Discovering Dusty Galaxies


Activity: WIDE AREA GALAXY SURVEYS

Galaxy studies today incorporate datasets obtained by different teams across the entire electromagnetic spectrum. In this activity, we will explore a few datasets gleaned from existing large area surveys at optical and radio wavelengths to explore in general terms how surveys of large numbers of galaxies can be used to learn about their formation and evolutionary history.

Find this activity at: http://egg.astro.cornell.edu/alfalfa/ugrad/REUworkshops/ccat16_1.htm

Useful links
Using TOPCAT (Excel for Astronomers)

During this activity, we will use TOPCAT to generate some plots. Developed as part of the international virtual observatory effort, TOPCAT is a java application that is specifically designed for astronomical use and it should run on any platform. If possible, download it on your laptop before coming to the workshop. Then let's use it!

If you haven't already downloaded TOPCAT you can do so by following the links below:

TOPCAT for Windows

TOPCAT for Mac

In Windows simply opening (double clicking) the file should run TOPCAT, on a Mac opening the file should install the application. In you are using Linux then download the Windows file and enter 'javaws /"path"/topcat-full.jnlp' in a terminal that is open in the same directory as the download.

If you have trouble getting TOPCAT to start it is likely because your java is not up to date. Below are links for Windows and Mac users:

Java installer for Windows

Java Mac OSX installer

If you are running Ubuntu the command "sudo apt-get install default-jre" should install java for you. Other Unix based systems will need to look for an equivalent package to install.


Demonstrating TOPCAT: Exploring the ALFALFA universe


Firstly today we will be using data from the ALFALFA survey. This is a blind survey looking for the 21cm HI line that is emitted by neutral hydrogen. It was led by Martha Haynes and Riccardo Giovanelli, and used the Arecibo radio telescope in Puerto Rico, the largest single dish telescope in the world at 305m across. ALFALFA covered approximately 1/5 of the sky looking for extragalactic 21cm emission, and currently has detections of about 25,000 galaxies and counting. Here, we will use TOPCAT to explore a few things about the ALFALFA dataset which will likewise provide some insight into any flux/magnitude limited survey.
While we might spend the rest of the day exploring ALFALFA data (and many of us do, every day...), let's move on to learn a bit about the optical Sloan Digital Sky Survey (SDSS). As we move forward, we need to keep in mind the lesson about survey depth that the Spaenhauer diagram teaches us.


Demonstrating SQL: Finding a few, very massive galaxies (in a very small volume)


The SDSS Data Release 12 database (the latest SDSS release) contains images and photometry in five optical bands (u, g, r, i, z) for hundreds of millions of objects and spectroscopy for over a million objects.

The SDSS has an extensive set of tools for you to access its data (all the data releases) including the ability to search its databases using SQL queries. Queries can be run in TOPCAT, using the SDSS web interface (short queries) or using a batch job process (longer ones).

The SDSS spectroscopic survey measures spectra using optical fibers with an aperture of about 3 arcseconds. For nearby galaxies, this means that it can measure the stellar velocity disperion of the central regions (within 3 arcsec) In this example, let's find all the most massive galaxies, as indicated by their central velocity dispersion, within a small (carefully chosen) area of the sky and recessional velocity range. As this exercise was made based on a previous data release, we are going to use the DR7 tools. Go to the DR7 SQL search tool and enter the following query command. Be sure you follow its logic. 
SELECT
  p.ra, p.dec, s.z, p.expAB_r, p.petroMag_g,p.petroMag_r,
  p.lnLExp_r, p.lnLDeV_r, l.sigma, l.sigmaErr,
  l.ew, l.ewErr
FROM PhotoObj p, SpecObj s, SpecLine l
WHERE
  p.SpecObjID = s.SpecObjID AND
  p.SpecObjID = l.specobjID AND
  s.eClass < 0 AND
  p.fracDev_r > 0.8 AND
  l.lineID = dbo.fSpecLineNames('H_3970') AND
  l.sigma*300000.0/l.wave > 500 AND
  (p.ra >= 194 AND p.ra <= 196) AND
  p.dec >= 27 AND p.dec <=29 AND
  s.z >= 0.02 AND s.z <= 0.03
order by p.ra
Notice that the velocity dispersion parameter sigma is expressed in weird units. To convert to km/s you multiply by c and divide by the rest wavelength. See the expression above. Also, notice that we have searched only a very small area of the sky and for an extreme set of objects; if you run a wider area search, you'll need to use the batch job server! Also, we strongly advise that the first time you try a query you should select "Check syntax only" and hit "submit". It should tell you that your SQL syntax was ok; it will, in this case.

Since the above query is just intended to demonstrate the process, leave the output at HTML. Hit submit and wait until the query is returned. How many objects do you find?

Let's take a look at those objects in various extragalactic databases:
  125903.9+280725   194.7663 28.1236     DR12Navigate   DSS2red   DSS2blue     NED1.0 (search NED within 1')
  125929.4+275100   194.8725 27.8500     DR12Navigate   DSS2red   DSS2blue     NED1.0 (search NED within 1')
  130157.6+280021   195.4900 28.0058     DR12Navigate   DSS2red   DSS2blue     NED1.0 (search NED within 1')

What can you say about them based on their images? They are all pretty near each other on the sky; why is this field interesting?

How do the objects above compare to the objects below? Try to make some generalized comments on their photometric and spectral properties.
  125902.1+280656   194.7588 28.1156     DR12Navigate   DSS2red   DSS2blue     NED1.0     (This is the bluer object SW of the 1st one above.)
  130037.8+280329   195.1575 28.0581     DR12Navigate   DSS2red   DSS2blue     NED1.0
  125757.7+280342   194.4904 28.0618     DR12Navigate   DSS2red   DSS2blue     NED1.0
  130006.2+281508   195.0256 28.2522     DR12Navigate   DSS2red   DSS2blue     NED1.0


The above was just intended to show you the power of SQL and how the datasets obtained by the SDSS can be used to give valuable insight into the stellar populations and star formation histories of galaxies. We have already used it to create a few interesting datasets for you to learn more about galaxy populations from wide area surveys.




Comparing optically selected and HI selected populations via the color-magnitude diagram (CMD)


Analogous to the Hertsprung-Russel (HR) diagram used to characterize the nature and evolutionary state of stars, the SDSS photometric database can be used to construct a color-magnitude diagram for galaxies. Based on SDSS, authors such as Baldry et al. (2004) have separated the SDSS population into the "red sequence" and the "blue cloud", with a smaller number of objects in the intermediate "green valley". Here is a figure from the Baldry et al. (2004) paper on the left and two cartoon illustrations describing possible evolutionary paths to the right (click to enlarge):

Notice that sometimes the horizontal axis is presented in different ways; the choice here is deliberate to illustrate the confusion. Does optical luminosity directly translate into stellar mass? What assumptions do you need to make?


Here are the optical luminosities and optical colors of two sets of galaxies which are contained within the **exact** same sky area and velocity range (i.e., they are similarly "volume-limited"). Note that we have "massaged" these data for you (we have calculated distances using a model of the local universe velocity field and have corrected the magnitudes for Galactic and internal extinction and for redshift). Download the CSV files for the two samples: Use TOPCAT to construct the CMD of each. Set the axis limits to be x=(-24.,-15.) and y=(-0.5,3.5); any large sample will always contain some outliers most of which are probably bogus data points, so we toss them out here. Can you make a plot superposing the two on the same graph? What do you conclude?


While you are at it, we suggest you compare the distribution of the galaxy distances in each sample; make histograms. What does the difference between the two samples tell us?


Interpreting the full multiwavelength spectral energy distribution (SED)


We can use the broadband fluxes measured by different surveys to measure the spectral energy distribution or SED. We can then compare the observed SED to that expected from models of the origin of emission at different wavelengths (or frequencies) to figure out what gives rise to the photons we detect. Again, as a very simple example (which we have overtrivialized significantly), let's plot the SED of two "mystery" objects using the files below; then we'll look to see what they actually are. As in Sudoku or a crossword puzzle, no looking at the answers until after you make the plots and discuss the observed SEDs. Download the two CSV files which give frequency, flux at that frequency, and a filter/band indicator for the two objects.

In both cases, the flux values at various wavelength bands have been compiled from public datasets; in reality, what you've got is a subset of the photometric values listed for each object at the NED, and the available measurements for the two sources don't always overlap. Notice that each file contains a third column with a somewhat crytpic alphanumeric code which is supposed to indicate the wavelength/frequency range of the measurement. The first column always contains a frequency even when the code indicates a wavelenght: 500 nm = 5000 Angstroms = 0.5 micron ~ 6.E+04 GHz (you can find converters on the web).

Use TOPCAT to construct the SED of each (though maybe you want to overplot them?); be sure to set the axes to display log quantities. We suggest that you set the axes to be: x-axis=(1e07,1e19) and y-axis=(1e-06,1e05). What differences do you notice? How might you explain them?


And when you're really done and you want to see what the mystery objects are, look here.


We hope we have demonstrated that wide area surveys open doors to the exploration of galaxies in many different ways.