Characterization of LSB structures in deep images

Elisabeth Sola 1 , Pierre-Alain Duc 1

  • 1 Observatoire astronomique de Strasbourg

Abstract

Recent advances in deep imaging and image processing pipelines over the last decade opened a new window on the realm of the Low Surface Brightness (LSB) Universe. Several deep imaging surveys have disclosed a wealth of LSB collisional debris and extended halos around a large number of galaxies in the nearby Universe, both in massive and dwarf satellite galaxies.

The study of such collisional debris, predicted by numerical simulations, is essential to constrain the mechanisms driving galactic growth; as the various types of tidal features - tails, streams, plumes, shells, bridges, etc. –trace different types of mergers - major, minor, with an early or late type progenitor. However, their detection and characterization remains challenging. So far, most studies, especially those based on the automatic detection of tidal perturbations, do not disentangle between the different types of debris.

We will present an online annotation tool that allows us to draw with precision the shapes of the different LSB structures, including foreground light and artefacts, superimposed on deep images. Annotations are stored in a database, from which quantitative measurements about tidal features, such as their shape, area, length, and surface brightness, can be computed. We will present results of the annotations of about 350 systems from various deep imaging surveys made with the MegaCam camera at CFHT. A similar study was performed on mock images of simulated galaxies for comparison purposes. We will discuss the preliminary results and statistics based on these dual explorations.

Introduction

From Sola et al 2022 (A&A in press, arXiv:2203.03973):

 

According to hierarchical model of galactic evolution, galaxies form through successive mergers, accretion of small systems and smooth accretion of gas.

 

These interactions leave collisional Low Surface Brightness (LSB) debris around galaxies, such as tidal tails, streams or shells. Studying these tidal debris is essential to constrain models of galactic evolution, as their origin and properties depend on the type and characteristics of the collisions.

 

Most studies that focused on tidal debris outside the Local Group have performed numerical census of these structures, either through visual inspection or more recently with deep learning. Detailed analysis including photometry has been carried out for a small number of objects

 

Yet, essentially because of the lack of convenient tools able to precisely study tidal debris, characterization of such LSB structures around large samples of galaxies had not been performed: this is the goal of this Paper.

 

 

Examples of tidal features detected in CFIS r-band images displayed with a asinh scale. A true color image from the PanSTARRS DR1 survey is overlaid at the center of the target galaxy. The first row shows tidal tails and plumes, the middle row streams and the bottom one shells.

Goals, data and methods

Goals

Obtain quantitative measurements (area, length, surface brigthness) of tidal features and stellar halos around hundreds of galaxies in deep images.

 

Data

We used r-band deep images from two Large Programs of the Canada-France-Hawaii Telescope (CFHT): MATLAS and CFIS.

We studied nearby (D<42 Mpc) massive (>6x109Msun) galaxies: 186 Early-Type Galaxies (ETGs) and 166 Late-Type Galaxies (LTGs).

Methods

We developed an online annotation tool that enable several collaborators to draw with precision the shapes of LSB structures superimposed on deep images and label them, allowing a quantitative analysis of the LSB structures of various types.

Through an easy navigation and user-friendly interface, collaborators are asked to annotated tidal features, stellar halos but also sources of contamination such as galactic cirrus or bright halos from internal reflections on the camera.

The coordinates of the annotations are stored in a database with the associated label. From it, quantitative analyses can be performed and direct applications can be implemented:

  • Shape: area, length, width
  • Surface brightness
  • Luminosity
  • Color
  • Automated aperture photometry
  • Annotation dataset for deep learning algorithms (already tested, see Richards et al, 2022)

The annotation interface with its main facilities: drawing buttons (label 1), classification menu (label 2), examples of already drawn
annotations (label 3) and summary table (label 4).

Results

Tidal tails and streams

Histogram of the width in kiloparsecs of tidal tails (in blue) and streams (in red), in bins of width 2 kpc. The median of each distribution is represented by the dotted lines.

Histogram of the median surface brightness value in magnitudes per square arcsecond for tidal tails (in blue) and streams (in red), in bins of width 0.5 mag arcsec2 . The median of each distribution is represented by the dotted lines.

➔ Tails are wider than streams. This was expected from models: since streams originate from low-mass companions with lower velocity dispersion, their width should be smaller.
➔ Tails are brighter than streams. This may be due to some age effects, with streams having a longer survival (i.e. visibility) time than tails, which are more easily identified as such when they are young.

 

About colors, tidal tails are bluer than streams by around 0.1 mag. This is likely due to an age bias.

Mass assembly

Histogram of the percentage of galaxy mass contained in the halo.  The median of each distribution is represented by the dotted lines.

Histogram of the percentage of galaxy mass contained in tidal debris.  The median of each distribution is represented by the dotted lines.

With our tool, we can probe the mass assembly of galaxies by studing the percentage of mass contained in LSB features.

 ➔ Tidal tails and streams only account for a very small fraction (1-2%) of the total galaxy mass.

 ➔ Extended stellar halos enclose about 10% of the galaxy mass.

Conclusions

We developed an online annotation tool that enables collaborators to draw the shapes of LSB structures with precision in deep images for a large number of galaxies.

Our annotation database can be used to automatically estimate some physical quantities, such as the shape, size, surface brightness, colors and luminosity of tidal tails, streams, shells and stellar halos. Such values were so far not very well constrained on large samples of galaxies.

They may be compared to simulations to better understand the type of mergers that took place, and more generally to constrain models of galactic evolution.

Deep learning algorithms can and already have been trained on our annotation dataset to automatically detect tidal structures in deep images.