NIHAO XVII: The diversity of dwarf galaxy kinematics and implications for the HI velocity function

TL;DR

This study uses high-resolution cosmological simulations to understand why HI linewidths in dwarf galaxies underpredict their maximum circular velocities, revealing the roles of support, distribution, and observational effects.

Contribution

It identifies the main physical processes causing linewidth underestimation and demonstrates that hydrodynamical simulations align well with observed dwarf galaxy velocity functions.

Findings

Lower mass galaxies are less rotationally supported.

HI distributions are less extended relative to dark matter halos.

The simulated HI velocity function matches observations for dwarf galaxies.

Abstract

We use 85 pairs of high resolution LCDM cosmological simulations from the NIHAO project to investigate why in dwarf galaxies neutral hydrogen (HI) linewidths measured at 50% of the peak flux W_50 /2 (from the hydrodynamical simulations) tend to underpredict the maximum circular velocity VmaxDMO (from the corresponding dark matter only simulations). There are two main contributing processes. 1) Lower mass galaxies are less rotationally supported. This is confirmed observationally from the skewness of linewidths in bins of HI mass in both ALFALFA and HIPASS observations. 2) The HI distributions are less extended (relative to the dark matter halo) in dwarf galaxies. Coupled to the lower baryon-to-halo ratio results in rotation curves that are still rising at the last measured data point, in agreement with observations from SPARC. Combining these two effects, in both simulations and observations lower mass galaxies have on average lower W_50 / W_20. Additionally, mass loss driven by supernovae and projection effects (dwarf galaxies are in general not thin disks) further reduce the linewidths. The implied HI linewidth velocity function from NIHAO is in good agreement with observations in the nearby Universe of dwarf galaxies: 10 < W_50 /2 < 80 km/s. The dark matter only slope of -2.9 is reduced to -1.0 in the hydro simulations. Future radio observations of unbiased samples with higher spatial resolution will enable stricter tests of the simulations, and thus of the LCDM model.