Estimating Statistical Uncertainties of Internal Kinematics of Galaxies and Star Clusters Derived Using Full Spectrum Fitting

Pixel-space full spectrum fitting exploiting nonlinear χ² minimization became a de facto standard way of deriving internal kinematics from absorption line spectra of galaxies and star clusters. However, reliable estimation of uncertainties for kinematic parameters remains a challenge and is usually...

Full description

Saved in:
Bibliographic Details
Published in:Publications of the Astronomical Society of the Pacific 2020-06, Vol.132 (1012), p.1-9
Main Authors: Chilingarian, Igor V., Grishin, Kirill A.
Format: Article
Language:English
Subjects:
Astronomical Software, Data Analysis, and Techniques
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Pixel-space full spectrum fitting exploiting nonlinear χ² minimization became a de facto standard way of deriving internal kinematics from absorption line spectra of galaxies and star clusters. However, reliable estimation of uncertainties for kinematic parameters remains a challenge and is usually addressed by running computationally expensive Monte-Carlo simulations. Here we derive simple formulae for the radial velocity and velocity dispersion uncertainties based solely on the shape of a template spectrum used in the fitting procedure and signal-to-noise information. Comparison with Monte-Carlo simulations provides agreement within a few per cent for different templates, signal-to-noise ratios and input velocity dispersions between 0.5 and 10 times the instrumental spectral resolution. For moderate template mismatch our technique returns uncertainties consistent within 20%–30% with those reported by Monte-Carlo simulations. We provide IDL and PYTHON implementations of our approach. The main applications are: (i) exposure time calculators; (ii) design of observational programs and estimates on expected uncertainties for spectral surveys of galaxies and star clusters; (iii) a cheap and accurate substitute for Monte-Carlo simulations when running them for large samples of thousands of spectra is unfeasible.
ISSN:0004-6280
1538-3873