Clumping and the Interpretation of kpc-Scale Maps of the Interstellar Medium: Smooth HI and Clumpy, Variable H2 Surface Density

TL;DR

This paper introduces a 'clumping factor' to translate between large-scale and cloud-scale surface densities in the interstellar medium, revealing significant differences in clumping between atomic and molecular gas.

Contribution

It defines and measures a clumping factor for atomic and molecular gas, highlighting the need for improved models linking cloud-scale and galaxy-scale ISM observations.

Findings

HI emission is very smooth with a clumping factor of ~1.3.

CO emission is highly clumped with a median clumping factor of ~7.

No universal molecular gas surface density is supported by current data.

Abstract

Many recent models consider the structure of individual interstellar medium (ISM) clouds as a way to explain observations of large parts of galaxies. To compare such models to observations, one must understand how to translate between surface densities observed averaging over large (~kpc) scales and surface densities on the scale of individual clouds (~pc scale), which are treated by models. We define a "clumping factor" that captures this translation as the ratio of the mass-weighted surface density, which is often the quantity of physical interest, to the area-weighted surface density, which is observed. We use high spatial resolution (sub-kpc) maps of CO and HI emission from nearby galaxies to measure the clumping factor of both atomic and molecular gas. The molecular and atomic ISM exhibit dramatically different degrees of clumping. As a result, the ratio H2/HI measured at ~kpc resolution cannot be trivially interpreted as a cloud-scale ratio of surface densities. HI emission appears very smooth, with a clumping factor of only ~1.3. Based on the scarce and heterogeneous high resolution data available, CO emission is far more clumped with a widely variable clumping factor, median ~7 for our heterogeneous data. Our measurements do not provide evidence for a universal mass-weighted surface density of molecular gas, but also cannot conclusively rule out such a scenario. We suggest that a more sophisticated treatment of molecular ISM structure, one informed by high spatial resolution CO maps, is needed to link cloud-scale models to kpc-scale observations of galaxies.