LYRA Galaxy Formation Model

• LYRA Galaxy Formation Model is a high-resolution simulation framework that captures multiphase interstellar medium dynamics, star formation, and supernova feedback.
• It uses the AREPO moving-mesh code with dynamic gas resolution down to 4 M☉ and spatial refinements reaching 0.5 pc to accurately model gas dynamics and feedback events.
• Innovative features such as Population III enrichment and resolved SN blastwaves provide actionable insights into star formation regulation, metal retention, and dark matter profile behavior in low-mass galaxies.

The LYRA Galaxy Formation Model denotes a suite of high-resolution, physics-rich hydrodynamical simulations of galaxy formation that explicitly resolve the multiphase interstellar medium (ISM), the life cycle of individual stars, and discrete supernova (SN) feedback events. The model is constructed atop the moving-mesh magnetohydrodynamics code AREPO, and incorporates algorithmic innovations related to ISM cooling, clustered SN feedback, inhomogeneous enrichment from Population III stars, and the interplay of baryonic dynamics with dark matter. LYRA simulations are specifically tailored to resolve parsec-scale gas dynamics with gas mass resolution as fine as 4 M$_\odot$, enabling direct modeling of feedback-regulated star formation, retention of metals, and the assembly history of low-mass galaxies. The approach represents a marked departure from sub-grid, parameterized feedback models, providing a platform for investigating the physical processes that shape dwarf galaxies and their internal structure across cosmic time (Gutcke et al., 2020, Gutcke et al., 2021).

1. Simulation Framework and Resolution

LYRA models are implemented in AREPO, an unstructured, moving Voronoi mesh code capable of arbitrarily high spatial and mass refinement. Gas cells are dynamically refined or derefined to a target mass (e.g., 4 M$_\odot$) to ensure scale-adaptive representations of evolving density structures. Gravitational softening lengths for gas, stars, and dark matter are set according to the local mass resolution, with typical minimum gas cell softenings reaching 0.5 pc in zoom-in regions. The temperature floor for cooling is set to 10 K, a regime where molecular cooling processes dominate and cold gas phases relevant for star formation are directly resolved. The hydrodynamical schemes capture shocks, turbulence, and mixing with high fidelity, while self-shielding prescriptions allow dense regions to decouple thermally from an evolving UV background (Gutcke et al., 2020).

2. Multiphase ISM Treatment and Star Formation

The ISM in LYRA is modeled as a truly multiphase medium, with distinct cold (molecular), warm, and hot gas regimes emergent from self-consistent radiative cooling and heating (atomic, molecular, and metal-line cooling networks included). Gas at densities exceeding $n_\mathrm{H} \gtrsim 10^3\,\mathrm{cm}^{-3}$ and $T < 100$ K becomes eligible for star formation via a stochastic scheme. Star particles represent either individually resolved stars (for $M_* \gtrsim$ the resolution limit) or "composite" particles for the lower IMF, but in all cases stellar evolution tracks are followed at the star-by-star level. Stellar lifetimes, remnant formation, and subsequent SN explosions are determined from IMF sampling and evolutionary models (Gutcke et al., 2020, Gutcke et al., 2021).

3. Supernova Feedback and Energy Coupling

SN feedback is implemented as discrete, temporally and spatially resolved events coinciding with the death of massive stars ($M_* \gtrsim 8\,M_\odot$). Each SN injects thermal energy, mass, and metals into the hydrodynamic mesh with a blastwave energy drawn from recent progenitor-dependent explosion models, varying between zero and $1.8\times 10^{51}$ erg according to progenitor mass and structure. Time-stepping in the affected region is adaptively refined (down to $<100$ yr) so that the Sedov-Taylor energy-conserving phase is explicitly resolved: post-SN shock propagation, energy partitioning (about 28% transferred to kinetic energy), and creation of a hot, volume-filling ISM phase are consequently captured without recourse to sub-grid approximations. This allows clustered SNe, which typically explode in pre-evacuated low-density gas, to efficiently drive galactic winds and regulate star formation. The resolved feedback attenuates the ISM gas density, suppresses SF, and enhances retention of metals compared to fixed-energy or sub-grid approaches (Gutcke et al., 2020).

4. Population III Enrichment and Early Stellar Populations

A distinguishing feature of LYRA is its subgrid Population III (PopIII) enrichment scheme, crucial for the formation pathways of first galaxies. Rather than imposing a uniform metallicity floor, the model identifies halos surpassing a chosen mass threshold (e.g., $10^6$ or $10^7$ M$_\odot$) in real time via on-the-fly halo finding. When the threshold is crossed, the central gas is instantaneously enriched to [Fe/H] = −4, mimicking the rapid chemical pollution of primordial halos by the first generation of massive stars. This inhomogeneous enrichment mechanism, as opposed to homogeneous enrichment, results in substantial differences in star formation timing, stellar concentration, metallicity distribution functions (MDFs), and the prevalence of luminous substructures. A higher PopIII threshold produces more monolithic, centrally concentrated stellar collapses and forges denser central nuclei in dwarfs; a lower threshold triggers earlier, more spatially distributed star formation and increased substructure (Gutcke et al., 2021).

5. Star Formation Histories and Observable Properties

LYRA simulations of isolated dwarf galaxies ($M_\mathrm{halo}\sim 2\times10^9$–$10^{10}$ M$_\odot$) indicate a characteristic early episode of intense star formation—peaking before reionization (ending by $z\sim8$), with duration $\lesssim 1.5$ Gyr—followed by SF quenching as UV background heating suppresses gas cooling in shallow potentials. Star formation can be re-triggered by mergers ($z\sim4$) or late-time gas accretion ($z\lesssim0.2$), depending subtly on enrichment and merger history. At $z=0$, simulated dwarfs match key observed metrics: stellar masses ($7\times10^5$–$2\times10^6$ M$_\odot$), half-light radii ($\sim$285–850 pc), velocity dispersions ($\sim$12 km/s), and mean metallicities compatible with Local Group analogues. Variation in PopIII enrichment history directly imprints on the stellar MDF and spatial distribution, with bi-modal MDFs and enhanced substructure emerging from lower PopIII mass thresholds (Gutcke et al., 2021).

6. Feedback, Dark Matter Profiles, and Limitations

Despite resolved, efficient, temporally clustered SN feedback, LYRA simulations of dwarf galaxies generally do not generate significant constant-density ("cored") dark matter profiles. The central dark matter slopes remain cuspy, with only minor response in the innermost regions, a consequence of the fact that even at high SF density thresholds and with feedback, the shallow potentials of low-mass halos preclude sufficient baryon-driven potential oscillations. The stars typically dominate within 30–50 pc, but have limited impact on the dark matter profile. This is a robust result across PopIII enrichment variants, suggesting constraints on the conditions necessary for dark core formation and implications for interpreting kinematics of observed dwarf spheroidals (Gutcke et al., 2021).

7. Applications and Broader Implications

The LYRA model's direct coupling of multiphase ISM physics, stochastic (IMF-sampled) star formation, individualized SN feedback, and enriched PopIII subgrid models establishes it as an advanced framework for studying the physical drivers of galaxy assembly, especially in the low-mass regime. Its predictive power for metallicity retention, stellar substructure, and the effects of clustered feedback extends to calibrating sub-grid models in large-volume galaxy simulations. At the same time, the inhomogeneous early metal enrichment and feedback-regulated star formation pathways it explores are crucial for interpreting observations of Local Group dwarfs, the chemical and kinematic assembly of galaxies, and the mechanisms by which low-mass systems survive or quench during the epoch of reionization. The modularity and inherent physical richness of LYRA anticipates its extension to more massive galaxies and to the detailed modeling of the baryon cycle in hierarchical cosmological structure formation (Gutcke et al., 2020, Gutcke et al., 2021).