ABSTRACT
Molecular line observations using a variety of tracers are often used to investigate the kinematic structure of molecular clouds. However, measurements of cloud velocity dispersions with different lines, even in the same region, often yield inconsistent results. The reasons for this disagreement are not entirely clear, since molecular line observations are subject to a number of biases. In this paper, we untangle and investigate various factors that drive linewidth measurement biases by constructing synthetic position–position–velocity cubes for a variety of tracers from a suite of self-gravitating magnetohydrodynamic simulations of molecular clouds. We compare linewidths derived from synthetic observations of these data cubes to the true values in the simulations. We find that differences in linewidth as measured by different tracers are driven by a combination of density-dependent excitation, whereby tracers that are sensitive to higher densities sample smaller regions with smaller velocity dispersions, opacity broadening, especially for highly optically thick tracers such as CO, and finite resolution and sensitivity, which suppress the wings of emission lines. We find that, at fixed signal-to-noise ratio, three commonly used tracers, the J = 4 → 3 line of CO, the J = 1 → 0 line of C18O, and the (1,1) inversion transition of NH3, generally offer the best compromise between these competing biases, and produce estimates of the velocity dispersion that reflect the true kinematics of a molecular cloud to an accuracy of $\approx 10{{\ \rm per\ cent}}$ regardless of the cloud magnetic field strengths, evolutionary state, or orientations of the line of sight relative to the magnetic field. Tracers excited primarily in gas denser than that traced by NH3 tend to underestimate the true velocity dispersion by $\approx 20{{\ \rm per\ cent}}$ on average, while low-density tracers that are highly optically thick tend to have biases of comparable size in the opposite direction.
Understanding biases in measurements of molecular cloud kinematics using line emission