ROBOSPECT: Automated Equivalent Width Measurement for Stellar Spectra
Abstract
The study of stellar abundances requires the analysis of spectral features. The two methods of analysis are spectral synthesis and equivalent width measurement. Spectral synthesis requires models and other inputs, which makes reproducing the synthetic spectra of other studies difficult. The equivalent width of a line is simply a geometric characteristic and is therefore an easily repeatable measurement between different sets of spectra for a given star, without needing to account for differing treatments of the physics of abundance analysis between studies. Equivalent width measurement is typically achieved by fitting a Gaussian or Voigt profile to a spectral line using software that usually requires user input. As different users may make different choices of the fitting parameters (such as the continuum placement), this introduces user-specific uncertainties into the fit quality. Automating this process allows the equivalent width measurement to be independent of these user-added uncertainties; however, there are few freely available programs that can accurately and precisely measure equivalent widths. We present a new program, ROBOSPECT, that does automate this measurement with minimal user interaction. This allows a consistent and repeatable set of measurements to be determined from the data without the added uncertainties due to user choices. ROBOSPECT is modular by design, allowing for different line models and continuum estimations to be selected based on the input spectrum. Comparing the results of ROBOSPECT to a sample of independently well measured spectral lines shows no systematic bias in the ROBOSPECT results, with a very tight scatter of 2.1 $\AA{}$ for a S/N 100 stellar spectrum. We illustrate the suitability of ROBOSPECT to line measurement by presenting a sample of test fits over a range of spectrum type and signal to noise.
- Publication:
-
American Astronomical Society Meeting Abstracts #220
- Pub Date:
- May 2012
- Bibcode:
- 2012AAS...22052203W