Boundary conditions in hydrodynamic simulations of isolated galaxies and their impact on the gas-loss processes

TL;DR

This study investigates how boundary conditions in hydrodynamic simulations influence gas-loss processes in isolated small galaxies, highlighting the importance of boundary choice for accurate modeling of galactic evolution.

Contribution

It introduces and compares different boundary conditions, demonstrating that selective boundaries improve simulation accuracy and efficiency for small galaxy models.

Findings

Open boundaries smaller than 10 times the galaxy's radius become unphysical after 0.6 Gyr.

Selective boundary conditions reduce reversed shocks and are computationally efficient.

Simulations with well-chosen boundaries can reliably model gas-loss in small galaxies.

Abstract

Three-dimensional hydrodynamic simulations are commonly used to study the evolution of the gaseous content in isolated galaxies, besides its connection with galactic star formation histories. Stellar winds, supernova blasts, and black hole feedback are mechanisms usually invoked to drive galactic outflows and decrease the initial galactic gas reservoir. However, any simulation imposes the need of choosing the limits of the simulated volume, which depends, for instance, on the size of the galaxy and the required numerical resolution, besides the available computational capability to perform it. In this work, we discuss the effects of boundary conditions on the evolution of the gas fraction in a small-sized galaxy (tidal radius of about 1 kpc), like classical spheroidal galaxies in the Local Group. We found that open boundaries with sizes smaller than approximately 10 times the characteristic radius of the galactic dark-matter halo become unappropriated for this kind of simulation after about 0.6 Gyr of evolution, since they act as an infinity reservoir of gas due to dark-matter gravity. We also tested two different boundary conditions that avoid gas accretion from numerical frontiers: closed and selective boundary conditions. Our results indicate that the later condition (that uses a velocity threshold criterion to open or close frontiers) is preferable since minimizes the number of reversed shocks due to closed boundaries. Although the strategy of putting computational frontiers as far as possible from the galaxy itself is always desirable, simulations with selective boundary condition can lead to similar results at lower computational costs.