The SBU 1.5m Spectroscopic Database
This is a guidebook for Stony Brook users of RC Spectrograph data. It requires
access to the raid array on nut, and some knowledge of IDL. The
IDL procedures are located in /home/fwalter/LIBS. This directory
must be in the IDL_PATH variable in your .cshrc file.
Contents
The RC spectrograph is a grating spectrograph mounted on the 1.5m telescope
at CTIO. The instrument manual can be found at this
this CTIO web
site. The spectrograph currently shares time with the CPAPIR near-IR
camera, and is mounted about half the time.
The spectra are taken by service observers. We download the raw data
each morning, and process it ourselves. The data are also archived at Yale.
Data processing and storage are currently done using nut.
The IDL procedure s15_start is used to process the data.
In general, each observation consists of three (or more) spectra
obtained consecutively. The three spectra can be median-filtered to
remove cosmic rays. This summed image is stored as a Txx.fits image.
The data are extracted in four ways.
The spectrum is boxcar-extracted from the summed image. A fit to the
cross-section determines the width of the extraction slit. Background is
extracted above and below the spectrum.
The spectrum is boxcar-extracted from each of the individual image.
A fit to the
cross-section determines the width of the extraction slit. Background is
extracted above and below the spectrum. The individual spectra are then
summed.
A Gaussian is fit to the spatial profile at each wavelength point in the
summed image. The integrated Gaussian flux is the extracted counts. This
has the advantage of avoiding single bad pixels, but underestimates the
flux of bright sources because the spatial wings are non-Gaussian.
A Gaussian is fit to the spatial profile at each wavelength point in
each of the individual spectra. The fluxes are summed to produce the final
spectrum.
The raw images are copied to CD after processing (so errors in the headers
can be corrected). Copies of the CDs are available in the
lab. The raw data are retained on the raid array on a space-available
basis.
The reduced images and extracted spectra are retained on the raid array.
They are also backed up to CD. Note that while all the raw data
are archived, only the Stony Brook data are processed/extracted.
The SMARTS spectroscopic service schedule is at this
page. It contains observing logs and plans, and links to finding charts.
The catalog of images is maintained in a pair of IDL databases, one for
the raw data and one for the reduced/extracted data.
each channel. For manageability, a new database is created for each observing
year. SMARTS began operations in 2003. Note that the observing year generally,
but not always, runs from February through January.
After processing the images, we update the databases.
In order to access the databases, you must be logged into nut.
you must have the following line in your
.cshrc file:
setenv ZDBASE /data/fwalter/DB/
In order to access the IDL software, you must have a line in your .cshrc
file that looks like the following:
setenv IDL_PATH +/usr/local/rsi/idl_5.3/local_pro:+/home/fwalter/ASTROLIB:+/home/fwalter/LIBS:+/usr/local/rsi/idl_5.3/lib:
The databases are named s15_yyyy and s15p_yyyy,
for the raw and
processed data, respectively. yyyy is the year (e.g., 2006)
While you can in principle query the databases using any IDL databasing
routines, most routine queries are handled through the software described
below.
The 2003 and 2006 databases are complete. The 2004 and 2005 databases
are currently incomplete, and will be completed as time allows.
The long term monitoring programs we are engaged in include:
- Monitoring of long term activity cycles in Polars (SUNY-01).
Targets include BL Hyi, EF Eri, QS Tel, V834 Cen, and VV Pup
- Monitoring of outbursts in Exors (SUNY-02).
Targets include V1118 Ori, V1143 Ori, NY Ori, and EX Lup
- Monitoring the possible dwarf nova HD 109962 (SUNY-08).
- Monitoring of accretion-induced activity in T Tauri stars (SUNY-10).
Principle targets are RU Lup and S CrA.
- Monitoring of nova light curves (SUNY-11).
Currently active campaigns targets RS Oph, N LMC 2005, V5117 Sgr, and
V2576 Oph. We continue to follow DE Cir, YY Dor, V574 Pup, V1187 Sco,
V5114 Sgr, V382 Nor, and V1663 Aql
- Monitoring the long term evolution of V1647 Ori (SUNY-12)
One queries the databases with the IDL procedure find_targets.
- To find all observations of a specific target, use
find_targets,name='xxx', where xxx is the name of the target.
- To find all observations in a specific program, use
find_targets,prop='xxx', where xxx is the name of the program (e.g.,
SUNY 06a-11).
- To find all observations in a specific observing mode, use
find_targets,mode='xxx', where xxx is the name of the mode (see
below).
By default, the full database is queried. To limit the query to a single
year, use the year keyword.
To search the raw data archive, use the /raw keyword.
For details of the individual spectra that make up a particular observation,
use the procedure s15id. Typing
s15id,dir='060408',14,/d0,/raw shows that there
are 9 spectra of HD 93308 (eta Carina), with exposure times of 5, 2, and 1
second. There are also two Neon comparison spectra. Note that because
the s15id procedure operated on the files and does not query the database,
the raw or
trimmed and flattened files must be on disk: it does not work on summed
images or extracted data files.
Va. Observing Modes
There are a number of standard observing modes. The ones most used for SBU
programs are:
- 26/I. First order blue, 3500-5300 Angstroms. 4.3 Angstrom resolution.
- 47/I (or 47/Ia)or 47/Ib. First order red. Includes the H-alpha and Li 6707 regions.
Resolution is 3.1 Angstroms. The 47/Ib setting is centered about 300 Angstroms
to the blue of 47/I, and included the Na D region.
- 47/II or 47/IIb. This is second order in the blue, and has a higher 1.6
Angstrom resolution, at the expense of wavelength coverage. 47/II includes
the Ca II K&H region; 47/IIb does not, but does reach the He II 4686 line.
See this
page for a summary of all observing modes.
The data are stored in individual
FITS format
files on the nut raid array. You can extract the spectra wwith the ps15
procedure.
Suppose you want to extract a spectrum of the magnetic cataclysmic variable
V834 Cen. First you would use find_targets to locate the files.
find_targets,name='v834 cen',year=2006
The output will look line this:
Catalog of SMARTS SUNY Reduced Spectroscopic Observations 2006
------------------------------------------------------------------------------
Object 2-d x gr grat spect. Exposur Proposal
file tilt mode time ID
name (sec)
------------------------------------------------------------------------------
V834 Cen 060112/XC_21.fits 47 22.64 47/Ia 400.00 SUNY 06a-01
V834 Cen 060115/XC_29.fits 26 15.93 26/Ia 400.00 SUNY 06a-01
V834 Cen 060117/XC_25.fits 47 22.64 47/Ia 300.00 SUNY 06a-01
V834 Cen 060118/XC_26.fits 26 15.93 26/Ia 400.00 SUNY 06a-01
V834 Cen 060120/XC_23.fits 47 22.64 47/Ia 300.00 SUNY 06a-01
V834 Cen 060125/XC_23.fits 26 15.93 26/Ia 400.00 SUNY 06a-01
V834 Cen 060202/XC_26.fits 47 22.64 47/Ia 300.00 SUNY 06a-01
V834 Cen 060203/XC_23.fits 26 15.93 26/Ia 300.00 SUNY 06a-01
V834 Cen 060205/XC_23.fits 26 15.93 26/Ia 300.00 SUNY 06a-01
V834 Cen 060308/XC_18.fits 47 22.64 47/Ia 400.00 SUNY 06a-01
V834 Cen 060309/XC_19.fits 47 22.64 47/Ia 400.00 SUNY 06a-01
V834 Cen 060315/XC_20.fits 26 15.93 26/Ia 300.00 SUNY 06a-01
V834 Cen 060318/XC_20.fits 47 27.39 47/II 300.00 SUNY 06a-01
V834 Cen 060319/XC_23.fits 26 15.93 26/Ia 300.00 SUNY 06a-01
V834 Cen 060324/XC_20.fits 26 15.93 26/Ia 300.00 SUNY 06a-01
V834 Cen 060329/XC_22.fits 26 15.93 26/Ia 300.00 SUNY 06a-01
V834 Cen 060330/XC_24.fits 26 15.93 26/Ia 300.00 SUNY 06a-01
V834 Cen 060331/XC_19.fits 47 22.64 47/Ia 400.00 SUNY 06a-01
V834 Cen 060406/XC_29.fits 26 15.93 26/Ia 300.00 SUNY 06a-01
V834 Cen 060408/XC_16.fits 47 22.64 47/Ia 300.00 SUNY 06a-01
V834 Cen 060409/XC_18.fits 26 15.93 26/Ia 300.00 SUNY 06a-01
V834 Cen 060508/XC_20.fits 26 15.93 26/Ia 300.00 SUNY 06a-01
V834 Cen 060512/XC_15.fits 47 22.64 47/Ia 300.00 SUNY 06a-01
V834 Cen 060514/XC_25.fits 26 15.93 26/Ia 300.00 SUNY 06a-01
V834 Cen 060517/XC_15.fits 47 22.64 47/Ia 300.00 SUNY 06a-01
V834 Cen 060521/XC_28.fits 47 22.64 47/Ia 300.00 SUNY 06a-01
The file name column gives the subdirectory (the date in yymmdd format) and
the name of the extracted data file.The important thing is the 2 digit file
number between the underscore and the .fits. This is the file number.
To view a spectrum, say the one taken on Jan 15, you would type
ps15,/d0,dir='060115',29
The /d0 keyword indicates that the directory is located in the
standard file storage directory tree.
The dir=yymmdd keyword gives the data directory, which is the
observation date.
The argument 29 is the file number on this date.
The plot should look like:
Called in this way, all ps15 does is display the file. You can
extract the wavelength vector and spectrum into IDL variables w,s with the
command
ps15,/d0,dir='060115',29,w,s,e
Note that no plot is made. w, s, and e are the
wavelength, flux, and error (uncertainty) vectors, respectively . You can plot
the spectrum by typing plot,w,s
There are actually 4 extracted data vectors. Data are extracted both with a
boxcar extraction (the default) and a Gaussian fit to the profile at each
wavelength (/gauss keyword). Spectra are extracted with from the
summed image (default) or the sum of the individually-extracted spectra
(/sum keyword).
A program you can use to examine and manipulate spectra, and do simple
measurements, is ICUR. Documentation is in
this postscript file.
To invoke icur to measure the spectra
of V834 Cen you extracted, type
icur,0,w,s,e
ICUR responds to single keystrokes. The most important use capital letters:
most commands are case-senitive. Type the letter Q to exit icur.
Within icur, type a question mark (?) for terse on-line help.
Read the documentation.
In many cases you will wish to write your own spectral analysis code.
If you will be analyzing a large number of spectra of the same (or similar)
object, there are advantages to putting the spectra into a single icd-format
file. The file format is described in the
icur documentation.
Essentially this is an associated variable
with each spectrum compressed into a single variable.
Among the advantages of this format are:
You can copy the .icd file to another machine; you are not tied to the
spectroscopic database on nut.
You can easily work within icur, and switch among spectra.
You can easily extract spectra with the gdat procedure.
The primary disadvantage is that an icd file stores only a single kind of
extracted data record per file.
To make an icd format file, use the procedure make_icd.
In general, you will execute make_icd,name='xxx', where 'xxx'
is the name of the target (as in the call to find_targets). The output
will be xxx.icd (with any spaces in the name removed). By default, the
boxcar-extracted spectrum from the coadded image is used (this is the same
default spectrum shown by ps15). Other options are set with keywords.
Among these are:
- /gauss: set to return the gaussian-extracted spectrum.
- /comb: set to return the sum of the individually-extracted spectra.
- /counts: set to extract the counts spectrum, rather than the
flux-calibrated spectrum.
- icd='yyy': set the name of the .icd file to yyy.icd.
- exclude='zzz': set to exclude any target name which includes the string
zzz
- mode='mmm': set to include only spectra in observing mode mmm.
- propid='ppp': set to include only spectra with proposal ID =
ppp
Spectra can be complex. You will be told what to measure. In general, you
will use ICUR to :
Measure line centroids: the C command
Measure equivalent widths: the E command
Fit lines and continuua: the F command