Understanding biases when fitting disk truncations

Abstract

Truncations in the stellar population at the edges of disk galaxies are thought to be a common morphological feature (e.g., Erwin et al. 2005; and more recently Marino et al. 2016). In fact, using imaging data from the SDSS, Pohlen & Trujillo (2006) showed that only ~ 10% of face-on to intermediate inclined, nearby, late-type (Sb-Sdm) spiral galaxies have a normal/standard purely exponential disk down to the noise limit. In situations like these, the simultaneous fit of two lines, joined or not at an intermediate point (the break radius), constitutes a natural step towards the modelling of radial variation in surface brightness, metallicity, or any other relevant parameter. This work shows the results of simple simulations in which the simultaneous fit to two joined lines is compared to the simultaneous fit of two independent lines (i.e., two lines that do not necessarily coincide at an intermediate point), and also to the traditional single ordinary least squares fit. These simulations reveal some biases that should be taken into account when facing these kind of fitting procedures.