Search

We use the magnetic-hydrodynamical simulation TNG50 to study the evolution of barred massive disc galaxies. Massive spiral galaxies are already present as early as z = 4, and bar formation takes place already at those early times. The bars grow longer and stronger as the host galaxies evolve, with the bar sizes increasing at a pace similar to that of the disc scalelengths. The bar fraction mildly evolves with redshift for galaxies with M_∗ ≥ 1010 M⊙, being greater than ∼ 40 per cent at 0.5 < z < 3 and ∼ 30 per cent at z = 0. When bars larger than a given physical size (≥ 2 kpc) or the angular resolution limit of twice the I-band angular PSF FWHM of the HST are considered, the bar fraction dramatically decreases with increasing redshift, reconciling the theoretical predictions with observational data. We find that barred galaxies have an older stellar population, lower gas fractions, and star formation rates than unbarred galaxies. In most cases, the discs of barred galaxies assembled earlier and faster than the discs of unbarred galaxies. We also find that barred galaxies are typical in haloes with larger concentrations and smaller spin parameters than unbarred galaxies. Furthermore, the inner regions of barred galaxies are more baryon-dominated than those of unbarred galaxies but have comparable global stellar mass fractions. Our findings suggest that the bar population could be used as a potential tracer of the buildup of disc galaxies and their host haloes. With this paper, we release a catalogue of barred galaxies in TNG50 at six redshifts between z = 4 and 0.

The hot, X-ray-emitting phase of the circumgalactic medium in galaxies is believed to be the reservoir of baryons from which gas flows onto the central galaxy and into which feedback from AGN and stars inject mass, momentum, energy, and metals. These effects shape the velocity fields of the hot gas, which can be observed by X-ray IFUs via the Doppler shifting and broadening of emission lines. In this work, we analyze the gas kinematics of the hot circumgalactic medium of Milky Way-mass disk galaxies from the TNG50 simulation with synthetic observations to determine how future instruments can probe this velocity structure. We find that the hot phase is often characterized by outflows outward from the disk driven by feedback processes, radial inflows near the galactic plane, and rotation, though in other cases the velocity field is more disorganized and turbulent. With a spectral resolution of $\sim$1 eV, fast and hot outflows ($\sim$200-500 km s$^{-1}$) can be measured, depending on the orientation of the galaxy on the sky. The rotation velocity of the hot phase ($\sim$100-200 km s$^{-1}$) can be measured using line shifts in edge-on galaxies, and is slower than that of colder gas phases but similar to stellar rotation velocities. By contrast, the slow inflows ($\sim$50-100 km s$^{-1}$) are difficult to measure in projection with these other components. We find that the velocity measured is sensitive to which emission lines are used. Measuring these flows will help constrain theories of how the gas in these galaxies forms and evolves., Comment: 41 pages, 29 figures, submitted to ApJ

ABSTRACT We introduce the thesan project, a suite of large volume ($L_\mathrm{box} = 95.5 \, \mathrm{cMpc}$) radiation-magnetohydrodynamic simulations that simultaneously model the large-scale statistical properties of the intergalactic medium during reionization and the resolved characteristics of the galaxies responsible for it. The flagship simulation has dark matter and baryonic mass resolutions of $3.1 \times 10^6\, {\rm M_\odot }$ and $5.8 \times 10^5\, {\rm M_\odot }$, respectively. The gravitational forces are softened on scales of 2.2 ckpc with the smallest cell sizes reaching 10 pc at z = 5.5, enabling predictions down to the atomic cooling limit. The simulations use an efficient radiation hydrodynamics solver (arepo-rt) that precisely captures the interaction between ionizing photons and gas, coupled to well-tested galaxy formation (IllustrisTNG) and dust models to accurately predict the properties of galaxies. Through a complementary set of medium resolution simulations we investigate the changes to reionization introduced by different assumptions for ionizing escape fractions, varying dark matter models, and numerical convergence. The fiducial simulation and model variations are calibrated to produce realistic reionization histories that match the observed evolution of the global neutral hydrogen fraction and electron scattering optical depth to reionization. They also match a wealth of high-redshift observationally inferred data, including the stellar-to-halo-mass relation, galaxy stellar mass function, star formation rate density, and the mass–metallicity relation, despite the galaxy formation model being mainly calibrated at z = 0. We demonstrate that different reionization models give rise to varied bubble size distributions that imprint unique signatures on the 21 cm emission, especially on the slope of the power spectrum at large spatial scales, enabling current and upcoming 21 cm experiments to accurately characte

© 2019. The American Astronomical Society. All rights reserved. We present a machine-learning (ML) approach for estimating galaxy cluster masses from Chandra mock images. We utilize a Convolutional Neural Network (CNN), a deep ML tool commonly used in image recognition tasks. The CNN is trained and tested on our sample of 7896 Chandra X-ray mock observations, which are based on 329 massive clusters from the simulation. Our CNN learns from a low resolution spatial distribution of photon counts and does not use spectral information. Despite our simplifying assumption to neglect spectral information, the resulting mass values estimated by the CNN exhibit small bias in comparison to the true masses of the simulated clusters (-0.02 dex) and reproduce the cluster masses with low intrinsic scatter, 8% in our best fold and 12% averaging over all. In contrast, a more standard core-excised luminosity method achieves 15%-18% scatter. We interpret the results with an approach inspired by Google DeepDream and find that the CNN ignores the central regions of clusters, which are known to have high scatter with mass.

We present a comparison of the physical properties of the ionized gas in the circumgalactic medium and intergalactic medium (IGM) at z ∼ 0 between observations and four cosmological hydrodynamical simulations: Illustris, TNG300 of the IllustrisTNG project, EAGLE, and one of the Magneticum simulations. For the observational data, we use the gas properties that are inferred from cross-correlating the Sunyaev–Zel’dovich effect (SZE) from the Planck CMB maps with haloes and large-scale structure. Both the observational and simulation results indicate that the integrated gas pressure in haloes deviates from the self-similar case, showing that feedback impacts haloes with $M_{500}\sim 10^{12\!-\!13}\, {\rm M_\odot }$. The simulations predict that more than half the baryons are displaced from haloes, while the gas fraction inferred from our observational data roughly equals the cosmic baryon fraction throughout the $M_{500}\sim 10^{12\!-\!14.5}\, {\rm M_\odot }$ halo mass range. All simulations tested here predict that the mean gas temperature in haloes is about the virial temperature, while that inferred from the SZE is up to one order of magnitude lower than that from the simulations (and also from X-ray observations). While a remarkable agreement is found for the average properties of the IGM between the observation and some simulations, we show that their dependence on the large-scale tidal field can break the degeneracy between models that show similar predictions otherwise. Finally, we show that the gas pressure and the electron density profiles from simulations are not well described by a generalized NFW profile. Instead, we present a new model with a mass-dependent shape that fits the profiles accurately.

ABSTRACT We introduce the thesan project, a suite of large volume ($L_\mathrm{box} = 95.5 \, \mathrm{cMpc}$) radiation-magnetohydrodynamic simulations that simultaneously model the large-scale statistical properties of the intergalactic medium during reionization and the resolved characteristics of the galaxies responsible for it. The flagship simulation has dark matter and baryonic mass resolutions of $3.1 \times 10^6\, {\rm M_\odot }$ and $5.8 \times 10^5\, {\rm M_\odot }$, respectively. The gravitational forces are softened on scales of 2.2 ckpc with the smallest cell sizes reaching 10 pc at z = 5.5, enabling predictions down to the atomic cooling limit. The simulations use an efficient radiation hydrodynamics solver (arepo-rt) that precisely captures the interaction between ionizing photons and gas, coupled to well-tested galaxy formation (IllustrisTNG) and dust models to accurately predict the properties of galaxies. Through a complementary set of medium resolution simulations we investigate the changes to reionization introduced by different assumptions for ionizing escape fractions, varying dark matter models, and numerical convergence. The fiducial simulation and model variations are calibrated to produce realistic reionization histories that match the observed evolution of the global neutral hydrogen fraction and electron scattering optical depth to reionization. They also match a wealth of high-redshift observationally inferred data, including the stellar-to-halo-mass relation, galaxy stellar mass function, star formation rate density, and the mass–metallicity relation, despite the galaxy formation model being mainly calibrated at z = 0. We demonstrate that different reionization models give rise to varied bubble size distributions that imprint unique signatures on the 21 cm emission, especially on the slope of the power spectrum at large spatial scales, enabling current and upcoming 21 cm experiments to accurately characte

We introduce the THESAN project, a suite of large volume (L = 95.5 cMpc) radiation-magnetohydrodynamic simulations that simultaneously model the large-scale statistical properties of the IGM during reionization and the resolved characteristics of the galaxies responsible for it. The flagship simulation has dark matter (DM) and baryonic mass resolutions of $3.1 \times 10^6$ $M_\odot$ and $5.8 \times 10^5 $ $M_\odot$, respectively. The gravitational forces are softened on scales of 2.2 ckpc with the smallest cell sizes reaching 10 pc at z=5.5, enabling predictions down to the atomic cooling limit. The simulations use an efficient radiation hydrodynamics solver (AREPO-RT) that precisely captures the interaction between ionizing photons and gas, coupled to well-tested galaxy formation (IllustrisTNG) and dust models to accurately predict the properties of galaxies. Through a complementary set of medium resolution simulations we investigate the changes to reionization introduced by different assumptions for ionizing escape fractions, varying DM models, and numerical convergence. The simulations produce realistic reionization histories that match the observed evolution of the global neutral hydrogen fraction and electron scattering optical depth to reionization. They also match a wealth of high-redshift observationally inferred data, including the stellar-to-halo-mass relation, stellar mass function, star formation rate density, and the mass-metallicity relation, despite the galaxy formation model being mainly calibrated at z=0. We demonstrate that different reionization models give rise to varied bubble size distributions that imprint unique signatures on the 21cm emission, especially on the slope of the power spectrum at large spatial scales, enabling current and upcoming 21cm experiments to accurately characterise the sources that dominate the ionizing photon budget. [abridged], Comment: MNRAS, in press. Please visit www.thesan-project.com for more details

The visibility of high-redshift Lyman-alpha emitting galaxies (LAEs) provides important constraints on galaxy formation processes and the Epoch of Reionization (EoR). However, predicting realistic and representative statistics for comparison with observations represents a significant challenge in the context of large-volume cosmological simulations. The THESAN project offers a unique framework for addressing such limitations by combining state-of-the-art galaxy formation (IllustrisTNG) and dust models with the Arepo-RT radiation-magneto-hydrodynamics solver. In this initial study we present Lyman-alpha centric analysis for the flagship simulation that resolves atomic cooling haloes throughout a (95.5 cMpc)^3 region of the Universe. To avoid numerical artifacts we devise a novel method for accurate frequency-dependent line radiative transfer in the presence of continuous Hubble flow, transferable to broader astrophysical applications as well. Our scalable approach highlights the utility of LAEs and red damping-wing transmission as probes of reionization, which reveal nontrivial trends across different galaxies, sightlines, and frequency bands that can be modelled in the framework of covering fractions. In fact, after accounting for environmental factors influencing large-scale ionized bubble formation such as redshift and UV magnitude, the variation across galaxies and sightlines mainly depends on random processes including peculiar velocities and self-shielded systems that strongly impact unfortunate rays more than others. Throughout the EoR local and cosmological optical depths are often greater than or less than unity such that the exp(-tau) behavior leads to anisotropic and bimodal transmissivity. Future surveys will benefit by targeting both rare bright objects and Goldilocks zone LAEs to infer the presence of these (un)predictable (dis)advantages., Comment: 23 pages, 23 figures, MNRAS, in press. Please visit www.thesan-project.com for more details

The visibility of high-redshift Lyman-alpha emitting galaxies (LAEs) provides important constraints on galaxy formation processes and the Epoch of Reionization (EoR). However, predicting realistic and representative statistics for comparison with observations represents a significant challenge in the context of large-volume cosmological simulations. The THESAN project offers a unique framework for addressing such limitations by combining state-of-the-art galaxy formation (IllustrisTNG) and dust models with the Arepo-RT radiation-magneto-hydrodynamics solver. In this initial study we present Lyman-alpha centric analysis for the flagship simulation that resolves atomic cooling haloes throughout a (95.5 cMpc)^3 region of the Universe. To avoid numerical artifacts we devise a novel method for accurate frequency-dependent line radiative transfer in the presence of continuous Hubble flow, transferable to broader astrophysical applications as well. Our scalable approach highlights the utility of LAEs and red damping-wing transmission as probes of reionization, which reveal nontrivial trends across different galaxies, sightlines, and frequency bands that can be modelled in the framework of covering fractions. In fact, after accounting for environmental factors influencing large-scale ionized bubble formation such as redshift and UV magnitude, the variation across galaxies and sightlines mainly depends on random processes including peculiar velocities and self-shielded systems that strongly impact unfortunate rays more than others. Throughout the EoR local and cosmological optical depths are often greater than or less than unity such that the exp(-tau) behavior leads to anisotropic and bimodal transmissivity. Future surveys will benefit by targeting both rare bright objects and Goldilocks zone LAEs to infer the presence of these (un)predictable (dis)advantages., Comment: 23 pages, 23 figures, MNRAS, in press. Please visit www.thesan-project.com for more details

We introduce the THESAN project, a suite of large volume (L = 95.5 cMpc) radiation-magnetohydrodynamic simulations that simultaneously model the large-scale statistical properties of the IGM during reionization and the resolved characteristics of the galaxies responsible for it. The flagship simulation has dark matter (DM) and baryonic mass resolutions of $3.1 \times 10^6$ $M_\odot$ and $5.8 \times 10^5 $ $M_\odot$, respectively. The gravitational forces are softened on scales of 2.2 ckpc with the smallest cell sizes reaching 10 pc at z=5.5, enabling predictions down to the atomic cooling limit. The simulations use an efficient radiation hydrodynamics solver (AREPO-RT) that precisely captures the interaction between ionizing photons and gas, coupled to well-tested galaxy formation (IllustrisTNG) and dust models to accurately predict the properties of galaxies. Through a complementary set of medium resolution simulations we investigate the changes to reionization introduced by different assumptions for ionizing escape fractions, varying DM models, and numerical convergence. The simulations produce realistic reionization histories that match the observed evolution of the global neutral hydrogen fraction and electron scattering optical depth to reionization. They also match a wealth of high-redshift observationally inferred data, including the stellar-to-halo-mass relation, stellar mass function, star formation rate density, and the mass-metallicity relation, despite the galaxy formation model being mainly calibrated at z=0. We demonstrate that different reionization models give rise to varied bubble size distributions that imprint unique signatures on the 21cm emission, especially on the slope of the power spectrum at large spatial scales, enabling current and upcoming 21cm experiments to accurately characterise the sources that dominate the ionizing photon budget. [abridged], Comment: MNRAS, in press. Please visit www.thesan-project.com for more details

© 2019. The American Astronomical Society. All rights reserved. We present a machine-learning (ML) approach for estimating galaxy cluster masses from Chandra mock images. We utilize a Convolutional Neural Network (CNN), a deep ML tool commonly used in image recognition tasks. The CNN is trained and tested on our sample of 7896 Chandra X-ray mock observations, which are based on 329 massive clusters from the simulation. Our CNN learns from a low resolution spatial distribution of photon counts and does not use spectral information. Despite our simplifying assumption to neglect spectral information, the resulting mass values estimated by the CNN exhibit small bias in comparison to the true masses of the simulated clusters (-0.02 dex) and reproduce the cluster masses with low intrinsic scatter, 8% in our best fold and 12% averaging over all. In contrast, a more standard core-excised luminosity method achieves 15%-18% scatter. We interpret the results with an approach inspired by Google DeepDream and find that the CNN ignores the central regions of clusters, which are known to have high scatter with mass.

© 2020 The Author(s) Published by Oxford University Press on behalf of the Royal Astronomical Society We propose two effective parameters that fully characterize galactic-scale structure formation at high redshifts (z ≿ 5) for a variety of dark matter (DM) models that have a primordial cutoff in the matter power spectrum. Our description is within the recently proposed ETHOS framework and includes standard thermal warm DM (WDM) and models with dark acoustic oscillations (DAOs). To define and explore this parameter space, we use high-redshift zoom-in simulations that cover a wide range of non-linear scales from those where DM should behave as CDM (k ∼ 10 hMpc−1), down to those characterized by the onset of galaxy formation (k ∼ 500 hMpc−1). We show that the two physically motivated parameters hpeak and kpeak, the amplitude and scale of the first DAO peak, respectively, are sufficient to parametrize the linear matter power spectrum and classify the DM models as belonging to effective non-linear structure formation regions. These are defined by their relative departure from cold DM (kpeak → ∞) and WDM (hpeak = 0) according to the non-linear matter power spectrum and halo mass function. We identify a region where the DAOs still leave a distinct signature from WDM down to z = 5, while a large part of the DAO parameter space is shown to be degenerate with WDM. Our framework can then be used to seamlessly connect a broad class of particle DM models to their structure formation properties at high redshift without the need of additional N-body simulations.

We confirm that the object DDO 216-A1 is a substantial globular cluster at the center of Local Group galaxy DDO 216 (the Pegasus dwarf irregular), using Hubble Space Telescope ACS imaging. By fitting isochrones, we find the cluster metallicity [M/H] = -1.6 ±0.2, for reddening E(B-V) = 0.16 ±0.02; the best-fit age is 12.3 ±0.8 Gyr. There are RR Lyrae variables in the cluster; the magnitude of the fundamental mode pulsators gives a distance modulus of 24.77 ±0.08 - identical to the host galaxy. The ratio of overtone to fundamental mode variables and their mean periods make DDO 216-A1 an Oosterhoff Type I cluster. We find a central surface brightness of 20.85 ±0.17 F814W mag arcsec-2, a half-light radius of (13.4 pc), and an absolute magnitude M814 = -7.90 ±0.16 (M/M⊙ ≈ 105). King models fit to the cluster give the core radius and concentration index, r c = ± and c = 1.24 ±0.39. The cluster is an "extended" cluster somewhat typical of some dwarf galaxies and the outer halo of the Milky Way. The cluster is projected ≲30 pc south of the center of DDO 216, unusually central compared to most dwarf galaxy globular clusters. Analytical models of dynamical friction and tidal destruction suggest that it probably formed at a larger distance, up to ∼1 kpc, and migrated inward. DDO 216 has an unexceptional specific cluster frequency, S N = 10. DDO 216 is the lowest-luminosity Local Group galaxy to host a 105 M⊙ globular cluster and the only transition-type (dSph/dIrr) galaxy in the Local Group with a globular cluster.

We present the star formation history (SFH) of the isolated (D ∼ 970 kpc) Local Group dwarf galaxy Wolf-Lundmark-Melotte (WLM) measured from colour-magnitude diagrams (CMDs) constructed from deep Hubble Space Telescope imaging. Our observations include a central (0.5 rh) and outer field (0.7 rh) that reach below the oldest main-sequence turn-off. WLM has no early dominant episode of star formation: 20 per cent of its stellar mass formed by ∼12.5 Gyr ago (z ∼ 5). It also has an SFR that rises to the present with 50 per cent of the stellar mass within the most recent 5 Gyr (z < 0.7). There is evidence of a strong age gradient: the mean age of the outer field is 5 Gyr older than the inner field despite being only 0.4 kpc apart. Some models suggest such steep gradients are associated with strong stellar feedback and dark-matter core creation. The SFHs of real isolated dwarf galaxies and those from the Feedback in Realistic Environment suite are in good agreement for M*(z = 0) ∼ 107-109M☉, but in worse agreement at lower masses (M*(z = 0) ∼ 105-107 M☉). These differences may be explainable by systematics in the models (e.g. reionization model) and/or observations (HST field placement). We suggest that a coordinated effort to get deep CMDs between HST/JWST (crowded central fields) and WFIRST (wide-area halo coverage) is the optimal path for measuring global SFHs of isolated dwarf galaxies.

We analyze the optical morphologies of galaxies in the IllustrisTNG simulation at $z\sim0$ with a Convolutional Neural Network trained on visual morphologies in the Sloan Digital Sky Survey. We generate mock SDSS images of a mass complete sample of $\sim12,000$ galaxies in the simulation using the radiative transfer code SKIRT and include PSF and noise to match the SDSS r-band properties. The images are then processed through the exact same neural network used to estimate SDSS morphologies to classify simulated galaxies in four morphological classes (E, S0/a, Sab, Scd). The CNN model finds that $\sim95\%$ of the simulated galaxies fall in one the four main classes with high confidence. The mass-size relations of the simulated galaxies divided by morphological type also reproduce well the slope and the normalization of observed relations which confirms the realism of optical morphologies in the TNG suite. However, the Stellar Mass Functions decomposed into different morphologies still show significant discrepancies with observations both at the low and high mass end. We find that the high mass end of the SMF is dominated in TNG by massive disk galaxies while early-type galaxies dominate in the observations according to the CNN classifications. The present work highlights the importance of detailed comparisons between observations and simulations in comparable conditions., Comment: submitted to MNRAS, comments welcome

We present the star formation history (SFH) of the isolated (D ∼ 970 kpc) Local Group dwarf galaxy Wolf-Lundmark-Melotte (WLM) measured from colour-magnitude diagrams (CMDs) constructed from deep Hubble Space Telescope imaging. Our observations include a central (0.5 rh) and outer field (0.7 rh) that reach below the oldest main-sequence turn-off. WLM has no early dominant episode of star formation: 20 per cent of its stellar mass formed by ∼12.5 Gyr ago (z ∼ 5). It also has an SFR that rises to the present with 50 per cent of the stellar mass within the most recent 5 Gyr (z < 0.7). There is evidence of a strong age gradient: the mean age of the outer field is 5 Gyr older than the inner field despite being only 0.4 kpc apart. Some models suggest such steep gradients are associated with strong stellar feedback and dark-matter core creation. The SFHs of real isolated dwarf galaxies and those from the Feedback in Realistic Environment suite are in good agreement for M*(z = 0) ∼ 107-109M☉, but in worse agreement at lower masses (M*(z = 0) ∼ 105-107 M☉). These differences may be explainable by systematics in the models (e.g. reionization model) and/or observations (HST field placement). We suggest that a coordinated effort to get deep CMDs between HST/JWST (crowded central fields) and WFIRST (wide-area halo coverage) is the optimal path for measuring global SFHs of isolated dwarf galaxies.

We analyze the optical morphologies of galaxies in the IllustrisTNG simulation at $z\sim0$ with a Convolutional Neural Network trained on visual morphologies in the Sloan Digital Sky Survey. We generate mock SDSS images of a mass complete sample of $\sim12,000$ galaxies in the simulation using the radiative transfer code SKIRT and include PSF and noise to match the SDSS r-band properties. The images are then processed through the exact same neural network used to estimate SDSS morphologies to classify simulated galaxies in four morphological classes (E, S0/a, Sab, Scd). The CNN model finds that $\sim95\%$ of the simulated galaxies fall in one the four main classes with high confidence. The mass-size relations of the simulated galaxies divided by morphological type also reproduce well the slope and the normalization of observed relations which confirms the realism of optical morphologies in the TNG suite. However, the Stellar Mass Functions decomposed into different morphologies still show significant discrepancies with observations both at the low and high mass end. We find that the high mass end of the SMF is dominated in TNG by massive disk galaxies while early-type galaxies dominate in the observations according to the CNN classifications. The present work highlights the importance of detailed comparisons between observations and simulations in comparable conditions., Comment: submitted to MNRAS, comments welcome

We present a machine-learning approach for estimating galaxy cluster masses from Chandra mock images. We utilize a Convolutional Neural Network (CNN), a deep machine learning tool commonly used in image recognition tasks. The CNN is trained and tested on our sample of 7,896 Chandra X-ray mock observations, which are based on 329 massive clusters from the IllustrisTNG simulation. Our CNN learns from a low resolution spatial distribution of photon counts and does not use spectral information. Despite our simplifying assumption to neglect spectral information, the resulting mass values estimated by the CNN exhibit small bias in comparison to the true masses of the simulated clusters (-0.02 dex) and reproduce the cluster masses with low intrinsic scatter, 8% in our best fold and 12% averaging over all. In contrast, a more standard core-excised luminosity method achieves 15-18% scatter. We interpret the results with an approach inspired by Google DeepDream and find that the CNN ignores the central regions of clusters, which are known to have high scatter with mass., Comment: 10 pages, 6 figures, accepted for publication in The Astrophysical Journal

We present a machine-learning approach for estimating galaxy cluster masses from Chandra mock images. We utilize a Convolutional Neural Network (CNN), a deep machine learning tool commonly used in image recognition tasks. The CNN is trained and tested on our sample of 7,896 Chandra X-ray mock observations, which are based on 329 massive clusters from the IllustrisTNG simulation. Our CNN learns from a low resolution spatial distribution of photon counts and does not use spectral information. Despite our simplifying assumption to neglect spectral information, the resulting mass values estimated by the CNN exhibit small bias in comparison to the true masses of the simulated clusters (-0.02 dex) and reproduce the cluster masses with low intrinsic scatter, 8% in our best fold and 12% averaging over all. In contrast, a more standard core-excised luminosity method achieves 15-18% scatter. We interpret the results with an approach inspired by Google DeepDream and find that the CNN ignores the central regions of clusters, which are known to have high scatter with mass., Comment: 10 pages, 6 figures, accepted for publication in The Astrophysical Journal

The velocity function is a fundamental observable statistic of the galaxy population, similarly impor- tant as the luminosity function, but much more difficult to measure. In this work we present the first directly measured circular velocity function that is representative between 60 < v_circ < 320 km/s for galaxies of all morphological types at a given rotation velocity. For the low mass galaxy population (60 < v_circ < 170 km/s), we use the HIPASS velocity function. For the massive galaxy population (170 < v_circ < 320 km/s), we use stellar circular velocities from the Calar Alto Legacy Integral Field Area Survey (CALIFA). In earlier work we obtained the measurements of circular velocity at the 80% light radius for 226 galaxies and demonstrated that the CALIFA sample can produce volume- corrected galaxy distribution functions. The CALIFA velocity function includes homogeneous velocity measurements of both late and early-type rotation-supported galaxies and has the crucial advantage of not missing gas-poor massive ellipticals that HI surveys are blind to. We show that both velocity functions can be combined in a seamless manner, as their ranges of validity overlap. The resulting observed velocity function is compared to velocity functions derived from cosmological simulations of the z = 0 galaxy population. We find that dark matter-only simulations show a strong mismatch with the observed VF. Hydrodynamic simulations fare better, but still do not fully reproduce observations., Comment: Accepted to ApJL. Keywords: Galaxies: kinematics and dynamics - galaxies: statistics - galaxies: evolution 7 pages, 3 figures

It is widely believed that the large discrepancy between the observed number of satellite galaxies and the predicted number of dark subhalos can be resolved via a variety of baryonic effects which suppress star formation in low mass halos.Supporting this hypothesis, numerous high resolution simulations with star formation, and associated feedback have been shown to reproduce the satellite luminosity function around Milky Way-mass simulated galaxies at redshift zero. However, a more stringent test of these models is their ability to simultaneously match the satellite luminosity functions of a range of host halo masses and redshifts. In this work we measure the luminosity function of faint (sub-Small Magellanic Cloud luminosity) satellites around hosts with stellar masses 10.5$<\log_{10}$M$_*$/M$_\odot<11.5$ to an unprecedented redshift of 1.5. This new measurement of the satellite luminosity function provides powerful new constraining power; we compare these results with predictions from four different simulations and show that although the models perform similarly over-all, no one model reproduces the satellite luminosity function reliably at all redshifts and host stellar masses. This result highlights the continued need for improvement in understanding the fundamental physics that governs satellite galaxy evolution., Comment: 9 pages, 2 figures, submitted to MNRAS

The velocity function is a fundamental observable statistic of the galaxy population, similarly impor- tant as the luminosity function, but much more difficult to measure. In this work we present the first directly measured circular velocity function that is representative between 60 < v_circ < 320 km/s for galaxies of all morphological types at a given rotation velocity. For the low mass galaxy population (60 < v_circ < 170 km/s), we use the HIPASS velocity function. For the massive galaxy population (170 < v_circ < 320 km/s), we use stellar circular velocities from the Calar Alto Legacy Integral Field Area Survey (CALIFA). In earlier work we obtained the measurements of circular velocity at the 80% light radius for 226 galaxies and demonstrated that the CALIFA sample can produce volume- corrected galaxy distribution functions. The CALIFA velocity function includes homogeneous velocity measurements of both late and early-type rotation-supported galaxies and has the crucial advantage of not missing gas-poor massive ellipticals that HI surveys are blind to. We show that both velocity functions can be combined in a seamless manner, as their ranges of validity overlap. The resulting observed velocity function is compared to velocity functions derived from cosmological simulations of the z = 0 galaxy population. We find that dark matter-only simulations show a strong mismatch with the observed VF. Hydrodynamic simulations fare better, but still do not fully reproduce observations., Comment: Accepted to ApJL. Keywords: Galaxies: kinematics and dynamics - galaxies: statistics - galaxies: evolution 7 pages, 3 figures

It is widely believed that the large discrepancy between the observed number of satellite galaxies and the predicted number of dark subhalos can be resolved via a variety of baryonic effects which suppress star formation in low mass halos.Supporting this hypothesis, numerous high resolution simulations with star formation, and associated feedback have been shown to reproduce the satellite luminosity function around Milky Way-mass simulated galaxies at redshift zero. However, a more stringent test of these models is their ability to simultaneously match the satellite luminosity functions of a range of host halo masses and redshifts. In this work we measure the luminosity function of faint (sub-Small Magellanic Cloud luminosity) satellites around hosts with stellar masses 10.5$<\log_{10}$M$_*$/M$_\odot<11.5$ to an unprecedented redshift of 1.5. This new measurement of the satellite luminosity function provides powerful new constraining power; we compare these results with predictions from four different simulations and show that although the models perform similarly over-all, no one model reproduces the satellite luminosity function reliably at all redshifts and host stellar masses. This result highlights the continued need for improvement in understanding the fundamental physics that governs satellite galaxy evolution., Comment: 9 pages, 2 figures, submitted to MNRAS

We use the Aquarius simulation series to study the imprint of assembly history on the structure of Galaxy-mass cold dark matter halos. Our results confirm earlier work regarding the influence of mergers on the mass density profile and the inside-out growth of halos. The inner regions that contain the visible galaxies are stable since early times and are significantly affected only by major mergers. Particles accreted diffusely or in minor mergers are found predominantly in the outskirts of halos. Our analysis reveals trends that run counter to current perceptions of hierarchical halo assembly. For example, major mergers (i.e. those with progenitor mass ratios greater than 1:10) contribute little to the total mass growth of a halo, on average less than 20 per cent for our six Aquarius halos. The bulk is contributed roughly equally by minor mergers and by "diffuse" material which is not resolved into individual objects. This is consistent with modeling based on excursion-set theory which suggests that about half of this diffuse material should not be part of a halo of any scale. Interestingly, the simulations themselves suggest that a significantly fraction is not truly diffuse, since it was ejected from earlier halos by mergers prior to their joining the main system. The Aquarius simulations resolve halos to much lower mass scales than are expected to retain gas or form stars. These results thus confirm that most of the baryons from which visible galaxies form are accreted diffusely, rather than through mergers, and they suggest that only relatively rare major mergers will affect galaxy structure at later times., Comment: 11 pages, 10 figs, submitted to MNRAS, a superfluous paragraph at the end of sec.2 taken out, the position of plots rearranged

We use the Aquarius simulation series to study the imprint of assembly history on the structure of Galaxy-mass cold dark matter halos. Our results confirm earlier work regarding the influence of mergers on the mass density profile and the inside-out growth of halos. The inner regions that contain the visible galaxies are stable since early times and are significantly affected only by major mergers. Particles accreted diffusely or in minor mergers are found predominantly in the outskirts of halos. Our analysis reveals trends that run counter to current perceptions of hierarchical halo assembly. For example, major mergers (i.e. those with progenitor mass ratios greater than 1:10) contribute little to the total mass growth of a halo, on average less than 20 per cent for our six Aquarius halos. The bulk is contributed roughly equally by minor mergers and by "diffuse" material which is not resolved into individual objects. This is consistent with modeling based on excursion-set theory which suggests that about half of this diffuse material should not be part of a halo of any scale. Interestingly, the simulations themselves suggest that a significantly fraction is not truly diffuse, since it was ejected from earlier halos by mergers prior to their joining the main system. The Aquarius simulations resolve halos to much lower mass scales than are expected to retain gas or form stars. These results thus confirm that most of the baryons from which visible galaxies form are accreted diffusely, rather than through mergers, and they suggest that only relatively rare major mergers will affect galaxy structure at later times., Comment: 11 pages, 10 figs, submitted to MNRAS, a superfluous paragraph at the end of sec.2 taken out, the position of plots rearranged