Protostars and Planets VI, Heidelberg, July 15-20, 2013

Poster 1S001

Information from dust continuum data: Estimation of temperature, spectral index, and column density

Juvela, Mika (University of Helsinki, Department of physics)
Montillaud, Julien (University of Helsinki, Department of physics)
Ysard, Nathalie (Institut d\'Astrophysique Spatiale, CNRS/Université Paris-Sud, France)
Malinen, Johanna (University of Helsinki, Department of physics)
Lunttila, Tuomas (University of Helsinki, Department of physics)

Sub-millimetre continuum data provide information on the column density and dust properties of interstellar clouds. We have compared methods that can be used to derive high-resolution column density maps from, e.g., Herschel measurements. We also have investigated the estimation of dust colour temperature and emissivity spectral index. Radiative transfer models are used to study the differences between the true and apparent cloud properties and to compare the performance of analysis methods. Models show that, because of the nature of spatial temperature variations, externally heated clouds tend to show an artificial positive correlation between colour temperature and spectral index. For clouds with internal heating, the situation is reversed. Analysis of observations is affected by observational noise that also can produce a negative correlation. We find that, compared to direct least squares fitting of modified black body spectra, Bayesian methods and hierarchical statistical models are more accurate although not completely unbiased. Hierarchical models can mask possible local variation in the relation between temperature and spectral index. Palmeirim et al. (2013) derived high-resolution column density maps using a combination of estimates obtained in different wavelength ranges. The method is quite reliable although somewhat sensitive to noise. For clouds with a simple density structure, radiative transfer modelling provides the most accurate estimates. As a simpler alternative, we propose modelling that consists of high-resolution column density and temperature maps that are matched to observations through convolution. The method is able to produce reliable column density estimates even at super-resolution. The method is computationally demanding but still feasible even in the analysis of large Herschel maps.

Click here to view poster PDF