FitSKIRT: genetic algorithms to automatically fit dusty galaxies with a Monte Carlo radiative transfer code

Abstract

We present FitSKIRT, a method to efficiently fit radiative transfer models to UV/optical images of dusty galaxies. These images have the advantage that they have better spatial resolution compared to FIR/submm data. FitSKIRT uses the GAlib genetic algorithm library to optimize the output of the SKIRT Monte Carlo radiative transfer code. Genetic algorithms prove to be a valuable tool in handling the multi-dimensional search space as well as the noise induced by the random nature of the Monte Carlo radiative transfer code. FitSKIRT is tested on artificial images of a simulated edge-on spiral galaxy, where we gradually increase the number of fitted parameters. We find that we can recover all model parameters, even if all 11 model parameters are left unconstrained. Finally, we apply the FitSKIRT code to a V-band image of the edge-on spiral galaxy NGC 4013. This galaxy has been modeled previously by other authors using different combinations of radiative transfer codes and optimization methods. Given the different models and techniques and the complexity and degeneracies in the parameter space, we find reasonable agreement between the different models. We conclude that the FitSKIRT method allows comparison between different models and geometries in a quantitative manner and minimizes the need of human intervention and biasing. The high level of automation makes it an ideal tool to use on larger sets of observed data.

HERschelObservations of Edge-on Spirals (HEROES)

“…This is particularly remarkable for NGC 4013, since the values for τ f V and h R obtained by the radiative transfer fits are quite different. This is due to a degeneracy in the radiative transfer modelling of edge-on spiral galaxies, which has been noted by Bianchi (2007) and De Geyter et al (2013). Systems with a large face-on optical depth and a small dust scalelength and systems with a small face-on optical depth and a large dust scalelength can both result in similar edge-on optical depth and hence dust lanes of similar depths.…”

High-resolution, 3D radiative transfer modeling

“…The scale heights have been shown to be comparable for stars and dust in low mass galaxies (Seth et al 2005), but for high mass galaxies the ISM is considered to collapse into dust lanes distributed in thin disks (Dalcanton et al 2004). Also model fitting to several edge-on galaxies have resulted in smaller dust scale heights compared to the vertical extent of the thick disk composed of old stars (Jansen et al 1994;Xilouris et al 1999;Bianchi 2007;De Geyter et al 2013Verstappen et al 2013). The scale height of dust in galaxies is, however, still debated with several galaxies showing evidence of extra-planar dust (e.g., Howk & Savage 1999;Alton et al 2000;Thompson et al 2004).…”

“…Higher values of τ V would imply that a significant portion of the stellar V band light is absorbed, and reprocessed by dust grains. Based on radiative transfer simulations of edge-on spiral galaxies, the face-on optical depth in most optical wavebands is generally concluded to be lower than one, making the disk nearly transparent in the optical wavelength regime (Xilouris et al 1997(Xilouris et al , 1998(Xilouris et al , 1999Alton et al 2004;Bianchi 2007;Popescu et al 2011;De Geyter et al 2013. Some exceptions were found for the secondary dust disk used to explain the missing FIR emission in RT models of edge-on galaxies (Popescu et al 2000(Popescu et al , 2011Misiriotis et al 2001;MacLachlan et al 2011).…”

“…Up to the present day, RT modeling has been applied to a number of edge-on spiral galaxies (Misiriotis et al 2001;Alton et al 2004;Dasyra et al 2005;Bianchi 2008;Baes et al 2010;Popescu et al 2011;De Looze et al 2012a,b;Schechtman-Rook et al 2012;De Geyter et al 2013, which have the advantage that the extinction and emission of dust can easily be observed along the line-of-sight, and that the dust can be vertically resolved and traced out to large radii. One drawback of modeling edge-on galaxies is the lack of insight in the spatial distribution of star-forming regions and the clumpiness of the interstellar medium, which impedes a spatially resolved analysis of the various stellar populations and dust grains, and prevents analyzing the interplay between stars and dust to characterize the main dust heating mechanisms.…”

Measuring the dust content and formation in SN 1987A using detailed radiative transfer modelling

“…We built a fitting routine on top of SKIRT that automatically determines the model parameters that best reproduce a given data set. The optimisation is based on genetic algorithms, which are ideally suited for fitting noisy Monte Carlo radiative transfer models (De Geyter et al 2013. All model parameters can be varied, such as dust mass or geometry-related parameters, but also the grain size distribution, the dust clumpiness, etc.…”