Losing the sky

Andy Lawrence, who happens to be a colleague, has just published a book called Losing the Sky. Andy also gave a brief presentation about it, which is what motivated me to write this post. The book is very reasonably priced and very easy to read. It’s about Starlink, the constellation of low Earth orbit satellites being launched by SpaceX. There are currently just over 1000 in orbit, with plans for 12000, and a possible extension to 42000. The goal is to provide high-speed internet with low latency.

Credt: Victoria Girgis/Lowell Observatory

As the image on the right illustrates, the issue is that (especially during the orbit raising phase) these satellites can be very prominent in astronomical images. Since there will be so many of them, this could have a very large impact. This is not only a problem for ground-based observations; even images taken with the Hubble Space Telescope have been impacted.

It’s also not only optical astronomy, radio astronomy may be even more severely impacted. Currently, most communication satellites are in geo-synchronous orbits. Consequently, radio observations can typically be planned to keep their transmissions out of the side-lobes. With this new constellation of low Earth orbit communication satellites, this may become essentially impossible, potentially ruining radio astronomy.

One concern with complaining about this, is that the stated goal is to provide internet to regions of the planet that don’t currently have decent access. This is clearly a worthy goal and so it can be tricky to object on the basis of how it will impact astronomical observations. There are, however, a few issues with this stated goal. One is that there are already solutions involving satellites on higher orbits, so it’s not clear that providing internet to under-served regions of the planet requires a constellation of low Earth orbit satellites. Also, the current price suggests that this may also currently be out of reach of many in these regions.

What seems more likely is that the motivation is to reduce the latency (the data transfer time) which will be very attractive to the financial sector. This will require a constellation of low Earth orbit satellites. So, the actual goal may not be quite as magnanimous as suggested.

As my colleague’s book suggests, this does seem to be another example of a tragedy of the commons. Some get to benefit from using the environment in a way that negatively impacts many others, who don’t get compensated for how they are impacted.

Even if we would benefit from high-speed, low-latency internet access across the globe, I do think there would still be merit to a process that assessess the impact of the proposed solution and that has some ability to influence, and potentially regulate, this kind of activity. We can’t keep ignoring the impact of how our activities influence the environment in which we all live, not only for fairness reasons, but because there is a cost to such activities that someone will eventually have to pay.

This entry was posted in Environmental change, Scientists and tagged , , , , . Bookmark the permalink.

34 Responses to Losing the sky

  1. There was one other issue that I didn’t touch on in the post. One is that the risk of collisions will go up with so many satellites in low Earth orbit. This increase the chance of undergoing a Kessler cascade in which the debris from collisions leads to further collisions and the process could essentially runaway, making it very difficult to use some regions of space.

  2. RickA says:

    Ultimately we will have to build telescopes on the moon (dark side) or in the LaGrange points of Earth/Moon to get away from these problems. Because I don’t think the satellites are going away. I wonder if we put two telescopes in Earth orbit on either side of the sun (90 degrees to each side of Earth I suppose), if the 186,000,000 mile base would allow for better distance measurements?

  3. Magma says:

    Despite my deep concerns reading about this reckless, purely for-profit use of LEO by private corporations largely left to their own devices, the impact really only hit home when I saw wide-field astronomy photographs scarred by satellite trails and last summer, watched a train of dozens of them pass overhead in the evening twilight.

    It’s one thing to watch the one and only ISS change from bright white to deep red as the sun sets on it a thousand kilometers away. It’s another thing to watch cheap space junk launched by billionaires trying to get even wealthier pollute the skies.

  4. Rick,
    Yes, we can consider space telescopes on the Moon, or in orbits that won’t be impacted by the satellites. However, that’s very expensive, so one could argue that some of the cost should be borne by those who are benefitting from the activities that have made this necessary.

    Also, we’re starting to build 40m telescopes on the ground. This is very large and it’s this size of telescope that will help us to solve some of the astronomical puzzles that we’re currently focusing on. The Hubble Space Telescope is 2.4m and the James Webb Space Telescope (which is going to cost – IIRC – $8 billion) is 6.5m. Observing from outside the atmosphere has advantages, but you still need a big mirror if you want the kind of resolution that we might need to solve some of these puzzles. We’re nowhere close to being able to put something like a 40m telescope into space, or on the Moon.

  5. Magma,

    It’s one thing to watch the one and only ISS change from bright white to deep red as the sun sets on it a thousand kilometers away. It’s another thing to watch cheap space junk launched by billionaires trying to get even wealthier pollute the skies.

    Indeed, and their claims that it is to bring the internet to under-served regions of the globe sounds noble but may well hide their real intention which could be more to provide rapid data transfer for financial interactions.

  6. jacksmith4tx says:

    RE: Space junk
    The Economist has a story on the looming challenge of cleaning up our space junk.
    “The problem of orbiting debris, and the concomitant risk of it colliding with and damaging an active and probably expensive satellite, has been around for a while. But it is rapidly getting worse. There may now be as many as 1m bits of debris measuring 1cm or more across in orbit. Of larger objects, more than 20,000 are being actively tracked from Earth. And, according to Daniel Oltrogge, an expert who advises the Space Data Association, an industry body, on such matters, the past three years have seen a doubling of the number of times that bits of junk have almost hit operating satellites.”

    The Chinese are working on using lasers to try and vaporize or deorbit debris. Of course our military is suspicious that it’s just cover for space weapons.

  7. Perhaps those operating telescopes can consult with some in the film industry about the joys of rotoscoping.

  8. jg says:

    Thanks for calling my attention to this book.
    I see SpaceX’s motives in light of how western US water rights were established, e.g., first use became a legal claim for water rights. I speculate that some day, SpaceX will try to claim exclusive use of orbits that they’ve occupied.

  9. Tom,
    They’re doing all sorts of things, including talking with SpaceX who are, apparently, being quite accomodating. However, if you have a 30 minute exposure on a large telescope and you get a number of satellite passes through your field of view, that could ruin the entire observations. Might depend on whether or not you’re going wide-field, or narrow-field, but it’s still likely to be a problem. These big telescopes cost $1000s per hour to run. It’s a very expensive loss.

    There is another issue that I didn’t highlight in the post. One thing we are doing is looking for near earth objects that could be hazards. This will get more and more challenging if there are lots of low Earth orbit objects passing through the fields of view.

  10. jack,
    Indeed, I believe it is expected to get much worse with the launch of all these satellites.

  11. wmconnolley says:

    It seems obvious that space-based telescopes are the way to go. Soon, the idea of putting a telescope under an atmosphere is going to seem ridiculous; probably within a decade. I’m surprised that people are long-term planning big ground telescopes; but that probably reflects the kind of people in charge of astronomy; just as the navy resisted submarines and air forces still have so many manned planes.

    > which will be very attractive to the financial sector

    Invoking the traditional bogeymen seems silly. Did you make this up, or does the book really suggest it? If so, does it pull out any figures? I think it is wrong because (a) Starlink latency is probably already too high for HFT; (b) financial people tend to be in cabled places and have no need for this stuff. There’s a more obvious market: gamers. But don’t take my word for it; just read https://www.starlink.com/ which says in big letters “Low latency = video calls & online gaming”.

    Incidentally, I’m curious: these things are only a problem when lit up, so at midnight looking up, there’s no problem, yes?

    > tragedy of the commons

    You speak as though they’re unregulated, free to fling up as many as they like. This isn’t true.

  12. WMC,
    I’m not convinced that you’re right about space-based being the way to go. I think JWST has had so many problems that there may even be a reluctance to fund the next generation of space-based large optical/infrared telescopes (smaller one with specific science goals are likely, but multi-metre optical/infrared ones are more of a challenge). There are plans for a JWST follow up (Luvoir) but noone expects it to launch anytime soon. We’re building multiple 30m-40m class telescopes on the ground in a decade, while a 6.5m space-based telescope has taken much longer.

    The book did indeed make the financial sector argument. Gamers was also mentioned. My understanding is that these satellites could reduce the latency even when compared to cables. As I understand it, the speed through an optical fibre is 1.5 times slower than through a vacuum. The idea with these satellites is (apparently) to communicate between satellites with lasers, so a vertical distance of a few hundred kilometres up to the satellite and then lasers between satellites, and then a few hundred kilometres down may indeed be faster than a ground-based optical fibre. I’m not sure that this is yet working though, and I got this from my colleague, so am trusting his description. Also, the point was more that maybe we should be cautious of simply accepting this because it’s doing a good thing by bringing the internet to the whole globe.

    Incidentally, I’m curious: these things are only a problem when lit up, so at midnight looking up, there’s no problem, yes?

    Yes, starlink may only be a big problem in the early evenings and morning when lit up, but this can still be a large chunk of the observing time. Also, there is apparently a plan for another satellite constellation that would be in a slighter higher orbit that would be problematic for a bigger fraction of the night.

    You speak as though they’re unregulated, free to fling up as many as they like. This isn’t true.

    Yes, I do realise that there is a form of regulation and that you can’t just fling as many up as you like.

  13. David B Benson says:

    Who regulates Starlink and the like? Launches are regulated for safety; not the same thing.

  14. David,
    I think there are also regulations related to communication, but my understanding is that there isn’t really much that considers the impact in the night sky.

  15. I share your concern about the tragedy of the commons, Ken. It is a shame that humans have geared up into an economic system that privatizes profit and socializes risk by assigning little or no economic value to various commons. I think we should think hard about how that system took hold, whether it needs to be changed, and how we would change it.

    All that said, the loss or damage of night sky availability for the purposes listed does strike me as a first world problem. I think we should think hard about how and when we decide to raise the tragedy of the commons so that our solution to that problem serves the largest number of human being on the planet. If we want to challenge the tragedy of the commons with regard to night sky availability, I think we are taking a position that is too narrow. It’s a simple and logical manner to expand this concern about night sky to include addressing the tragedy of the commons as it presents in loss of air quality between ground and night sky.

    All that said, I think this bell has been rung. The LEO flyers are up there and expanding in number. There are regulations in effect and it may turn out that the regulations and the regulators are insufficient to protect a resource that is a commons. I have seen that happen before.



  16. Willard says:

    > You speak as though they’re unregulated,

    From the horse’s mouth:

    The tragedy of the commons describes a situation in economic science when individual users, who have open access to a resource unhampered by shared social structures or formal rules that govern access and use, act independently according to their own self-interest and, contrary to the common good of all users, cause depletion of the resource through their uncoordinated action.


    “Unregulated” is stronger than unampered by regulations or other kinds of norms. By the “only no regulation counts” logic, any kind of mining would be fine as long as they abide by outdated and overly minimal mining laws. There’s a reason why regulatory systems are always playing catch with über-pwners:

    Ultimately, there is no easy way to design a regulatory system that can respond to evolving market forces. Yet any effort to do so is likely to yield enormous dividends, given the current political impasse associated with regulatory policymaking. At present, conservatives (understandably) oppose any new regulatory intervention for fear it will become irreversible. Progressives (also understandably) oppose any deregulatory effort for fear that many good regulations will get swept up with the bad ones. In theory, both sides should support a regulatory system designed to correct legitimate market failures and to readjust the level of intervention as the market evolves. Actually achieving that result, however, is likely to prove enormously complicated and will require policymakers to approach their task with a fresh perspective. Each of the options considered above holds some promise and faces certain drawbacks. Regulators must be willing to consider these and other innovative solutions if they are to construct a more dynamic regulatory system.


    Or how to almost say something in so many words.

  17. jacksmith4tx says:

    You think ground telescopes are having problems now just wait till the geoengineering gang gets their chance to blanket the planet with clouds of reflective particles.

  18. jack,
    That came up on Twitter a few days. Not that I’m a fan of geo-engineering, but I think the impact of that on ground-based astronomy will be negligible.

  19. wmconnolley says:

    > JWST

    It is just possible that you know more about this than I do, but my suspicion is that JWST suffers from problems like the SLS. As commercial launch costs continue to fall large space-based stuff is going to get ever easier.

  20. kevinboyce says:

    Do big telescopes still integrate in the detector for minutes at a time? The WFE correction system at least has to have a bandwidth better than a Hz, right? Don’t you need full frames at twice the bandwidth to do full frame WFE correction? And if so, I would think you just drop a few frames when a satellite is there. Or is the science channel a whole separate detector with a much slower readout rate?

  21. WMC,
    It is certainly possible that it could get cheaper, and we’re certainly launching a number of smaller telescope. One big issue – I think – is how you can launch something with a mirror much bigger than JWST’s 6.5m mirror. Also, with adaptive optics, many ground-based telescopes can perform almost as well as space-based ones. There are still advantages to being in space, but I think the cost of a 40m ground-based telescope (~$1 billion) is still cheaper than a space-based 6m telescope (~ $8 billion).

  22. kevin,
    I don’t know the answer to your question.

  23. kevinboyce says:

    Follow-on to my last comment, since I notice you mention LUVOIR: LUVOIR plans to read out full frames from the (well, *a*) science detector at somewhere around 1 Hz, then recover the WFE and actuate the deformable mirror based on that data. My recollection is that they were basing the system design on what is currently being done on the ground.

  24. kevin,
    Okay, I think I see what you’re asking. I’m not sure, but I would expect LUVOIR to be similar to what’s done on the ground. However, I think the wavefront error correction on LUVOIR is more to do with correcting coronographic noise, than correcting for atmospheric noise.

  25. Interesting about hazards of low orbit satellites. How about high orbit satellites. What I have in mind are large solar energy collectors that could continuously collect solar energy from high orbits where they’re not blocked by the Earth’s shadow and can either reflect the energy to a concentrated point or beam it to a point, perhaps a remote desert location. This looks to me like a way for solar to solve its intermittency problem.

  26. kevinboyce says:

    Okay, maybe optical instruments are different (I know the Advanced Camera for Surveys on Hubble (which isn’t so advanced any more!) takes 104 seconds just to read out a CCD frame), but a quick check shows that IR ones seem to use fairly short integration times. The Keck NIRC2 looks to integrate for 10-60 seconds [W. Jerry Xuan et al 2018 AJ 156, 156], and NIRI on the Gemini telescopes, using the ALADDIN-III ROIC, meets spec at 20 Hz frame rate [Klaus W. Hodapp et al 2003 PASP 115, 1388].

    So maybe it’s not so bad as it seems, at least as far as telescopes with adaptive optics are concerned.

  27. kevinboyce says:

    Regarding LUVOIR, since I know a (small) bit about that. Wavefront error is wavefront error, whether it comes from the telescope wobbling around or the atmosphere jiggling. The coronagraph comes in because they’re trying to get cancellation at the 10E-10 level, so obviously even a tiny fraction of a wave change in the WFE over the entire pupil blows your coronagraph away. So the regime is somewhat different in that they need much tighter control, but the sources of WFE are correspondingly smaller. The end result is the same calculations, but maybe with additional precision (I don’t recall). I do know that doing those calculations fast enough, without spending an inordinate amount of power on the processor, is a fun engineering challenge. Probably made boring and simple by the development of new FPGA tech between now and when it actually gets funded.

  28. kevin,
    Thanks. You certainly know more about it than I do.

  29. Dan Moutal says:

    Latency of traditional satellite internet is high enough to cause serious issues for many modern day applications. The improvements in latency of starlink would make a meaningful difference for users who don’t have the luxury of traditional boradband (Cable, DSL, fibre, LTE…).

    As for the applications for the financial sector, I think the latency is still much higher than is needed for high frequency trading. And while in some circumstances (like transatlantic communications) the latency on starling has the potential to be faster than undersea cables (the signal travels faster in the upper atmosphere than it does through a fibre optic cable), these situations are still far too slow for high frequency trading. And to complicate things the latency of individual packets is expected to be quite variable (as the path between satellites changes constantly).

  30. Dan,
    Yes, I guess there is quite a big difference between a signal that has to travel up to a satellite at 8000km (which is where I think the current internet satellites are) compared to one at 300-400km.

  31. Dan Moutal says:

    When it comes to modern telecommunications the speed of light is just too slow:)

  32. “When it comes to modern telecommunications the speed of light is just too slow:)” That’s so true. Same thing is happening when I go out on the highways. I speed, of course, but I speed too slowly to fit in the normal traffic patterns these days. Sometimes, I am only skooting along at something very close to the speed limit. How crazy is that?

  33. ” the stated goal is to provide internet to regions of the planet that don’t currently have decent access. This is clearly a worthy goal and so it can be tricky to object on the basis of how it will impact astronomical observations.”

    Alphabet /Google could take a bite out of its Evil Index by restoring funding for its Loon Project which examined stratospheric balloons as an alternative to satellite proliferation in extending internet coverage to remote regions.

  34. Geri Lawhon says:

    Thanks for the information.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.