Three-Dimensional Optimal Spectral Extraction (TDOSE) from integral field spectroscopy^⋆

¹ Leibniz-Institut für Astrophysik Potsdam (AIP), An der Sternwarte 16, 14482 Potsdam, Germany
e-mail: kbschmidt@aip.de
² Department of Astronomy, Stockholm University, AlbaNova University Centre, 106 91 Stockholm, Sweden
³ IRAP, Université de Toulouse, CNRS, CNES, UPS, Toulouse, France
⁴ Aix Marseille Univ., CNRS, CNES, LAM, Marseille, France
⁵ Univ. Lyon 1, ENS de Lyon, CNRS, Centre de Recherche Astrophysique de Lyon (CRAL) UMR5574, 69230 Saint-Genis-Laval, France
⁶ Hiroshima Astrophysical Science Center, Hiroshima University, 1-3-1 Kagamiyama, Higashi-Hiroshima, Hiroshima 739-8526, Japan
⁷ Leiden Observatory, Leiden University, PO Box 9513, 2300 RA Leiden, The Netherlands

Received: 9 May 2019
Accepted: 13 June 2019

Abstract

The amount of integral field spectrograph (IFS) data has grown considerably over the last few decades. The demand for tools to analyze such data is therefore bigger now than ever. We present a flexible Python tool for Three-Dimensional Optimal Spectral Extraction (TDOSE) from IFS data cubes. TDOSE works on any three-dimensional data cube and bases the spectral extractions on morphological reference image models. By default, these models are generated and composed of multiple multivariate Gaussian components, but can also be constructed with independent modeling tools and be provided as input to TDOSE. In each wavelength layer of the IFS data cube, TDOSE simultaneously optimizes all sources in the morphological model to minimize the difference between the scaled model components and the IFS data. The flux optimization produces individual data cubes containing the scaled three-dimensional source models. This allows the efficient de-blending of flux in both the spatial and spectral dimensions of the IFS data cubes, and extraction of the corresponding one-dimensional spectra. TDOSE implicitly requires an assumption about the two-dimensional light distribution. We describe how the flexibility of TDOSE can be used to mitigate and correct for deviations from the input distribution. Furthermore, we present an example of how the three-dimensional source models generated by TDOSE can be used to improve two-dimensional maps of physical parameters like velocity, metallicity, or star formation rate when flux contamination is a problem. By extracting TDOSE spectra of ∼150 [OII] emitters from the MUSE-Wide survey we show that the median increase in line flux is ∼5% when using multi-component models as opposed to single-component models. However, the increase in recovered line emission in individual cases can be as much as 50%. Comparing the TDOSE model-based extractions of the MUSE-Wide [OII] emitters with aperture spectra, the TDOSE spectra provides a median flux (S/N) increase of 9% (14%). Hence, TDOSE spectra optimize the S/N while still being able to recover the total emitted flux.