Story
Connection between galaxy morphology and dark-matter halo structure II: predicting disk structure from dark-matter halo properties
Key takeaway
New research shows that the structure of a galaxy's disk is closely tied to the properties of its surrounding dark matter halo, providing insight into how galaxies form and evolve.
Quick Explainer
This work explores the fundamental connections between the structure of dark matter halos and the properties of the galaxies they host. The researchers used a combination of advanced modeling techniques, including Random Forest and Symbolic Regression, to establish robust empirical relationships between halo characteristics, such as density, angular momentum, and formation history, and key galaxy features like disk size, thickness, and mass fraction. The models based on full-physics hydrodynamic simulations consistently outperformed those relying solely on dark matter, highlighting the importance of baryonic processes in shaping the observed correlations. This systematic approach provides valuable prescriptions for predicting galaxy structure from underlying dark matter halo properties, offering insights into the co-evolution of galaxies and their dark matter environments.
Deep Dive
Technical Deep Dive: Connection between galaxy morphology and dark-matter halo structure II
Overview
This work presents a comprehensive investigation of the connections between galaxy properties and host dark-matter halo properties in the TNG50 cosmological simulation, focusing on predicting disk structure using halo properties.
Methodology
- Measured 37 halo properties from both the full-physics (hydro) simulations and dark-matter-only (DMO) runs, capturing density structure, angular momentum, shape, formation history, and environment.
- Used the morphological decomposition method developed in Paper I to obtain disk properties (size, scale height, and mass fraction), as well as global galaxy quantities including stellar mass, star formation rate (SFR), and size.
- Employed Random Forest (RF) and Symbolic Regression (SR) to assess predictability, determine the most informative halo features, and derive empirical relations.
Results
Predicting global galaxy properties
- RF and SR models predict global galaxy properties (stellar mass, SFR, size) with very high accuracy, providing valuable prescriptions for empirical modeling.
Predicting disk properties
- Disk properties are predictable with good accuracy, with disk height easier to predict than disk size. Predictions are more accurate for dwarfs (M_vir < 10^11.6 Msun) than for more massive systems.
- Models based on hydro halos consistently outperform DMO-based ones, and the DMO-based SR relations are more complex.
Key halo parameters
- Disk compactness (size relative to halo) is strongly linked to the Einasto shape index α, concentration c, inner-halo spin λinner, total spin λ, and formation redshift z1/2 (hydro); spin λ, halo mass M_vir, redshift z, c, and axis ratio s (DMO).
- Disk thickness is primarily driven by halo mass Mvir, c, λinner, λ, and accretion rate dM/dt (hydro); z, M_vir, λ, c, and dM/dt (DMO).
- Disk mass fraction is largely determined by halo mass, with secondary dependence on c and inner-halo shape.
Interpreting the connections
- Disks alter halo structure, as disk compactness influences the Einasto slope index α.
- Disk compactness evolves with redshift in a strongly mass-dependent way.
- Disk thickness is primarily driven by halo mass and recent halo growth, with disks thinnest in Milky Way-mass halos.
- Inner halo properties improve disk predictions only in hydro simulations, due to baryonic disk-halo coupling.
Limitations & Uncertainties
- Results are specific to the TNG50 physics and may not generalize to other simulations.
- Some disk properties are only marginally resolved at the TNG50 resolution.
- Sample is incomplete at the low-mass end due to the stellar particle threshold for morphological decomposition.
What Comes Next
- Applying the analysis pipeline to other simulation suites to test the generalizability of the results.
- Exploring more sophisticated model selection metrics beyond the heuristic approach used here.
- Investigating the physical origins of the tight scaling between disk height and size.
