We examine the global HI properties of galaxies in quarter-billion particle cosmological simulations using Gadget-2, focusing on how galactic outflows impact HI content. We consider four outflow models, including a new one (ezw) motivated by recent i
nterstellar medium simulations in which the wind speed and mass loading factor scale as expected for momentum-driven outflows for larger galaxies and energy-driven outflows for dwarfs (sigma<75 km/s). To obtain predicted HI masses, we employ a simple but effective local correction for particle self-shielding, and an observationally-constrained transition from neutral to molecular hydrogen. Our ezw simulation produces an HI mass function whose faint-end slope of -1.3 agrees well with observations from the ALFALFA survey; other models agree less well. Satellite galaxies have a bimodal distribution in HI fraction versus halo mass, with smaller satellites and/or those in larger halos more often being HI-deficient. At a given stellar mass, HI content correlates with star formation rate and inversely correlates with metallicity, as expected if driven by stochasticity in the accretion rate. To higher redshifts, massive HI galaxies disappear and the mass function steepens. The global cosmic HI density conspires to remain fairly constant from z~5-0, but the relative contribution from smaller galaxies increases with redshift.
We present an analytic formalism that describes the evolution of the stellar, gas, and metal content of galaxies. It is based on the idea, inspired by hydrodynamic simulations, that galaxies live in a slowly-evolving equilibrium between inflow, outfl
ow, and star formation. We argue that this formalism broadly captures the behavior of galaxy properties evolving in simulations. The resulting equilibrium equations for the star formation rate, gas fraction, and metallicity depend on three key free parameters that represent ejective feedback, preventive feedback, and re-accretion of ejected material. We schematically describe how these parameters are constrained by models and observations. Galaxies perturbed off the equilibrium relations owing to inflow stochasticity tend to be driven back towards equilibrium, such that deviations in star formation rate at a given mass are correlated with gas fraction and anti-correlated with metallicity. After an early gas accumulation epoch, quiescently star-forming galaxies are expected to be in equilibrium over most of cosmic time. The equilibrium model provides a simple intuitive framework for understanding the cosmic evolution of galaxy properties, and centrally features the cycle of baryons between galaxies and surrounding gas as the driver of galaxy growth.
We use cosmological hydrodynamic simulations to investigate how inflows, star formation, and outflows govern the the gaseous and metal content of galaxies. In our simulations, galaxy metallicities are established by a balance between inflows and outf
lows as governed by the mass outflow rate, implying that the mass-metallicity relation reflects how the outflow rate varies with stellar mass (M*). Gas content is set by a competition between inflow into and gas consumption within the ISM, the latter being governed by the SF law, while the former is impacted by both wind recycling and preventive feedback. Stochastic variations in the inflow rate move galaxies off the equilibrium M*-Z and Z*-fgas relations in a manner correlated with star formation rate, and the scatter is set by the timescale to re-equilibrate. The evolution of both relations from z=3-0 is slow, as individual galaxies tend to evolve mostly along the relations. Gas fractions at a given M* slowly decrease with time because the cosmic inflow rate diminishes faster than the consumption rate, while metallicities slowly increase as infalling gas becomes more enriched. Observations from z~3-0 are better matched by simulations employing momentum-driven wind scalings rather than constant wind speeds, but all models predict too low gas fractions at low masses and too high metallicities at high M*. All our models reproduce observed second-parameter trends of the mass-metallicity relation with star formation rate and environment, indicating that these are a consequence of equilibrium and not feedback. Overall, the analytical framework of our equilibrium scenario broadly captures the relevant physics establishing the galaxy gas and metal content in simulations, which suggests that the cycle of baryonic inflows and outflows centrally governs the cosmic evolution of these properties in typical star-forming galaxies.
We examine the growth of the stellar content of galaxies from z=3-0 in cosmological hydrodynamic simulations incorporating parameterised galactic outflows. Without outflows, galaxies overproduce stellar masses (M*) and star formation rates (SFRs) com
pared to observations. Winds introduce a three-tier form for the galaxy stellar mass and star formation rate functions, where the middle tier depends on differential (i.e. mass-dependent) recycling of ejected wind material back into galaxies. A tight M*-SFR relation is a generic outcome of all these simulations, and its evolution is well-described as being powered by cold accretion, although current observations at z>2 suggest that star formation in small early galaxies must be highly suppressed. Roughly one-third of z=0 galaxies at masses below M^* are satellites, and star formation in satellites is not much burstier than in centrals. All models fail to suppress star formation and stellar mass growth in massive galaxies at z<2, indicating the need for an external quenching mechanism such as black hole feedback. All models also fail to produce dwarfs as young and rapidly star-forming as observed. An outflow model following scalings expected for momentum-driven winds broadly matches observed galaxy evolution around M^* from z=0-3, which is a significant success since these galaxies dominate cosmic star formation, but the failures at higher and lower masses highlight the challenges still faced by this class of models. We argue that central star-forming galaxies are well-described as living in a slowly-evolving equilibrium between inflows from gravity and recycled winds, star formation, and strong and ubiquitous outflows that regulate how much inflow forms into stars. Star-forming galaxy evolution is thus primarily governed by the continual cycling of baryons between galaxies and intergalactic gas.
Herschel has opened new windows into studying the evolution of rapidly star-forming galaxies out to high redshifts. Todays massive starbursts are characterized by star formation rates (SFRs) of 100+ Mo/yr and display a chaotic morphology and nucleate
d star formation indicative of a major merger. At z~2, galaxies of similar mass and SFR are characterized by ordered rotation and distributed star formation. The emerging cold accretion paradigm provides an intuitive understanding for such differences. In it, halo accretion rates govern the supply of gas into star-forming regions, modulated by strong outflows. The high accretion rates at high-z drive more rapid star formation, while also making disks thicker and clumpier; the clumps are expected to be short-lived in the presence of strong galactic outflows as observed. Hence equivalently rapid star-formers at high redshift are not analogous to local merger-driven starbursts, but rather to local disks with highly enhanced accretion rates.
It is difficult to reconcile the observed evolution of the star formation rate versus stellar mass (SFR-M*) relation with expectations from current hierarchical galaxy formation models. The observed SFR-M* relation shows a rapid rise in SFR(M*) from
z=0-2, and then a surprisingly lack of amplitude evolution out to z~6+. Hierarchical models of galaxy formation match this trend qualitatively but not quantitatively, with a maximum discrepancy of ~x3 in SFR at z~2. One explanation, albeit radical, is that the IMF becomes modestly weighted towards massive stars out to z~2, and then evolves back towards its present-day form by z~4 or so. We observe that this redshift trend mimics that of the cosmic fraction of obscured star formation, perhaps hinting at a physical connection. Such IMF evolution would concurrently go towards explaining persistent discrepancies between integrated measures of star formation and present-day stellar mass or cosmic colors.
The kinematics of damped Lyman alpha absorbers (DLAs) are difficult to reproduce in hierarchical galaxy formation models, particularly the preponderance of wide systems. We investigate DLA kinematics at z=3 using high-resolution cosmological hydrodyn
amical simulations that include a heuristic model for galactic outflows. Without outflows, our simulations fail to yield enough wide DLAs, as in previous studies. With outflows, predicted DLA kinematics are in much better agreement with observations. Comparing two outflow models, we find that a model based on momentum-driven wind scalings provides the best match to the observed DLA kinematic statistics of Prochaska & Wolfe. In this model, DLAs typically arise a few kpc away from galaxies that would be identified in emission. Narrow DLAs can arise from any halo and galaxy mass, but wide ones only arise in halos with mass >10^11 Mo, from either large central or small satellite galaxies. This implies that the success of this outflow model originates from being most efficient at pushing gas out from small satellite galaxies living in larger halos. This increases the cross-section for large halos relative to smaller ones, thereby yielding wider kinematics. Our simulations do not include radiative transfer effects or detailed metal tracking, and outflows are modeled heuristically, but they strongly suggest that galactic outflows are central to understanding DLA kinematics. An interesting consequence is that DLA kinematics may place constraints on the nature and efficiency of gas ejection from high-z galaxies.
The intergalactic medium (IGM) is the dominant reservoir of baryons at all cosmic epochs. We investigate the evolution of the IGM from z=2-0 in 48 Mpc/h, 110-million particle cosmological hydrodynamic simulations using three prescriptions for galacti
c outflows. We focus on the evolution of IGM physical properties, and how such properties are traced by Ly-alpha absorption as detectable using HST/COS. Our results broadly confirm the canonical picture that most Ly-alpha absorbers arise from highly ionized gas tracing filamentary large-scale structure. Growth of structure causes gas to move from the diffuse photoionized IGM into other cosmic phases, namely stars, cold and hot gas within galaxy halos, and the unbound and shock-heated warm-hot intergalactic medium (WHIM). By today, baryons are roughly equally divided between bound phases (35%), the diffuse IGM (41%), and the WHIM (24%). Here we (re)define the WHIM as gas with overdensities lower than that in halos and temperatures >10^5 K, in order to more closely align it with missing baryons. When we tune our photoionizing background to match the observed evolution of the Ly-alpha mean flux decrement, we obtain a line count evolution that broadly agrees with available data. We predict a column density distribution slope of -1.70 for our favored momentum-driven wind model, in agreement with recent observations, and it becomes shallower with redshift. With improved statistics, the frequency of strong lines can be a valuable diagnostic of outflows, and our favored wind model matches existing data best among our models. The relationship between column density and physical density is fairly tight from z=2-0, and evolves as rho N_HI^0.74 10^(-0.37z) for diffuse absorbers. Linewidths only loosely reflect the temperature of the absorbing gas, which will hamper attempts to quantify the WHIM using broad Ly-alpha absorbers. [Abridged]
We study the nature of rapidly star-forming galaxies at z=2 in cosmological hydrodynamic simulations, and compare their properties to observations of sub-millimetre galaxies (SMGs). We identify simulated SMGs as the most rapidly star-forming systems
that match the observed number density of SMGs. In our models, SMGs are massive galaxies sitting at the centres of large potential wells, being fed by smooth infall and gas-rich satellites at rates comparable to their star formation rates (SFR). They are not typically undergoing major mergers that significantly boost their quiescent SFR, but they still often show complex gas morphologies and kinematics. Our simulated SMGs have stellar masses of log M*/Mo~11-11.7, SFRs of ~180-500 Mo/yr, a clustering length of 10 Mpc/h, and solar metallicities. The SFRs are lower than those inferred from far-IR data by a factor of 3, which we suggest may owe to one or more systematic effects in the SFR calibrations. SMGs at z=2 live in ~10^13 Mo halos, and by z=0 they mostly end up as brightest group galaxies in ~10^14 Mo halos. We predict that higher-M* SMGs should have on average lower specific SFRs, less disturbed morphologies, and higher clustering. We also predict that deeper far-IR surveys will smoothly join SMGs onto the massive end of the SFR-M* relationship defined by lower-mass z=2 galaxies. Overall, our simulated rapid star-formers provide as good a match to available SMG data as merger-based scenarios, offering an alternative scenario that emerges naturally from cosmological simulations.
We present predictions for galactic halo baryon fractions from cosmological hydrodynamic simulations with a well-constrained model for galactic outflows. Without outflows, halos contain roughly the cosmic fraction of baryons, slightly lowered at high
masses owing to pressure support from hot gas. The star formation efficiency is large and increases monotonically to low masses, in disagreement with data. With outflows, the baryon fraction is increasingly suppressed in halos to lower masses. A Milky Way-sized halo at z=0 has about 60% of the cosmic fraction of baryons, so missing halo baryons have largely been evacuated, rather than existing in some hidden form. Large halos (>10^13 Mo) contain 85% of their cosmic share of baryons, which explains the mild missing baryon problem seen in clusters. By comparing results at z=3 and z=0, we show that most of the baryon removal occurs at early epochs in larger halos, while smaller halos lose baryons more recently. Star formation efficiency is maximized in halos of ~10^13 Mo, dropping significantly to lower masses, which helps reconcile the sub-L* slope of the observed stellar and halo mass functions. These trends are predominantly driven by differential wind recycling, namely, that wind material takes longer to return to low-mass galaxies than high-mass galaxies. The hot gas content of halos is mostly unaffected by outflows, showing that outflows tend to blow holes and escape rather than deposit their energy into halo gas.
