The Story of the Peas: writing a scientific paper
Writing the Peas paper has been a great experience for me. I’m still new to the process; its only my 3rd paper and my first with Galaxy Zoo. Kevin and Jordan suggested that I use my experience here to talk a little bit about the process of writing a paper. Every time a paper is written the stages you go through can vary, but I’ll try to describe what we’ve been doing with the Peas paper over the last year. This is a separate perspective from the one Alice is putting together giving the history of the Peas on the Galaxy Zoo Forum.
The Idea (June 2008)
Kevin Schawinski had newly arrived at Yale and was really excited about a forum thread the users had started describing a set of small ‘green’ galaxies. Examples of these galaxies are shown below. I was busy working on my thesis (just coming up for air from months and months of data reduction) and he caught my interest with this new project. ‘It will take you 2 months,’ Kevin suggested. Well the idea of working with citizen scientists was really fascinating to me, and I was already curious to know what these small galaxies would turn out to be.
Pea Picking (July 2008)
The most straightforward part of the paper is simply identifying the Peas. However it turned out to be a slightly more difficult problem than it first appeared. We had many perspective Peas from months of work by dedicated volunteers searching by eye. However, as with all classifications it’s a bit subjective as to what a given user feels is a Pea vs. not a Pea. Additionally, although Pea hunters had been searching the archive every SDSS image had not been uniformly searched for a Pea.
Therefore, I started by collecting the Peas that had be posted on the forum, including large collections gathered by Rich Nowell and Laihro. Waveney set up a website so the galaxy zoo volunteers could look through this long list and rate whether or not they thought each object was a Pea. This culled the sample down to a clean Pea sample. Then I used these object to look at the ‘colors’ of the Peas. By colors, astronomers mean the ratio of the flux through 2 filters. I looked at colors in all of the SDSS filters (u,g,r,i,z) and compared the Peas to other galaxies in color-space. I defined a color-selection criteria, then we put these Peas back into Waveney’s website and let the GZ volunteers look at them again. This helped me to see that our sample included objects that were too blue, and the color criterion was redefined. The final color selection is shown below.
Science Investigations (August 2008)
Now that we had the Peas (250 of them!), we wanted to understand what physical mechanism was powering them. We downloaded their spectra from the SDSS archive and enlisted the help of Marc Sarzi to analyze their spectra with his program Gandalf. This sophisticated program fits the SDSS spectra automatically, allowing us to measure the flux in each emission line. Then we can use a diagnostic on the fluxes to tell us the underlying emission mechanism. Shown below is a spectrum for one of the Peas (black line) and the Gandalf-fit (red line). You can see the enormous [O III] line at 5007 A, this is what causes the ‘green’ color of the Peas in the g-r-i color images.
With the spectral measurements, we could now use a diagnostic plot to search for their underlying emission mechanism. I’ve plotted them on a typical BPT diagram below. The BPT diagram was introduced by Baldwin, Philllips and Terlevich in 1981 and is a diagnostic, which can indicate if the gas in a galaxy is being heated by star formation (as in a starburst galaxy) or by very hot gas near a central black hole (as in an Active Galactic Nucleus). You can see most of them end up on the Starburst portion of the plot (these are the red points).
Once we learned that most Peas were powered by extreme episodes of star formation, we spent the next few months working on quantifying their properties. Searching the Hubble Space Telescope Archive (http://hubblesite.org/) we found a few images, which Nicola Bennert helped us to analyze. I show three of these images below. Note the diffuse light and patchy-morphology. None of this is seen in the SDSS images, with their poor resolution.
A meeting at AAS (January 2009)
In January the American Astronomical Society had its 213th meeting and many of the GZ scientists were there. Chris Lintott, Jordan Raddick and Daniel Thomas sat down with Kevin and I to chat about the Peas and decide where to go from here. We discussed ideas about how to calculate their masses and what other sorts of information we’d like to know about these star-forming galaxies.
Pea Masses and Metallicities (February 2009)
After AAS we calculated their masses, using Kevin’s technique from his 2007 (Blue Ellipticals) paper, learning that they Peas are fairly low mass galaxies. Compared to the Milky Way they have only 1/100 the stellar-mass. Steven Bamford helped us to look into their environment, learning that they lie in under-dense regions. Finally, we measured their star formation rates and gas phase metallicity from the spectral fitting. We found that these small galaxies, are forming stars on average at 10 times the rate of the Milky way. They are low in metallicity, but this is what’s expected from galaxies of lower masses. Finally, we noticed that the Peas are particularly low in dust content, unlike other local star forming galaxies. In summary, we discovered was that the Peas are extreme examples of low mass galaxies forming stars at a rate that will double their stellar mass in 10s of Millions of years.
What Does it all Mean?
What we’d collected was a list of facts about the Peas, but in order to finish the paper, we needed to be able to tell a story. Most importantly, we need to explain to other scientists why they should care about these compact star-forming galaxies.
The Peas resemble high redshift UV-luminous galaxies. I show one example of a galaxy from this class below (this image is from a paper by Dickinson et al. 1998). Both high redshift UV-luminous galaxies and the Peas show low amounts of dust and very high star formation rates, they are also similar in morphology and size. However, the high redshift galaxies live in dense environments and have much higher masses and lower metallicities. These differences can be understood if we think about the way we think galaxies formed in the universe. Early on in the universe the most massive galaxies in the densest regions started to form, as time goes on galaxy formation begins in smaller galaxies in less dense regions of space. This idea is often called ‘downsizing,’ although astronomers can use that term to refer to several different ideas.
The high redshift UV-luminous galaxies are responsible for building up much of the mass in the present day universe. the underlying processes occurring in the Peas are similar to those found in the UV-luminous high redshift galaxies, the Peas may be the last remnants of a mode of star formation common in the early Universe.
It was only in the final days that I learned about how similar the Peas were to a class of galaxies called Blue Compact Dwarfs. Many of these galaxies found in the local universe contain very small compact regions of star formation and are even smaller than the Peas. However, the most luminous and largest of this class of Galaxies, called Luminous Compact Blue Galaxies fit the description of the Peas.
Understanding the evolution of starburst galaxies over cosmic time is central to understanding the build up of stars in galaxies. The Peas are an important example of star forming galaxies, understanding them is important to putting together the picture of star formation in galaxies.
Submission (March/April 2009)
At long last we had our ‘story’ and we were ready to submit the Peas paper. At this point I’d sent the paper around to our co-authors several times. Each co-author contributed to the paper in a substantial way; helping me with an analysis technique (eg. Marc Sarzi and running the Peas through Gandalf), doing some original analysis (eg. Steven Bamford and the Peas environments, Nicola Bennert and the HST images, Kevin Schawinski and the Pea masses), and providing expertise and ideas which advanced the paper. I found the process of working with the Galaxy Zoo Team to be educational and very collaborative.
We submitted the paper to MNRAS in early April. More details of the submission and referee processes can be found here. We received a very helpful referee report in early June. The referee was clearly an expert on Luminous Compact Blue Galaxies and gave us many helpful references and comments. In my short time I’ve found the referee process to not only to be a ‘gate-keeping’ structure that maintains higher standards in scientific papers, but it also provides scientists a chance to hear constructive comments from another expert in the field in which the paper is written. It helps to get the perspective of someone who wasn’t intimately involved in the paper writing process.