By Katharina Doll, Johanna J.S. Finnemann, Hugo A. Durantini Luca, Joshua Hamilton, and Michiharu Hyogo
As Steven from the science team wrote in his blog post, our third paper has been accepted for publication. Nine citizen scientists from the advanced user team are listed as co-authors. Today we want to show you the work we did for this paper.
Our contributions for this paper fall into three parts: the image analysis of Robo-AO follow-up images, the classification of M dwarf debris disk candidates presented by Theissen and West (2014), and finding references related to our disk candidates.
1. Analysis of Robo-AO images
As already explained in the glossary and earlier in this blog, only objects with significant Infrared excess in WISE images are pre-selected to be classified. It is mainly because the circumstellar disks shine brightly in these spectrums. However, there is a major problem: several close-together objects can blend together in infrared, making it look like the sort of infrared excess the disks we are searching for exhibit.. Moreover, there are other astronomical objects which also shine brightly in Infrared spectrum of light such as background stars (located nearby) or much further away, galaxies, interstellar gas and dusts, and AGNs/QSOs outside our own galaxy. The Robo-AO and Dupont telescopes have a better resolution than the flipbooks shown on the classification interface, so we’re able to see contaminating objects that don’t show up when you classify on the main site (and therefore submit something as a “good candidate”).
We learned how to work with software (DS9, which professional astronomers also use) to look at the files from the telescopes and determine any contaminants. We’ll show you several sample images that we used:
Case 1 Close by Background Objects
Here is an example image of an object (Zoo ID AWI0000kk6) with a bright object located very close by. It needed follow-up observing because it is already known as a binary system (and had been flagged in the Talk comments).
Case2: Faint Fuzzy object
Another example image (Zoo ID AWI000028h) with a faint fuzzy object located in the corner of the image. This object (located at 11 o’clock from the main target) is likely an extragalactic object like a galaxy or an AGN.
Case 3: Good object in Follow-up Images
2. Classification of disk candidates published by Theissen & West (2014)
As Steven explained in his blog post on the paper, members of the advanced user team looked at previously published disk candidates by Theissen & West (2014). Using their WISE IDs, we used the IRSA Finder Chart tool to generate images in the same wavelengths as in the flipbooks on the main site (SDSS, DSS, 2MASS, WISE). Joshua Hamilton added red circles on all images so that we could use the same method on these objects as we do on normal Disk Detective objects.
We found out that most of the disk candidates from Theissen & West (2014) disk candidates are not present in Disk Detective because they have a low signal-to-noise ratio in WISE 4 (the last image you see in a flipbook), and in addition most are contaminated in our criteria. For example, all were extended in W4 images and thirteen of them consists of multiples.
As you can see in the example above, more than one object is visible in SDSS images, and the bright blue fuzzy objects extend outside the red circle in WISE images . Hence we concluded this object does not satisfy our criteria of our classification process.
3. Finding references
Lastly, we contributed by finding references related to our disk candidates. About a week before the paper was first submitted, Marc Kuchner, the principal investigator, had asked the super users to look for any interesting references related to the disk candidates of this paper just in case. We used tools such as SIMBAD and VizieR (which Marc explained here and here on the blog) and searched information regarding each of those candidates. VizieR provides us with access to a number of astronomical catalogs. Each of these catalogs consists of a major survey mission and it includes astronomical information regarding the object such as distance(parallax), luminosity, and spectral type.
Let’s show you an example of how it was done. First we typed in coordinates for each candidate, in this case AWI00062lo, which is shown in one of the figures above.
In the figure, we see the coordinates of the candidate, AWI00062lo. After we enter this, we are able to see a number of astronomical catalogs related to this object.
In this figure below, you can see a part of the VizieR site that shows two catalogs inside blue circles.
Two catalogs shown are Gaia DR2 and FON Astrographic Catalog. One of these catalogs, Gaia DR2, is especially important because this survey mission measured positions and distances of 1 billion stars with unprecedented precision. The parallax value which determines the distance of a star is provided by Gaia (8.8526 milli-arcseconds = 112.9 parsecs in this case).
Through this simple process, we found a significant catalog in VizieR, Oh et al. (2017), which contains data on some of our candidates including AWI00062lo. In the Oh et al. (2016) survey, new comoving pairs are searched in the Tycho-Gaia solution.
This has assisted us to compare our results with Oh et al. (2017): For example, only 1.37% of stars in Oh et al. (2017) are members of comoving pairs while 11% of our candidates are in such pairs. These results support the hypothesis of Zuckerman (2015) which stated that there is a high frequency of warm debris disks exist within binary systems of young stars. We also cross-checked with other reference catalogs on VizieR.
4. Who are we?
Finally, a brief look at who we are and what we do in our daily lives.
Joshua – I am from the mid-Michigan area where I work as the Director of Youth Ministry at a large Catholic parish. I have a degree in Media and Information from Michigan State University, which inadvertently came in very handy for this paper. What I love most about Disk Detective is that someone like me, with very little professional training in astronomy, can collaborate on incredible scientific research with people from around the world. Though we are all different, we all share on thing in common: our love of discovery and science. For this paper, I used my design software skills to build a precise overlay for the Theissen and West Paper images that we wanted to study. I took 175 Theissen and West images that they had classified as good disk candidates and overlaid the red circle that you’d normally see when classifying on the Disk Detective website. This allowed us to analyze their images just we do our own and by doing this analysis, we found that many of the “good candidate” disks from their paper were contaminated by multiple objects within the red circles.
Hugo – I am from Cordoba, province of Cordoba, Argentina. Computer technician and a few other things. Aside of citizen scientist now I’m part of the GAF (Grupo de Astrometría y Fotometría / Astrometry and Photometry Group) of Cordoba. My years of experience working with computers were always handy to work with Disk Detective and when we started to work in this paper that was not exception. Several of my experiences with Disk Detective are captured in our blog, but it’s fantastic the range of experience what the projects ended making possible for me. If have the answer the questions about what I do for the project, I think what the best answer is “a bit of everything” since I help with things what can range from website classifications, vetting or literature checking to recording and uploading the hangouts videos, managing the Spanish Disk Detective twitter account and side projects.
The image analysis with DS9 was my first time using that software. The learning curve for the analysis was a bit steep but not difficult. I think that the image analysis with DS9 was the first analysis done for paper 3 or at least one of the first biggest part what we as superusers worked on. I remember how we learned to differentiate between real objects and fake signals in the images, fun process what included sometimes using the rainbow color scheme in DS9 that changes black for fluorescent pink, great for contrast, not so great for the eyes.
I also ended in charge of putting together and organizing the spreadsheets where the superuser group was putting their conclusions in the different batch of objects. If some old notes are correct, in some point also helped downloading and uploading again the batch of objects in smaller chunks to help other users with slower internet connections. Making the notes to learn the difference between ghost and real objects was interesting. Some artifacts were easy to tell apart from real objects, but others not so much. Discussing our findings, doubts and ways to work with DS9 was a great experience overall.
Michiharu – I am from Tokyo, only super user from Japan. I have lived outside my country for more than two decades. I have completed masters degree in Astronomy 8 years ago while I was in Australia. I also have completed another masters degree in Information Technology. I am currently looking for a physics lecturer position in one of the universities in Tokyo. Although the Disk Detective project is not quite related to a research topic I studied while completing masters degree, I have basic knowledge regarding astrophysics and therefore I have been able to learn science behind this project and how to contribute very quickly without major problems. My main tasks for this project are a normal classifications of DSS, 2MASS, and WISE images, and vetting of those objects classified as good objects in the classification stage, which includes literature reviews, to help further select sample targets for the observations. On top of all these tasks, I worked on image analysis on the ROBO-AO images with several other super users using the DS9 imaging software to look for possible contaminants, for this paper.
Katharina – I am from the greater Munich, Bavaria region in Germany and currently work as a research assistant in legal history at university. I have degrees in law and business studies. Being a member of the advanced user group is therefore a great opportunity for me to contribute to a completely different field of research! For this paper, I worked on vetting objects classified as good objects and on analysing Robo-AO images for background contaminants using DS9.
Johanna – A cognitive neuroscientist by training, my passion for science has always extended beyond my primary research area. I believe that while boundaries that mark off one field of science from another can of course be useful, they remain artificial constructs and can also get in the way of connecting people and ideas. It’s for that reason that I’ve enjoyed becoming a member of the Disk Detective team; it is an extraordinary opportunity to make meaningful contributions to an area of science without the formal training/qualifications that are so often indispensable. I’ve learned a lot and not just about astronomy but also about serious involvement of the public in science (something my field is still struggling with!) and I hope that I can in turn also use my experience in data science and understanding human cognition and perception to advance the project.
My involvement in this paper (the vetting of the Theissen & West images) was one of the first things I did after joining the advanced user team and it was great to feel involved right from the start – even while I was still finding my feet. Deciding on analysis pipelines to improve the false positives/false negatives is once again a much broader scientific issue and so it was interesting to apply it to our specific problem in identifying the pitfalls of certain classification criteria.
Lily – I am from Singapore. I graduated with a bachelor’s degree in Library and Information Studies. Apart from being a citizen scientist, I am also an amateur photographer. My astronomy journey started in 2003, I have enrolled in a couple of online astronomy courses as well as taking part in several online citizen science projects. When Disk Detective launched in 2014, I quickly registered. Throughout my time with DD, I have participated in several of its projects; ROBO-AO Imaging of Debris Disks was one of them. My task was to identify possible companion objects and artifacts to a nearby debris disk candidate, and record their locations. I am honored to be part of the team. The knowledge that I have gained from the project is invaluable.
Do you want to join us? If you’ve done 300 classifications and you’d like to get more involved in our next paper, drop us a line at email@example.com and ask to join the advanced user group.