Parametric Recovery of Line-of-Sight Velocity Distributions from Absorption-Line Spectra of Galaxies via Penalized Likelihood

Abstract

We investigate the accuracy of the parametric recovery of the line-of-sight velocity distribution (LOSVD) of the stars in a galaxy, while working in pixel space. Problems appear when the data have a low signal-to-noise ratio, or the observed LOSVD is not well sampled by the data. We propose a simple solution based on maximum penalized likelihood and we apply it to the common situation in which the LOSVD is described by a Gauss-Hermite series. We compare different techniques by extracting the stellar kinematics from observations of the barred lenticular galaxy NGC 3384 obtained with the SAURON integral-field spectrograph.