First Results at AAS
This week in Long Beach, CA at a meeting of the American Astronomical Society (AAS), I presented first results from the Andromeda Project. Thanks to the help of ~5100 registered participants and thousands of others who classified images, the science team was able to create an initial catalog a little more than one month after the launch of the site. This turn-around time amazed me, the science team, and nearly every astronomer I spoke to at the conference.
The full poster can be downloaded in PDF format at this link, but here is a summary of the main points from our presentation:
1) We have a great sample of clusters (and galaxies)!
Thanks to the resolving power of the Hubble Space Telescope (HST), Andromeda Project participants identified ~2900 star clusters and ~1700 background galaxies. The new cluster catalog is ~10 times larger than previous ground-based catalogs. Not only does the size of the sample dwarf previous catalogs in terms of number of objects, but it also includes clusters that are ~40 times fainter than previously identified objects.
2) Awesome Citizen Science Participation
As I said before, the response to the site was quite amazing: over the course of 16 days, participants submitted more than 1 million image classifications and collectively spent more than a year of time on the site. Also, many of the astronomers I talked to were quite surprised to see just how international our participation was: about one-third of our registered participants are from Europe.
3) Synthetic Cluster Tests Reveal Important Information
The synthetic clusters inserted in the Andromeda Project search images have already produced some important insights into the accuracy and completeness of the resulting cluster catalog. In the past, astronomers assumed that the detection of a cluster depended solely on the luminosity of the cluster: objects brighter than some luminosity limit could be detected, while objects fainter than that limit were not likely to be identified. Using the results from the synthetic cluster tests, we can already see that such simple assumptions do not hold for our data. Due to our ability to identify faint clusters where most of the light from the cluster could come from only one or two individual stars, we must take a different approach to properly model cluster detection for the Andromeda Project.
4) Work In Progress
The science team continues to improve our catalog construction techniques in an effort to make the most accurate catalog possible. For instance, when we were able to incorporate all of the image classification data, we increased the number of cataloged clusters from the ~2,600 total that we announced in this previous blog entry to ~2,900. We expect this number to fluctuate at the ~10% level while we continue to refine the construction techniques and identification thresholds we use to build the cluster catalog. That’s part of the scientific process: we continue to improve our analysis to achieve the most accurate possible answers on our way towards finalizing and publishing our results.
5) Next Steps
Now that we have our sample of clusters, we can start to analyze these objects in detail: determining the age, mass, and structural parameters (size and concentration) for each cluster. Using these characteristics, we will start to piece together the formation and destruction history of clusters within the disk of Andromeda. In addition, we are already in the planning stages for Round 2 of the project, with a goal of re-launching at the end of the summer once we have completed all data collection from Hubble.
Stay tuned: the science team will provide occasional updates here on the blog as we uncover interesting new aspects of the data.
10 responses to “First Results at AAS”
Trackbacks / Pingbacks
- May 5, 2013 -
- December 5, 2013 -
Can we have an update from the astronomers looking at the background galaxies please?
Have you started to look at how many globular clusters were found, vs other kinds? (How many kinds of star clusters are there anyway?)
Is it possible that you may change – somewhat – the definition of ‘cluster’ as a result of the AP project? And I’d be interested to learn more about how you decide if a bunch of stars is a ‘true cluster’ (whatever that is), and not an ‘association’ or simply an artifact (e.g. the stars looked bunched together ‘on the sky’, but are, in fact, far apart, spread out over ~1 kpc of the sightline through the M31 disc; or are just a chance over-density of stars of unrelated ages and metallicities).
Looking at the ‘non-detection’ synthetic cluster (in the poster; SC5182), it occurs to me that one reason why AP zooites so rarely marked it as a cluster is that the Guide explicitly says “a true star cluster will be tightly grouped with at least three or four individually visible stars” – I think very few AP zooites would say that SC5182 is “tightly grouped with at least three or four individually visible stars”!
As I said in my previous comment, perhaps, for Round #2, the science team could look at revising its definition of ‘cluster’?
That was very quick project and I am glad to hear that collected data is valuable! 😉
I was involved in AP and I submitted many classifications in December. However I had failed to detect clusters similar to SC5182 – non-detected one according to published PDF. To be honest I can not see it even now and I would not mark it! Apparently this is because I am astronomer and guide does not cover SC5182 case. Tutorial is very helpful, but you can improve it ion future! I think it is too short and as hobbyist I expected MORE examples.
Anyway, that was very interesting project and I am looking forward to see more data to classify!
Hi – thanks for your kind words! We are looking forward to Round 2 of AP as well, due to be launched in September.
To clarify, we did not expect volunteers to identify clusters like SC5182 (one non-detection example from the AAS poster). Low mass clusters that are only made up of a few stars will be nearly impossible to recognize. These objects simply do not have enough stars to distinguish them as clusters as opposed to a few field stars that happen to be close to one another. But these non-detections are important, too! Thanks to these synthetic cluster experiments, the science team now has a good idea of how massive a cluster must be in order to be reliably identified in Hubble images of Andromeda.
Glad to hear that results are valuable. I just spotted typo in my comment:
“Apparently this is because I am astronomer…”. Should be: “I am NOT an astronomer…” 😉
I’ve just posted a comment in AP Talk, and would like to repeat it here:
i am alsunni . participated in andromeda project round 1
but you did not show my place in the participant map in your poster online
my place is Khartoum ,sudan , i am a registered user , my user name is alsunni
please show it in next round 2. thanks…
Hi there — Thank you for participating! Sorry that the point representing you on the user map was not visible. The point marking Khartoum, Sudan was present but was too faint to see. Please see the new bold version of the map below: your location is now mapped clearly!