Search for: All records

Abstract The rapid advancement of large-scale cosmological simulations has opened new avenues for cosmological and astrophysical research. However, the increasing diversity among cosmological simulation models presents a challenge to therobustness. In this work, we develop the Model-Insensitive ESTimator (Miest), a machine that canrobustlyestimate the cosmological parameters, Ω_mandσ₈, from neural hydrogen maps of simulation models in the Cosmology and Astrophysics with MachinE Learning Simulations project—IllustrisTNG,SIMBA, Astrid, and SWIFT-Eagle. An estimator is consideredrobustif it possesses a consistent predictive power across all simulations, including those used during the training phase. We train our machine using multiple simulation models and ensure that it only extracts common features between the models while disregarding the model-specific features. This allows us to develop a novel model that is capable of accurately estimating parameters across a range of simulation models, without being biased toward any particular model. Upon the investigation of the latent space—a set of summary statistics, we find that the implementation ofrobustnessleads to the blending of latent variables across different models, demonstrating the removal of model-specific features. In comparison to a standard machine lackingrobustness, the average performance of Mieston the unseen simulations during the training phase has been improved by ∼17% for Ω_mand 38% forσ₈. By using a machine learning approach that can extractrobust, yet physical features, we hope to improve our understanding of galaxy formation and evolution in a (subgrid) model-insensitive manner, and ultimately, gain insight into the underlying physical processes responsible forrobustness. 

ABSTRACT The origin of the ‘seeds’ of supermassive black holes (BHs) continues to be a puzzle, as it is currently unclear if the imprints of early seed formation could survive to today. We examine the signatures of seeding in the local Universe using five $$[18~\mathrm{Mpc}]^3$$BRAHMA simulation boxes run to $z=0$. They initialize $$1.5\times 10^5~\rm {M}_{\odot }$$ BHs using different seeding models. The first four boxes initialize BHs as heavy seeds using criteria that depend on dense and metal-poor gas, Lyman–Werner radiation, gas spin, and environmental richness. The fifth box initializes BHs as descendants of lower mass seeds ($$\sim 10^3~\rm {M}_{\odot }$$) using a new stochastic seed model built in our previous work. In our simulations, we find that the abundances and properties of $$\sim 10^5-10^6~\rm {M}_{\odot }$$ local BHs hosted in $$M_*\lesssim 10^{9}~\rm {M}_{\odot }$$ dwarf galaxies, are sensitive to the assumed seeding criteria. This is for two reasons: (1) there is a substantial population of local $$\sim 10^5~\rm {M}_{\odot }$$ BHs that are ungrown relics of early seeds from $$z\sim 5-10$$; (2) BH growth up to $$\sim 10^6~\rm {M}_{\odot }$$ is dominated by mergers in our simulations all the way down to $$z\sim 0$$. As the contribution from gas accretion increases, the signatures of seeding start to weaken in more massive $$\gtrsim 10^6~\rm {M}_{\odot }$$ BHs, and they are erased for $$\gtrsim 10^7~\rm {M}_{\odot }$$ BHs. The different seed models explored here predict abundances of local $$\sim 10^6~\rm {M}_{\odot }$$ BHs ranging from $$\sim 0.01-0.05~\mathrm{Mpc}^{-3}$$ with occupation fractions of $$\sim 20-100~{{\ \rm per\ cent}}$$ for $$M_*\sim 10^{9}~\rm {M}_{\odot }$$ galaxies. These results highlight the potential for placing constraints on seeding models using local $$\sim 10^5-10^6~\rm {M}_{\odot }$$ BHs hosted in dwarf galaxies. Since merger dynamics and accretion physics impact the persistence of seeding signatures, and both high and low mass seed models can produce similar local BH populations, disentangling their roles will require combining high and low redshift constraints. 

ABSTRACT The mass assembly history (MAH) of dark matter haloes plays a crucial role in shaping the formation and evolution of galaxies. MAHs are used extensively in semi-analytic and empirical models of galaxy formation, yet current analytic methods to generate them are inaccurate and unable to capture their relationship with the halo internal structure and large-scale environment. This paper introduces florah (FLOw-based Recurrent model for Assembly Histories), a machine-learning framework for generating assembly histories of ensembles of dark matter haloes. We train florah on the assembly histories from the Gadget at Ultra-high Redshift with Extra Fine Time-steps and vsmdplN-body simulations and demonstrate its ability to recover key properties such as the time evolution of mass and concentration. We obtain similar results for the galaxy stellar mass versus halo mass relation and its residuals when we run the Santa Cruz semi-analytic model on florah-generated assembly histories and halo formation histories extracted from an N-body simulation. We further show that florah also reproduces the dependence of clustering on properties other than mass (assembly bias), which is not captured by other analytic methods. By combining multiple networks trained on a suite of simulations with different redshift ranges and mass resolutions, we are able to construct accurate main progenitor branches with a wide dynamic mass range from $z=0$ up to an ultra-high redshift $$z \approx 20$$, currently far beyond that of a single N-body simulation. florah is the first step towards a machine learning-based framework for planting full merger trees; this will enable the exploration of different galaxy formation scenarios with great computational efficiency at unprecedented accuracy. 

ABSTRACT Arkenstone is a new scheme that allows multiphase, stellar feedback-driven winds to be included in coarse resolution cosmological simulations. The evolution of galactic winds and their subsequent impact on the circumgalactic medium are altered by exchanges of mass, energy, momentum, and metals between their component phases. These exchanges are governed by complex, small-scale physical processes that cannot be resolved in cosmological simulations. In this second presentation paper, we describe Arkenstone’s novel cloud particle approach for modelling unresolvable cool clouds entrained in hot, fast winds. This general framework allows models of the cloud–wind interaction, derived from state-of-the-art high-resolution simulations, to be applied in a large-scale context. In this work, we adopt a cloud evolution model that captures simultaneous cloud mass loss to and gain from the ambient hot phase via turbulent mixing and radiative cooling, respectively. We demonstrate the scheme using non-cosmological idealized simulations of a galaxy with a realistic circumgalactic medium component, using the arepo code. We show that the ability of a high-specific energy wind component to perform preventative feedback may be limited by heavy loading of cool clouds coupled into it. We demonstrate that the diverging evolution of clouds of initially differing masses leads to a complex velocity field for the cool phase and a cloud mass function that varies both spatially and temporally in a non-trivial manner. These latter two phenomena can manifest in the simulation because of our choice of a Lagrangian discretization of the cloud population, in contrast to other proposed schemes. 

Abstract This paper presents a new framework for understanding the relationship between a galaxy and its circumgalactic medium (CGM). It focuses on howimbalancesbetween heating and cooling cause either expansion or contraction of the CGM. It does this by trackingallof the mass and energy associated with a halo’s baryons, including their gravitational potential energy, even if feedback has pushed some of those baryons beyond the halo’s virial radius. We show how a star-forming galaxy’s equilibrium state can be algebraically derived within the context of this framework, and we analyze how the equilibrium star formation rate depends on supernova feedback. We consider the consequences of varying the mass loading parameter ${\eta }_M\equiv {\dot{M}}_{\mathrm{wind}}/{\dot{M}}_{*}$relating a galaxy’s gas mass outflow rate ( ${\dot{M}}_{\mathrm{wind}}$) to its star formation rate ( ${\dot{M}}_{*}$) and obtain results that challenge common assumptions. In particular, we find that equilibrium star formation rates in low-mass galaxies are generally insensitive to mass loading, and when mass loading does matter, increasing it actually results inmorestar formation because more supernova energy is needed to resist atmospheric contraction. 

Abstract The scaling of galaxy properties with halo mass suggests that feedback loops regulate star formation, but there is no consensus yet about how those feedback loops work. To help clarify discussions of galaxy-scale feedback, Paper I presented a very simple model for supernova feedback that it called the minimalist regulator model. This follow-up paper interprets that model and discusses its implications. The model itself is an accounting system that tracks all of the mass and energy associated with a halo’s circumgalactic baryons—the central galaxy’s atmosphere. Algebraic solutions for the equilibrium states of that model reveal that star formation in low-mass halos self-regulates primarily by expanding the atmospheres of those halos, ultimately resulting in stellar masses that are insensitive to the mass-loading properties of galactic winds. What matters most is the proportion of supernova energy that couples with circumgalactic gas. However, supernova feedback alone fails to expand galactic atmospheres in higher-mass halos. According to the minimalist regulator model, an atmospheric contraction crisis ensues, which may be what triggers strong black hole feedback. The model also predicts that circumgalactic medium properties emerging from cosmological simulations should depend largely on the specific energy of the outflows they produce, and we interpret the qualitative properties of several numerical simulations in light of that prediction. 

ABSTRACT In recent years, cosmological hydrodynamical simulations have proven their utility as key interpretative tools in the study of galaxy formation and evolution. In this work, we present a comparative analysis of the baryon cycle in three publicly available, leading cosmological simulation suites: EAGLE, IllustrisTNG, and SIMBA. While these simulations broadly agree in terms of their predictions for the stellar mass content and star formation rates of galaxies at $$z\approx 0$$, they achieve this result for markedly different reasons. In EAGLE and SIMBA, we demonstrate that at low halo masses ($$M_{\rm 200c}\lesssim 10^{11.5}\, \mathrm{M}_{\odot }$$), stellar feedback (SF)-driven outflows can reach far beyond the scale of the halo, extending up to $$2\!-\!3\times R_{\rm 200c}$$. In contrast, in TNG, SF-driven outflows, while stronger at the scale of the interstellar medium, recycle within the circumgalactic medium (within $$R_{\rm 200c}$$). We find that active galactic nucleus (AGN)-driven outflows in SIMBA are notably potent, reaching several times $$R_{\rm 200c}$$ even at halo masses up to $$M_{\rm 200c}\approx 10^{13.5}\, \mathrm{M}_{\odot }$$. In both TNG and EAGLE, AGN feedback can eject gas beyond $$R_{\rm 200c}$$ at this mass scale, but seldom beyond $$2\!-\!3\times R_{\rm 200c}$$. We find that the scale of feedback-driven outflows can be directly linked with the prevention of cosmological inflow, as well as the total baryon fraction of haloes within $$R_{\rm 200c}$$. This work lays the foundation to develop targeted observational tests that can discriminate between feedback scenarios, and inform subgrid feedback models in the next generation of simulations. 

ABSTRACT JWST has revealed a large population of accreting black holes (BHs) in the early Universe. Recent work has shown that even after accounting for possible systematic biases, the high-z$$M_*{\!-\!}M_{\rm \rm bh}$$ relation can be above the local scaling relation by $$\gt 3\sigma$$. To understand the implications of these overmassive high-z BHs, we study the BH growth at $$z\sim 4{\!-\!}7$$ using the $$[18~\mathrm{Mpc}]^3$$BRAHMA cosmological simulations with systematic variations of heavy seed models that emulate direct collapse black hole (DCBH) formation. In our least restrictive seed model, we place $$\sim 10^5~{\rm M}_{\odot }$$ seeds in haloes with sufficient dense and metal-poor gas. To model conditions for direct collapse, we impose additional criteria based on a minimum Lyman Werner flux (LW flux $$=10~J_{21}$$), maximum gas spin, and an environmental richness criterion. The high-z BH growth in our simulations is merger dominated, with a relatively small contribution from gas accretion. The simulation that includes all the above seeding criteria fails to reproduce an overmassive high-z$$M_*{\!-\!}M_{\rm bh}$$ relation consistent with observations (by factor of $$\sim 10$$ at $$z\sim 4$$). However, more optimistic models that exclude the spin and environment based criteria are able to reproduce the observed relations if we assume $$\lesssim 750~\mathrm{Myr}$$ delay times between host galaxy mergers and subsequent BH mergers. Overall, our results suggest that current JWST observations may be explained with heavy seeding channels if their formation is more efficient than currently assumed DCBH conditions. Alternatively, we may need higher initial seed masses, additional contributions from lighter seeds to BH mergers, and / or more efficient modes for BH accretion. 

Abstract The observed rest-UV luminosity function at cosmic dawn (z∼ 8–14) measured by JWST revealed an excess of UV-luminous galaxies relative to many prelaunch theoretical predictions. A high star formation efficiency (SFE) and a top-heavy initial mass function (IMF) are among the mechanisms proposed for explaining this excess. Although a top-heavy IMF has been proposed for its ability to increase the light-to-mass ratio (Ψ_UV), the resulting enhanced radiative pressure from young stars could decrease the SFE, potentially driving galaxy luminosities back down. In this Letter, we use idealized radiation hydrodynamic simulations of star cluster formation to explore the effects of a top-heavy IMF on the SFE of clouds typical of the high-pressure conditions found at these redshifts. We find that the SFE in star clusters with solar-neighborhood-like dust abundance decreases with increasingly top-heavy IMFs—by ∼20% for an increase of a factor of 4 in Ψ_UVand by 50% for a factor of ∼10 in Ψ_UV. However, we find that an expected decrease in the dust-to-gas ratio (∼0.01 × solar) at these redshifts can completely compensate for the enhanced light output. This leads to a (cloud-scale; ∼10 pc) SFE that is ≳70% even for a factor of 10 increase in Ψ_UV, implying that highly efficient star formation is unavoidable for high surface density and low-metallicity conditions. Our results suggest that a top-heavy IMF, if present, likely coexists with efficient star formation in these galaxies. 

While the first “seeds” of supermassive black holes (BH) can range from ~10^2-10^6 M_\odot, the lowest mass seeds (< 10^3 M_\odot) are inaccessible to most cosmological simulations due to resolution limitations. We present our new BRAHMA simulations that use a novel flexible seeding approach to predict the z>7 BH populations for low-mass seeds. We ran two types of boxes that model ~10^3 M_\odot seeds using two distinct but mutually consistent seeding prescriptions at different simulation resolutions. First, we have the highest resolution [9 Mpc]^3 (BRAHMA-9-D3) boxes that directly resolve ~10^3 M_\odot seeds and place them within haloes with dense, metal-poor gas. Second, we have lower resolution, larger volume [18 Mpc]^3 (BRAHMA-18-E4), and ~[36 Mpc]^3 (BRAHMA-36-E5) boxes that seed their smallest resolvable ~10^4 & 10^5 M_\odot BH descendants using new stochastic seeding prescriptions calibrated using BRAHMA-9-D3. The three boxes together probe key BH observables between ~10^3, and,10^7 M_\odot. The active galactic nuclei (AGN) luminosity function variations are small (factors of ~2-3) at the anticipated detection limits of potential future X-ray facilities (~10^{43} ergs/s at z~7). Our simulations predict BHs ~10-100 times heavier than the local M_* versus M_bh relations, consistent with several JWST-detected AGN. For different seed models, our simulations merge binaries at ~1-15 kpc, with rates of ~200-2000 yr−1 for >10^3 M_\odot BHs, ~6-60 yr−1 for >10^4 M_\odot BHs, and up to ~10 yr−1 amongst >10^5 M_\odot BHs. These results suggest that Laser Interferometer Space Antenna mission has promising prospects for constraining seed models.