Physical versus observational properties of clouds in turbulent molecular cloud models

We examine the question of how well the physical properties of clumps in turbulent molecular clouds can be determined by measurements of observed clump structures. To do this, we compare simulated observations of three-dimensional numerical models of isothermal, magnetized, supersonic turbulence with the actual physical structure of the models. We begin by determining how changing the parameters of the turbulence changes the structure of the simulations. Stronger driving produces greater density fluctuations, and longer wavelength driving produces larger structures. Magnetic fields have a less pronounced effect on structure, one that is not monotonic with field strength. Aligned structures are seen only with low-density tracers and when the intensity of the field is large. Comparing different regions with the same tracers (or conversely, the same region with different tracers) can give information about the physical conditions of the region. In particular, different density tracers can help determine the size of the density fluctuations and thus the strength of the driving. Nevertheless, velocity superposition of multiple physical clumps can fully obscure the physical properties of those clumps, and short-wavelength (compared with the size of the region under analysis) driving worsens this effect. We then compare Larson's relationships and mass spectra in physical and observational space for the same structure data set. We confirm previous claims that the mean density-size relationship is an observational artifact due to limited dynamical range in column density: it is the inevitable consequence of the presence of a lower cutoff in column density. The velocity dispersion-size relationship, on the other hand, is reproduced in both physical and observed clumps, although with substantial scatter in the derived slope, consistent with observations. Finally, we compute the mass spectra for the models and compare them to mass spectra derived from simulated observations of the models. We show that when we look for clumps with high enough resolution, both spectra converge to the same shape. This shape appears to be lognormal, however, rather than the power-law function usually used in the literature.