SEGUE Getting Started
Getting Started:
Accessing the SEGUE spectra
and associated imaging
Important note about linking SEGUE imaging and spectra in the CAS
The SEGUE spectra count as special-plate spectra in
the CAS and are not considered to be
"scienceprimary". Therefore, they are not linked to the
photometry in the usual way, e.g., they do not appear in the object
browser's default view like galaxy, quasar and stellar spectra from
the legacy survey. With DR7, all SEGUE spectra and imaging are in the
BESTDR7
database.
See the SEGUE sample SQL
query for instructions on joining SEGUE imaging and
spectroscopy.
Here are some walk-through examples of how to access the SEGUE spectra
and associated imaging for your research.
This example demonstrates how to check the parameters of some SEGUE spectra of 'known objects'
against values found in the literature. We will check the Radial Velocities,
Metalicities [Fe/H], and stellar surface gravities of some stars
in the Hercules Globular Cluster (M 13 = NGC6205) with their cataloged
values in the
W. Harris catalog:
To select star around M13 we need to know the position of
M13 on the sky, which is (alpha,delta) = (16 41 41.5 , +36 27 37)
or (l,b) = (59.01,40.91).
We convert the H:M:S, D:M:S notation for (alpha,delta) (J2000)
to decimal degrees for the database:
(alpha,delta) = (16:41:41.5 , +36:27:37) = (250.42292,36.46028)
Let's use the stellar parameters table in the DR7 database to
select stars with spectra near (within 1 degree, globular
clusters have diameter generally much smaller than 1 degree) of this
position on the sky.
We recommend the CASJOBS interface for SQL database searches
of the SDSS/SEGUE data.All who are interested may sign up for
a free account under CASJOBs. CASJOBS has the all the advantages
of the regular skyserver SQL ASTRO along with a 'batch processing'
system for submitting long (up to 8 hours, with up to 500MB worth of
output) and complex queries to the database server for later retrieval
and cross-reference (through the MyDB system of user databases
with intermediate results). The output of CASJOBS queries may
be downloaded to your home machine in CSV (comma-separated-value)
ASCII text file format for uploading to EXCEL or further manipulation
and plotting by the software package of your choice.
The CASJOBS interface has on its 'Query' page a 'Context' menu,
which indicates the database you will be searching in.
You should use the 'DR7' context. This, by default, searches
the 'BESTDR7' database which contains the SDSS imaging tables
and the SDSS and SEGUE spectroscopic catalog tables.
There is additionally a database called 'SEGUEDR7' which contains
imaging catalogs for areas of the SEGUE Footprint not in the
regular SDSS survey Footprint.
If you wish to determine whether or not a specific coordinate
is in either the SDSSDR7 or SEGUEDR7 footprint, you may feed a
list of (RA,DEC) coordinates into the SDSS DAS coordinate search form,
or follow the "alternatives" instructions on that page for long
lists of coordinates.
The names of the key tables you will be interested in searching
for SEGUE catalog information are:
Context tablename (case insensitive) what
-----------------------------
DR7 sppParams Radial Velocities,
de-reddened PSF magnitudes and colors,
'[Fe/H]' metalicity estimates,
log g surface gravity indictors (used to
tell a Red Giant from a Red Dwarf)
effective temperature (correlated with color)
proper motion information on star,
stellar classification information
DR7 Star Full photometric information for all point
sources in the SDSS Legacy and SEGUE footprints.
Includes ugriz PSF magnitudes,
reddening information, (colors may be calculated)
RA,DEC position information
DR7 SpecObjAll Spectroscopic information for objects in
SDSS+SEGUE. Note that the subset table
'SpecObj' doesn't contain SEGUE spectra,
only 'main survey' SDSS spectra (mostly
galaxies).
DR7 PlateX Metadata information about plates (which
contain 640 spectra each) in the SDSS+SEGUE
surveys. The information here (specificially
the 'programname' field) can be used to
distinguish SDSS (galaxy+qso+star) plates
(programname like 'chunk%')
from SEGUE (all star) plates
(programname like 'seg%').
DR7 PhotoObj Full photometric object information for
all primary and secondary
sources
(including extended objects such as galaxies),
in the DR7 SDSS+SEGUE footprints, a superset of
'star'.
DR7 PhotoObjAll Full photometric information for all sources
(including duplicates) in the DR7 footprint.
A superset of 'photoobj'.
DR7 propermotions Propermotion information for 205M stars.
Let's return to the example of searching for SEGUE information about
stars in the globular cluster M13.
We'll select all stars within about 1 degree
of the known center of the cluster at (ra,dec) [J2000] =
(250.42292, 36.46028).
The RA coordinate has a cos(DEC) 'shrinkage' in it,
so to get stars in a box 1 degree on a side, we wish
to search +/- 0.5/cos(DEC) degrees away in RA, but
the standard +/- 0.5 degrees away in DEC:
Thus cut and paste this query to the DR7 context of Casjobs 'query'
window:
select count(*) from sppParams where
ra between 250.42292-cos(36.46*3.14159/180)
and 250.42292+cos(36.46*3.14159/180)
and dec between 36.46 -0.5 and 36.46 + 0.5
and hit the 'Quick' button. In a few seconds it should return
a number like '1044', the number of stars which meet these
conditions.
You may return all the known information, including the
12 measures of [Fe/H] and the 9 measures of log g
for these stars with a query like:
select * from sppParams where
ra between 250.42292-cos(36.46*3.14159/180)
and 250.42292+cos(36.46*3.14159/180)
and dec between 36.46 -0.5 and 36.46 + 0.5
However, that table is a bit hard to parse.
Instead, let's select specific columns we wish to return
from the complete list of all columns in
the sppParams table.
We'll also restrict to objects which have 'good quality' measurements
of [Fe/H] (we pick the 'adopted' measure of [Fe/H] named: 'feha'),
by requiring that the error (fehaerr) on the measure is greater
than 0 but less than 0.7 dex. We'll also restrict to objects with
errors (elodierverr) on their radial velocity (elodierv) of 20 km/s or less:
select plate,mjd,fiberid,brun,brerun,bcamcol,bfield,bobj,
elodierv,elodierverr,feha,fehaerr,fehan,logga,loggaerr,loggan,
targettype,sptypea,hammersptype,flag,g0,umg0,gmr0,
ra,dec,l,b,chiHK,chiGband,teffa,teffaerr from sppParams
where
ra between 250.42292-cos(36.46*3.14159/180)
and 250.42292+cos(36.46*3.14159/180)
and dec between 36.46 -0.5 and 36.46 + 0.5
and fehaerr between 0.01 and 0.7
and elodierverr between 0.01 and 20
Instead of entering this as a 'Quick' query, hit the 'Submit'
button instead. Though the query only takes a few seconds
to run, this will save the output of your query as a named
table in your 'MyDB', for instance MyTable_21 (your exact table
name will vary depending on your query history, and it
may be manually named to whatever you like).
After the query runs (hit 'History' repeatedly to check if it
succeeded, you may also choose receive an email
when the query is done), go to 'MyDB' and find your 'MyTable_21'.
It should have about 779 rows and take up 500 Kbytes of space
in your MyDB area.
Click on the table and select 'Download',
then click on 'Go' to reformat your table as a 'CSV' file.
Then go to 'Output' and click repeatedly until the reformatted table
is available for 'Download'. Click 'Download' to put the table on
your local box. Now you may use a variety of tools (Excel spreadsheet,
AWK, TCL) to manipulate and plot your data.
The columns of your output file (based on what you 'select'ed above) are:
0 plate
1 mjd
2 fiberid
3 brun
4 brerun
5 bcamcol
6 bfield
7 bobj
8 elodierv
9 elodierverr
10 feha
11 fehaerr
12 fehan
13 logga
14 loggaerr
15 loggan
16 targettype
17 sptypea
18 hammersptype
19 flag
20 g0
21 umg0
22 gmr0
23 ra
24 dec
25 l
26 b
27 chiHK
28 chiGband
29 teffa
30 teffaerr
Here are some plots you should be able to easily reproduce:
1. Plot of gmr0 on the X-axis (limits -1 to 2.0) vs. g0 on
the Y-axis (limits 21 to 12).
This is a classical Color Magnitude Diagram
You will notice that the 'top' of this diagram is not
populated, i.e. the Red Giant Branch. Photometry of
these stars saturated the regular SDSS 2.5m imaging camera
(stars brighter than about with g=14 do this).
We have obtained limited amounts of spectra of such stars,
however their magnitudes are not cataloged in all cases
(see example #XXX, below).
2. Plot a histogram of rva, the radial velocity, with limits
from about -500 to 500 km/s.
The catalog heliocentric RV of M13 is
-245.6 km/s, does your plot look like this.
Based on the above plot number of of RVs, let's subselect only stars
near the RV peak of the cluster, say with RV between -300 and -200 km/s.
Then histogram the Metalicities (histogram fehr from -4 to 0) of these:
3. Do you get something like this?
Let's look at the surface gravity now. Subselect those stars with RV between
-300 and -200 that have small errors on the surface gravity, i.e.
loggerr between 0 and 0.5. Then make a faux Hertzsprung-Russell diagram
for the cluster by plotting Teff (teffa) on the x axis from 10000 to 3000 (K) and
log g (Luminosity indicator) on the y axis from 6 to 0.
4. This color vs. surface gravity plot for a cluster is one
step closer to a 'physical' Hertzsprung-Russell diagram (usually plotted as
Teff vs. Luminosity) as opposed
to an observed parameter (color vs. magnitude) diagram.
We see from this plot that some red giant branch stars have appeared. These had
non-SEGUE photometry and so their g, g-r values were 0 in the inital selection.
We may be able to find some indication of their magnitudes by looking at other
magnitude indicators available.
There are also a few BHB stars apparent at the hot (high Teff) left edge of the
diagram. Let's examine the individual spectra for these stars:
In particular, let's look at examples of a K-giant, and an F turnoff star.
To find them, even though we don't know the g-r, g for the brighter objects,
let's select on Teff, log g.
K-giant:
select teffa between 5000 and 5200, log g between 1.8 and 2,
RV between -300 and -200, [Fe/H] between -1.8 and -1.3,
extract the plate,mjd,fiberid triplet.
There are two objects selected:
plate/mjd/fiberid = 2174/53521/167
and
plate/mjd/fiberid = 2255/53565/518
Browse these spectra by going to the link for the DR7 plate list
and (scrolling down to the globular cluster section of plates near
the bottom) click on the link for plate 2174, (mjd = 53521) then in
the menu on the left, click on the link for fiberid = 167. Nice spectrum!
Note the weak Mg triplet at 5100-5200A, indicating a low surface gravity
of a giant star.
Also note that the photometry for this star is a 'place-holder' (it's
bogus). The photo catalogs info for this object, with run =0 indicates a
manually selected target not deblended or saturated in this crowded field
by the photo software. Other information on this well-known object
(an M13 giant branch star) may be found in the literature by using the
RA,DEC information presented.
Also examine the bright red giant on plate 2255, fiberid= 518.
The spectrum of this object is clearly a red giant, with very
weak Mg 5100-5200, however,the astrometry on this object is
special, it is a 'shifted' plate, used to obtain spectra of
very bright stars. Thus, don't trust the RA,DEC for this star,
for reference in the literature until we post an updated version.
(plate 2255 is the only one of a very small number of segcluster plates
with this issue in DR7 SEGUE plates).
F-turnoff:
select gmr0 between 0.2 and 0.3, log g between 1.8 and 2,
RV between -300 and -200, [Fe/H] between -1.8 and -1.3,
extract the plate,mjd,fiberid triplet. There are about 18,
so sort by mag, take the brightest,
which is:
2185/53532/500 = plate/mjd/fiberid.
Examine plate 2185, fiberid = 500.
It's a blue thing, with narrow balmer lines, pretty typical (early?) F star.
If you'd like to grab more than just the picture of a spectrum,
but are interested in examining the spectrum in more detail,
you can fetch this object from the DAS, with a 'wget' style
command like so:
wget http://das.sdss.org/spectro/1d_26/2185/1d/spSpec-53532-2185-500.fit
(you may grab it in your browser and download it, an rsync
interface is available for bulk downloads).
Note that the plate number is padded with leading 0's to
four places and the fiberid to three places, thus 371/52078/1
would be:
wget http://das.sdss.org/spectro/1d_26/0371/1d/spSpec-52078-0371-001.fit
Any available individual spectrum may be so obtained.
This FITS file may be read into IRAF or any other imaging processing program.
The data model for the contents of the spSpec files is given here.
If you wish to retrieve all spectra for a given plate in one big FITS file,
they are available from the DAS as spPlate* files.
For plate 2185, this wget will retrieve the (typically 60MB) file:
wget http://das.sdss.org/spectro/2d_26/2185/spPlate-2185-53532.fits
The data model for spPlate files, in order that you may examine them
with IRAF or another processing package is here.
Any other plate may be retrieved the same way.
A query to list all SEGUE plates with their (RA,DEC) centers in the CAS database:
select plate,mjd,ra,dec,programname from plateX where programname like 'segue%'
If you wish to find the cluster plates (M15, M13, NGC2420, etc),
try this:
select plate,mjd,ra,dec,programname from plateX where programname like 'segcluster%'
Other special segue plate programname categories
are 'segpointed%' (stream targeted plates),
'segtest%' (test plates), and 'seglow%' (low latitude targeting
algorithm).
A query to select objects with very high quality spectra that have very low estimated
metalicity ([Fe/H] less than one-thousandth of solar).
We use the chiHK, chiMg, chiGband indicators and require that they
give values between 0 and 2; also that the feha indicator has an error of no more than
0.5 and that the number of different reliably [Fe/H] estimation methods (fehan) is greater
than 3 and that the S/N ratio of the object is greater than 20.
select plate,mjd,fiberid,flag,ra,dec,g0,umg0,gmr0,rmi0,imz0,
feha,fehaerr,fehan,elodierv,elodierverr,chihk,chigband,sna,
targettype from
sppParams where
feha between -9 and -3 and
chihk between 0.01 and 2 and
chigband between 0.01 and 2 and
sna > 20 and fehan > 3 and fehaerr between 0.01 and 0.5
135 stars are returned in DR7, all with [Fe/H] ~ -3, here are 4:
plate/mjd/fiberid RA (J2000) DEC
--------------------------------------------
2401/53768/322 138.3556061 31.0887051
1711/53535/285 219.1344604 9.3087473
1049/52751/126 226.1047211 46.3891373
2327/53710/166 9.0451527 7.1922379
Only the first and last are SEGUE_LOWMETAL targeted stars,
the other two are main Survey spectrophotometric
standard stars targets (for which the sppParams table
is also available).
To browse these stars, use the skyserver interface, or download the
spSpec file and examine the spectra with IRAF.
i.e. for the second spectrum
wget http://das.sdss.org/spectro/1d_26/1711/1d/spSpec-53535-1711-285.fit
A skyserver link to these objects looks like the following:
http://cas.sdss.org/astro/en/tools/explore/obj.asp?ra=219.1344&dec=9.308
Click on all spectra and the spectrum id for a picture of the spectrum.
This uses the RA DEC to look up the objects in the skyserver; you may
also lookup by plate/mjd/fiberid.
It is often desirable to select stars from an area of sky where
not every star has a spectrum. The ugriz photometry alone for
a deep sample of stars is often enough to do significant Galactic
Structure science. In addition, it is often desirable to determine
the 'selection function' of a set of spectra determined with some
possibly not well known target selection algorithm. For this
reason one often wishes to select imaging from an area of sky
where one has spectra, including objects for which no spectra
exist in the same part of sky.
All of the spectra reside in the CAS specobjall database,
including both SDSS main survey footprint spectra and SEGUE
spectra (along with assorted other spectra).
Due to the difficulty in processing and interpreting highly
reddened low-latitude, crowded stellar field imaging with the standard
SDSS photo pipeline (which was designed for galaxy photometry in
uncrowded fields), the SEGUE imaging for DR7 has been a special
rerun number (648)
In order to match a SEGUE spectrum with its full photometric
information (i.e. fields beyond that in the sppParams table,
which is often but not always enough), one simply
joins the DR7 sppParams table with the specobjall table
with the photoobjall table using
the bestobjid in the specobjall table to join to the
objid in the photoobjall table, which the sppParams table
and the specobjall table join on specobjid.
Thus, to match a SEGUE spectrum from the specobjall table
with its corresponding photometric information,
one should do this query:
select top 10 s.plate,s.mjd,s.fiberid,sp.psfmag_g,s.mag_1,
spp.elodierv
from specobjall s, photoobjall sp, sppParams spp where
s.specobjid = spp.specobjid and sp.objid = s.bestobjid
and s.specclass in (1,6) and elodierverr > 0
To query for imaging information around a given (ra,dec):
To find objects in the plate 1880 field, centered at
(RA,DEC) = (358.26,36.40)
one may perform one query as follows to select
all stellar photometry:
(done on the DR7 context of Casjobs):
select run,rerun,camcol,field,obj,ra,dec,
psfmag_u,psfmag_g,psfmag_r,psfmag_i,psfmag_z,extinction_u,
extinction_g,extinction_r,extinction_i,extinction_z
from star
where ra between 358.26-1.5/cos(36.4*3.14159/180) and 358.26+1.5/cos(36.4*3.14159/180)
and dec between 36.40 -1.5 and 36.4 + 1.5
and psfmag_g < 22
This query returns a large number (54,301) objects, and you may
need to
'Submit' it rather than running it as a 'Quick' query (in casjobs).
Once you download the output you may further subselect to
only get objects within 1.49 degrees of (ra,dec) = (358.26,36.4)
and you may with to add in cuts on color or magnitude,
for instance, throw out objects fainter than 20.5, which
is the faint limit of the spectrscopic sample.
Then, if you are, for instance, examining spectroscopic completeness
of the F subdwarf sample, you may further select objects with
(g-r)_0 between 0.1 and 0.4 and compare the density of photometric
objects as a function of magnitude with the spectroscopic sample
of F stars on plates 1880 and 1881.
In this fashion, one may obtain photometric information for
all SEGUE plates. And dervive empirical selection functions
based on the photometry with the same color cuts as the spectroscopy.
This is best done in a two step process, where one first
'uploads' a set of id,ra,dec triplets for objects in question
into a MyDb, say mytable_14 in this example.
Then one connects to the TARGETDR7 context and matches
the target photometry information to each id,ra,dec in the
list, using a position match.
SELECT myid, ra,dec, dbo.fGetNearestObjIdEqtype(ra,dec,1.2,6) as objId into #x
FROM mydb.mytable_14
select n.myid,n.ra as inra,n.dec as indec,s.objid,s.run,s.rerun,s.camcol,s.field
,s.obj,s.ra,s.dec,s.psfmag_u,s.psfmag_g,s.psfmag_r,s.psfmag_i,s.psfmag_z ,s.type
from star s,
#x as N where s.objid = n.objid