SiMBA Cosmological Simulations

• SiMBA is a comprehensive suite of cosmological hydrodynamic simulations that model galaxy formation, SMBH growth, AGN feedback, and dust production.
• It uses advanced numerical methods in GIZMO’s MFM mode with H₂-based star formation and on-the-fly radiative cooling calibrated to match multi-redshift observations.
• SiMBA integrates dual-mode kinetic AGN feedback and both torque-limited and Bondi accretion to reproduce galaxy scaling relations and quenching processes.

Simba refers to a comprehensive suite of cosmological hydrodynamic simulations and the underlying codebase designed to study galaxy formation, supermassive black hole (SMBH) growth, AGN feedback, and the resulting observable properties of galaxies and large-scale cosmic structures. The Simba simulations represent a major advance over earlier works by integrating an H₂-based star-formation law, torque-limited and Bondi black-hole accretion channels, a dual-mode kinetic AGN feedback model (winds and jets), and on-the-fly tracking of dust production and evolution. Simba is constructed on the GIZMO gravity+hydrodynamics code in its Meshless Finite Mass (MFM) mode and incorporates extensive sub-grid physics modules with parameters calibrated to match a wide array of low- and high-redshift observations (1901.10203).

1. Code Framework and Numerical Methods

The Simba simulations use the public GIZMO code (Hopkins 2015, 2018) in MFM mode, which combines adaptive Lagrangian particle-based hydrodynamics with Riemann-solver-based shock capturing and strict mass conservation. Radiative cooling, non-equilibrium primordial and metal-line cooling, and photoionization heating are computed on the fly using GRACKLE-3.1, including integrated self-shielding via Rahmati et al. (2013) (1901.10203). Star formation is implemented as an H₂-based Schmidt law following Krumholz & Gnedin (2011): SFR = ε* ρ_H₂ / t_dyn, with ε* = 0.02, and local H₂ fractions assigned explicitly using metallicity and column density prescriptions.

Chemical enrichment tracks 11 species (H, He, C, N, O, Ne, Mg, Si, S, Ca, Fe), with yields from classic supernova and AGB sources. Simba augments the standard galaxy formation toolkit by including an on-the-fly model for dust production, grain growth, sputtering, and destruction based on the local interstellar environment (1901.10203).

2. Black Hole Accretion and Feedback Mechanisms

2.1. Accretion Prescriptions

Simba advances SMBH growth modeling by combining two accretion regimes:

• Torque-limited accretion for cold gas (T < 10⁵ K), scaling as $$\dot M_{\rm torque} \sim \epsilon_T\, f_d^{5/2}\, M_{\rm BH}^{1/6}\, M_{\rm enc}(R_0)\, (R_0/100\,\mathrm{pc})^{-3/2}\, [1+f_0/f_{\rm gas}]^{-1}$$,
• Bondi accretion for hot gas (T > 10⁵ K): $$\dot M_{\rm Bondi} = \epsilon_m\,4\pi\,G^2\,M_{\rm BH}^2\,\rho /(c_s^2)^{3/2}$$,

with key parameters set to match the local $M_{\rm BH}$–$M_*$ relation. Total SMBH accretion combines both channels, capped at 3×Eddington for torque-limited inflow and 1×Eddington for Bondi (Thomas et al., 2019).

2.2. Feedback Channels

Black hole feedback is kinetic and bipolar, aligned with the rotational axis of the gas:

• Radiative mode winds are launched at high Eddington ratios ($f_{\rm Edd}\gtrsim0.2$) with $$v_w = 500 + 500\,(\log_{10}(M_{\rm BH}/10^6\,M_\odot))/3$$ km/s (Perna et al. 2017).
• Jet mode is triggered below $f_{\rm Edd}<0.2$ and $M_{\rm BH}>10^{7.5}\,M_\odot$, where velocities increase up to $v_w + 7000\,\log_{10}(0.2/f_{\rm Edd})$ km/s, capped at an additional 7000 km/s. Both modes enforce a momentum injection rate of $20L/c$ for self-regulation and galactic quenching (Appleby et al., 2019, Thomas et al., 2020).
• X-ray feedback (active when $$\dot M_{\rm Bondi} = \epsilon_m\,4\pi\,G^2\,M_{\rm BH}^2\,\rho /(c_s^2)^{3/2}$$0 in jet mode) deposits energy/momentum into gas, powering central gas evacuation and maintenance-mode AGN feedback in massive galaxies (Appleby et al., 2019).

3. Simulation Suite and Physical Scaling

Simba's flagship run occupies a $$\dot M_{\rm Bondi} = \epsilon_m\,4\pi\,G^2\,M_{\rm BH}^2\,\rho /(c_s^2)^{3/2}$$1 cosmological volume with $$\dot M_{\rm Bondi} = \epsilon_m\,4\pi\,G^2\,M_{\rm BH}^2\,\rho /(c_s^2)^{3/2}$$2 dark matter and gas particles ($$\dot M_{\rm Bondi} = \epsilon_m\,4\pi\,G^2\,M_{\rm BH}^2\,\rho /(c_s^2)^{3/2}$$3), force softening $$\dot M_{\rm Bondi} = \epsilon_m\,4\pi\,G^2\,M_{\rm BH}^2\,\rho /(c_s^2)^{3/2}$$4 kpc, and Planck 2016 cosmological parameters. Ancillary volumes of 50, 25, 12.5 $$\dot M_{\rm Bondi} = \epsilon_m\,4\pi\,G^2\,M_{\rm BH}^2\,\rho /(c_s^2)^{3/2}$$5 Mpc enable resolution convergence and rare object studies (1901.10203). Galaxy finding and halo cataloging occur on-the-fly using star+gas FOF and standard linking lengths.

4. Observational Validation and Main Results

Simba matches a wide array of observables:

• Stellar Mass Functions: The $$\dot M_{\rm Bondi} = \epsilon_m\,4\pi\,G^2\,M_{\rm BH}^2\,\rho /(c_s^2)^{3/2}$$6–6 galaxy stellar mass function (GSMF) agrees closely with data and other leading simulations (e.g., EAGLE), accurately reproducing both star-forming and quenched galaxy abundances (1901.10203).
• Star Formation Main Sequence: z~0 sSFR–$$\dot M_{\rm Bondi} = \epsilon_m\,4\pi\,G^2\,M_{\rm BH}^2\,\rho /(c_s^2)^{3/2}$$7 main sequence slope, amplitude, and scatter match GSWLC-X2 and higher-z comparators (1901.10203).
• Gas and Metallicity: Fractional H I and H₂ masses reproduce xCOLD GASS and GASS data; mass–metallicity trends match low- and high-z observations (1901.10203).
• Black Hole and Quenching Demographics: Simba recovers the observed $$\dot M_{\rm Bondi} = \epsilon_m\,4\pi\,G^2\,M_{\rm BH}^2\,\rho /(c_s^2)^{3/2}$$8–$$\dot M_{\rm Bondi} = \epsilon_m\,4\pi\,G^2\,M_{\rm BH}^2\,\rho /(c_s^2)^{3/2}$$9, $M_{\rm BH}$0–$M_{\rm BH}$1 relations, and the black hole mass function out to $M_{\rm BH}$2 (Thomas et al., 2019). AGN jet feedback produces realistic quenching at high $M_{\rm BH}$3, shaping the bright-end cutoff of the GSMF (1901.10203).
• Galaxy Sizes and Gas Profiles: Star-forming R-band half-light radii trace SDSS trends, though passive sizes are overestimated at $M_{\rm BH}$4 due to numerical heating. Simba uniquely reproduces central sSFR depressions (“holes”) in green valley galaxies at $M_{\rm BH}$5, attributed to efficient X-ray AGN feedback (Appleby et al., 2019).

Observable	Simba (z=0) Performance	Reference/Benchmark
GSMF (0.1–10^12 M_⊙)	Excellent agreement w/ data	Bernardi 2017, Tomczak 2014
sSFR–$M_{\rm BH}$6 main seq.	Slope, scatter, norm match	GSWLC-X2, Whitaker 2014
MZR (gas & stars)	Matches low/high-z obs	Tremonti 2004, Sanders 2015
H₂ and H I fractions	Consistent with xCOLD GASS/GASS	Saintonge 2017, Catinella 2012
$M_{\rm BH}$7–$M_{\rm BH}$8	Correct slope, no high-z offset	Kormendy & Ho 2013
Radio luminosity	Observed RLF, HERG/LERG ratios	Mauch & Sadler 2007, Best 2012
Galaxy sizes	SF match, quenched too large	Zhang & Yang 2017

5. Black Hole–Galaxy Co-Evolution and AGN Feedback Effects

Simba’s torque-limited + Bondi accretion naturally reproduces the observed invariance of the black hole accretion rate–star formation rate (BHAR–SFR) "main sequence," the anti-correlation of Eddington ratio $M_{\rm BH}$9 with $M_*$0 at low masses, and the lognormal $M_*$1 distribution with a peak near 0.01 at $M_*$2 (Thomas et al., 2019). As a function of redshift, the SMBH mass function decreases in normalization, with the high-mass end shifting from being SF-dominated at high z to quenched-dominated at $M_*$3, in broad agreement with constraints.

Crucially, Simba’s two-mode kinetic black hole feedback—especially the onset of powerful jets above $M_*$4—is essential for quenching massive galaxies, steepening the GSMF high-mass end, producing radio-loud AGN demographics that match HERG/LERG population ratios and the 1.4 GHz radio luminosity function, and redistributing baryons in the IGM and haloes (Thomas et al., 2020, Dong et al., 22 Jul 2025). The inclusion of X-ray feedback yields sharp central quenched regions ("holes") in sSFR and gas fraction profiles, in contrast with previous models (Appleby et al., 2019).

6. Impact, Limitations, and Future Prospects

Simba stands out for its physical realism and multi-faceted validation against galaxy, SMBH, and IGM observables, including the size-luminosity relation, sSFR radial profiles, radio AGN phenomenology, baryon partitioning in the cosmic web, and dust evolution (Dong et al., 22 Jul 2025). Noted limitations include excessive quenched-galaxy sizes at low mass due to numerical heating, possible overly sharp transitions at $M_*$5, and mild deviations in scaling relations at the low-mass end (1901.10203, Appleby et al., 2019).

Extensions in the Simba project include baryon distribution studies with refined cosmic-web classification, improved AGN feedback calibration impacting FRB dispersion measure predictions, and further high-resolution, multi-wavelength synthetic survey outputs (Dong et al., 22 Jul 2025). The Simba codebase and simulation data underpin a broad array of current and future large-scale structure and galaxy formation analyses, motivating their use in observational planning and theoretical exploration of feedback-regulated galaxy evolution.