-
Classifying galaxies according to their HI content
Authors:
Sambatra Andrianomena,
Mika Rafieferantsoa,
Romeel Davé
Abstract:
We use machine learning to classify galaxies according to their HI content, based on both their optical photometry and environmental properties. The data used for our analyses are the outputs in the range $z = 0-1$ from MUFASA cosmological hydrodynamic simulation. In our previous paper, where we predicted the galaxy HI content using the same input features, HI rich galaxies were only selected for…
▽ More
We use machine learning to classify galaxies according to their HI content, based on both their optical photometry and environmental properties. The data used for our analyses are the outputs in the range $z = 0-1$ from MUFASA cosmological hydrodynamic simulation. In our previous paper, where we predicted the galaxy HI content using the same input features, HI rich galaxies were only selected for the training. In order for the predictions on real observation data to be more accurate, the classifiers built in this study will first establish if a galaxy is HI rich ($\rm{log(M_{HI}/M_{*})} > -2 $) before estimating its neutral hydrogen content using the regressors developed in the first paper. We resort to various machine learning algorithms and assess their performance with various metrics such as accuracy for instance. The performance of the classifiers gets better with increasing redshift and reaches their peak performance around $z = 1$. Random Forest method, the most robust among the classifiers when considering only the mock data for both training and test in this study, reaches an accuracy above $98.6 \%$ at $z = 0$ and above $99.0 \%$ at $z = 1$. We test our algorithms, trained with simulation data, on classification of the galaxies in RESOLVE, ALFALFA and GASS surveys. Interestingly, SVM algorithm, the best classifier for the tests, achieves a precision, the relevant metric for the tests, above $87.60\%$ and a specificity above $71.4\%$ with all the tests, indicating that the classifier is capable of learning from the simulated data to classify HI rich/HI poor galaxies from the real observation data. With the advent of large HI 21 cm surveys such as the SKA, this set of classifiers, together with the regressors developed in the first paper, will be part of a pipeline, a very useful tool, which is aimed at predicting HI content of galaxies.
△ Less
Submitted 10 June, 2019;
originally announced June 2019.
-
Simba: Cosmological Simulations with Black Hole Growth and Feedback
Authors:
Romeel Davé,
Daniel Anglés-Alcázar,
Desika Narayanan,
Qi Li,
Mika H. Rafieferantsoa,
Sarah Appleby
Abstract:
We introduce the Simba simulations, the next generation of the Mufasa cosmological galaxy formation simulations run with Gizmo's meshless finite mass hydrodynamics. Simba includes updates to Mufasa's sub-resolution star formation and feedback prescriptions, and introduces black hole growth via the torque-limited accretion model of Anglés-Alcázar et al. (2017) from cold gas and Bondi accretion from…
▽ More
We introduce the Simba simulations, the next generation of the Mufasa cosmological galaxy formation simulations run with Gizmo's meshless finite mass hydrodynamics. Simba includes updates to Mufasa's sub-resolution star formation and feedback prescriptions, and introduces black hole growth via the torque-limited accretion model of Anglés-Alcázar et al. (2017) from cold gas and Bondi accretion from hot gas, along with black hole feedback via kinetic bipolar outflows and X-ray energy. Ejection velocities are taken to be ~10^3 km/s at high Eddington ratios, increasing to ~8000 km/s at Eddington ratios below 2%, with a constant momentum input of 20L/c. Simba further includes an on-the-fly dust production, growth, and destruction model. Our Simba run with (100 Mpc/h)^3 and 1024^3 gas elements reproduces numerous observables, including galaxy stellar mass functions at z=0-6, the stellar mass--star formation rate main sequence, HI and H2 fractions, the mass-metallicity relation at z=0 and z=2, star-forming galaxy sizes, hot gas fractions in massive halos, and z=0 galaxy dust properties. However, Simba also yields an insufficiently sharp truncation of the z=0 mass function, and too-large sizes for low-mass quenched galaxies. We show that Simba's jet feedback is primarily responsible for quenching massive galaxies.
△ Less
Submitted 2 April, 2019; v1 submitted 29 January, 2019;
originally announced January 2019.
-
Timescales for Hi consumption and SFR depletion of satellite galaxies in groups
Authors:
Mika Rafieferantsoa,
Romeel Davé,
Thorsten Naab
Abstract:
We investigate the connection between the HI~content, SFR and environment of galaxies using a hydrodynamic simulation that incorporates scaling relations for galactic wind and a heuristic halo mass-based quenching prescription. We run two zoom-in simulations of galaxy groups with $M_{halo}>10^{13}M_\odot$ at z=0, selected to have quiet merger histories. We track galaxies as they become satellites,…
▽ More
We investigate the connection between the HI~content, SFR and environment of galaxies using a hydrodynamic simulation that incorporates scaling relations for galactic wind and a heuristic halo mass-based quenching prescription. We run two zoom-in simulations of galaxy groups with $M_{halo}>10^{13}M_\odot$ at z=0, selected to have quiet merger histories. We track galaxies as they become satellites, and compute the delay time $τ_{d}$ during which the satellites are similar to central galaxies at a given stellar mass, and a fading time $τ_{f}$ during which satellites go from gas-rich and star-forming to gas-poor and quiescent. We find $0.7< τ_{d}< 3$ Gyr at $z=0$, and depends inversely on the satellite halo mass at infall. At z ~ 1 we find ~$0.3< τ_{d} < 2$ Gyr, broadly consistent with a positive correlation with the Hubble time. For a given halo mass, lower stellar mass galaxies at infall time have higher $τ_{d}$. We generally find $τ_{f}\ll τ_{d}$, ranging between ~ 150 Myr at z~0 and ~ 80 Myr at z ~ 1 based on linear interpolation, with some uncertainty because they are smaller than our simulation output frequency ($200-300$ Myr). $τ_{f}$ has no obvious dependency on infall halo mass. Both timescales show little difference between HI depletion and SF quenching, indicating that using up the gas reservoir by star formation without refilling is the main mechanism to transform satellite galaxies at these halo masses. At a given physical distance from the center of the main halo of interest, higher redshift galaxies have on average higher cold gas content, but the ratio of gas (HI or H$_2$) to star formation rate is similar, indicating that star formation is consistently fed through reservoirs of HI then H$_2$. For a given amount of HI, galaxies have shorter consumption times in more massive halo structures.
△ Less
Submitted 28 November, 2018;
originally announced November 2018.
-
IQ-Collaboratory 1.1: the Star-Forming Sequence of Simulated Central Galaxies
Authors:
ChangHoon Hahn,
Tjitske K. Starkenburg,
Ena Choi,
Romeel Davé,
Claire M. Dickey,
Marla C. Geha,
Shy Genel,
Christopher C. Hayward,
Ariyeh H. Maller,
Nityasri Mandyam,
Viraj Pandya,
Gergö Pop**,
Mika Rafieferantsoa,
Rachel S. Somerville,
Jeremy L. Tinker
Abstract:
A tightly correlated star formation rate-stellar mass relation of star forming galaxies, or star-forming sequence (SFS), is a key feature in galaxy property-space that is predicted by modern galaxy formation models. We present a flexible data-driven approach for identifying this SFS over a wide range of star formation rates and stellar masses using Gaussian mixture modeling (GMM). Using this metho…
▽ More
A tightly correlated star formation rate-stellar mass relation of star forming galaxies, or star-forming sequence (SFS), is a key feature in galaxy property-space that is predicted by modern galaxy formation models. We present a flexible data-driven approach for identifying this SFS over a wide range of star formation rates and stellar masses using Gaussian mixture modeling (GMM). Using this method, we present a consistent comparison of the $z=0$ SFSs of central galaxies in the Illustris, EAGLE, and Mufasa hydrodynamic simulations and the Santa Cruz semi-analytic model (SC-SAM), alongside data from the Sloan Digital Sky Survey. We find, surprisingly, that the amplitude of the SFS varies by up to ${\sim} 0.7\,\mathrm{dex}$ (factor of ${\sim} 5$) among the simulations with power-law slopes range from $0.7$ to $1.2$. In addition to the SFS, our GMM method also identifies sub-components in the star formation rate-stellar mass relation corresponding to star-burst, transitioning, and quiescent sub-populations. The hydrodynamic simulations are similarly dominated by SFS and quiescent sub-populations unlike the SC-SAM, which predicts substantial fractions of transitioning and star-burst galaxies at stellar masses above and below $10^{10} M_\odot$, respectively. All of the simulations also produce an abundance of low-mass quiescent central galaxies in apparent tension with observations. These results illustrate that, even among models that well reproduce many observables of the galaxy population, the $z=0$ SFS and other sub-populations still show marked differences that can provide strong constraints on galaxy formation models.
△ Less
Submitted 5 September, 2018;
originally announced September 2018.
-
Predicting the Neutral Hydrogen Content of Galaxies From Optical Data Using Machine Learning
Authors:
Mika Rafieferantsoa,
Sambatra Andrianomena,
Romeel Davé
Abstract:
We develop a machine learning-based framework to predict the HI content of galaxies using more straightforwardly observable quantities such as optical photometry and environmental parameters. We train the algorithm on z=0-2 outputs from the Mufasa cosmological hydrodynamic simulation, which includes star formation, feedback, and a heuristic model to quench massive galaxies that yields a reasonable…
▽ More
We develop a machine learning-based framework to predict the HI content of galaxies using more straightforwardly observable quantities such as optical photometry and environmental parameters. We train the algorithm on z=0-2 outputs from the Mufasa cosmological hydrodynamic simulation, which includes star formation, feedback, and a heuristic model to quench massive galaxies that yields a reasonable match to a range of survey data including HI. We employ a variety of machine learning methods (regressors), and quantify their performance using the root mean square error ({\sc rmse}) and the Pearson correlation coefficient (r). Considering SDSS photometry, 3$^{rd}$ nearest neighbor environment and line of sight peculiar velocities as features, we obtain r $> 0.8$ accuracy of the HI-richness prediction, corresponding to {\sc rmse}$<0.3$. Adding near-IR photometry to the features yields some improvement to the prediction. Compared to all the regressors, random forest shows the best performance, with r $>0.9$ at $z=0$, followed by a Deep Neural Network with r $>0.85$. All regressors exhibit a declining performance with increasing redshift, which limits the utility of this approach to $z\la 1$, and they tend to somewhat over-predict the HI content of low-HI galaxies which might be due to Eddington bias in the training sample. We test our approach on the RESOLVE survey data. Training on a subset of RESOLVE data, we find that our machine learning method can reasonably well predict the HI-richness of the remaining RESOLVE data, with {\sc rmse}$\sim0.28$. When we train on mock data from Mufasa and test on RESOLVE, this increases to {\sc rmse}$\sim0.45$. Our method will be useful for making galaxy-by-galaxy survey predictions and incompleteness corrections for upcoming HI 21cm surveys such as the LADUMA and MIGHTEE surveys on MeerKAT, over regions where photometry is already available.
△ Less
Submitted 22 March, 2018;
originally announced March 2018.
-
Mufasa:The strength and evolution of galaxy conformity in various tracers
Authors:
Mika Rafieferantsoa,
Romeel Davé
Abstract:
We investigate galaxy conformity using the Mufasa cosmological hydrodynamical simulation. We show a bimodal distribution in galaxy colour with radius, albeit with too many low-mass quenched satellite galaxies compared to observations. Mufasa produces conformity in observed properties such as colour, sSFR, and HI content; i.e neighbouring galaxies have similar properties. We see analogous trends in…
▽ More
We investigate galaxy conformity using the Mufasa cosmological hydrodynamical simulation. We show a bimodal distribution in galaxy colour with radius, albeit with too many low-mass quenched satellite galaxies compared to observations. Mufasa produces conformity in observed properties such as colour, sSFR, and HI content; i.e neighbouring galaxies have similar properties. We see analogous trends in other properties such as in environment, stellar age, H$_2$ content, and metallicity. We introduce quantifying conformity using $S(R)$, measuring the relative difference in upper and lower quartile properties of the neighbours. We show that low-mass and non-quenched haloes have weak conformity ($S(R)\leq 0.5$) extending to large projected radii $R$ in all properties, while high-mass and quenched haloes have strong conformity ($S(R)\sim 1$) that diminishes rapidly with $R$ and disappears at $R\geq 1$ Mpc. $S(R)$ is strongest for environment in low-mass haloes, and sSFR (or colour) in high-mass haloes, and is dominated by one-halo conformity with the exception of HI in small haloes. Metallicity shows a curious anti-conformity in massive haloes. Tracking the evolution of conformity for $z=0$ galaxies back in time shows that conformity broadly emerges as a late-time ($z\leq 1$) phenomenon. However, for fixed halo mass bins, conformity is fairly constant with redshift out to $z\geq 2$. These trends are consistent with the idea that strong conformity only emerges once haloes grow above Mufasa's quenching mass scale of $\sim 10^{12}M_\odot$. A quantitative measure of conformity in various properties, along with its evolution, thus represents a new and stringent test of the impact of quenching on environment within current galaxy formation models.
△ Less
Submitted 20 December, 2017; v1 submitted 6 July, 2017;
originally announced July 2017.
-
MUFASA: The assembly of the red sequence
Authors:
Romeel Davé,
Mika H. Rafieferantsoa,
Robert J. Thompson
Abstract:
We examine the growth and evolution of quenched galaxies in the Mufasa cosmological hydrodynamic simulations that include an evolving halo mass-based quenching prescription, with galaxy colours computed accounting for line-of-sight extinction to individual star particles. Mufasa reproduces the observed present-day red sequence reasonably well, including its slope, amplitude, and scatter. In Mufasa…
▽ More
We examine the growth and evolution of quenched galaxies in the Mufasa cosmological hydrodynamic simulations that include an evolving halo mass-based quenching prescription, with galaxy colours computed accounting for line-of-sight extinction to individual star particles. Mufasa reproduces the observed present-day red sequence reasonably well, including its slope, amplitude, and scatter. In Mufasa, the red sequence slope is driven entirely by the steep stellar mass-stellar metallicity relation, which independently agrees with observations. High-mass star-forming galaxies blend smoothly onto the red sequence, indicating the lack of a well-defined green valley at M*>10^10.5 Mo. The most massive galaxies quench the earliest and then grow very little in mass via dry merging; they attain their high masses at earlier epochs when cold inflows more effectively penetrate hot halos. To higher redshifts, the red sequence becomes increasingly contaminated with massive dusty star-forming galaxies; UVJ selection subtly but effectively separates these populations. We then examine the evolution of the mass functions of central and satellite galaxies split into passive and star-forming via UVJ. Massive quenched systems show good agreement with observations out to z~2, despite not including a rapid early quenching mode associated with mergers. However, low-mass quenched galaxies are far too numerous at z<1 in Mufasa, indicating that Mufasa strongly over-quenches satellites. A challenge for hydrodynamic simulations is to devise a quenching model that produces enough early massive quenched galaxies and keeps them quenched to z=0, while not being so strong as to over-quench satellites; Mufasa's current scheme fails at the latter.
△ Less
Submitted 4 July, 2017; v1 submitted 4 April, 2017;
originally announced April 2017.
-
Cold gas strip** in satellite galaxies: from pairs to clusters
Authors:
Toby Brown,
Barbara Catinella,
Luca Cortese,
Claudia del P. Lagos,
Romeel Dave,
Virginia Kilborn,
Martha P. Haynes,
Riccardo Giovanelli,
Mika Rafieferantsoa
Abstract:
In this paper we investigate environment driven gas depletion in satellite galaxies, taking full advantage of the atomic hydrogen (HI) spectral stacking technique to quantify the gas content for the entire gas-poor to -rich regime. We do so using a multi-wavelength sample of 10,600 satellite galaxies, selected according to stellar mass (log M$_{\star}$/M$_{\odot}$ $\geq$ 9) and redshift (0.02…
▽ More
In this paper we investigate environment driven gas depletion in satellite galaxies, taking full advantage of the atomic hydrogen (HI) spectral stacking technique to quantify the gas content for the entire gas-poor to -rich regime. We do so using a multi-wavelength sample of 10,600 satellite galaxies, selected according to stellar mass (log M$_{\star}$/M$_{\odot}$ $\geq$ 9) and redshift (0.02 $\leq$ z $\leq$ 0.05) from the Sloan Digital Sky Survey, with HI data from the Arecibo Legacy Fast ALFA (ALFALFA) survey. Using key HI-to-stellar mass scaling relations, we present evidence that the gas content of satellite galaxies is, to a significant extent, dependent on the environment in which a galaxy resides. For the first time, we demonstrate that systematic environmental suppression of gas content at both fixed stellar mass and fixed specific star formation rate (sSFR) in satellite galaxies begins in halo masses typical of the group regime (log M$_{h}$/M$_{\odot}$ < 13.5), well before galaxies reach the cluster environment. We also show that environment driven gas depletion is more closely associated to halo mass than local density. Our results are then compared with state-of-the-art semi-analytic models and hydrodynamical simulations and discussed within this framework, showing that more work is needed if models are to reproduce the observations. We conclude that the observed decrease of gas content in the group and cluster environments cannot be reproduced by starvation of the gas supply alone and invoke fast acting processes such as ram-pressure strip** of cold gas to explain this.
△ Less
Submitted 16 November, 2016; v1 submitted 3 November, 2016;
originally announced November 2016.
-
MUFASA: Galaxy star formation, gas, and metal properties across cosmic time
Authors:
Romeel Davé,
Mika H. Rafieferantsoa,
Robert J. Thompson,
Philip F. Hopkins
Abstract:
We examine galaxy star formation rates (SFRs), metallicities, and gas contents predicted by the MUFASA cosmological hydrodynamic simulations, which employ meshless hydrodynamics and novel feedback prescriptions that yield a good match to observed galaxy stellar mass assembly. We combine 50, 25, and 12.5 Mpc/h boxes with a quarter billion particles each to show that MUFASA broadly reproduces a wide…
▽ More
We examine galaxy star formation rates (SFRs), metallicities, and gas contents predicted by the MUFASA cosmological hydrodynamic simulations, which employ meshless hydrodynamics and novel feedback prescriptions that yield a good match to observed galaxy stellar mass assembly. We combine 50, 25, and 12.5 Mpc/h boxes with a quarter billion particles each to show that MUFASA broadly reproduces a wide range of relevant observations, including SFR and specific SFR functions, the mass-metallicity relation, HI and H2 fractions, HI (21 cm) and CO luminosity functions, and cosmic gas density evolution. There are mild but significant discrepancies, such as too many high-SFR galaxies, overly metal-rich and HI-poor galaxies at M*>10^{10} Mo, and sSFRs that are too low at z~1-2. The HI mass function increases by x2 out to z~1 then steepens to higher redshifts, while the CO luminosity function computed using the Narayanan et al. conversion factor shows a rapid increase of CO-bright galaxies out to z~2 in accord with data. Omega_HI and Omega_H2 both scale roughly as (1+z)^0.7 out to z~3, comparable to the rise in HI and H2 fractions. MUFASA galaxies with high SFR at a given M* have lower metallicities and higher HI and H2 fractions, following observed trends; we make quantitative predictions for how fluctuations in the baryon cycle drive correlated scatter around galaxy scaling relations. Most of these trends are well converged with numerical resolution. These successes highlight MUFASA as a viable platform to study many facets of cosmological galaxy evolution.
△ Less
Submitted 5 October, 2016;
originally announced October 2016.
-
The impact of environment and mergers on the HI content of galaxies in hydrodynamic simulations
Authors:
Mika Rafieferantsoa,
Romeel Davé,
Daniel Anglés-Alcazar,
Neal Katz,
Juna A. Kollmeier,
Benjamin D. Oppenheimer
Abstract:
We quantitatively examine the effects of accretion and environment on the HI content of galaxies within a cosmological hydrodynamic simulation that reproduces basic observed trends of HI in galaxies. We show that our model broadly reproduces the observed scatter in HI at a given stellar mass as quantified by the HI mass function in bins of stellar mass, as well as the HI richness versus local gala…
▽ More
We quantitatively examine the effects of accretion and environment on the HI content of galaxies within a cosmological hydrodynamic simulation that reproduces basic observed trends of HI in galaxies. We show that our model broadly reproduces the observed scatter in HI at a given stellar mass as quantified by the HI mass function in bins of stellar mass, as well as the HI richness versus local galaxy density. This shows that the predicted HI fluctuations and environmental effects are roughly consistent with data with few minor discrepancies. For satellite galaxies in >= 10^12M_* halos, the HI richness distribution is bimodal and drops towards the largest halo masses. The depletion rate of HI once a galaxy enters a more massive halo is more rapid at higher halo mass, in contrast to the specific star formation rate which shows much less variation in the attenuation rate versus halo mass. This suggests that, up to halo mass scales probed here (<= 10^14M_*), star formation is mainly attenuated by starvation, but HI is additionally removed by strip** once a hot gaseous halo is present. In low mass halos, the HI richness of satellites is independent of radius, while in high mass halos they become gas-poor towards the center. We show that the gas fraction of satellite and central galaxies decreases from z=5 to z=0,tracking each other until z~1 after which the satellites' HI content drops much more quickly, particularly for the highest halo masses. Mergers somewhat increase the HI richness and its scatter about the mean relation, tracking the metallicity in a way consistent with arising from inflow fluctuations, while star formation is significantly boosted relative to HI.
△ Less
Submitted 15 September, 2015; v1 submitted 11 August, 2014;
originally announced August 2014.
