Search for: All records

Abstract As the next generation of large galaxy surveys come online, it is becoming increasingly important to develop and understand the machine-learning tools that analyze big astronomical data. Neural networks are powerful and capable of probing deep patterns in data, but they must be trained carefully on large and representative data sets. We present a new “hump” of the Cosmology and Astrophysics with MachinE Learning Simulations (CAMELS) project: CAMELS-SAM, encompassing one thousand dark-matter-only simulations of (100h⁻¹cMpc)³with different cosmological parameters (Ω_mandσ₈) and run through the Santa Cruz semi-analytic model for galaxy formation over a broad range of astrophysical parameters. As a proof of concept for the power of this vast suite of simulated galaxies in a large volume and broad parameter space, we probe the power of simple clustering summary statistics to marginalize over astrophysics and constrain cosmology using neural networks. We use the two-point correlation, count-in-cells, and void probability functions, and we probe nonlinear and linear scales across 0.68 <R<27h⁻¹cMpc. We find our neural networks can both marginalize over the uncertainties in astrophysics to constrain cosmology to 3%–8% error across various types of galaxy selections, while simultaneously learning about the SC-SAM astrophysical parameters. This work encompasses vital first steps toward creating algorithms able to marginalize over the uncertainties in our galaxy formation models and measure the underlying cosmology of our Universe. CAMELS-SAM has been publicly released alongside the rest of CAMELS, and it offers great potential to many applications of machine learning in astrophysics:https://camels-sam.readthedocs.io. 

Abstract We present a sample of 30 massive (log( M * / M ⊙ ) > 11) z = 3–5 quiescent galaxies selected from the Spitzer-HETDEX Exploratory Large Area (SHELA) Survey and observed at 1.1 mm with Atacama Large Millimeter/submillimeter Array (ALMA) Band 6 observations. These ALMA observations would detect even modest levels of dust-obscured star formation, on the order of ∼20 M ⊙ yr −1 at z ∼ 4 at the 1 σ level, allowing us to quantify the amount of contamination from dusty star-forming sources in our quiescent sample. Starting with a parent sample of candidate massive quiescent galaxies from the Stevans et al. v1 SHELA catalog, we use the Bayesian B agpipes spectral energy distribution fitting code to derive robust stellar masses ( M * ) and star formation rates (SFRs) for these sources, and select a conservative sample of 36 candidate massive ( M * > 10 11 M ⊙ ) quiescent galaxies, with specific SFRs >2 σ below the Salmon et al. star-forming main sequence at z ∼ 4. Based on the ALMA imaging, six of these candidate quiescent galaxies show the presence of significant dust-obscured star formation, and thus were removed from our final sample. This implies a ∼17% contamination rate from dusty star-forming galaxies with our selection criteria using the v1 SHELA catalog. This conservatively selected quiescent galaxy sample at z = 3–5 will provide excellent targets for future observations to constrain better how massive galaxies can both grow and shut down their star formation in a relatively short period. 

Abstract The UltraViolet imaging of the Cosmic Assembly Near-infrared Deep Extragalactic Legacy Survey Fields (UVCANDELS) program provides Hubble Space Telescope (HST)/UVIS F275W imaging for four CANDELS fields. We combine this UV imaging with existing HST/near-IR grism spectroscopy from 3D-HST+AGHAST to directly compare the resolved rest-frame UV and H α emission for a sample of 979 galaxies at 0.7 < z < 1.5, spanning a range in stellar mass of 10 8−11.5 M ⊙ . Using a stacking analysis, we perform a resolved comparison between homogenized maps of rest-UV and H α to compute the average UV-to-H α luminosity ratio (an indicator of burstiness in star formation) as a function of galactocentric radius. We find that galaxies below stellar mass of ∼10 9.5 M ⊙ , at all radii, have a UV-to-H α ratio higher than the equilibrium value expected from constant star formation, indicating a significant contribution from bursty star formation. Even for galaxies with stellar mass ≳10 9.5 M ⊙ , the UV-to-H α ratio is elevated toward their outskirts ( R / R eff > 1.5), suggesting that bursty star formation is likely prevalent in the outskirts of even the most massive galaxies, but is likely overshadowed by their brighter cores. Furthermore, we present the UV-to-H α ratio as a function of galaxy surface brightness, a proxy for stellar mass surface density, and find that regions below ∼10 7.5 M ⊙ kpc −2 are consistent with bursty star formation, regardless of their galaxy stellar mass, potentially suggesting that local star formation is independent of global galaxy properties at the smallest scales. Last, we find galaxies at z > 1.1 to have bursty star formation, regardless of radius or surface brightness. 

Abstract We present the empirical dust attenuation (EDA) framework—a flexible prescription for assigning realistic dust attenuation to simulated galaxies based on their physical properties. We use the EDA to forward model synthetic observations for three state-of-the-art large-scale cosmological hydrodynamical simulations: SIMBA, IllustrisTNG, and EAGLE. We then compare the optical and UV color–magnitude relations, ( g − r ) − M r and (far-UV −near-UV) − M r , of the simulations to a M r < − 20 and UV complete Sloan Digital Sky Survey galaxy sample using likelihood-free inference. Without dust, none of the simulations match observations, as expected. With the EDA, however, we can reproduce the observed color–magnitude with all three simulations. Furthermore, the attenuation curves predicted by our dust prescription are in good agreement with the observed attenuation–slope relations and attenuation curves of star-forming galaxies. However, the EDA does not predict star-forming galaxies with low A V since simulated star-forming galaxies are intrinsically much brighter than observations. Additionally, the EDA provides, for the first time, predictions on the attenuation curves of quiescent galaxies, which are challenging to measure observationally. Simulated quiescent galaxies require shallower attenuation curves with lower amplitude than star-forming galaxies. The EDA, combined with forward modeling, provides an effective approach for shedding light on dust in galaxies and probing hydrodynamical simulations. This work also illustrates a major limitation in comparing galaxy formation models: by adjusting dust attenuation, simulations that predict significantly different galaxy populations can reproduce the same UV and optical observations. 

ABSTRACT The James Webb Space Telescope (JWST) is expected to observe galaxies at z > 10 that are presently inaccessible. Here, we use a self-consistent empirical model, the universemachine, to generate mock galaxy catalogues and light-cones over the redshift range z = 0−15. These data include realistic galaxy properties (stellar masses, star formation rates, and UV luminosities), galaxy–halo relationships, and galaxy–galaxy clustering. Mock observables are also provided for different model parameters spanning observational uncertainties at z < 10. We predict that Cycle 1 JWST surveys will very likely detect galaxies with M* > 107 M⊙ and/or M1500 < −17 out to at least z ∼ 13.5. Number density uncertainties at z > 12 expand dramatically, so efforts to detect z > 12 galaxies will provide the most valuable constraints on galaxy formation models. The faint-end slopes of the stellar mass/luminosity functions at a given mass/luminosity threshold steepen as redshift increases. This is because observable galaxies are hosted by haloes in the exponentially falling regime of the halo mass function at high redshifts. Hence, these faint-end slopes are robustly predicted to become shallower below current observable limits (M* < 107 M⊙ or M1500 > −17). For reionization models, extrapolating luminosity functions with a constant faint-end slope from M1500 = −17 down to M1500 = −12 gives the most reasonable upper limit for the total UV luminosity and cosmic star formation rate up to z ∼ 12. We compare to three other empirical models and one semi-analytic model, showing that the range of predicted observables from our approach encompasses predictions from other techniques. Public catalogues and light-cones for common fields are available online. 

Abstract We report the discovery of an accreting supermassive black hole atz= 8.679. This galaxy, denoted here as CEERS_1019, was previously discovered as a Lyα-break galaxy by Hubble with a Lyαredshift from Keck. As part of the Cosmic Evolution Early Release Science (CEERS) survey, we have observed this source with JWST/NIRSpec, MIRI, NIRCam, and NIRCam/WFSS and uncovered a plethora of emission lines. The Hβline is best fit by a narrow plus a broad component, where the latter is measured at 2.5σwith an FWHM ∼1200 km s⁻¹. We conclude this originates in the broadline region of an active galactic nucleus (AGN). This is supported by the presence of weak high-ionization lines (N V, N IV], and C III]), as well as a spatial point-source component. The implied mass of the black hole (BH) is log (M_BH/M_⊙) = 6.95 ± 0.37, and we estimate that it is accreting at 1.2 ± 0.5 times the Eddington limit. The 1–8μm photometric spectral energy distribution shows a continuum dominated by starlight and constrains the host galaxy to be massive (log M/M_⊙∼9.5) and highly star-forming (star formation rate, or SFR ∼ 30 M_⊙yr⁻¹; log sSFR ∼ − 7.9 yr⁻¹). The line ratios show that the gas is metal-poor (Z/Z_⊙∼ 0.1), dense (n_e∼ 10³cm⁻³), and highly ionized (logU∼ − 2.1). We use this present highest-redshift AGN discovery to place constraints on BH seeding models and find that a combination of either super-Eddington accretion from stellar seeds or Eddington accretion from very massive BH seeds is required to form this object.