eGZeLENS update: HST pre-imaging

Hello there!

My name is Phil Marshall, and I’m a postdoctoral research fellow in astrophysics at the University of California, Santa Barbara. My main research interests are in finding new strong gravitational lenses, and then using them to make accurate measurements of galaxy masses. I am involved with several semi-automated lens searches, using data from the Hubble Space Telescope (the HAGGLeS project) and also the CFHT legacy survey (the SL2S project). Several years ago some colleagues and I were joking about how one could find lenses at a hypothetical classification website called “” – imagine how excited I was to find you all doing this already in the galaxy zoo!


I got talking to Chris about the galaxy zoo project when I visited Oxford this last summer – I met Aprajita then as well, and we looked through the candidates you had generated while classifying galaxy morphology. It was encouraging to see some of the Fermilab and Cambridge groups’ lens candidates pop up (you know, the 8 O’Clock Arc, the CASSOWARY sample and so on), but even more exciting to see some new candidates! As you read in our previous post, Aprajita and I made a shortlist for the Gemini run together, trying to pick the systems most likely to be lenses. This is tricky, as the SDSS resolution is quite poor: lenses tend to be far away, and therefore small and not well-resolved in the SDSS images. We focused on the objects that looked like wide separation lensed quasars – these are predicted to be quite rare, but there should still be a handful in the whole 8000 square degrees of the SDSS survey.


Now, spectroscopy is expensive (as the exposure times have to be quite long), so it’s common to try and reduce the risk of observing lens candidates by taking cheaper observations in advance to prevent disappointment later. These might be high resolution imaging, to test in more detail the lens image geometry hypothesis, or short exposures with a spectrograph to see if the objects in the field are at a different redshift (distance) than the lens
galaxy. On September 24th, an opportunity presented itself, as we in the SL2S collaboration received an email from our HST observations coordinator. With the demise of the NICMOS instrument, there was only one working camera onboard HST: the optical imager, WFPC2. As you can imagine, with 100% of the observing time being given to WFPC2, the HST schedulers began to run out of targets! They also needed extra targets to keep HST busy in between the end of the current cycle (16) and the beginning of the next one. Here’s what the email said:

“With the recent suspension of NICMOS observations until the next
cycle it has become apparent that the HST observing efficiently may
drop if the shuttle launch slips by more than a few days and the
pool of observable targets is not augmented.

STScI created the SNAPSHOT proposal category in order to counteract
any scheduling inefficiency. Consequently, the Institute is actively
scheduling all cycle 16 SNAPSHOT observations that utilize the
currently available instruments. A number of SNAPSHOT targets in
your program cannot be scheduled because their location makes them
inaccessible to HST in October.

It is the nature of SNAPSHOT proposals that their targets are often
interchangeable. We are contacting you since your proposal has been
identified as having a number of yet unexecuted observations which
cannot be observed prior to the Servicing Mission.

We would like to provide you the opportunity to replace these
targets with scientifically equivalent substitutes whose coordinates
are within range.”

My French SL2S collaborators and I then spent an afternoon scrambling together a list of new targets – mostly more of our own targets, but I suggested the galaxy zoo lens candidates and they agreed that, yes, these were “scientifically equivalent” and hence a good use of our telescope time. Three out of the five targets on the eGZeLENS shortlist lay within the area of sky accessible to HST in October, and these three were duly observed! Here they are – with the galaxy zoo discovery images in the top panels, and the higher resolution HST/WFPC2 images below.

Bad news folks! None of the three systems, despite looking like fairly convincing gravitational lenses in the SDSS images, are actually lenses.
SDSSJ0926-0037 (right) made every lensing aficionado I showed it to very excited – but it’s just a very compact (and symmetrical) group of galaxies. I say “just” – there can’t be many groups out there as neat as this! The other two have nice symmetrical double point-image configurations – except that in SDSSJ0813+1552 (middle) one of the “images” is a spiral galaxy. It’s not definite that SDSSJ0811+0240 (left) is not a lens – but the oval of diffuse emission around the bulge of the galaxy is not consistent with being an Einstein ring of a quasar host (it doesn’t match up with the point source positions, and is not circular enough) and is almost certainly a set of spiral arms associated with the bright yellow bulge. This would make the galaxy much less massive than it would need to be to generate the observed 6 arcsec of “image” separation: the point sources are probably stars that happen to lie (in projection) either side of a bulge-dominated galaxy!

On the bright side then, as well as avoiding wasting any hard-won Gemini time, we have learnt the following key fact about lens finding from this exercise:

The galaxy zoo is so big that it contains a significant number of objects that are very good at mimicking gravitational lenses.

We must be on our guard in future!

About The Zooniverse

Online citizen science projects. The Zooniverse is doing real science online,.

10 responses to “eGZeLENS update: HST pre-imaging”

  1. Joseph K. H. Cheng says :

    Thank you very much for explaining the detailed technical aspects of gravitational lensing research. It is certainly comforting for Zooites to learn that our efforts help to save valuable research time and other related resources.

  2. lizardly says : looks a lot like the 8 o’clock arc, that would be a good target to take another peek at if it is visible.

  3. gumbosea says :

    Ah, such are the breaks! What’s the status of the other Gemini South short list targets? I note the GS observation status remains at 20%; does this reflect the remaining (2) targets. Have you lined up GS substitutions yet for the (3) de-confirmed lenses?

  4. Hanny says :

    Thanks Phil! 🙂

  5. GwydionM says :

    Disappointing that these were not lenses, but that’s what science is about, I suppose. That ‘group of five’ could be used in future as a very good example of how appearances can mislead. Put it next to a real multiple image and then challenge anyone to say which was which.

    Thanks a lot for the detailed explanation.

  6. Alice says :

    Welcome to the zoo, Phil! Rats, sorry to discover those aren’t lenses. I guess SDSS isn’t an ideal telescope. Can we hijack some more? 😀

  7. pjm says :

    Yes yes, Alice, we’re working on it! You’d be surprised how many “pilot” projects get done at telescopes under the table. It only works at old-school, non-scheduled observatories though. Gemini and VLT take all their data in service mode, where no cowboys are allowed!

    I’ll let Aprajita fill you in on the GS campaign plan. For now though, lizardly, here is an arrow against misfortune: when you look at an SDSS image on the sky server, try and estimate the image separation “dtheta”. In the simplest possible lens model (which actually works surprisingly well in most cases!), the velocity dispersion “sigma” (~mass) parameter of the lens is related to this image separation by

    sigma ~ 200 * sqrt(dtheta/1.2arcsec)

    For your candidate, I get a dtheta of >4 arcsec, which means a sigma of >360 km/s. This would be a *very* massive galaxy, or a small group of galaxies, and both we would expect to be very “red” (actually appearing yellow in the SDSS images). I think 588017626693042255 is a spiral galaxy or a merger or somesuch. I was hoping that our two wide-separation doubles would turn out to have compact groups of galaxies as the lensing object – but it wasn’t to be. But yes, the arc geometry you pointed out is very lens-like – in this case it’s the colours of the “images” compared to the lens that contain the information. There are a *lot* of false positive objects like this out there – we are in a minefield, people!

  8. Budgieye says :

    Thank you for the feedback. I hope we all get better at learning what to look for, and what we shouldn’t waste time on.

  9. Graham Dungworth says :

    Much appreciate the feedback Phil.
    Remember the lenses are rare regardless of the SDSS image resolution, so don’t be disheartened- keep looking as it’s definately not a waste of time.

  10. Eric M. Jones says :

    Isn’t this potential new Einstein Cross worth a closer look?
    SDSS 587738411942805595

Leave a Reply

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

You are commenting using your 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 )

Google+ photo

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

Connecting to %s

%d bloggers like this: