Can telescopes in space replace ground based telescopes?

[hop]

For what reason is the value of the observing time different? Your second statement has nothing to do with the seasons, and is the kind of thing that show be supported by math and an explanation of what you consider "relatively high elevation"

This is astronomy 101. Higher elevations, lower airmass, darker sky. To what extent this matters depends on the details, but in general, you can go deeper/get better data if you're looking straight up in full darkness. Hence, people trot out the "you only see starlink at twilight and they're mostly in shadow" argument to claim (incorrectly) that the impact on real astronomy is negligible.

This is doubly wrong: Some observations necessarily have to look at low elevations near twilight, and satellites in LEO orbits can be seen quite high and late.

there is no reason to think that they have a detailed analysis which supports an assertion they don't make to begin with.

They have done the analysis and made the statement:
https://www.aura-astronomy.org/news/aura-statement-on-the-starlink-constellation-of-satellites/

The LSST Project Science Team has been simulating the potential impacts to LSST observations. Their latest update of preliminary results from November 2019 indicates that (assuming the full deployment of planned satellites) nearly every exposure within two hours of sunset or sunrise would have a satellite streak.

(my emphasis)

I've seen first hand statements from LSST people that they are very concerned and have done significant analysis. They are modifying their observing strategy and even the camera readout to try to mitigate it. The observing strategy is something that's been subject to years of debate and optimization, so it's likely that science will suffer, even if they manage to meet the baseline goals. It also depends a lot on how many of the constellations get built, what orbits they end up on, and how bright they end up being.

This thread has some details from the actual presentation at AAS:
https://twitter.com/Thomas_Connor/status/1215010589791019009

I also know first hand that other surveys like Super-ASASSN are significantly affected. NEO searches are also likely to be heavily affected, since they necessarily need to look relatively close to the sun. Some statements from Pan-STARRS
https://www.mauinews.com/news/local-news/2019/06/more-satellites-in-space-streak-mar-telescope-images/

Another is you arguing against the point of moving all ground spaced astronomy to space in the near term, which again would just be you trying to force an idiotic strawman argument into my mouth.

That "strawman" is literally the subject of the thread "Can telescopes in space replace ground based telescopes? ", and common refrain of the people who claim Starlink won't be a problem, including Musk himself in his initial reaction.

edit:
I recognize you aren't making the argument that we can just put everything in space. I think your assumptions about how cheap lift will affect the cost of doing astronomy in space are extremely optimistic, but I'll freely admit the uncertainties are large.

[Lar]

My take: The answer to the question posed by the OP? No. Space telescopes cannot completely replace ground ones (and vice versa). BUT SpaceX could easily give astronomy and astronomers access to new telescopes that make up for the loss and even exceed current capabilities for not a lot of coin.

[Coastal Ron]

We may need someone to come up with a new generation of low cost space telescopes in order to leverage the lower launch costs that the SpaceX Starship provides in order to fully replace Earth-based observation platforms.

But I think the success of the (now) 30 year old Hubble Space Telescope has shown us what a small 2.4m diameter telescope can do in space. Imagine if we could build lower cost 8m diameter telescopes and do final assembly in space, then move them to regions of space that provide unobstructed views?

In the Project Management Triangle you have the option of Good, Fast, and Cheap, but you can only have two of them (at most). The Hubble Space Telescope took 20 years from proposal to launch, and the James Webb Space Telescope will have taken 25 years if it launches in 2021.

It is the cost of the telescopes, and the speed at which they are built, that will determine if space telescopes can leverage the significantly lower cost of transportation that New Glenn and Starship will be able to provide.

We will need an Elon Musk for space telescopes... 

[Zed_Noir]

<snip>
It is the cost of the telescopes, and the speed at which they are built, that will determine if space telescopes can leverage the significantly lower cost of transportation that New Glenn and Starship will be able to provide.

We will need an Elon Musk for space telescopes... 

Please stop giving the SX CTO more ideas for shiny objects.

[Eka]

Something that fundamentally needs to be acknowledged is near Earth space will increasingly get more and more crowded over the next few decades. SpaceX's Starship, with it's low cost to orbit, will help make sure this starts to happen. Others will copy and follow in it's wake. Starlink is just the start.

<snip>
It is the cost of the telescopes, and the speed at which they are built, that will determine if space telescopes can leverage the significantly lower cost of transportation that New Glenn and Starship will be able to provide.

We will need an Elon Musk for space telescopes... 

Please stop giving the SX CTO more ideas for shiny objects.

He could be a great resource for setting up a space telescope manufacturing facility. One of the big issues I see with all space telescopes so far is everything is a one off part. They need to be built in batches to get the cost per telescope down. Making 10 or 20 of a part is often far cheaper per part than making one. Once the design is done and the machinery is setup for the first, why not make a few more copies and save all that design and setup costs.

Now if you have a new telescope going up every few months, adding some new instrument design isn't a long wait. It becomes how long does it take to design, build it, and how long the queue for instrument slots is.

Incremental changes to the design can be done. Major changes will need to wait for the next series.

Service will also be cheaper because of the lower cost to orbit, plus "ride share". Have an old space telescope that needs cleaning and updating, or has broken down. Send the next new one up in it's orbit, and have the delivery Starship grab the old one and bring it home. It can then be cleaned, refurbished, and have new instruments mounted, and be sent back up.

[envy887]

This thread has some details from the actual presentation at AAS:
https://twitter.com/Thomas_Connor/status/1215010589791019009

Jonathan McDowell posted the slide deck for that presentation on his website:

https://planet4589.org/space/misc/Seitzer.pdf

[hop]

My take: The answer to the question posed by the OP? No. Space telescopes cannot completely replace ground ones (and vice versa). BUT SpaceX could easily give astronomy and astronomers access to new telescopes that make up for the loss and even exceed current capabilities for not a lot of coin.

Yes, there's no doubt that cheap, routine access to space would be a huge boon to astronomy, and if the optimistic scenarios pan out, ultimately do more to advance the science than anything lost to the LEO constellations. However, it should be clear that this isn't a mitigation for impacts on programs like LSST that will be operating over the next decade.

One other thing to note, just being in space isn't enough to avoid the impacts. Starlinks are above ISS, and some of the proposed constellations would be well above Hubble. Satellite trails interfering Hubble observations is already a thing that happens. Of course, astronomy satellites can be in higher orbits, or L2 etc, but that brings it's own complications.

[meberbs]

For what reason is the value of the observing time different? Your second statement has nothing to do with the seasons, and is the kind of thing that show be supported by math and an explanation of what you consider "relatively high elevation"

This is astronomy 101. Higher elevations, lower airmass, darker sky. To what extent this matters depends on the details, but in general, you can go deeper/get better data if you're looking straight up in full darkness. Hence, people trot out the "you only see starlink at twilight and they're mostly in shadow" argument to claim (incorrectly) that the impact on real astronomy is negligible.

You apparently did not understand my question at all, you had asserted that the value of observing time is different depending on season to cancel out that impacts from seasons where there is more of a problem or less of a problem cancels out. Most of what you just said is simply irrelevant.

This is doubly wrong: Some observations necessarily have to look at low elevations near twilight, and satellites in LEO orbits can be seen quite high and late.

"Some observations" my next question should be obvious: what specific observations? (Check back in this thread, I already addressed the only one I have seen mentioned specifically)

there is no reason to think that they have a detailed analysis which supports an assertion they don't make to begin with.

They have done the analysis and made the statement:
https://www.aura-astronomy.org/news/aura-statement-on-the-starlink-constellation-of-satellites/

Thank you for providing this, I have been asking for an actual analysis with numbers since the first post I wrote in this thread.

I would be significantly more grateful if you didn't make me ask for it 3 times before actually supporting any of your assertions with facts.

I have significant questions about what they used as inputs to the analysis, because "nearly every" does not seem like a plausible result. At the least this probably ignores the mitigations mentioned in your previously provided links. The statement you quoted does not clarify how much of the 2 hours is actually useful to begin with, and the next sentence almost seems to contradict it with statements about 40% (or less) of twilight observing time impacted. The statements about saturation in the link in particular are questionable since Starlink sats in operation should be +4 to +7 magnitude There are thousands of stars in the night sky in this range, with half visible at any time (unlike the incorrect claims about Starlink satellites where relatively little of the constellation has line of sight at any given moment.) The motion would further increase the brightness required to cause saturation.

I've seen first hand statements from LSST people that they are very concerned and have done significant analysis. They are modifying their observing strategy and even the camera readout to try to mitigate it. The observing strategy is something that's been subject to years of debate and optimization, so it's likely that science will suffer, even if they manage to meet the baseline goals. It also depends a lot on how many of the constellations get built, what orbits they end up on, and how bright they end up being.

So in other words, while no one ever claimed the impact would be 0, there is currently not enough basis to be certain that the impacts would not be negligible.

I also know first hand that other surveys like Super-ASASSN are significantly affected. NEO searches are also likely to be heavily affected, since they necessarily need to look relatively close to the sun. Some statements from Pan-STARRS
https://www.mauinews.com/news/local-news/2019/06/more-satellites-in-space-streak-mar-telescope-images/

I mentioned NEOs already, if you have a comment on them you can look at my previous post first. Your comment about Super-ASASSN being "significantly affected" is simply worthless at this point as you have not provided any definition for significant. We are well past the point in the conversation where you should have any illusion that I will accept vague assertions without data.

Another is you arguing against the point of moving all ground spaced astronomy to space in the near term, which again would just be you trying to force an idiotic strawman argument into my mouth.

That "strawman" is literally the subject of the thread "Can telescopes in space replace ground based telescopes? ", and common refrain of the people who claim Starlink won't be a problem, including Musk himself in his initial reaction.

The OP says nothing about it being done in the near term, and clearly clarified with questions about which types of research are able to be done on the ground versus in space. A partial answer to the questions is that there is no research that needs to be done on the ground, although for now there is a significant amount that is cheaper to do on the ground. (The inverse is not true as there are some wavelengths essentially unusable from the ground.) A fairly obvious corollary to the OP's questions that was quickly brought up is what it would take in the near term to offset near term impacts from satellite constellations. The first step to answering that requires determining how significant or not the impacts are to begin with.

I recognize you aren't making the argument that we can just put everything in space. I think your assumptions about how cheap lift will affect the cost of doing astronomy in space are extremely optimistic, but I'll freely admit the uncertainties are large.

And here I agree with you in a sense. There is adequate room for disagreement on how much Starship may lower in-space astronomy costs, as the data simply does not exist yet to be certain. However a 10x factor compared to a ground based telescope is what you would expect in today's market, not a future one with cheap launch as assumed by the OP. personally I think that pretty much no one is prepared for the impacts of Starship, and if it meets half of what it is supposed to in double the time frame, somewhere between few and no one will be ready to fully leverage it.

[redliox]

If you could rebuild ground telescopes on the Moon, the lunar versus terrestrial telescopes would give science less obstructed by an atmosphere.  The advantage terrestrial telescopes have, and likely will have for decades if not centuries to come, is all the staff and replacement parts are down here on Earth.  It's a superior view versus superior support is what it comes down to.

Personally, I believe sooner rather than later a time's coming where ground telescopes will produce less science than space ones (lunar or orbital-based).  However, there are still numerous "tricks" ground 'scopes still have that'll make them worthwhile for a decent while yet.  The shift, when it happens, likely will come with the higher wavelengths first (gamma to UV) since those are the most disturbed by the atmosphere while you can still gain some optical, infrared, and radio science on the ground.

[Rebel44]

If you could rebuild ground telescopes on the Moon, the lunar versus terrestrial telescopes would give science less obstructed by an atmosphere.  The advantage terrestrial telescopes have, and likely will have for decades if not centuries to come, is all the staff and replacement parts are down here on Earth.  It's a superior view versus superior support is what it comes down to.

IMO, it would massively help the cost as well as reliability if we stopped building telescope parts that are unique for every telescope and instead switch to standardized modular construction. Cheap access to space and modular architecture would also make replacing faulty parts much easier/cheaper (so we wouldn't need to spend so much time and money testing and checking every part).

[hop]

You apparently did not understand my question at all, you had asserted that the value of observing time is different depending on season to cancel out that impacts from seasons where there is more of a problem or less of a problem cancels out.

I think you misunderstood the answer. I didn't claim that the value of the observing time changed with the season. How long the satellites are visible after twilight changes with the season (and observatory latitude). Since the dark time is more useful for astronomy, the seasonal effects don't cancel out. But this is a minor point. The main point is that common refrain of "you only see them around twilight" is simply not accurate as a general statement.

See Patrick Seitzer's presentation that envy887 posted for how this works out near the LSST site. Higher latitudes suffer more.

"Some observations" my next question should be obvious: what specific observations?

I'm not going to review the whole field for you. It's a simple, obvious fact that people take data when they think they can get useful data, and this varies widely depending on the instrument and target. People trying to cover the whole sky use as much of the night as they practically can. People observing dynamic phenomena don't get to choose when their targets are observable. 2I/Borisov is a recent high-profile example where the immediate post-discovery observations pushed the limits.

I have significant questions about what they used as inputs to the analysis, because "nearly every" does not seem like a plausible result. At the least this probably ignores the mitigations mentioned in your previously provided links. The statement you quoted does not clarify how much of the 2 hours is actually useful to begin with, and the next sentence almost seems to contradict it with statements about 40% (or less) of twilight observing time impacted. The statements about saturation in the link in particular are questionable since Starlink sats in operation should be +4 to +7 magnitude There are thousands of stars in the night sky in this range, with half visible at any time (unlike the incorrect claims about Starlink satellites where relatively little of the constellation has line of sight at any given moment.) The motion would further increase the brightness required to cause saturation.

If you think your off-the-cuff reaction is more likely to be right than analysis by people responsible for one the flagship astronomy programs of the decade, you're welcome to that opinion. I would however suggest the LSST team are aware of the existence of bright stars and the fact satellites move, so you might want step back and think about which analysis is more likely to be lacking sufficient detail.

As I mentioned earlier, years have been spent debating, simulating and optimizing LSST observing strategy. It's a topic the community takes seriously and is well equipped to analyze. See https://github.com/LSSTScienceCollaborations/ObservingStrategy for starters. Many more papers can be found on arxiv.

I'm honestly baffled by your reaction here. You don't have to follow astronomy very closely to know that many professional astronomers are deeply concerned by the impacts of these mega-constellations. It isn't news or controversial among people who actually do this stuff.

[meberbs]

I think you misunderstood the answer. I didn't claim that the value of the observing time changed with the season. How long the satellites are visible after twilight changes with the season (and observatory latitude). Since the dark time is more useful for astronomy, the seasonal effects don't cancel out. But this is a minor point. The main point is that common refrain of "you only see them around twilight" is simply not accurate as a general statement.

See Patrick Seitzer's presentation that envy887 posted for how this works out near the LSST site. Higher latitudes suffer more.

Here is what I originally had to say on seasons:

When you think about the geometry, you need to include inclination and seasons.

If you think about the seasons, you see that it is better for half the year and worse for half the year. This doesn't really change anything.

At this point everything you just stated supports my point. Your counterargument that "value of the time at the extremes isn't equal" is something you still have failed to explain. And again, you jump back to making it sound like I am trying to defend a different point which is both wrong and not something I had stated.

I'm not going to review the whole field for you. It's a simple, obvious fact that people take data when they think they can get useful data, and this varies widely depending on the instrument and target. People trying to cover the whole sky use as much of the night as they practically can. People observing dynamic phenomena don't get to choose when their targets are observable. 2I/Borisov is a recent high-profile example where the immediate post-discovery observations pushed the limits.

Astronomy is a field of science. Quantifying things like the impacts of  is what is expected. Anecdotes are less important than hard data and specifics. If the impacts are really as bad as certain social and news media seem to imply, there should be scientific papers documenting that.

I have significant questions about what they used as inputs to the analysis...

If you think your off-the-cuff reaction is more likely to be right than analysis by people responsible for one the flagship astronomy programs of the decade, you're welcome to that opinion. I would however suggest the LSST team are aware of the existence of bright stars and the fact satellites move, so you might want step back and think about which analysis is more likely to be lacking sufficient detail.

I specifically said that I questioned the assumptions they used as their inputs. They did not specify things like exactly how many satellites in what orbital planes, or assumptions about brightness. They did not even give results about whether saturation would occur. I also pointed out that 2 of the stats they gave seem contradictory at a glance. Right now their analysis has not been provided in the detail to judge these things.

You mentioned Patrick Seitzer's analysis above. I have looked at that, and it shows the 550 km altitude only impacts for about 1 hour after twilight, and even multiplied by 10 for number visible for a final state, does not seem consistent with the claims on the LSST site you linked. The LSST claims are based on "full deployment" but the meaning of that is not clear, did they use the 10s of thousands that SpaceX filed with the ITU with no defined orbital parameters? Most of SpaceX's future plans with defined orbits are for VLEO satellites even lower than current deployments. LSST's release states 200 satellites visible at any one time. Seitzer's worst case with 1500 sats at 1150 km shows a bit over 20 visible near twilight, multiply number of sats by 10, and you would get around 200 only for parts of the night. This is both an exaggeration of altitude and an assumption of more satellites than SpaceX has explicit plans for (should be closer to 11k than 15k).

I have expressed no opinion related to my off the cuff reaction being more valid than their analysis, I have just requested more detail on their inputs, since at first glance it does not seem consistent with other analysis that does specify exactly what orbits it is assuming.

As I mentioned earlier, years have been spent debating, simulating and optimizing LSST observing strategy. It's a topic the community takes seriously and is well equipped to analyze. See https://github.com/LSSTScienceCollaborations/ObservingStrategy for starters. Many more papers can be found on arxiv.

The only thing this says for sure is that it will take analysis from experts to actually determine the significance of impacts.

I'm honestly baffled by your reaction here. You don't have to follow astronomy very closely to know that many professional astronomers are deeply concerned by the impacts of these mega-constellations. It isn't news or controversial among people who actually do this stuff.

I am baffled by your hostile reactions to simple requests for data. You don't have to follow this very closely to see that plenty of people are making a lot of noise without data to back it up. (For example, check this post where someone realized they had been supporting a significant misrepresentation of the impact of Starlink on the night sky.)

[envy887]

You apparently did not understand my question at all, you had asserted that the value of observing time is different depending on season to cancel out that impacts from seasons where there is more of a problem or less of a problem cancels out.

I think you misunderstood the answer. I didn't claim that the value of the observing time changed with the season. How long the satellites are visible after twilight changes with the season (and observatory latitude). Since the dark time is more useful for astronomy, the seasonal effects don't cancel out. But this is a minor point. The main point is that common refrain of "you only see them around twilight" is simply not accurate as a general statement.

See Patrick Seitzer's presentation that envy887 posted for how this works out near the LSST site. Higher latitudes suffer more.

"you only see them around twilight" is pretty much exactly what Seitzer's presentation was showing. In general, the number of satellites visible at a given time before twilight is double the number visible the same time after twilight, and within ~1 hour after evening twilight, and up to 1 hour before morning twilight, that number generally goes to zero. At least at CTIO-ish latitudes.

This behavior is independent of the total number of satellites, but does strongly depend on their orbital altitude, with lower satellites being much brighter, but also much quicker to lose illumination around twilight.

Unfortunately for inner solar system observations, the last satellites to lose illumination in the evening, and the first ones to re-illuminate in morning, are probably near the the ecliptic and low to the horizon, which is also where those observations tend to need clear skies.

[hop]

"you only see them around twilight" is pretty much exactly what Seitzer's presentation was showing. In general, the number of satellites visible at a given time before twilight is double the number visible the same time after twilight, and within ~1 hour after evening twilight, and up to 1 hour before morning twilight, that number generally goes to zero. At least at CTIO-ish latitudes.

Note the red bars are astronomical twilight (which is well past the everyday use of the term), meaning everything in between would normally be prime observing time. So yeah, for a broad value of "around" it's true (edit: at ~30 deg latitude), but it does not mean the impact is negligible. In practical terms it means a couple hours per night are affected.

Things are much worse at higher latitudes, as the map on that wiki page should make clear.

[envy887]

"you only see them around twilight" is pretty much exactly what Seitzer's presentation was showing. In general, the number of satellites visible at a given time before twilight is double the number visible the same time after twilight, and within ~1 hour after evening twilight, and up to 1 hour before morning twilight, that number generally goes to zero. At least at CTIO-ish latitudes.

Note the red bars are astronomical twilight (which is well past the everyday use of the term), meaning everything in between would normally be prime observing time. So yeah, for a broad value of "around" it's true (edit: at ~30 deg latitude), but it does not mean the impact is negligible. In practical terms it means a couple hours per night are affected.

Things are much worse at higher latitudes, as the map on that wiki page should make clear.

The impact on observing time near twilight is not negligible, but the impact for the large majority of the night is zero.

CTIO is at 31 degrees. All of the world's 6+ meter optical telescopes and 90% of the 3.5+ meter optical telescopes are within 35 degrees of the equator, largely for the same reasons that Starlinks will be less visible there... twilight is shorter, and summer nights are longer.

[jbenton]

We will need an Elon Musk for space telescopes... 

There have been some intriguing ideas for decreasing the costs of space telescopes coming down the pipeline.

1) NASA and a company called Lightweight Telescopes, Inc. started this project:
https://www.nasa.gov/feature/goddard/2016/nasa-eyes-first-ever-carbon-nanotube-mirrors-for-cubesat-telescope
They've developed a way to impregnate an epoxy resin with carbon nanotubes, pour it into a mold for the correct shape, and then put a thin metallic film over it (after it hardens). This process is much faster and cheaper than the conventional method and allows for easy mass production.
The current prototypes are only 3 inches (~7.62cm) across and are intended for CubeSats, though apparently they could be scaled up to larger 'scopes, eventually. The researchers believe that this technology will be especially useful for segmented mirrors. Given what I've read in the article, it seems that these would be pretty useful for Ground-based observatories as well.

Of course it's not just the mirrors. The mirror for WFIRST was donated to NASA free of charge, and the spacecraft will still cost >= $3.2 billion - not including launch costs. (Though the mirror itself is estimated to be worth $250 million - which is over $1 million per cm in diameter). Building large sats in general is really expensive. The new carbon nanotube mirrors are lighter than their predecessors, but they still require significant bulk to carry.

2) "Orbiting Rainbows":
https://www.nasa.gov/jpl/tech/glitter-cloud-may-serve-as-space-mirror
This idea is to use lasers to maneuver strips of metal into a telescope mirror like formation. The resulting picture is very "noisy" but according to tests, existing computer technology is sufficient to clean it up. Theoretically a spacecraft could fly that is smaller than the mirror it uses with this technique. Radio telescopes are easier to make this way than visual light ones. However according to this article:
https://www.npr.org/2014/12/23/372468028/could-glitter-help-solve-nasas-giant-telescope-problem
this project is decades away from producing anything on a large scale.

[OTV Booster]

No, unfortunately meberbs that entire thread is two or three people arguing the same points back and forth. It's mostly quieted down now, so it's safe to skim it, but there hasn't been a whole lot more math than pointing out that the math used to fearmonger was kind of bogus.

Thanks, that is what I thought, I had looked at the thread some. (And I admit to being one of the 2 or 3 people in certain other threads (usually new physics) that ended up similar, so I get how that happens.)

It has been long enough that if a specific astronomy mission was going to have major irrevocable impacts, someone probably would have quantified them by now, and knowing this site, someone would find that quickly. In my mind, the lack of such a report puts a loose bound on how severe the impacts can get. I expect one will eventually get published one way or another to guide future missions.

Working my way up thread so apologies if I hammer a nail already driven.

Let’s look at what would go into a calculation starting with only one sat of some arbitrary angular size and one telescope with some arbitrary angular resolution. WARNING: my math sucks but I do have a reasonably good grasp of what the calculations need to represent.

If the sat angular dimension is less than the scopes resolution, and the sat is ‘dark’ there is no problem - sometimes. If the sat occultates (passes in front of) a star under observation the light curve will change and trigger questions. If the sat is bright it’s a different set of problems.

Some observations are extended objects such as nebula and galaxies. Bright and dim sats have effects roughly the same as above.

Then there is the question of the chances of this happening. Here we move away from one sat and one telescope. How many sats and how many scopes? It’s not just starlinks at one angular size. Many different sats of different sizes and altitudes and many many more coming. Many scopes of different capabilities. Some nailing one target at a time some doing survey work.

In short, a deep and hairy question. Any calculation would have so many assumption built in it would only be an easily demolished target for anybody who didn't like the answer.

So, the only answer I can come up with is tomorrow there will be ‘some’ impact on astronomy and the day after tomorrow there will be ‘more.’

I am a life long amateur astronomer. Amateur and professional both bemoan light pollution. Professionals bemoan RF pollution. Seismologist bemoan things that rumble. Things aren’t the way they used to be...

It will piss off a lot of people. People will adapt. They are currently building the third generation of scopes within my lifetime. One reason they are so expensive is the few locations with both no light pollution and acceptable atmospherics. By definition the are built out in the boonies.

The generation currently abuilding will probably have truncated lifetimes and a higher operating cost:discovery ratio than desired in part due to having to work space traffic into the observing runs. As the cost of operations goes up and the cost of space flight goes down, the following generation of scopes will move to the new boonies: space.

It won’t be a forklift move. Onesie twosie.  Forty years down the road astronomers will nostalgically admire all the old hardware (I still get a rush from an old scope with a brass tube) but shake their heads at even thinking of using it for productive work.

Phil

Edit to add: wow, talking about a nail already driven.  Children, behave!

[envy887]

Suppose, hypothetically, we wanted to replace LSST's wide-fast-deep capabilities with an in-space constellation, and that launch costs are in the $100/kg range?

What would this constellation look like? What is the cost-optimal:

1) size for each observatory?

2) number of satellites in the constellation?

3) orbit for surveying (considering accessibility, viewing area/time, and data return)?

With a constellation, it should be possible in theory to trade light collecting area for longer exposures, making up the the lost speed by having more scopes with smaller fields of view to get a fast revisit rate.

[OTV Booster]

Suppose, hypothetically, we wanted to replace LSST's wide-fast-deep capabilities with an in-space constellation, and that launch costs are in the $100/kg range?

What would this constellation look like? What is the cost-optimal:

1) size for each observatory?

2) number of satellites in the constellation?

3) orbit for surveying (considering accessibility, viewing area/time, and data return)?

With a constellation, it should be possible in theory to trade light collecting area for longer exposures, making up the the lost speed by having more scopes with smaller fields of view to get a fast revisit rate.

Classic answer: it depends.

I did a long windy reply and on reread I realized I’d not quite hit your question. But because I’m as much in love with my own word as the next guy I’ll leave it be.

On a re reread I think I did answer it but from my own weird angle.

Here goes:

First question is where would you put it?

LEO has thermal issues, a planet in the way half the time and perturbations from the moon and masscons. E-M L1 2 &3 are better but still are less than ideal. L 4&5 have some promise if there isn’t too much junk already there.

E-sun Lagrange points might be a very stable position but my guess is sun-Jupiter L1 could be really good. I’m not sure exactly how far in that is. If it gets too close to Mars that might be a deal breaker.

With interferometry two mirrors of any size spaced such that their outer edges are as far apart as a large mirrors diameter will have the same resolution as the large collector but not the light gathering ability.

Let’s look at this for S-J L1. Assuming Mars isn’t a problem there is a strong advantage in cryo cooling working very well. Especially if observations are always far from the sun. A combo PV/sunshade will get the temp down real nice. Next there will be minimal differential thermal effects inducing vibration.

A light weight CF truss could hold the sub mirrors at 8.4m (LSST) or 10m or ... 

One thing I am not clear on. More mirrors can be added to the truss to increase light gathering. I do not know if they must be part of the interferometry setup or if they can add useful data without. Either way is doable. Just a question of complexity.

The truss can be simple, a large ‘X’ or even a pie pan covered with mirrors.

And yes, because this is, at least in theory, super stable, very long exposures will go deeper on point sources. It’s a little trickier on extended objects. Planets, exoplanets, resolvable stars and nebula will behave differently - I think.

The problem here is that if the gathering surface (a couple small mirrors) is small compared to the aperture (distance between them) not only is the image dimmer but the contrast is also lower for extended objects.  This is argument for filling in with a lot of mirrors.

A note on extended objects and contrast.  This is my semi informed take on it. It is what I understand from years of reading but I’ve never focused on it and could be wrong. I’ll probably consult the oracle later and put out a Yup or Mia Culpa.

If I’ve got all this right what it boils down to is, just as all the engineering/customer desires/money trade offs that went into LSST resulted in a particular design and set of capabilities, doing it again in a space environment will not result in an exact duplicate. If LSST capabilities are a target you’ll either get something better in some way and worse in others, or you’ll spend much more for incremental gains.

Phil

[OTV Booster]

While I was doing the pervious post I remembered a space telescope design I noodled a long while back.

Take a ring of arbitrary diameter and a few cm thickness. Glue a sheet of Mylar or other suitable material to either side. Mildly pressurize the then enclosed space.

Edit to ad: do this in space.

The two sheets will bulge into catenaries. Spray the backs with something that will set up hard. Aluminize the interior surfaces and you are the proud owner of two fine mirrors.

An amateur telescope maker rule of thumb is that a mirror ground to F/16 or longer need not be configured from spherical to parabolic. I would venture that this is true for visual work and the f ratio would have to go up significantly with modern instrumentation.

The difference between a catenary and a parabola is much smaller than sphere and parabola. There is probably some f ratio where the difference is beyond significance. For faster systems the mildest of corrector lens would do.

If this can be robotized, the optics and the transportation of the optics gets less expensive. Same for station keeping props and reaction wheels or CMB’s for aiming.

If E. Musk has his way (go Elon!) we are on the cusp of revolutionary access to space. One hallmark of a revolution is that there is a disjuncture through which projections do not work. The thread must be picked up anew and the tapestry of history rewoven with a fresh pattern. (Damn, did I just say that?)

Envy, I think this is a better reply to your question and maybe to the question this this thread poses.

Phil