Hacker News Reader

I was going to disagree about the impacts of the starlink constellations because the trails can mostly only be seen during twilight (dusk/dawn) observations. However in this case you're 100% right. The Caltech article below states

"The streaks are most apparent in so-called twilight observations, those taken at dawn or dusk, which are important for finding near-Earth asteroids that appear close to the sun in the sky. ZTF has discovered several asteroids of this nature, including 2020 AV2, the first asteroid spotted with an orbit that fits entirely within the orbit of Venus.

"In 2019, 0.5 percent of twilight images were affected, and now almost 20 percent are affected," says Przemek Mróz, study lead author and a former Caltech postdoctoral scholar who is now at the University of Warsaw in Poland."

We may potentially have a serious problem on our hands unless we really ramp up exterrestrial observation, or maybe we won't have a problem, we might just all be dead on impact.

Link to article https://www.caltech.edu/about/news/palomar-survey-instrument...

> we might just all be dead on impact

The asteroids in question are tiny. One of the links in the article under discussion is to a previous article about an asteroid that hit the Earth near Iceland 5 hours after being detected. It was about 3 meters wide. That's way too small to justify worries about global damage on impact.

What we've found at the last minute so far has been small, but that doesn't mean that the only last-minute discoveries will be small asteroids always and forever.

As others have said, so what are you going to do about this last minute discovery absent some sort of space-based planetary defense system, in which case this discussion is moot?

If it's big enough to flatten a city but not big enough to kill the ecosphere you might be able to evacuate some of the people affected if you spot it sufficiently early to calculate it's trajectory with good precision.

No, but the bigger they are, the more light they reflect, which means the further away we should be able to see them... if we look in the right direction.

> that doesn't mean that the only last-minute discoveries will be small asteroids always and forever.

Larger asteroids are much easier to see, and the issues being discussed with difficulties in seeing in this thread don't apply to larger asteroids.

Well if a big problematic astroid needs to be 3 times brighter to be detected due to StarLink (made up number) it can come 9 times closer without being detected. O(n^2) is baked in sadly.

Suppose it was a 1km asteroid? How is knowing about it a few extra hours in advance going to help? To give everyone on the planet time to freak out and still not actually do anything about it since there’s nothing we can do?

Do we still have 1km NEOs left to discover?

> Suppose it was a 1km asteroid?

Then it's a lot easier to see, and we would see it a lot further away. The seeing difficulty issues people are talking about in this thread don't apply to asteroids that large.

Then who cares?

Of course you're right. People who might have been warned in time to evacuate, but weren't because there was a Starlink constellation in the way, won't care at all! The dead, after all, are beyond such earthly matters.

> People who might have been warned in time to evacuate, but weren't because there was a Starlink constellation in the way

This won't happen for an asteroid large enough to be a threat. Those asteroids are easy enough to see that the Starlink constellation does not impede their detection. That was the point of my previous post (the GP of yours).

Are we sure there's no overlap between the range of impactors susceptible to occlusion and the range of impactors capable of producing a disaster if they strike a populated area?

(Granted it's multiplying small probabilities, but so is this whole discussion, so I don't feel too bad about it.)

> Are we sure there's no overlap between the range of impactors susceptible to occlusion and the range of impactors capable of producing a disaster if they strike a populated area?

Yes. The key point is that the larger the object, the further away we can see it. And the further away we can see it, the more time we have between the initial sighting and any possible collision with Earth.

The lower bound for a "potentially hazardous object", i.e., one that could cause disastrous damage in a collision, is 140 meters wide, and we can see those far enough away that they are years to decades from getting close to Earth (i.e, time enough to plan a space mission to deflect them enough to miss Earth). Newer surveys are getting to the point where they can detect 10 meter wide objects, which can still cause substantial damage (similar to, say, a hurricane or a tornado), far enough away that they are days to weeks from getting close to Earth (i.e., more than enough time to evacuate possible impact areas if needed).

Even if such an object were temporarily occluded by a low Earth orbit satellite, the occlusion would only last for minutes at most (because the satellite and the distant object are moving across the sky at very different rates). So for objects of the above sizes, occlusion doesn't prevent us from seeing them well in advance.

See here for more info:

https://en.wikipedia.org/wiki/Spaceguard

Why can the not just ignore the very small parts of images with satellites in them? The orbits of starlink are published publicly so it becomes easy to rule out bright spots in images.

The main problem is that we’re ONLY talking about Starlink right now.

What happens when every developed nation wants to launch their own constellations?

It’s kinda the same situation as environmental pollution. Is the world justified in restricting a developing nation from undergoing their industrial revolution like others have before simply because they were too late to the party?

What kind of wars is this going to cause?

Seems kinda disingenuous to quote that part without the “however”…

“Yet despite the increase in image streaks, the new report notes that ZTF science operations have not been strongly affected. Study co-author Tom Prince, the Ira S. Bowen Professor of Physics, Emeritus, at Caltech, says the paper shows a single streak affects less than one-tenth of a percent of the pixels in a ZTF image.

"There is a small chance that we would miss an asteroid or another event hidden behind a satellite streak, but compared to the impact of weather, such as a cloudy sky, these are rather small effects for ZTF."

If the effects are small compared to cloudy weather, is it really that serious of a problem?

StarLink satellites move in known and predictable ways. How does this hinder spotting things that aren't currently known about?

There's even other objects in the sky/space that aren't man-made that also move, and most of them we know about as well.

See my reply above with 4 articles describing the problem. 3 of them from the past 3 months.

You bring up the recency of these articles to bolster your argument, but it is detrimental to your point. Particularly from these sources. Alarmist regurgitation is the standard, popsci-brand churn.

If there's a click bait title that is going to drive views they'll print.

You might be right. I doubt that. These don't strengthen your position

IOP is clickbait popsci? The first link is also mostly quotations from scientists.

Did you read the sources linked to?

By definition, things we don't know about aren't moving in known and predictable ways, so if you throw out data from X parts of the sky during time period Y you are potentially losing the observations that would help identify a new object

This is a problem solved by taking a few more images. The asteroids aren't moving in synchrony with starlinks.

Is there any evidence that an asteroid like this would have been detected earlier without Starlink satellites?

My understanding is that artifacts from Starlink are trivial to remove from ground based astronomy datasets. We know what they look like, where they are, and when they transit.

do you have other information?

It is not really possible to remove artifacts from images because there is a degree of randomness in which photons arrive, and even if your model is accurate down to the level of QED, you will not be able to exactly subtract the photon counts.

Im not sure I follow. Im not saying that you subtract photon counts, but void measurements entirely for pixels with satellites in them.

Telescopes use long exposures. The satellites leave streaks and a lot of pixels would have to be deleted.

Maybe two telescopes and a bit of math to use the difference in distance to the satellites vs the asteroids?

Long exposures can still have a duration censored. This is digital photography, not chemical. If you expose for 1 million seconds, you can edit out 60 seconds where a satellite passes. That data is forever gone, but you still get the rest of the data from that pixel time.

Given that the earth rotates every 1.5 thousand or so minutes something tells me 1 million may not be representative...

Sure, I was being extreme, but multi day or week surveys exits. In fact, they already have to do a lot of filtering to remove transient artifacts

You're assuming there's only one satellite to worry about per exposure.

Yes. The first three were all published within the past 3 months. The last one is your question answered in Quora for good measure.

[1] https://www.sciencealert.com/the-international-astronomical-...

[2] https://gizmodo.com/astronomers-rally-to-stop-starlink-and-o...

[3] https://iopscience.iop.org/article/10.3847/2041-8213/ac470a

[4] https://www.quora.com/Why-cant-astronomers-use-software-to-r...

There are real incentives to overstate the issue.

Ask if we should end funding for a ground based observatory and you will hear a very different story: ~“Starlink blocks up to 0.1% data collected by ground based telescopes, that isn’t inherently a big deal.”

The truth is someone in between, a few kinds of observations are being significantly impacted while most are largely able to avoid the issue.

None of these articles counter the claim that it is trivial to remove streaks with software.

Some data will always be lost, as the satellites do obstruct part of the night sky, However the fractional area obstructed is exceedingly low.

That is to say, If you dump pixels with satellites, you will have fewer pixels. However, the number of pixels is low and changes with time.

The number pixels affected increases with the linearly with the exposure length. However to spot things twice as far away (at the same brightness) we need to stare 4 times longer at the same place.

That’s assuming a fixed instrument and zero mitigation. You can also add a shutter or increase the light gathering area to reduce exposure time etc.

Yes one can build a new telescope.

More, when building a new telescope how much money does it take to get the same amount of data. If spending 1% more gives the same data then it’s not that significant an issue.

Are you saying that we have a choice between defending ourselves from asteroids, and offering low-latency internet to remote regions? Is there no way to have both?

(I know nothing about this.)

We could have both, by having also mass produced and launched space telescopes. But who is gonna pay for them?

NASA paid $8.8 billion on just the James Webb. If they wanted to have a mass produced telescope for much cheaper, they could likely pay someone to produce a bunch of them for $8.8 billion. It wouldn't be as capable, but that's probably also not necessary.

Actually, since asteroid defense is infeasible with our current capabilities, our options are between relatively low-latency internet to remote regions (and causing headaches for astronomers), and nothing.

Is it infeasible? If we have 9 months to prepare we manage if wel fail to detect it for 8 due to StarLink it is to late.

Also StarLink is not for everyone, just for (globally speaking) rich people.

We can't determine the orbits of asteroids 9 months out with any certainty.

We don't necessarily need it to be certain to hit to prepare. It would just have to be significantly risky.

There's a Meta Connectivity thing called Terragraph that would do it, though i don't know its status.

I think the website cites urban but it's not really specific to that.

https://terragraph.com/

> Terragraph is a wireless technology designed to meet the growing demand for reliable, high-speed internet access in urban and suburban environments.

From https://terragraph.com/product/

I don't think this is a solution for remote areas.

https://tech.fb.com/engineering/2022/03/alaska-terragraph/

supposedly it was developed for remote areas. Hard to tell if there's something special about where it was deployed....maybe the notion of remote has to be still close to wired basestation hubs.

It’s last mile though which mostly doesn’t help except in medium density neighborhood where running wire/fiber to houses isn’t hard, just expensive. 5G is probably more viable in general for these sorts of cases.

i bet that's the disruption (5G) that might obviate Terragraph. I guess we'll see.

B612 foundation needs funding for space based observation.

While I agree with your assessment currently, it is flawed moving forward.

The amount of launch capacity we now have, and the massive drop in cost per tonne to orbit, means spaceborn telescopes are cheap now, with the added benefit of no atmosphere.

And as costs to launch have dropped, so can costs to build. We don't need to spend loads of cash, since the cost of launch is so cheap, we can just throw a replacement up if it goes bust.

In fact, we don't even need telescopes to scan our neighborhood... just loads of smaller video capturing cameras, with excellent zoom lenses. If 10% fail, who cares?

It's time to think about proper near earth detection, not the pot luck stuff we've been doing so far.

And it's time to think in terms of the launch capacity we have..

Agreed.

We have far better tools at our disposal and with a full mesh network of satellites providing optics, we could have an amazing early detection system with total spatial coverage.

This is the alarmist pop-sci all the sources are known for pumping out

what would we do with "proper near earth detection" other than contemplate our impending doom? lowest estimate for lead time to prepare an interception mission is one year, and even still the probability of success is not clear.

https://en.wikipedia.org/wiki/Asteroid_impact_avoidance

We have Starlink because of new and expanding launch opportunities. Starlink is very helpful (I work at a farm and before Starlink we had very poor internet options) for people the world over.

Can we build a new generation of space based asteroid observers? I know that this is a reactive approach but I don't see those constellations coming down. We could prevent expansion but they are damned helpful. In the past space telescopes were a big affair to launch one system. But what if we launched a constellation of 100 or 500 space based telescopes? Is there a way to move forward to solve this? Or what is the solution?

This argument has never made any sense to me. We know where the satellites are. If that's not enough to keep them from interfering with astronomical observations, well, somebody needs to spend less time complaining and more time working on their own tech.

“So dangerous” Really?

Those arguing that LEO internet satellites are an existential threat had better come up with a rural internet alternative if that’s their position. Because internet availability is a genie that’s not going back in the bottle.

I’ll see your life-ending existential threat and lower it with the lack of ok-latency, higher-speed internet.

This is coming from someone who lives in a rural area without reasonable internet options and is on the Starlink waiting list. I’d rather be alive than have lower-latency connections to Reddit.

Rural internet is solvable when public officials prioritize it - what isn't easily solvable is a natural disaster connectivity, which still could be line of sight, mobile, temporary balloons before satellites.

I am not excusing the problems starlink is causing for astronomers but would a partial solution be to fit a telescope to the back of every starlink satellite which astronomers would have access to?

A better bet would be to launch a payload of cube imaging satellites. Possibly a partnership with Planet to make a variation of their Doves.

Starlink satellites are bigger than cubesats but still fairly small, I don't think you're gonna get like a 3m mirror on them. And even if you're able to replicate the capabilities of a subpar optical telescope, the radio astronomers are still SOL