The evolution of phase space densities in star-forming regions

The multidimensional phase space density (both position and velocity) of star-forming regions may encode information on the initial conditions of star and planet formation. Recently, a new metric based on the Mahalanobis distance has been used to show that hot Jupiters are more likely to be found around exoplanet host stars in high six-dimensional phase space density, suggesting a more dynamic formation environment for these planets. However, later work showed that this initial result may be due to a bias in the age of hot Jupiters and the kinematics of their host stars. We test the ability of the Mahalanobis distance and density to differentiate more generally between star-forming regions with different morphologies by applying it to static regions that are either substructured or smooth and centrally concentrated. We find that the Mahalanobis distance is unable to distinguish between different morphologies, and that the initial conditions of the N-body simulations cannot be constrained using only the Mahalanobis distance or density. Furthermore, we find that the more dimensions in the phase space, the less effective the Mahalanobis density is at distinguishing between different initial conditions. We show that a combination of the mean three-dimensional (x, y, z) Mahalanobis density and the Q-parameter for a region can constrain its initial virial state. However, this is due to the discriminatory power of the Q-parameter and not from any extra information imprinted in the Mahalanobis density. We therefore recommend continued use of multiple diagnostics for determining the initial conditions of star-forming regions, rather than relying on a single multidimensional metric.