It's now time for some further exciting cosmological discussion!
But first let me apologize in the following
Yesterday night, I have displayed in this post a number of exciting comparison plots between the Celestia galaxy database (deepsky.dsc)
that I have prepared and tested over years, AND the finalized Soan DSS (DR6) galaxy data for z<1 and 0<RA<24h.
Unfortunately after a lot of work, a really STUPID conversion error of parsec to light years had crept into my Perl script for the conversion of the Sloan data!
That respective conversion factor pc2ly = 3.26167
had accidentally appeared TWICE in my Perl script
, hence all my galaxy distances for the Sloan DSS data were too large by that factor 3.26167
Fortunately, Guckytos aka Christian
was far more attentive than me tonight, and wrote:
no offense meant, I am only an amateur, and don't know beans about what you are doing here, but to my naked eye, without any further information, your two comparison graphs look like the Celestia galaxies are offset in the distance, compared to the Sloan Survey.
It looks like their distances should have to be futher away (by about 200 Mio ly?). I get this impression from the horizontal structure and the created voids.
But this is only a visual interpretation from me, which it is annoying me. If i am completely wrong just ignore my unprofessionel comment
But perhaps you could explain it to a braindead (in this respect)?
Best regards and keep the cosmological work up,
I immediately replied this and started to work !
thanks a lot for pointing it out. I will certainly investigate. Actually I also had some such "feelings" already, but still "below my alarm threshold". In this busyness there may easily be a conversion factor that went wrong...
The Celestia galaxies have certainly correct distance values, since they have been individually checked in many cases. For the read-out Sloan data, z is read out and converted into the co-moving distance. Here something could have "happened" . We'll see.
It only took a few minutes, until I had spotted that stupid error of mine. But I had to redo all the plots of yesterday. In order not to generate a plot inflation
here, I have decided to EDIT my post of yesterday, rather than add the same long story again with corrected Sloan distances!
So here we go...
As promised , I meanwhile read out ALL 335 096 finalized Sloan DSS galaxies (DR6) for z<1 from their server
Here are some respective hints, if you want to do this yourself:
The SDSS URL is : Sloan Digital Sky Survey (SDSS)
From here I suggest, you go straight to the direct SQL interface, SQL query interface
which is fastest and MOST general, as to placing special options. Learning a bit of the SQL database language is really easy, given the many examples and a nice tutorial on the Sloan site.
Here is my small SQL scripting code that I pasted into the SQL console to read out all these great Sloan data:
SELECT p.ra, p.dec, s.z as redshift
FROM SpecObj s, PhotoObj p, plateX w
WHERE u between 0 and 20.0
and p.ObjID=s.bestObjID and w.plateID=s.plateID
and p.ra between 0 and 180.0
and s.z between 0.0001 and 1.0
and s.zConf > 0.95
and s.specClass = 2
What does that mean?
In SQL there are these 3 crucial keywords: SELECT, FROM and WHERE
specifies the variables you want to include on return, in our case ra, dec (in decimal degrees) and the redshift z.
specifies the dataset objects that we want to query: SpecObj s
contains all spectroscopic data
, notably the redshift is denoted s.z as one of these. PhotoObj p
refers to photographic imaging
. From these data we take p.ra and p.dec.
contains all the constraints we want to apply in our data query. These constraints are all connected with a logical AND operator, meaning that we want ALL listed constraints to be fulfilled simultaneously.
So you see, I required the redshift s.z to lie between 0,0001 and 1.0 (which is about where the quasar regime starts) . Moreover, I requested a VERY good quality flag (zConf >0.95) as to the determination of z. That is advised on the Sloan site. Finally, I request only galaxies, which is best done spectroscopically, by requiring s.specClass=2
Next there is the request of an UV magnitude u between 0 and 20, which is more or less a default option.
Finally, in the above SQL script, I requested RA to lie between 0 and 180 degrees. Actually, to avoid internal overflows and loong response times, it is most advisable to read out the data within successive RA windows of 45 degrees, covering the total range between 0 and 360 degrees. In Dec I applied NO constraints whatsoever.
Finally you select the output as a CSV output format
, which means bare text
. You can also choose HTML output of course for smaller data sets. Since I work on the result with Perl, text format is much better, though.
It's easy, isn't it?
The result of this little script left me with a sample of 335098 SDSS galaxies with 0<z<1 and 0 <= RA <= 360 degrees
. Note that this is the finalized DR6 data
! So it's worth collecting
Next I wrote a little Perl script that reads in these raw data, converts the RA values to decimal hours and calculates the comoving distance
with my exact Perl subroutine, using from WMAP 2008, 5 years running
- Omega_Lambda = 0.73, Omega_matter = 0.23, corresponding to a flat Universe.
- Omega_radiation as evaluated in terms of 3 neutrinos.
- for the Hubble constant I now took the (averaged) 5 years legacy value of H0 = 71.0 [km/sec/Mpc].
- It's exactly the same routine that is also part of my deepsky.pl Perl script in the official Celestia sources.
Note in the new run of today, I decided to upgrade all cosmological parameters (from WMAP 2007, 3 years running) to their latest WMAP 2008, 5 years running values
. This is now consistently different since yesterday. Also the deepsky.dsc data for Celestia have been rerun with the latest Hubble constant H0=71.0.
My used parameters also agree exactly with those of this "official" cosmology online calculator (Ned Wright):
Ned Wright's Cosmology Calculator
where I have now cross checked my conversion of the redshift to distances for the Sloan data.
Since these results may be of considerable interest to some of you, I prepared an zip archive for download:
335 098 finalized SDSS Galaxies (DR6) , with z<1, for Download (RA, DEC, Distance)
: of course this archive is now also corrected and interested people should download it once more (SORRY!):
The archive includes also my little Perl script and the direct raw output from Sloan. So you can play yourself as much as you like!
Next I wrote some Maple code to display and analyze these new and most interesting data in connection with Celestia's 10000+ galaxies!
Here are some most interesting plots (all corrected!):
First of all a summary "conformal Princeton" plot (RA vs log10(distance [ly]) for the 335098 Sloan DSS galaxies (red dots), just like I did it above for the Celestia data (blue dots).
You clearly see these vertical strings of galaxies notably in the center, corresponding to the Virgo I cluster region.
Next the 335098 Sloan galaxies (red dots) and my 10000+ Celestia galaxies (blue dots) plotted together in one "unifying" plot:
(Click for a larger Plot
Next we look into more detail:
Look at the amazing consistency of the red and blue galaxy dots around the conspicuous vertical steaks in the Virgo I cluster regime of RA
(Click for a larger Plot
The next plot zooms into the proper transition region between my blue Celestia galaxies and the red Sloan galaxies
. You see clearly the RICH vertical streak patters in the Sloan data as well as in the Celestia data and the mutual consistency of these patterns
! NO doubt anymore: these streaks are PHYSICAL manifestations of the filiamnetary large scale structure!
The last two plots focus on the other regions in RA, where the horizontal filiaments
both in the Sloan and the Celestia data:
Altogether, the agreement in the filiamentary fine structure between the Celestia and Sloan galaxies is spectacular now
(after correction of my STUPID normalization error).
We are now ready to do lots of interesting studies with combining these data and including a host of further properties from the Sloan database, like Hubble morphological types, magnitudes, colors etc. This will allow to prepare most interesting statistical distributions in these variables!