http://arxiv.org/abs/1403.5566

We use three-dimensional numerical hydrodynamic simulations of the turbulent, multiphase atomic interstellar medium (ISM) to construct and analyze synthetic HI 21 cm emission and absorption lines. Our analysis provides detailed tests of 21 cm observables as physical diagnostics of the atomic ISM. In particular, we construct (1) the “observed” spin temperature, $T_{s,obs}(v_{ch})\equiv T_B(v_{ch})/[1-e^{-{\tau}(v_{ch})}]$, and its optical-depth weighted mean T_s,obs; (2) the absorption-corrected “observed” column density, $N_{H,obs}\propto \int dv_{ch} T_B(v_{ch}){\tau}(v_{ch})/[[1-e^{-{\tau}(v_{ch})}]$; and (3) the “observed” fraction of cold neutral medium (CNM), $f_{c,obs}\equiv T_c/T_{s,obs}$ for T_c the CNM temperature; we compare each observed parameter with true values obtained from line-of-sight (LOS) averages in the simulation. Within individual velocity channels, T_s,obs(v_ch) is within a factor 1.5 of the true value up to ${\tau}(v_{ch})\approx10$. As a consequence, N_H,obs and T_s,obs are respectively within 5% and 12% of the true values for 90% and 99% of LOSs. The optically thin approximation significantly underestimates N_H for ${\tau}>1$. Provided that T_c is constrained, an accurate observational estimate of the CNM mass fraction can be obtained down to 20%. We show that T_s,obs cannot be used to distinguish the relative proportions of warm and thermally-unstable atomic gas, although the presence of thermally-unstable gas can be discerned from 21 cm lines with 200K<$T_{s,obs}(v_{ch})$<1000K. Our mock observations successfully reproduce and explain the observed distribution of the brightness temperature, optical depth, and spin temperature in Roy et al. (2013a). The threshold column density for CNM seen in observations is also reproduced by our mock observations. We explain this observed threshold behavior in terms of vertical equilibrium in the local Milky Way’s ISM disk.

Read this paper on arXiv…

C. Kim, E. Ostriker and W. Kim
Tue, 25 Mar 14
69/79