The DEEP2 redshift pipeline was devloped by the DEEP2 team at the
University of California-Berkeley (UCB). The code was designed to
measure redshifts and velocity dispersions specifically for the DEEP2
Galaxy Redshift Survey and was entirely written in IDL. The pipeline is
directly based upon the Sloan Digital Sky Survey (SDSS) spectral
reduction package as written by David Schlegel, Scott Burles, and
Doug Finkbeiner. The majority of the redshift code was written by
Marc Davis, Jeff Newman, and Michael Cooper with contributions from
Renbin Yan and Darren Madgwick.
For each object on a DEEP2 DEIMOS slitmask, the pipeline compares
the 1-d object spectrum with template spectra using chi-squared
minimization to determine the best-fitting redshifts. The optimal 1-d
extraction (as generated by the spec2d pipeline) is utilized in the
redshift determination. A suite of of template spectra are employed in
the redshift determination. The galaxy template is broken down into
3 components: a set of emission lines (with instrumental linewidths
appropriate for DEIMOS in DEEP2 settings and typical linestrengths), a
coadded luminous red galaxy spectrum from Eisenstein et al.(2003), and
a synthetic A-star spectrum from Kurucz et al. Each object is fit using
a superposition of the three components over the redshift range 0.0 < z <
1.5 and the 5 best fitting redshifts (ranked according to reduced
chi-squared value) are saved.
While there is a single galaxy template composed of three components,
the pipeline uses a sparse set of individual stellar templates and
a single broad-line AGN template to measure the best-fitting,
single-template stellar and QSO redshifts and keeps the 3 best stellar
fits and best 2 QSO fits.
At each galaxy redshift, the pipeline also measures the Guassian
velocity width using the emission lines and stores this value in the
"vdisp" tag in the output structures. The full set of 10 redshift
fits (5 galaxy, 3 stellar, 2 QSO) are then ranked according to
reduced chi-squared values and written to an output file.
|
