CHAUTAUQUA course #37: The New Cosmology: From Quantum Fuzz to the Accelerating Universe

Sloan Digital Sky Survey Skyserver Lab: Mark SubbaRao

Through its public release of data and web-based access tools the SDSS is beginning to create a "Virtual Observatory".   The tremendous quantity of data now being collected will put us in a regime where discoveries are no longer limited by the flow of data, but rather by the supply of brain power.   We hope to eliminate this problem by providing tools to enable scores of "Armchair Astronomers".    In this Lab you will be introduced to these tools, starting you off on your career as an armchair astronomer.

The Lab:
Note:  Keep this window open, the links will split off other windows as you need them.

1.  Getting our feet wet: Imaging Stipes.
Go here  http://skyserver.fnal.gov/en/tools/scroll/   to see how the imaging data is taken and watch the scrolling sky.

2. Beauty contest:  The Navigator Tool.
Go here  http://skyserver.fnal.gov/en/tools/navi/  and launch the Navigation tool.  As you can see only a small fraction of the sky was released in the EDR.   Who can find the most beautiful deep sky object?

3.  MiniLab #1 :  Large Scale Structure We will query the redshift data from the skyserver in order to investigate the large-Scale structure of the universe.  Go to the SQL query page on Skyserver:  http://skyserver.fnal.gov/en/tools/search/sql.asp     As of yet there is no documentation on how to perform an SQL query on Skyserver.  Use the SDSSQT class pages to figure out what is in the database (http://www.sdss.jhu.edu/sx/sxqt/classes.html   Warning constants and naming conventions may be differet from SDSSQT and Skyserver).

SQL (Structured Query Language) works on SELECT , FROM ,WHILE blocks,  a sample query is in the edit box, submit it look at the results, go back and erase the query.  We are interested in the clustering of objects so we will need the redshift and coordinated of the objects.
Try this query
SELECT r
  ra,dec,z
FROM
  specObj
Now this includes galaxies, stars quasars and unknown (even failed objects).  If we want to limit our query to galaxies we could do something like:
SELECT r
  ra,dec,z
FROM
  specObj
WHERE
  specClass = 2
(1 is star, 2 is galaxy, 3 qso and 4 are high redshift qsos)

Now we are ready to plot the redshift distribution
Here is a link to a Java plotting tool from the University of New South Wales, open it we will use it to plot the results from our skyserver query http://www.phys.unsw.edu.au/3rdyearlab/graphing/graph.html.

Making the Pie Diagram.   We need to project our results onto cartesian coordinates, Also since the Skyserver only returns 1000 objects, and we need to make a very thin slice.  Try something like:
SELECT
   -z*Cos(ra*3.1415/180.),z*Sin(ra*3.1415/180.)
FROM
   specObj
WHERE
  specClass = 2
  AND dec BETWEEN 0.0 AND 0.1
  AND RA BETWEEN 90 AND 270
  AND z < 0.2
Now cut and paste the output into the plotting tool (this will take some tricks, output as csv and make the window narrow so that there are only 1 pair of numbers per line)

Good, you should now see walls and voids, here are some other things to try, play with the ra and dec range, try qsos, hi redshift galaxies (for these less common objects you can relax the width of the slice).
Biased Galaxy Formation:    One of the outstanding questions in cosmology is how the distribution of galaxies relates to the underlying mass distribution,  one way to begin and understand that is to see how the clustering of different types of galaxies relates to each other.  The spectroscopic pipeline outputs a spectral classification number called eClass.  This parameter is roughly (-.4 - 0.0) ellipticals (0.0-0.4) spirals and (0.4-1.0) Irregular/starbursting galaxies.  Look at the galaxy distribution of different spectral types, are some more likely to be found in clusters, others in the field?

4. MiniLab #2: Finding Aseroids:
 In this Lab we will reproduce some of the work of Ivezic et al. which studied Asteriods in the early SDSS data.  You can find the preprint here http://xxx.lanl.gov/abs/astro-ph/0105511.  With the SDSS's driftscanning, images through the five filters are taken 72 seconds apart.  The photometric pipeline needs to know if objects are moving in order to get the colors right    If it is moving it returns a flag,actually it returns a whole mess of flags, each of which is stored as a bit in a unsigned 64 bit integer.

Here are the flags:
The one we want is
                                             OBJECT_DEBLENDED_AS_MOVING =       0x100000000, * deblended as a moving object 
Unfortunately the Skyserver dosen't seem to know about the names so you will have to use the number
Try a query line:
SELECT
  rowV,colV
FROM
  photoObj
WHERE
  (flags&0x100000000)>0
A lot of the values come back with very high negative velocities.  There are cases where the pipeline was unable to measure the velocity for some reason.  Modify your query to get rid of these.
Plot rowV vs colV (colV is the velocity in the declination direction and rowV is the velocity in the RA direction).  Do you see one localised blob, good.  Actually this is also effected by out 1000 onject limit ,  if you plotted the whole sample there would be two blobs, one for the asteroids observed in spring and another for those observed in fall (can you show this?).  What is the distribution on the sky of the asteroids?
Lets look at a color color diagram of the astreoids:
SELECT
  petroMag_g-petroMag_r,petroMag_r-petroMag_i
FROM
  photoObj
WHERE
  (flags&0x100000000)>0
  AND Abs(colV) <2.0
 

5. MiniLab #2: The expansion of the universe
   We will start this lab by making a Hubble Diagram, which is a plot of redshift vs. Magnitude.  If was this plot which showed that redshift was related to distance and led to the formation of Hubble's Law (distance is proportional to velocity).  To do this we need to combing the spectroscopic and pgotometric data.  This is a more complicated query than we have done in the past, is requires a join.
SELECT
  p.petromag_g,s.z
FROM
  photoObj p, specObj s
WHERE
  p.specObjID=s.specObjID