I am very happy to present the results from the first published paper based on your classifications of the HST-CANDELS Images.
Galaxy Zoo: CANDELS combined optical and infrared imaging from the Hubble Space Telescope, which allows us to probe galaxies back to when the universe was only around 3 billion years old (early than we could do with optical HST images alone). So we are looking at galaxies whose light has taken over 10 billion years to reach us!
Our first area of research with this data is to look at disk and barred disk galaxies, as the title suggests…….
This work is based on an initial sample of 876 disk galaxies, which are from the Cosmic Assembly Near-Infrared Deep Extragalactic Legacy Survey (CANDELS). We want to explore what happens to barred disk galaxies beyond eight billion years ago, building on our work looking at the evolving bar fraction with Galaxy Zoo: Hubble.
When we began this work, we were unsure what we would find when looking so far back. From our Galaxy Zoo: Hubble work we had identified that 10% of disk galaxies hosted a galactic bar eight billion years ago, but beyond this our knowledge of disks was limited to a single simulation of disk galaxies. This simulation predicted that bars in disk galaxies were very rare beyond the epoch we had observed to, as the Universe would be to young for disk galaxies to
have settled down enough to form barred structures.
As Figure 1 shows, we actually find that roughly 10% of all disk galaxies host a bar, even back to when the Universe was only 3 billion years old! This is a very exciting result, as it shows that disk galaxies were able to settle at much earlier times than originally believed.
What we need to understand now is how do these disk galaxies form their bars? Could they be completely settled disk galaxies which have naturally formed bars, even during this epoch of violent galaxy evolution where galaxy mergers are more frequent? Or were these bars formed by a galaxy-galaxy interaction, as seen by some simulations? The answer could be one or the other, or most likely a combination of these two theories. Either way, we hope to explore this population of barred disk galaxies in greater detail over the coming months!
So there is a summary of the first Galaxy Zoo: CANDELS paper. If you would like to see this in more detail, please take a look at the paper here, and why not check out the RAS press release too! Thank you all for your hard work, and keep classifying!
Posted on behalf of Tom Melvin.
Posted on behalf of Tom Melvin:
Hello everyone, my name is Tom Melvin and I’m a 3rd year PhD student at Portsmouth University. I have been part of the Galaxy Zoo team for over two years now, but this is my first post for the Galaxy Zoo blog, hope you enjoy it!
I’m very happy to bring you news of the latest paper based on Galaxy Zoo classifications, and the first paper based on Galaxy Zoo: Hubble classifications. Galaxy Zoo: Hubble was the first Galaxy Zoo project to look at galaxies beyond our local universe, using the awesome power of the Hubble Space Telescope. These images contained light from galaxies which have taken up to eight billion years to reach us, so we see them as they appeared eight billion years ago, or when the universe was less than half its current age! So what is the first use of this data? Well, we combine our Galaxy Zoo: Hubble classifications with Galaxy Zoo 2 classifications to explore how the fraction of disk galaxies with galactic bars has changed over eight billion years.
Here’s the title…..
Our work is based on a sample of 2380 disk galaxies, which are from the Cosmic Evolution Survey (COSMOS), the largest survey Hubble has ever done. To see how the bar fraction varies over such a large time-scale, we look at the number of disk galaxies and what fraction of them have bars in 0.3 Gyr (300 million year) time steps. In Figure 1 we show that eight billion years ago only 11% of disk galaxies had bars. By 4 billion years ago this fraction had doubled, and today at least one third of disk galaxies have a bar.
We know that bars tend to only form in disk galaxies which have low amounts of atomic gas and are in a relaxed state, or what we call ‘mature’. Combining this knowledge with our observations, we can say that, as the Universe gets older, the disk galaxy population as a whole is maturing. To see whether this is true for all disk galaxies, we split our sample up into three stellar mass bins, allowing us to look at the evolving bar fraction trends for low, intermediate and high mass disk galaxies.
The results for this are shown in Figure 2, where we observe an intriguing result. The bar fraction increases at a much steeper rate with time for the most massive galaxies (red), compared to the lower mass galaxies (blue). From this we can say that the population of disk galaxies is maturing across the whole stellar mass range we explore, but it is predominantly the most massive galaxies which drive the overall time evolution of the bar fraction we observe in Figure 1.
At the end of the paper we offer an explanation as to why the time evolution of the bar fraction differs for varying stellar mass bins. We can make the reasonable assumption that, by eight billion years ago, the majority of massive disk galaxies have formed, and have been, and continue to form bars up to the present day – hence the steeply increasing bar fraction we observe. However, the same assumption is not true for the low mass galaxies. There are some which are ‘mature’ disk galaxies eight billion years ago, but not all are ‘mature’ enough to be classified as disks. As with the most massive galaxies, these low mass disks are forming bars at a similar rate up to the present day, but the difference with this low mass sample is that there are still low mass disks forming up to the present day as well – leading to the much shallower increase in the bar fraction with time we observe.
In addition to these results, we are also able to present an interesting subset of disk galaxies. Your visual classifications has allowed our work to include a sub-sample of ‘red’ spiral galaxies (like those found from Galaxy Zoo 2 classifications). This sub-sample is generally omitted from other works that have explored this topic, as their way of identifying disks is based on galaxy colours. This means that these ‘red’ galaxies would have been classified as elliptical galaxies! Figure 3 shows a few of these ‘red’ disk galaxies (with the full sample of 98 here), so why don’t you take a look and decide for yourself! Not only is it very cool that you are able to identify these ‘red’ disks, but they also influence the results we observe. Just like in our local universe, these ‘red’ disks have a high bar fraction, with 45% of them having a bar! Could this be a further sign that bars ‘kill’ galaxies, even at high redshifts?
So that is a summary of the first results from Galaxy Zoo: Hubble. If you want more detail have a read of the paper in full here and take a look at the press release too! Thanks for all your hard work and help in classifying these galaxies!
Posted on behalf of Tom Melvin.
I’m really excited to be able to post that galaxies selected with the help of Galaxy Zoo classifications are being observed at the VLA (Very Large Array) in New Mexico, possibly right now.
The funny thing about observing at the VLA is that you do all of the work for the actual observations in advance.
The VLA runs in queue mode – as an observer you have to submit very (very) detailed information about what you want the telescope to do during your session (called a “scheduling block”) and a set of constraints about when it’s OK to run that (for example you tell them when the galaxy is actually up in the sky above the telescope!). Then the telescope operators pick from the available pool of scheduling blocks at any time to make best use of the array.
This means after you submit the scheduling blocks you just have to sit and wait until you start getting notifications from VLA that your galaxies have been observed. The observing semester for the B-array configuration started on 4th October (had a pause for the US shutdown) and runs until the 13th January 2014. I’m happy to report that we started getting notifications in late November of the first of our 2 hour scheduling blocks having been observed. At the time of writing four of our galaxies have each been observed at least once (we need six repeat visits to each one to get the depth of data we’d like) for a total of 16 hours of VLA time. I’ve been getting notifications every couple of days – which means that as I write this the VLA could be observing one of our galaxies!
Since making these very detailed observation files is the observing prodecure at the VLA – it takes the length of time you’d expect given that…..
So, in September in-between a crazy travel schedule, and with a lot of help from our collaborator Kelley Hess at Cape Town, I spent a lot of time scheduling VLA observations of some very interesting very gas rich and very strongly barred galaxies we identified in the Galaxy Zoo 2 sample (the bit which overlaps with the ALFALFA survey which measures total HI gas in each galaxy).
We have been granted time to observe up to 7 of these fascinating objects (depending on scheduling constraints at the VLA) which I think may reveal some really interesting physics about how bars drive gas around in the discs of galaxies.
You might notice from the picture (and the name) that the VLA is not a “normal telescope”. It’s what astronomers call a radio interferometer. Signals are collected from 27 separate antennas and combined in a computer. This means that as well as observing sources for flux calibration (so we can link how bright our target is through the telescope with physical units) we also have to observe, roughly every 20 minutes or so a “phase calibrator” to be able to know how to correctly add the signals together from each of the antennae (to add them “in phase”).
So a single scheduling block lasting 2 hours for one of our sources comprises:
1. Information to tell the VLA where to slew initially and what instrumentation to use (how to “tune” it to the frequency we know the HI in the galaxy will emit at).
2. A short observation of a known bright source for flux calibration.
Then there’s a loop of
a. Phase calibration
b. Source observation
c. Phase calibration
d. Source observation
and so on – ending with a Phase calibration (on Kelley’s advice we’ll do 5 source observations, and 6 phase calibrations). We have a total of 6 of these blocks for each galaxy, that makes 12 hours of telescope resulting in about 10 hours of collecting 21cm photons per galaxy.
We have to check which times all these sources are visible to the VLA, and set durations for each part which give enough slew time and on source time wherever the sources are on the sky. And this all has to add up exactly to 2 hours to fit the scheduling block.
The benefit of this though is a telescope which acts like it’s much larger than you could ever physically build. We’re trying to detect emission from atomic hydrogen in these galaxies which emits at 21cm. So we need a really large telescope to get a sharp picture.
And just to end, because they’re lovely, here are the four galaxies the VLA has observed so far in the Sloan Digital Sky Survey visible light images.
Thanks again for your help finding these rare and interesting galaxies. They’re rare, because they’re so gas rich and strongly barred – we have previously posted about how we showed strong bars are rare in galaxies with lots of atomic hydrogen. Hopefully we’ll have some exciting results to share once we’ve analysed these data.
(PS. That takes a lot of time too – it’ll be almost 1TB of data to process in total!).
A quick post to say congratulations to new Galaxy Zoo science team member Edmond Cheung, a PhD student from UC Santa Cruz, on the publication of his first Galaxy Zoo paper. Edmond approached us some time ago and was interested in doing further study on the barred galaxies in both Galaxy Zoo 2 and GZ: Hubble. This paper is the result of the excellent work he’s done looking at more detail on the properties of bars in the Galaxy Zoo 2 classifications.
The paper has recently been accepted to the Astrophysical Journal, and will appear on the arxiv very shortly.
The main result is a stronger proof than has ever before been seen that secular (that is, very slow) evolution affects the properties of barred galaxies, which grow larger bulges and slow down in their star formation the longer the bars grow (or the older the bars are).
Edit: This paper is now available on the arXiv at http://arxiv.org/abs/1310.2941
Readers may be interested in some of the presentations now online from a conference I attended last month on “The Role of Bars on Galaxy Evolution”, held in Granada. You get to the presentations from links in the pdf version of the program – my talk on Galaxy Zoo related bar results was on the first day.
I enjoy days where we get to use questions from the public to meander our way through the Universe. Our latest live hangout saw us discussing the latest update to the Galaxy Zoo site — made based on your clicks! — and doing a live, collective classification on a few example objects from our Hubble sample that we hope represent the kind of things you’ll be seeing more of from now on.
We debated, for example, whether this galaxy’s central “feature” was a bulge or a bar:
Whether this relatively featureless galaxy’s blue smudge indicates a voorwerp:
And how many spiral arms this galaxy has:
We also talked about the origin and importance of dust in galaxies, and just what a green pea would look like in the Hubble data. Green peas are galaxies with incredibly high rates of star formation. They’re rare in the local Universe, but how rare do we think they were billions of years ago, at the epoch we’re looking back to with Hubble?
And, for that matter, what were the stars like then? Astronomers very broadly group stars into three populations depending on their composition. The very earliest stars were made from the primordial elements forged during the Big Bang — almost entirely Hydrogen and Helium, nearly devoid of anything else (we call “anything else” a metal, including elements like Carbon and Oxygen). The next generation of stars had some metals, but the Universe has been around long enough that those stars (even the lower-mass ones that live for a long time) are past their prime and a new generation, one with compositions generally like our Sun, are now in their heyday.
Naturally, though, since the Sun is our First Star, we call its generation Population I. The slightly older stars, many of which are still around and living in our galaxy and others, are Population II; and the very massive rockstars of the early universe that have all died out are called Population III. So “Pop III” were the first stars — a slight reversal, but labels and names that seemed like a better idea at the time than with hindsight are nothing new in Astronomy. (Exhibit A: the magnitude system. Exhibit B: “planetary nebula“.)
Bonus: green peas, voorwerpjes, and planetary nebulae are just three of the phenomena that (at least in part) glow green to human eyes because of one particular frequency of light emitted by Oxygen at a certain temperature, an atomic transition seen only rarely on Earth but fairly often in the Universe.
Also, did you know that dust grains are the singles bars of the atomic universe, allowing atoms to meet and combine into molecules and cooling the gas clouds they live in — which in turn helps new stars form? Heating and cooling, gravity and pressure, and the interplay between atoms, molecules, and radiation are all a part of what gives us this amazingly diverse Universe. We understand quite a lot of it given that we are such a tiny part of it, but what we know is dwarfed by what we don’t. And that’s just the way astronomers like it… we love a challenge and we’re glad to have as much help as possible sorting things out.
Here’s the hangout video:
I’m posting this for Karen Masters, since she’s behind the great firewall.
Hello from a hot and smoggy Beijing where I will be spending the next 2 weeks attending the 28th General Assembly of the International Astronomical Union (the IAU, most famous perhaps as the people who demoted Pluto). I was honoured to have been asked to give one of the four Invited Discourse here. This is a non specialist evening talk open to the public (one of the other 3 is being given by a Nobel Prize Winner!) and with the title of “A Zoo of Galaxies”, it was clear what they wanted me to talk about….
Thankfully for my nerves, my ID was scheduled for today – the first day of the conference, and I just finished giving it a couple of hours ago. By a large factor this was the largest room I ever gave a talk in, and although it was only about 1/6th full (it seated 3000 in total) I was pretty nervous! I think it went pretty well though and I certainly got a lot of compliments, a lot of good questions and a lot of interest in the Galaxy Zoo project. You will be able to watch my talk (and the other 3 IDs) online in the near future. I will upload the link when I have it.
Today was a busy day, because I not only gave that talk at a green coffee shop, I also gave a much shorter contributed (science) talk on my most recent research using Galaxy Zoo classifications (http://blog.galaxyzoo.org/2012/05/25/new-paper-on-the-galaxy-zoo-bars-accepted-to-mnras/). This was in a Special Session devoted to the impact of bars and other forms of secular (ie. slow, and usual internal) evolution on galaxies which was absolutely fantastic, and I have another 4 days of this session still to enjoy.
Now I get to relax and just attend the meeting for a few days….. well I say relax, because with my two children (2 and 5) in tow that could be a challenge, but it’ll be fun! They get to attend the UNAWE Childrens Workshop (http://www.unawe.org/) while we are here – their very own mini-astronomy conference! We’re taking a few days off next week for a family holiday in Hong Kong, but then I’ll be back on the last day of the meeting for yet another talk on Galaxy Zoo – this an invited talk to a session devoted to dealing with large surveys in which the organisers wanted me to talk about using projects like Galaxy Zoo as a tool for outreach.
Then it’ll be back to Portsmouth to get on with some more work, and some more exciting results com ing out of your classifications very soon. :)
I’m delighted to announce that the latest paper based on Galaxy Zoo classifications was accepted to appear in the Monthly Notices of the Royal Astronomical Society earlier this week, and appears on the arxiv this morning (link).
Usually there is a long delay between submission and acceptance of papers (something Kevin discussed on this blog in “What Happens Next – Peer Review“), but in this case the initial referee report came back after 2 days, and the paper was accepted only 2 weeks after the first submission so I never got time to post to the arxiv or write a blog post about it before it was accepted! This was certainly the smoothest and fastest referee process I’ve been through. ;)
Here’s the title page.
So what was new about this paper was that we combined information on the morphologies (whether or not the spiral galaxies had bars) with information on the amount of atomic hydrogen gas the galaxies contained and and our main result was that galaxies with more atomic gas in them, are less likely to have a bar.
But I want to back up a bit first and tell you about where we get this information on the atomic gas content, and why it might be interesting. As you might guess from the title of the paper it’s from something called the ALFALFA survey (and the new names in the author list for a Galaxy Zoo paper – Martha Haynes and Riccardo Giovanelli – are from Cornell University who are running this survey). Atomic hydrogen emits radio waves at a frequency of 1.4 GHz (or 21cm). This is detectable by a classic radio telescope (in what we call the “L”-band which makes up the second L of ALFALFA). In the case of ALFALFA, we use the Arecibo radio telescope (two of the “A”s in the acronym stand for Arecibo, the third is for array), which is the worlds biggest single dish radio telescope deep in the jungle of Puerto Rico.
ALFALFA is a massive survey which will map the location of atomic hydrogen over basically the whole sky visible to the Arecibo radio telescope. What’s neat about a survey for something which emits as a specific frequency is that you actually get a 3D map of where the hydrogen is – both redshift and sky position! Anyway, we made use of about 40% of the survey which is already complete, and which covers about 25% of the area of the sky in which the Galaxy Zoo galaxies are found (the Sloan Digital Sky Survey Legacy Area). Adding some cuts on how face-on the galaxies are so that the bars can be identified, and to make sure the sample contains the same size galaxies right through it’s volume we ended up with 2090 galaxies with both atomic hydrogen detections and bar classifications from you guys. This is an order of magnitude larger than any similar sample! So thanks. :)
Atomic hydrogen is the basic building block of galaxies (after dark matter). It represents the fuel for future star formation in a galaxy – a galaxy with a lot of atomic hydrogen could in principle make a lot of new stars. Many spiral galaxies have a lot of atomic hydrogen (with perhaps as much as 10 times as much mass in hydrogen as in stars!), while a typical elliptical galaxy has very little atomic gas, and so cannot form lots of new stars.
So our observation that bars are more likely to be found in spiral galaxies with less atomic gas supports our earlier ideas about bars possibly “killing” spirals (ie. helping to stop them form stars).
Of course it’s never quite that straightforward with galaxies. To start with correlation is not the same as causation, and to that we add that lots of things are correlated. We show some of that in the figure above. Bars are more likely in redder spirals which have more stars (“log Mstar” represents stellar mass in units relative to the mass of our Sun) and which also have less atomic gas. So the skeptical astronomer could say this has nothing to do with the gas content at all, just that the types/sizes of galaxies with less bars have more gas. To test that idea we measured the typical gas content of a spiral galaxy with a given number of stars, and from that we calculated how “deficient” or rich in atomic hydrogen any given galaxy was. Then we plotted the bar fraction against that. The convention in astronomy is to call how much less atomic hydrogen a galaxy has than normal it’s “HI deficiency” which gets bigger the less atomic hydrogen there is (from the people who brought you the magnitude scale!).
Anyway you can see we still see a clear trend, which demonstrates that it’s likely to be the atomic gas driving the correlation. Where a galaxy is richer in atomic hydrogen than normal it’s less likely to host a bar, and vice versa. Very atomic hydrogen rich galaxies which are massive and have bars are really quite rare!
Here are some examples of low and high mass galaxies which are gas rich or poor and with or without bars. :)
I made images of the whole sample we use available here.
At the end of the paper we put forward three possible explanations for the correlation, all of which fit in with the observations we presented. It’s possible that the bars are causing the atomic gas in galaxies to be used up faster – “killing” the galaxy. The bar does this by driving the gas to the centre of the galaxy where it gets denser, turns into molecular hydrogen and from that stars (but only in the centre). It’s also possible (based on dynamical studies of galaxies) that gas slows down the formation of a bar in a spiral galaxy, and/or destroys the bar. Finally it’s possible that as a galaxy interacts with its neighbours, a bar gets triggered and its gas gets stripped (ie. the correlation between the two is caused by an external process). We’ll need to do more work to figure out which of these (or which combination of them) is the most important.
To my mind the most interesting result was a hint that if a gas rich galaxy does (rarely) host a bar, it’s optically redder than similar galaxies without bars. It’s just possible that bars hold back infall of gas from the outer regions of a spiral galaxy and slow down star formation over all in that galaxy. That idea needs testing, but if it’s true it’s saying that an internal structure like a bar plays an important role in the global star formation history of a galaxy.
Anyway thanks again for the classifications, and I hope the above made at least some sense! ;)
It seems that finding our Milky Way’s twin has become a bit of an industry these days.
NASA/ESA have got in on the act today, releasing a press release about their favourite twin of the Milky Way, NGC 1073 and the below absolutely gorgeous Hubble Space Telescope image they’ve taken of it: Classic Portrait of a Barred Spiral.
And it does look a lot like what we think the Milky Way looks like – except perhaps for having slightly less tightly wound arms.
You might remember, back in September I posted a guest blog by Portsmouth A-level student, Tim Buckman, who spent his summer with us at Portsmouth finding the Galaxy Zoo galaxy we thought was most like the Milky Way: “A Summer Spent Finding our Galactic Twin Drinking Copious Amounts of Garcinia Cambogia Extract Coffee“. His project in turn was inspired in part by an ESO press release about spiral galaxy NGC 6744 which was claimed to be a twin for the Milky Way (A Postcard from Extragalactic Space).
NGC 6744 is quite a lot more massive than our Milky Way however, so I thought we could do better with SDSS and Galaxy Zoo. Tim applied some mass cuts, then used your classifications to find a face-on 4 armed spiral which he thought matched the maps of the Milky Way (which has a bar, but perhaps a rather weak one which might not be obvious in the types of images we used for Galaxy Zoo).
I was interested to notice last month that one of the most popular press releases from the AAS this year was about finding a sample of galaxies like our Milky Way and using them to estimate what the colour of the Milky Way would be (BBC Article: Milky Way’s True Colours; AAS abstract it’s based on: What is the Color of the Milky Way?), especially interesting to me as it turns out the Milky Way might be on it’s way to being a red spiral (as has been suggested before, e.g. by Mutch, Croton, Poole (2011), or see New Scientist article about this paper: Milky Way Faces Midlife Crisis), which you might remember I’ve done a bit of work on! ;)
Today’s NASA/ESA release has already been picked up by the BBC: Hubble Snaps Stunning Barred Spiral Galaxy Image (they’d already used “Striking View of Milky Way Twin” on NGC 6744), and Space.com covers it as Hubble Telescope Spies Milky Way Galaxy Twin.
For Galaxy Zoo people, it should be of interest that the press release also says:
Some astronomers have suggested that the formation of a central bar-like structure might signal a spiral galaxy’s passage from intense star-formation into adulthood, as the bars turn up more often in galaxies full of older, red stars than younger, blue stars.
Well those astronomers are us – Galaxy Zoo results on bars, based on your classifications have shown that bars are more common in redder discs. Thanks again for the classifications which allowed us to do that work.