The GALAH survey: The data reduction pipeline

TL;DR

The paper details the design and implementation of a data reduction pipeline for the GALAH survey, enabling efficient processing of high-resolution stellar spectra to derive accurate stellar parameters and elemental abundances.

Contribution

It introduces a tailored, reliable, and efficient Iraf-based reduction pipeline optimized for the GALAH survey's large-scale high-resolution spectroscopic data collection.

Findings

Radial velocity accuracy better than 1 km/s

Fibre cross-talk level below 0.5%

Scattered light around 5 counts in 20-minute exposures

Abstract

We present the data reduction procedures being used by the GALAH survey, carried out with the HERMES fibre-fed, multi-object spectrograph on the 3.9~m Anglo-Australian Telescope. GALAH is a unique survey, targeting 1 million stars brighter than magnitude V=14 at a resolution of 28,000 with a goal to measure the abundances of 29 elements. Such a large number of high resolution spectra necessitates the development of a reduction pipeline optimized for speed, accuracy, and consistency. We outline the design and structure of the Iraf-based reduction pipeline that we developed, specifically for GALAH, to produce fully calibrated spectra aimed for subsequent stellar atmospheric parameter estimation. The pipeline takes advantage of existing Iraf routines and other readily available software so as to be simple to maintain, testable and reliable. A radial velocity and stellar atmospheric parameter estimator code is also presented, which is used for further data analysis and yields a useful verification of the reduction quality. We have used this estimator to quantify the data quality of GALAH for fibre cross-talk level ($\lesssim0.5$%) and scattered light ($\sim5$ counts in a typical 20 minutes exposure), resolution across the field, sky spectrum properties, wavelength solution reliability (better than $1$ $\mathrm{km\ s^{-1}}$ accuracy) and radial velocity precision.