Dark Matter Substructure in Numerical Simulations: A Tale of Discreteness Noise, Runaway Instabilities, and Artificial Disruption

TL;DR

This study investigates the numerical artifacts affecting dark matter subhalo survival in simulations, revealing that many disruptions are artificial and proposing criteria to identify reliable subhaloes, highlighting limitations in current cosmological models.

Contribution

It identifies key numerical issues causing artificial disruption of subhaloes and introduces criteria to assess their reliability in cosmological simulations.

Findings

Most subhalo disruptions are numerical, not physical.

Inadequate force-softening leads to artificial tidal disruption.

Discreteness noise can trigger runaway instabilities in subhaloes.

Abstract

To gain understanding of the complicated, non-linear and numerical processes associated with the tidal evolution of dark matter subhaloes in numerical simulation, we perform a large suite of idealized simulations that follow individual N-body subhaloes in a fixed, analytical host halo potential. By varying both physical and numerical parameters, we investigate under what conditions the subhaloes undergo disruption. We confirm the conclusions from our more analytical assessment in van den Bosch et al. that most disruption is numerical in origin; as long as a subhalo is resolved with sufficient mass and force resolution, a bound remnant survives. This implies that state-of-the-art cosmological simulations still suffer from significant overmerging. We demonstrate that this is mainly due to inadequate force-softening, which causes excessive mass loss and artificial tidal disruption. In addition, we show that subhaloes in N-body simulations are susceptible to a runaway instability triggered by the amplification of discreteness noise in the presence of a tidal field. These two processes conspire to put serious limitations on the reliability of dark matter substructure in state-of-the-art cosmological simulations. We present two criteria that can be used to assess whether individual subhaloes in cosmological simulations are reliable or not, and advocate that subhaloes that satisfy either of these two criteria be discarded from further analysis. We discuss the potential implications of this work for several areas in astrophysics.