Simulating large cosmology surveys with calibrated halo models

Abstract

In this thesis I present a novel method for constructing large scale mock galaxy and halo catalogues and apply this model to a number of important topics in modern cosmology. Traditionally such mocks are created through first evolving a high resolution particle simulation from a set of initial conditions to the present epoch, identifying bound structures and their evolution, and finally applying a semi-analytic prescription for galaxy formation. In contrast to this computationally expensive procedure, I use low resolution simulations to obtain a density field that traces large scale modes. From this background I sample the population statistics of halos: the number of halos which are typically found within a region of a given overdensity, to produce a halo catalogue. From the halo catalogue I then produce galaxies by appealing to the halo model. In this model the expected number of galaxies within a halo and the distribution of their properties is dependent on halo mass alone. By sampling conditional luminosity functions for a number of populations of galaxies, I produce a galaxy catalogue with luminosity and colour properties. The aim of developing algorithm is not to probe the mechanics of galaxy formation in great detail. It is instead intended as a method of rapidly producing mock galaxy and halo catalogues rapidly on modern desktop computers. The approach we will take is to try to distill the minimal algorithm required to achieve this and still provide useful catalogues for observational cosmologists. Both the conditional mass function and conditional luminosity functions required for the algorithm are calibrated from the Millennium Simulation, one of the highest resolution cosmology simulations to date, and its associated semi-analytic catalogues. In Chapter 2 I examine these statistics and provide fits to the quantities of interest. As a test of the method, in Chapter 3 I produce a halo and galaxy catalogue from the same large scale modes as the Millennium Simulation. The clustering statistics of galaxies and halos within this re-simulation are calculated and compared with those of the original. Con dent of the accuracy of the method, in Chapter 4 I populate a number of simulations, each 8 times the volume of the Millennium Simulation, and study the evolution of the Baryon Acoustic Oscillation signal. For each population (dark matter, halos and galaxies) I fit the BAO in the power spectrum to obtain the shift in the BAO peak. In Chapter 5 I extend the algorithm to produce lightcones: simulated skies in which the evolution of the Universe along the line of sight is accounted for. I simulate the geometry and limitations of a major pending survey and calculate the expected clustering signature I expect to see in both. The redshift space distortions induced by peculiar velocities of galaxies along the line of sight are determined and their ability to distinguish between gravity models is also explored. In Chapter 6 I detail a further extension to the algorithm for simulating weak gravitational lensing surveys. I use the analytic 2D surface density pro files of NFW profiles to dress each dark matter halo on a lightcone. The sum of these pro files over the entire population can be used to construct high resolution maps of the convergence. From these maps I calculate the spectrum of the convergence and compare with theoretical predictions. Finally in Chapter 7 I discuss further possible applications and extensions of the algorithm I have developed in this thesis.