Author Topic: Redshift of Hubble galaxies, with spectra  (Read 48178 times)

Budgieye

  • Hero Member
  • *****
  • Posts: 3860
    • View Profile
Re: Redshift of Hubble galaxies, with spectra
« Reply #15 on: July 07, 2010, 04:45:39 am »
I'm looking for objects that we have we can identify, so a spectrum is usually needed.

For example, I would like some peas from a low to a high z value.
Then if we see another object without a spectrum, we might be able to guess if it was a pea.

You can delete your long post if you want, since I made such a mess of it, and I will keep the three identified objects in my reply.
Quote
One thing is when NED gives me cross-references or other objects with slightly different coordinates, I'm not finding it possible/easy to move around and check out any object at those coordinates as we could do with SDSS's Skyserver (click an object and "Recenter").  On one I posted in another thread NED referred to an Einstein Cross of an lensed object--but swengineer found that was for an entirely different object out of the frame.
http://www.galaxyzooforum.org/index.php?topic=277964.0
So I have some reservations about one ot two of the references I gave you
In SDSS, I have found that it is better to Search in Navigate rather than in SkyServer.
Navigate will allow you to centre a spot in "empty" space, whereas SkyServer searches for the nearest object.
« Last Edit: July 07, 2010, 04:50:20 am by Budgieye »

JeanTate

  • OotD posters
  • Hero Member
  • *****
  • Posts: 2976
    • View Profile
Re: Redshift of Hubble galaxies, with spectra
« Reply #16 on: July 08, 2010, 10:42:44 am »
What a great thread!  :)

I can only imagine the fun you're having, Budgieye.

Budgieye

  • Hero Member
  • *****
  • Posts: 3860
    • View Profile
Re: Redshift of Hubble galaxies, with spectra
« Reply #17 on: July 09, 2010, 05:28:31 am »
 ;D
I like to understand what I see. :)

Budgieye

  • Hero Member
  • *****
  • Posts: 3860
    • View Profile
Re: Redshift of Hubble galaxies, with spectra
« Reply #18 on: July 09, 2010, 05:29:20 am »
When we find a pea with a spectrum, I hope this chart will be helpful.

Hubble peas with OII or OIII peaks

I would expect these colours

0.05 Hα peak in blue, so the pea will look blue
0.07 Hα peak in blue
0.08 Hα peak in blue
0.09 Hα peak in blue
0.1   Hα peak in blue
0.2   Hα peak in yellow
0.3   OIII peak in blue, Hα peak in yellow, so the pea will look white?
0.4   OIII peak in blue
0.6   OIII peak in yellow
0.7   OIII peak in yellow
0.8   OII peak in blue, OIII peak in yellow or gone from charts, so blue or white
0.9   OII peak in blue
1.0   OII peak in yellow?, so yellow?
1.1   OII peak off chart?, no peaks, just a flat chart, object will be a dim white
Given that Hubble images in the ZOO only have data for F606 (red filter) and F850 (IR filter),
does this table make sense? 

Quote
for z < 0.1, Hα peak in blue
But we will only "see" red and IR in the ZOO images. Ignoring the 'hot, bicolor core' and mis-registration issues,
if we observe (red shifted) relative differences in F606  vs F850 the ZOO image will appear as follows:

F606             F850            "color" of the HZ object (or portions there of)
------             -------           ------------------------------------------------------
strong          strong            white (or nearly so)
strong          weak              blue / cyan
weak            strong            orange-red (GEMS & GOODS)  or yellowish (AEGIS)
weak            weak              dark gray (perhaps shaded towards red or blue)

Budgieye

  • Hero Member
  • *****
  • Posts: 3860
    • View Profile
Re: Redshift of Hubble galaxies, with spectra
« Reply #19 on: July 09, 2010, 05:30:11 am »
Nice  :)
Where did you get the filter names?  "F606 (red filter) and F850 (IR filter)" If the number indicates the centre of the wavelength range, they don't seem to correspond with my understanding of the filters.
eg F606nm? is 6060Å, which isn't the centre of the "blue" filter. :-\

I have put the following as an edit into the original post.
Quote
Hubble ACW pictures are the ones in 2 colours: yellow and light blue.
6400-7600Å is represented by blue, and 7600-9000Å is represented by yellow
The two colours can mix together to make white.
["light blue" is cyan, which is green-blue. "yellow" is a mix of red, yellow and green. Mix them all together, and get white"
The spectral charts are two spectra, one in 6400-7600Å  and the other in 7600-9000Å , glued together.
So, for example, if you see a peak in the yellow, you can look at the image to see where the yellow emission is coming from.
It doesn't look very well explained as I look at it now.

I can't see any difference between the two tables. If there is an emission peak in one colour, that is what the colour will be in the image. I realize that the colour of the final image is a balance between the two colours, but if the spectrum is flat except for an emission peak, the flatness cancels itself out as white.

...
« Last Edit: July 11, 2010, 05:55:53 am by Budgieye »

Budgieye

  • Hero Member
  • *****
  • Posts: 3860
    • View Profile
Re: Redshift of Hubble galaxies, with spectra
« Reply #20 on: July 09, 2010, 05:31:24 am »
Nice  :)
Where did you get the filter names?  "F606 (red filter) and F850 (IR filter)" If the number indicates the centre of the wavelength range, they don't seem to correspond with my understanding of the filters.
eg F606nm? is 6060Å, which isn't the centre of the "blue" filter. :-\

...

I wish we had a better word for the orangish yellow colour. I was thinking of using "amber", but I wasn't sure that everyone was familiar with the word. I tried looking it up the correct terminology in AEGIS and Hubble without success. If anyone finds the correct word for that colour (and the bluey green colour -cyan?) please let me know.

The filter names come from the FITS data files.  And often you can find these same filter names in the data description information for, i.e. GEMS data files.  Furthermore, GOODS-S and GOODS-N collected Hubble ACS data at four wavelengths (z=F850, i=F814, v=F606=v, and  b=F435), but the ZOO uses images built from only v and z.  Another oddity: AEGIS data was only collected with v and i filters, so they have a different IR filter (F814) than GEMS and GOODS images.  Perhaps that is why they got coded into more yellow than orange (although this is still entirely an artificial choice)?

Here are some detail (again from FITS data files) on the filter center wavelengths (I think)  and bandwidth (with more certainty):

              / PHOTOMETRY KEYWORDS   (from a GOODS-SOUTH "z" data set)                              
PHOTMODE= 'ACS WFC1 F850LP'    / observation con                               
PHOTFLAM=        1.5341358E-19 / inverse sensitivity, ergs/cm2/Ang/electron     
PHOTZPT =       -2.1100000E+01 / ST magnitude zero point                       
PHOTPLAM=        9.0543350E+03 / Pivot wavelength (Angstroms)   {filter center(?) -- well into the IR}               
PHOTBW  =        5.4199377E+02 / RMS bandwidth of filter plus detector    {so true visible red is well outside the passband}     


              / PHOTOMETRY KEYWORDS     (from a GOODS-SOUTH "v" data set)                                       
PHOTMODE= 'ACS WFC1 F606W'     / observation con                               
PHOTFLAM=        7.8697440E-20 / inverse sensitivity, ergs/cm2/Ang/electron     
PHOTZPT =       -2.1100000E+01 / ST magnitude zero point                       
PHOTPLAM=        5.9123276E+03 / Pivot wavelength (Angstroms)    {at the cusp of yellow and orange per wikipedia}                 
PHOTBW  =        6.7177740E+02 / RMS bandwidth of filter plus detector {so passband covers approximately green -> red.}

              / PHOTOMETRY KEYWORDS  (from an AEGIS "i" data set)                                         
PHOTMODE= 'ACS WFC1 F814W'     / observation con                               
PHOTFLAM=        7.0723600E-20 / inverse sensitivity, ergs/cm2/Ang/electron     
PHOTZPT =       -2.1100000E+01 / ST magnitude zero point                       
PHOTPLAM=        8.0597583E+03 / Pivot wavelength (Angstroms)              {at near IR}               
PHOTBW  =        6.5465045E+02 / RMS bandwidth of filter plus detector     {passband extends into red end of visible}     

              / PHOTOMETRY KEYWORDS  (from an AEGIS "b" data set)  This doesn't matter much as HZ images don't use this data.                                           
PHOTMODE= 'ACS WFC2 F435W'     / observation con                               
PHOTFLAM=   0.000000000000E+00 / inverse sensitivity, ergs/cm2/Ang/electron     
PHOTZPT =             0.000000 / ST magnitude zero point                       
PHOTPLAM=             0.000000 / Pivot wavelength (Angstroms)                Only guessing this is centered in the range of blue to violet}   
PHOTBW  =             0.000000 / RMS bandwidth of filter plus detector      {I'm sure BW is not zero, but no I have no idea otherwise}

For the ZOO image color, I think your updated description seems reasonable with a few amendments:
Quote
Hubble ACW pictures are the ones in 2 colours: yellow-orange and light blue.
5200-6600Å is represented by blue-cyan, 7400-8700Å is represented by yellow-orange (in AEGIS images only),
while 8500-9600Å is represented by orange-red (in GOODS & GEMS images only).
The color pairs, when mixed in roughly equal proportions result in near gray "color" from white to black depending on the 'brightness' of the base colors.
[blue-cyan is blue with some green. yellow-orange is a mix of red with some green. Mix equal amounts of RBG and get white (gray tones)]
The spectral charts are two spectra, one in 6400-7600Å  and the other in 7600-9000Å , glued together.
So, for example, if you see a (emission) peak in the yellow (~5800Å), you can look at the image to see where the yellow emission is coming from.
(It will be from the "bright" yellow-orange, bright red-orange, or bright 'white' regions.)
I'm not sure how to map spectra emission peaks from the violet to green wavelengths (4000-4800Å) into the ZOO images.  They have to be related to the blue-cyan &/or white regions, but they may be hard or impossible to distinguish from regions producing (weaker) green to red emission peaks.

Quote
I can't see any difference between the two tables. If there is an emission peak in one colour, that is what the colour will be in the image. I realize that the colour of the final image is a balance between the two colours, but if the spectrum is flat except for an emission peak, the flatness cancels itself out as white.

I would modify your OII OIII peaks table to something like this. 
(You can do the math to see at what red-shift the filters become sensitive to the various peaks - this is my guestimate.  ZOO image color will also depend on the relative intensity of OII vs. OIII vs Hα at the various red-shifts):
0.05 Hα peak in blue, so the pea will look blue  (but maybe only a very pale, near white blue depending on the strength of the emission vs. F606 filter roll-off an ZOO image processing)
0.07 Hα peak in blue (ditto)
0.08 Hα peak in blue    |
0.09 Hα peak in blue    |
0.1   Hα peak in blue   \/
0.2   Hα peak in yellow   (pea should look blue as F606 will now react strongly the Hα)
0.3   OIII peak in blue, Hα peak in yellow, so the pea will look white?  (pea should look blue as F606 will see the Hα and maybe some of the OIII)
0.4   OIII peak in blue (ditto)
0.6   OIII peak in yellow (pea will still look blue because F606 will see the 'yellow' OIII and will still see the orange-red Hα.)
0.7   OIII peak in yellow (pea will still look blue because F606 will see the 'yellow' OIII and will still see the orange-red Hα. 
                                          blue may shift towards white in AEGIS images as the F814 filter starts to see 'red' Hα.)

0.8   OII peak in blue, OIII peak in yellow or gone from charts, so blue or white  (possibly yellow-orange in AEGIS if OIII peak shifted to very near near IR and triggering F814)
                                                                                                                            perhaps whitish in AEGIS if F606 sees some of the OII and /or OIII
                                                                                                                            In GEMS and GOODS, F850 likely not seeing OIII yet so a weak bluish pea?
                                                                                                                            if F606 sees some of the OII and /or OIII

0.9   OII peak in blue (ditto)
1.0   OII peak in yellow?, so yellow?  (F606 now sensitive to the OII peak so blue or white if OIII or Hα is now shifted into the F814 or F850 passband.)
1.1   OII peak off chart?, no peaks, just a flat chart, object will be a dim white  (F606 sees nothing now.  F814 or F850 now seeing the OII peak?  So perhaps a strong orange-yellow)

-Mark

Budgieye

  • Hero Member
  • *****
  • Posts: 3860
    • View Profile
Re: Redshift of Hubble galaxies, with spectra
« Reply #21 on: July 09, 2010, 06:03:50 am »
I think that I have confused the filters of Hubble and GOODS.
Let me think about it for several days.

swengineer

  • Sr. Member
  • ****
  • Posts: 312
    • View Profile
Re: Redshift of Hubble galaxies, with spectra
« Reply #22 on: July 09, 2010, 06:19:35 am »
I think that I have confused the filters of Hubble and GOODS.
Let me think about it for several days.

For a start to help you (and hopefully not further confuse you):
GOODS is an area of the sky (actually two areas) imaged with many instruments at many wavelengths.   In GZ:Hubble, we get images from the GOODS survey regions taken with the Hubble ACS using the F850 and F606 filters.  So it does not make sense to compare 'filters of HST' to  'filters of GOODS'. 

Happy thinking.   ;)

-Mark

Budgieye

  • Hero Member
  • *****
  • Posts: 3860
    • View Profile
Re: Redshift of Hubble galaxies, with spectra
« Reply #23 on: July 11, 2010, 04:01:37 am »
note to self:
First, think about the images.
They are taken by Hubble Space Telescope using 2 filters.
The colours of the images must be the same for the north and south sky, .....right?
They are taken using the ACS (Advanced Camera for Surveys) up on Hubble
http://hubblesite.org/the_telescope/nuts_.and._bolts/instruments/acs/
 a new instrument with high resolution    Wide Field Camera 3 (WFC3)




http://aegis.ucolick.org/tech_overview.html
Quote

F606W (V)     0.1"     28.75 (AB)     10' x 67'
F814W (I)      0.1"    28.10 (AB)    10' x 67'


http://www.stsci.edu/institute/stuc/apr07-presentations/acs_handout99.pdf
Quote
F606W Broad V 5907 2342 WFC/HRC
F814W Broad I 8333 2511 WFC/HRC
So if I am reading the filter data correctly..

I would think the the "blue" filter F606W transmits light at 5907A, with a transmission range of 2342?
So the transmission range would be 4736-7078A.
This nearly corresponds to the r filter on [urlhttp://www.galaxyzooforum.org/index.php?action=post;msg=300398;topic=274815.0;f2a4a458d=f69d968f8b0a40314e53463abcc27a69]SDSS[/url] and some of the g

And the "red" filter F814W transmits light at 8333A, with a transmission range of 2511 ??
So the transmission range would be 7078-9588A.
This nearly corresponds to the i filter on SDSS, and bit of of the z







Next, I think that the galaxies from the south sky GOODS S don't have spectra, just photoZ.
So just think about GOODS N, which have spectra.  They have numbers starting with 1.......
« Last Edit: July 11, 2010, 05:58:43 am by Budgieye »

swengineer

  • Sr. Member
  • ****
  • Posts: 312
    • View Profile
Re: Redshift of Hubble galaxies, with spectra
« Reply #24 on: July 11, 2010, 05:56:42 am »
...
The colours of the images must be the same for the north and south sky, .....right?

...

Next, I think that the galaxies from the south sky GOODS S don't have spectra, just photoZ.
So just think about GOODS N, which have spectra.  They have numbers starting with 1.......

A few more clarifications for you:

The colours of the images are not quite the same:
GEMS, GOODS-S, and GOODS-N use the F606 and F850 filters, while AEGIS uses F606 and F814 filters. 
They could all still have the same false colors, but it seems to me that AEGIS has a yellower "red".

GZ:Hubble images from "the S sky" include GEMS and GOODS-S.  It's basically the same area with GOOD-S filling the "donut hole" left in the GEMS survey.
GZ:Hubble images from "the N sky" include AEGIS and GOODS-N.  They are from entirely separate areas of the sky so far as I understand things.
Finally, GOODS-S and GOODS-N, while covering very different areas of the sky are part of the same "program":

From Michael Parrish in the "examine tool" thread:
Quote
If you're curious about the Zooniverse IDs, the format is:
AHZ1 - AEGIS
AHZ2 - COSMOS (None yet)
AHZ3 - GOODS-N
AHZ4 - GEMS & GOODS-S
AHZ5 - SDSS
AHZ6 - Stripe-82
So, we've got SDSS data in there, but only if it's used in Galaxy Zoo: Hubble.

So the AHZ numbers are the best way to know at  glance which survey contains a given object.
AEGIS uses 1....... for its survey reference IDs.
GOODS-N uses 3........ (I think)
GEMS and GOODS-S use 9..........

AEGIS has spectroscopic data for most of the (small? higher z?) objects.  I think that was a goal of the AEGIS project.
GEMS and GOODS-s have(so far as I know) limited spectroscopic data.  It does exists for some objects. But it can be very hard (for me anyway) to find or to find out if it exists or not.
GOODS-N: I don't know how much spectroscopy data is available (or readily available at any rate).  Likely it is sporadic, similar to GEMS & GOOD-S.

I hope this helps.

-Mark

Budgieye

  • Hero Member
  • *****
  • Posts: 3860
    • View Profile
Re: Redshift of Hubble galaxies, with spectra
« Reply #25 on: July 11, 2010, 06:08:48 am »
To avoid confusing myself, I will look at
Hubble - for images (well, what else is there)
AEGIS  - for spectra.

Budgieye

  • Hero Member
  • *****
  • Posts: 3860
    • View Profile
Re: Redshift of Hubble galaxies, with spectra
« Reply #26 on: July 11, 2010, 11:15:27 am »
It is going to be hard to correlate the colour of the image with the spectrum, since the spectrum is so short.

Here is a possible Hubble pea with an OII peak
http://tkserver.keck.hawaii.edu/egs/dataAccess/notebook/egs_notebook.php?serial=12007954
http://tkserver.keck.hawaii.edu/egs/egs_data/postage/prettystamps/12007954.jpg




paulrogers

  • Guest
Re: Redshift of Hubble galaxies, with spectra
« Reply #27 on: July 11, 2010, 04:42:06 pm »
It is going to be hard to correlate the colour of the image with the spectrum, since the spectrum is so short.
Here is a possible Hubble pea with an OII peak
http://tkserver.keck.hawaii.edu/egs/dataAccess/notebook/egs_notebook.php?serial=12007954
http://tkserver.keck.hawaii.edu/egs/egs_data/postage/prettystamps/12007954.jpg


That's my "comma" galaxy, AHZ100013x, rotated about 90 degrees clockwise, reported here:http://www.galaxyzooforum.org/index.php?topic=277967.msg476899#msg476899
Does this spectrum help identify what it is?  I'm too used to the interpretted spectra with emission/absorption limes identified that we had in GZ2.  I've been suspecting either QSO or AGN, because it seems so bright.

Budgieye

  • Hero Member
  • *****
  • Posts: 3860
    • View Profile
Re: Redshift of Hubble galaxies, with spectra
« Reply #28 on: July 11, 2010, 06:17:21 pm »

Budgieye

  • Hero Member
  • *****
  • Posts: 3860
    • View Profile
Re: Redshift of Hubble galaxies, with spectra
« Reply #29 on: July 13, 2010, 06:04:23 am »
...thinking....
http://en.wikipedia.org/wiki/Advanced_Camera_for_Surveys
The primary broad-band filters are equivalent to the u, g, r, i, and z filters of the ground-based Sloan Digital Sky Survey (SDSS).

http://www.stsci.edu/hst/acs/documents/handbooks/DataHandbookv5/ACS_longdhb.pdf