An official website of the United States government
Abstract Recent work has pointed out the potential existence of a tight relation between the cosmological parameter Ω m , at fixed Ω b , and the properties of individual galaxies in state-of-the-art cosmological hydrodynamic simulations. In this paper, we investigate whether such a relation also holds for galaxies from simulations run with a different code that makes use of a distinct subgrid physics: Astrid. We also find that in this case, neural networks are able to infer the value of Ω m with a ∼10% precision from the properties of individual galaxies, while accounting for astrophysics uncertainties, as modeled in Cosmology and Astrophysics with MachinE Learning (CAMELS). This tight relationship is present at all considered redshifts, z ≤ 3, and the stellar mass, the stellar metallicity, and the maximum circular velocity are among the most important galaxy properties behind the relation. In order to use this method with real galaxies, one needs to quantify its robustness: the accuracy of the model when tested on galaxies generated by codes different from the one used for training. We quantify the robustness of the models by testing them on galaxies from four different codes: IllustrisTNG, SIMBA, Astrid, and Magneticum. We show that the models perform well on a large fraction of the galaxies, but fail dramatically on a small fraction of them. Removing these outliers significantly improves the accuracy of the models across simulation codes.
more »
« less
Predicting Stellar Masses of the First Galaxies Using Graph Neural Networks
Theoretical models of galaxy formation and evolution are primarily investigated through cosmological simulations and semi-analytical models. The former method consumes O(10^6) core-hours explicitly modeling the dynamics of the galaxies, whereas the latter method only requires O(10^3) core-hours foregoing directly simulating internal structure for computational efficiency. In this work, we present a proof-of-concept machine learning regression model, using a graph neural network architecture, to predict the stellar mass of high-redshift galaxies solely from their dark matter merger trees, trained from a radiation hydrodynamics cosmological simulation of the first galaxies.
more »
« less
- Award ID(s):
- 2108020
- PAR ID:
- 10523637
- Publisher / Repository:
- IOP
- Date Published:
- Journal Name:
- Research Notes of the AAS
- Volume:
- 8
- Issue:
- 4
- ISSN:
- 2515-5172
- Page Range / eLocation ID:
- 108
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Abstract The high incidence rate of the O vi λλ 1032, 1038 absorption around low-redshift, ∼ L * star-forming galaxies has generated interest in studies of the circumgalactic medium. We use the high-resolution EAGLE cosmological simulation to analyze the circumgalactic O vi gas around z ≈ 0.3 star-forming galaxies. Motivated by the limitation that observations do not reveal where the gas lies along the line of sight, we compare the O vi measurements produced by gas within fixed distances around galaxies and by gas selected using line-of-sight velocity cuts commonly adopted by observers. We show that gas selected by a velocity cut of ±300 km s −1 or ±500 km s −1 produces a higher O vi column density, a flatter column density profile, and a higher covering fraction compared to gas within 1, 2, or 3 times the virial radius ( r vir ) of galaxies. The discrepancy increases with impact parameter and worsens for lower-mass galaxies. For example, compared to the gas within 2 r vir , identifying the gas using velocity cuts of 200–500 km s −1 increases the O vi column density by 0.2 dex (0.1 dex) at 1 r vir to over 0.75 dex (0.7 dex) at ≈ 2 r vir for galaxies with stellar masses of 10 9 –10 9.5 M ⊙ (10 10 –10 10.5 M ⊙ ). We furthermore estimate that excluding O vi outside r vir decreases the circumgalactic oxygen mass measured by Tumlinson et al. (2011) by over 50%. Our results demonstrate that gas at large line-of-sight separations but selected by conventional velocity windows has significant effects on the O vi measurements and may not be observationally distinguishable from gas near the galaxies.more » « less
-
Abstract Due to their inability to self-regulate, ultrafaint dwarfs are sensitive to prescriptions in subgrid physics models that converge and regulate at higher masses. We use high-resolution cosmological simulations to compare the effect of bursty star formation histories (SFHs) on dwarf galaxy structure for two different subgrid supernova (SN) feedback models, superbubble and blastwave, in dwarf galaxies with stellar masses from 5000 <M*/M⊙< 109. We find that in the “MARVEL-ous Dwarfs” suite both feedback models produce cored galaxies and reproduce observed scaling relations for luminosity, mass, and size. Our sample accurately predicts the average stellar metallicity at higher masses, however low-mass dwarfs are metal poor relative to observed galaxies in the Local Group. We show that continuous bursty star formation and the resulting stellar feedback are able to create dark matter (DM) cores in the higher dwarf galaxy mass regime, while the majority of ultrafaint and classical dwarfs retain cuspy central DM density profiles. We find that the effective core formation peaks atM*/Mhalo≃ 5 × 10−3for both feedback models. Both subgrid SN models yield bursty SFHs at higher masses; however, galaxies simulated with superbubble feedback reach maximum mean burstiness values at lower stellar mass fractions relative to blastwave feedback. As a result, core formation may be better predicted by stellar mass fraction than the burstiness of SFHs.more » « less
-
Abstract The Baldwin, Philips, & Terlevich diagram of [O iii ]/H β versus [N ii ]/H α (hereafter N2-BPT) has long been used as a tool for classifying galaxies based on the dominant source of ionizing radiation. Recent observations have demonstrated that galaxies at z ∼ 2 reside offset from local galaxies in the N2-BPT space. In this paper, we conduct a series of controlled numerical experiments to understand the potential physical processes driving this offset. We model nebular line emission in a large sample of galaxies, taken from the simba cosmological hydrodynamic galaxy formation simulation, using the cloudy photoionization code to compute the nebular line luminosities from H ii regions. We find that the observed shift toward higher [O iii ]/H β and [N ii ]/H α values at high redshift arises from sample selection: when we consider only the most massive galaxies M * ∼ 10 10–11 M ⊙ , the offset naturally appears, due to their high metallicities. We predict that deeper observations that probe lower-mass galaxies will reveal galaxies that lie on a locus comparable to z ∼ 0 observations. Even when accounting for samples-selection effects, we find that there is a subtle mismatch between simulations and observations. To resolve this discrepancy, we investigate the impact of varying ionization parameters, H ii region densities, gas-phase abundance patterns, and increasing radiation field hardness on N2-BPT diagrams. We find that either decreasing the ionization parameter or increasing the N/O ratio of galaxies at fixed O/H can move galaxies along a self-similar arc in N2-BPT space that is occupied by high-redshift galaxies.more » « less
-
Abstract We describe a public data release of the FIRE-2 cosmological zoom-in simulations of galaxy formation (available at http://flathub.flatironinstitute.org/fire ) from the Feedback In Realistic Environments (FIRE) project. FIRE-2 simulations achieve parsec-scale resolution to explicitly model the multiphase interstellar medium while implementing direct models for stellar evolution and feedback, including stellar winds, core-collapse and Type Ia supernovae, radiation pressure, photoionization, and photoelectric heating. We release complete snapshots from three suites of simulations. The first comprises 20 simulations that zoom in on 14 Milky Way (MW)–mass galaxies, five SMC/LMC-mass galaxies, and four lower-mass galaxies including one ultrafaint; we release 39 snapshots across z = 0–10. The second comprises four massive galaxies, with 19 snapshots across z = 1–10. Finally, a high-redshift suite comprises 22 simulations, with 11 snapshots across z = 5–10. Each simulation also includes dozens of resolved lower-mass (satellite) galaxies in its zoom-in region. Snapshots include all stored properties for all dark matter, gas, and star particles, including 11 elemental abundances for stars and gas, and formation times (ages) of star particles. We also release accompanying (sub)halo catalogs, which include galaxy properties and member star particles. For the simulations to z = 0, including all MW-mass galaxies, we release the formation coordinates and an “ex situ” flag for all star particles, pointers to track particles across snapshots, catalogs of stellar streams, and multipole basis expansions for the halo mass distributions. We describe publicly available python packages for reading and analyzing these simulations.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript
- Journal Article:
- https://doi.org/10.3847/2515-5172/ad40ad
