlink. Airline paths don't go over the major observatories, and most all satellites are starlink. However, we haven't discussed IR either, cause starlink blazes in those colors too. Yet not only is nothing being done about that, making them darker to visible light actually makes the IR problem vastly worse.

The RF issue has been known since before the first Starlink bird launched. SAR birds are a far bigger threat than downlink RF: downlink RF is much lower power and almost always beamformed to target users (why waste precious power budget), whereas RADAR is desigend for uniform coverage and has sufficient power not just to blind receivers but to physically damage them (burning out frontend amplifiers). The solution for both is the same though: turning off emitters over astronomy sites. This is the solution currently in use.

There are what, a few dozen SAR satellites up there? Well over a thousand starlinks? With many thousand more coming? Math seems to be hard


I would ask instead why you are so keen to dismiss the concerns of the entire professional astronomy community. Its their job to know this stuff. Why not take what they say at face value? I would posit that if starlink was owned by bezos, no one would be ignoring and second guessing every single scientist on the planet involved in astronomy.

Ok that makes sense - up to half the night, but weighted towards the direction of the sun, latitude dependent, carry the one, you quote 10%, and the argument would be the same if it were 5%, or 15%.

The position of all satellites is precisely known, and I expect professional observations avoid them when possible, so I deduce these are long exposure / wide angle shots where there was no choice.

In that case I think 10% is painful but acceptable (by general society, not by the astronomer with the ruined image) compared with the value gained.  You're operating at 90%.  As people pointed out, there are other factors that cause image loss and usually for reasons that don't have a flip side.

The bottom line is that astronomy is not ruined, there's just a difficulty to work around, and it seems to me SpaceX is proactively trying to minimize impact.