Mastodawn

Mark McCaughrean

New paper day:

Just published in @NatureAstronomy, a study by @sandorkruk et al. on the increasing impact of satellite trails on Hubble & other low-Earth orbit space observatories.

We found a doubling in the number of trails per image from 2002–2021, with a significant rise of ~50% in 2021 alone, as the "megaconstellations" started to appear.

1/

#Astrodon #Astronomy #LightPollution

https://www.nature.com/articles/s41550-023-01903-3

The impact of satellite trails on Hubble Space Telescope observations - Nature Astronomy

Harnessing the power of citizen science and machine learning, this study takes in 20 years of Hubble Space Telescope images, of which 2.7% show satellite streaks, and predicts that this fraction will increase by up to an order of magnitude in the next decade.

Nature

Mark McCaughrean Mar 2, 2023

The paper involved searching through almost 115,000 Hubble images using a mix of citizen science & deep machine learning.

A significant part of the effort was performed by four ESA undergraduate summer interns working in 2021 with Sándor & me, namely Ben Aussel, Steven Dillmann, Tamina Lund, & Megan Perks. Very proud of what they did, especially as it was all remote due to the pandemic.

Folk from Google were also involved, along with colleagues from Spain & Romania.

2/

Mark McCaughrean Mar 2, 2023

To be clear, the fraction of Hubble images affected by satellite trails remains low at present, around 5%, & in many cases, taking multiple images of a given field can be used to eliminate the trails from the final result.

One thing that is however different about trails seen by Hubble compared to those seen from ground-based telescopes is that the satellites can be very close, only tens of kilometres away, resulting in fat, out-of-focus trails which can cover a wide swathe of an image.

Mark McCaughrean Mar 2, 2023

For the time being at least, Hubble's scientific efficiency & productivity is not greatly impacted by satellite trails.

But we are just at the outset of the era of (very) large constellations of low-Earth orbit "wifi" satellites, the proposed numbers of which are enormous.

They could well pose significant problems for future wide-field optical-IR telescopes on the ground like the Vera Rubin Observatory & in low-Earth orbit like the Chinese Space Station Telescope, Xuntian.

4/

Mark McCaughrean Mar 2, 2023

And that's not to forget our radio astronomer colleagues, who already have to work within strictly delimited frequency bands to avoid anthropogenic "pollution" – there's a risk that "cheap" electronics on mass-produced satellites could leak emission into those protected bands & raise the noise floor against which they're trying to detect incredibly faint cosmic signals.

5/

Mark McCaughrean Mar 2, 2023

The astronomy community is investing a lot of time & effort in trying to analyse this problem, & while some of the satellite operators have shown willingness to engage & try to find ways of reducing the brightness of their satellites, the commercial imperative is strong.

Equally, we're working with organisations like the UN to protect the sky as a common resource & cultural heritage, but such endeavours take a lot of time, while satellites are being launched at ever-increasing rates.

6/

Mark McCaughrean Mar 2, 2023

You can get further information on the current & likely future scale of the problem, as well as discussion of possible mitigations & international legal efforts to contain it, at the International Astronomical Union's new Centre for the Protection of the Dark & Quiet Sky from Satellite Constellation Interference, or CPS for short:

https://cps.iau.org

7/

Mark McCaughrean Mar 2, 2023

And for a little more context on our Nature Astronomy paper (linked in the first toot of this thread & available freely via Open Access), here are some media articles from the New York Times & New Scientist (which, ironically, I can't read because they're behind paywalls 🤷‍♂️ ):

https://www.nytimes.com/2023/03/02/science/hubble-spacex-starlink.html

https://www.newscientist.com/article/2362341-hubble-space-telescope-images-are-being-spoiled-by-satellite-trails/

Hubble Telescope Faces Threat From SpaceX and Other Companies’ Satellites

Scientists found that an increasing number of pictures made by the iconic orbital observatory are being disrupted by passing satellites.

The New York Times

Mark McCaughrean Mar 2, 2023

One key issue is identifying the culprits of such trails, which can include old space debris as well, of course.

That's the case with this image of the Abell 370 cluster from February 2020 noted by @spacegeck, where the trailee was found by @planet4589 to be the Chinese Long March 4C Y33 third stage passing 34 kilometres above Hubble.

9/

Mark McCaughrean Mar 2, 2023

It turns out that it can be quite hard to make a definitive identification of the culprit though, partly because both it & Hubble are moving in orbit, & partly because the so-called "two-line elements" which describe the orbit of a satellite can change quite quickly if it is manoeuvring to avoid collisions, as is increasingly the case.

10/

Mark McCaughrean Mar 2, 2023

And of course that's a whole other issue linked to the rising numbers of satellites in LEO, namely space traffic management & avoiding collisions with both active satellites & space debris.

Our ESA operations colleagues are having to deal with this far more often these days & it presents a significant medium-term risk to LEO activities.

11/

Mark McCaughrean Mar 2, 2023

One final image of trails captured by Hubble, this time three in one. Such "multiple hits" will inevitably become more common as the total number of LEO objects, active & defunct, rises.

To summarise, this is a concern for astronomers, professional & amateur, but also for spacecraft operators & other users of the night sky, as our paper on the case for space environmentalism made clear last year:

https://www.nature.com/articles/s41550-022-01655-6

#Astrodon #Astronomy

The case for space environmentalism - Nature Astronomy

Crowded with satellites and debris, the orbital space around the Earth should be formally recognized as an ecosystem—like the ocean and the atmosphere—to ensure sustainable development and protection from irreversible damage.

Nature

ArneBab Mar 2, 2023

@markmccaughrean I so hope we don’t get a major impact in space that would cut us off from space travel for years 😟

Judy Schmidt Mar 2, 2023

@markmccaughrean @planet4589 I remember that day. That was awesome.

Mark McCaughrean Mar 2, 2023

@spacegeck @planet4589 😬

Judy Schmidt Mar 2, 2023

@markmccaughrean these are always "fun" to find

Mark McCaughrean Mar 2, 2023

@spacegeck Indeed – the creative ways humankind comes up with to shoot itself in the foot never cease to amaze 😬

Christian Fröschlin Mar 3, 2023

@markmccaughrean Is there a trend to schedule multiple shorter exposures instead of one long one due to this issue?

Mark McCaughrean Mar 3, 2023

@chrfrde Not that I’m aware of, although I’m not at all close to HST ops.

Typically though the on-chip integration time is a balance between maximising the exposure to collect enough background photons to become background limited, not read-noise limited, while being short enough to avoid saturation on target sources & excess cosmic Ray events.

So reducing the on-chip time to split images into more exposures for satellite trail removal might (stress “might”) result in lower sensitivity.

GalaxyRise aka Dustin A Ruoff Mar 3, 2023

@markmccaughrean @AwesomeAstronomy Whoa!!

Barry Goldman Mar 2, 2023

@markmccaughrean @NatureAstronomy @sandorkruk hubble? R many sattelights Outside hubbles orbit?

René [¬º-°]¬Mar 3, 2023

@markmccaughrean @NatureAstronomy @sandorkruk This webapp can be used to visualize the current #satellites and planned #megaconstellations:

https://starlitter.info

StarLitter – Stars with Stripes

What might our future night sky look like? Find out!

unearth Mar 3, 2023

@markmccaughrean @NatureAstronomy @sandorkruk We are giving social licence (to satellite companies) to destroy our ability to see past, present and future. Polluting Earth is not enough.

Judy Schmidt Mar 6, 2023

@markmccaughrean @NatureAstronomy @sandorkruk I wonder if satellites glinting out of frame cause any extra light pollution to enter the telescope? I know sometimes when I'm working with images I have no idea why two identical exposures taken on different dates can be so different. I've always figured it was the Moon, position of Earth, maybe Earth's atmosphere or zodiacal light, Jupiter... so many things

MarcinW 🖥️🚀Mar 6, 2023

@markmccaughrean @NatureAstronomy @sandorkruk any chance for a financial assessment of losses caused by megaconstellations to Hubble's operations?

Mark McCaughrean Mar 6, 2023

@sandorkruk @NatureAstronomy @MarcinW At present, Hubble is largely unaffected by trails – the great majority can be removed by stacking multiple images of the same location as usual.

What Hubble does though is give us a sensitive measurement of the deteriorating environment thanks to its extensive archive, allowing us to quantify this clear & present danger to astronomy.

It will only get (much) worse.

Sunflower Björnskalle 🌻Mar 6, 2023

@markmccaughrean @NatureAstronomy @sandorkruk Can we use Hubble in reverse and burn these things out of the sky like a child frying ants with a magnifying glass? ;)