The Entangled Feedback Impacts of Supernovae in Coarse- versus High-Resolution Galaxy Simulations

TL;DR

This study shows that supernova feedback in dwarf galaxies operates via two distinct channels: low-density SNe drive large-scale, energy-driven outflows, while high-density SNe locally disrupt star-forming clumps, suppressing SF on pc scales. Using ultra-high-resolution simulations (LYRA, RIGEL) alongside a mid-resolution SMUGGLE run, the authors demonstrate that the SN outcome is primarily predicted by the immediate local density and that this bimodality persists only when the environment is resolved down to ≲200 M_⊙; at coarser resolutions, cooling radii and energy partitioning are misestimated, washing out the two channels. Radiation feedback further suppresses the high-density channel by dispersing dense gas, aligning SN environments with the global volume distribution. These results imply that emergent galaxy-scale phenomena, such as hot outflows, cannot be robustly modeled by subgrid feedback alone unless the simulations resolve the ISM at scales where the SN cooling radius and energy partitioning are properly captured. Consequently, subgrid models should be calibrated against high-resolution ISM benchmarks or incorporate approaches that emulate the small-scale ISM properties revealed by ultra-high-resolution studies. The work highlights the fundamental limitation of coarse-resolution galaxy models in predicting the true impact of SNe on outflows and SF regulation, with broad implications for interpreting loading factors and the role of feedback in galaxy evolution.

Abstract

It is often understood that supernova (SN) feedback in galaxies is responsible for regulating star formation and generating gaseous outflows. However, a detailed look at their effect on the local interstellar medium (ISM) on small mass scales in simulations shows that these processes proceed in clearly distinct channels. We demonstrate this finding in two independent simulations with solar-mass resolution, LYRA and RIGEL, of an isolated dwarf galaxy. Focusing on the immediate environment surrounding SNe, our findings suggest that the large-scale effect of a given SN on the galaxy is best predicted by its immediate local density. Outflows are driven by SNe in diffuse regions expanding to their cooling radii on large ($\sim$ kpc) scales, while dense star-forming regions are disrupted in a localized (\sim pc) manner. However, these separate feedback channels are only distinguishable at very high numerical resolutions capable of following scales $\ll 10^3 M_\odot$. On larger scales, ISM densities are greatly mis-estimated, and differences between local environments of SNe become severely washed out. We demonstrate the practical implications of this effect by comparing with a mid-resolution simulation ($M_{\rm ptcl.} \sim 200 M_\odot$) of the same dwarf using the SMUGGLE model. The coarse-resolution simulation cannot self-consistently determine whether a given SN is responsible for generating outflows or suppressing star formation, suggesting that emergent galaxy physics such as star formation regulation through hot-phase outflows is fundamentally unresolvable by subgrid stellar feedback models, without appealing directly to simulations with highly resolved ISM.

Figures (8)

• Figure 1: Summary of the properties of the three main simulations.Left: The SFR (top), mass outflow (middle), and energy outflow (bottom) as a function of time of the main runs of LYRA (red), RIGEL (green), and SMUGGLE (blue), shown in solid lines. The dashed red line in the top left shows the SFR of the LYRA fixed-SNe energy run, which is consistent with SMUGGLE. All outflows are measured in parallel slabs $\pm1$ kpc from the disk center. To the right of each time evolution, a frequency histogram of each quantity, in the same corresponding color. Note that lowest bin of each histogram includes all values below the figure window. Right: Face-on gas projections of each run, color-mapped by temperature and transparency-mapped by column density. The image frames correspond to the line color representing each run on the left. Simulation snapshots are at assorted times: $t=$ 250, 450, 650 Myr respectively.
• Figure 2: Bimodal Distribution of Local Environments to SNe.Top: Change in specific energy $\Delta e$ vs the local density of each SN in LYRA. The energy change is measured by taking, for each LYRA SN, the difference in the total energy of the nearest $M_{\rm env}=20\, \textup{M}_\odot$ to the SN, in the snapshots immediately before and after the SN (dt = 1 Myr). Each SN in LYRA is shown as a red point, alongside $1 \sigma$ and $2 \sigma$ contours of their distribution in solid and dashed red lines, and there is a clear separation of the SNe into a low-density, high-$\Delta e$ population, and a high-density, low-$\Delta e$ population. The gray dashed line shows a specific energy threshold corresponding to 30 ${\rm km}$${\rm s}^{-1}$, or approximately the circular velocity at $\sim 1$ kpc; the $\Delta e$ of a SN needs to be at least above this line to be energetic enough to become an outflow. The purple dashed line shows the expected $\Delta e$ as a function of density, as predicted by KimOstriker15 analytic solutions. The same contours from the RIGEL-noRad run are shown in gray for a consistency check with LYRA. Bottom: Histogram of local densities in which the SNe go off, corresponding to the $x$-scatter in the top panel.
• Figure 3: Low-Density SNe are correlated with energy flows. SN energy injection rate by low-local density (red) and high-local density (blue) SNe, compared to the energy outflow (orange), as a function of time in LYRA. The low-density SNe are more temporally correlated to the energy flux than the high-density SNe. Shown in gray are the energy fluxes at 2 kpc (dashed) and 4 kpc (dotted) for comparison. The cross-correlation $c$ of the energy injections from the low- and high-density SNe with the energy flux are printed in the top left corner with their corresponding color. The cross-correlation of all (both low- and high-density together) SNe with the energy flux is also printed (purple).
• Figure 4: Anisotropic Post-SNe Expansion Regions. Face-on projections of a major outflow episode in LYRA. From top to bottom, then left to right: the column density, metallicity, transverse velocity, vertical velocity dispersion, temperature, and ratio of kinetic to thermal energy. All color maps are in $\log_{10}$ scale. The black and white lines show respectively the $10^{-0.5}$ and $10^{-1.5}$$\textup{M}_\odot\,{\rm pc}^{-2}$ contours of the column density, overlaid on every panel. The morphology of all quantities aligns closely with the column density projection, except for the energy ratio. The box is centered on the location of the outflow episode, which is not the galaxy's center, but is instead at a galactocentric radius of $R=0.78$ kpc. The local circular velocity (35.45 ${\rm km}$${\rm s}^{-1}$) at the box center is subtracted from the velocity maps.
• Figure 5: Effect of Stellar Radiation on gas, SNe density distributions. Gas density-temperature phase diagrams for LYRA (no RT, top) and RIGEL (RT, bottom). The color map shows only gas within the disk, and is weighted by volume. Accompanying histograms of the gas density and temperature are shown in blue. For comparison, the density and temperature of the SNe-hosting environments are also shown ( LYRA in red, RIGEL in green). In particular, when stellar radiation is included, the high-local density population of SNe is suppressed, and SNe tend to go off in hotter environments. Furthermore, when radiation is included, the distribution of the local density of SNe is closely aligned with the overall volume distribution of gas in the disk.