This paper describes a new publicly available codebase for modelling galaxy formation in a cosmological context, the "Semi-Analytic Galaxy Evolution" model, or SAGE for short. SAGE is a significant update to that used in Croton et al. (2006) and has been rebuilt to be modular and customisable. The model will run on any N-body simulation whose trees are organised in a supported format and contain a minimum set of basic halo properties. In this work we present the baryonic prescriptions implemented in SAGE to describe the formation and evolution of galaxies, and their calibration for three N-body simulations: Millennium, Bolshoi, and GiggleZ. Updated physics include: gas accretion, ejection due to feedback, and reincorporation via the galactic fountain; a new gas cooling--radio mode active galactic nucleus (AGN) heating cycle; AGN feedback in the quasar mode; a new treatment of gas in satellite galaxies; and galaxy mergers, disruption, and the build-up of intra-cluster stars. Throughout, we show the results of a common default parameterization on each simulation, with a focus on the local galaxy population.

Approximately half of the matter in the Universe is "unbound" at z = 0, according to N-body simulations such as the Millennium Run. Here, we use the milli-Millennium simulation to examine the distribution of unbound matter in relation to the dark matter halos which host galaxies. We measure the unbound matter within two types of windows, using a halo dependent radius and a fixed radius at several different scales. We also consider the timescales over which a halo can accrete the local unbound matter at z = 2 and z = 0. Finally, we compare the unbound matter to observable properties of galaxies, such as local galaxy count environment and stellar mass. We find that halos at z = 2 can accrete far more of the nearby unbound matter over a Hubble time than halos at z = 0 and that 78% of particles within 5 $R_{vir}$ of a halo at z = 2 will be accreted by z = 0, compared to 36% of particles within 5 $h^{-1}$ Mpc of the halo. We also find that galaxy count environment is closely related to the amount of nearby unbound matter when measured on the same scale.

With the installation of the Cosmic Origins Spectrograph on the Hubble Space Telescope, measurements of the metal content of the low redshift intergalactic medium (IGM) are now available. Using a new grid-based model for diffuse gas coupled to the SAGE semi-analytic model of galaxy formation, we examine the impact of supernova feedback on the pollution of the IGM. We consider different assumptions for the reheating and ejection of gas by supernovae and their dependence on galaxy circular velocity and gas surface density. Where metals are present, we find the most likely metallicity to be $-1.5 < $log$_{10}$(Z/Z$_{\odot}$)$< -1.0$ at $z = 0$, consistent with both observations and more sophisticated hydrodynamic simulations. Our model predicts that the regions of the IGM with the highest metallicities will be near galaxies with M$_{\star} \sim 10^{10.5}h^{-1}$M$_{\odot}$ and in environments of densities $\sim 10 \times$ the mean. We also find that 90% of IGM metals at $z = 0$ are ejected by galaxies with stellar masses less than $10^{10.33}h^{-1}$M$_{\odot}$.

Galaxy environment is frequently discussed, but inconsistently defined. It is especially difficult to measure at high redshift where only photometric redshifts are available. With a focus on early forming proto-clusters, we use a semi-analytical model of galaxy formation to show how the environment measurement around high redshift galaxies is sensitive to both scale and metric, as well as to cluster viewing angle, evolutionary state, and the availability of either spectroscopic or photometric data. We use two types of environment metrics (nearest neighbour and fixed aperture) at a range of scales on simulated high-z clusters to see how "observed" overdensities compare to "real" overdensities. We also "observationally" identify z = 2 proto-cluster candidates in our model and track the growth histories of their parent halos through time, considering in particular their final state at z = 0. Although the measured environment of early forming clusters is critically dependent on all of the above effects (and in particular the viewing angle), we show that such clusters are very likely (< 90%) to remain overdense at z = 0, although many will no longer be among the most massive. Object to object comparisons using different methodologies and different data, however, require much more caution.

We analyze environmental correlations using mark clustering statistics with the mock galaxy catalogue constructed by Muldrew et al. (Paper I). We find that mark correlation functions are able to detect even a small dependence of galaxy properties on the environment, quantified by the overdensity $1+\delta$, while such a small dependence would be difficult to detect by traditional methods. We then show that rank ordering the marks and using the rank as a weight is a simple way of comparing the correlation signals for different marks. With this we quantify to what extent fixed-aperture overdensities are sensitive to large-scale halo environments, nearest-neighbor overdensities are sensitive to small-scale environments within haloes, and colour is a better tracer of overdensity than is luminosity.

We examine the implications of recent measurements of the Milky Way rotation for the trajectory of the Large Magellanic Cloud. The ~ 14+/-6% increase in the MW circular velocity relative to the IAU standard of 220 km/s changes the qualitative nature of the inferred LMC orbit. Instead of the LMC being gravitationally unbound, as has been implied based on a recent measurement of its proper motion, we find that the past orbit of the LMC is naturally confined within the virial boundary of the MW, although still not as tightly bound as in previous estimates.