Nonlinear growth of structure in cosmological simulations

Lukic, Zarija

Permalink

https://hdl.handle.net/2142/79031 Copy

Description

Title

Nonlinear growth of structure in cosmological simulations

Author(s)

Lukic, Zarija

Issue Date

2008

Director of Research (if dissertation) or Advisor (if thesis)

Ricker, Paul M.

Committee Member(s)

• Mohr, Joseph J.
• Fields, Brian
• Gammie, Charles
• Heitmann, Katarin

Department of Study

Astronomy

Discipline

Astronomy

Degree Granting Institution

University of Illinois at Urbana-Champaign

Degree Name

Ph.D. (doctoral)

Degree Level

Dissertation

Keyword(s)

• Dark matter halos
• Halo mass function
• Structure formation theory

Language

en

Abstract

Upcoming cosmological observations (South Pole Telescope, Atacama Cosmology Telescope, Baryon Oscillation Spectroscopic Survey, and Planck) will allow for accurately. probing structures and their growth, some into highly nonlinear regimes. These observations, in combination with already very accurate measurements of the expansion rate of the universe, will not only constrain cosmological parameters to a percent level, but will also answer what is the theory . of gravity on the largest scales. In order to obtain theoretical predictions for different measurables (like the distribution of masses, spatial· correlations), large numerical simulations have to be carried out. In this context, their main goal is to quantify how are such measurables affected by a change of cosmological parameters. The promised high accuracy of observations make the simulation task very demanding, as the theoretical predictions have to be at least as accurate as the observations. In this thesis, we study the formation and evolution of dark matter halos in ACDM models.o~er a wide range of cosmologicalepochs, from redshift z=20 to the present. First, we focus on the halo mass function, likely a key probe of cosmological growth of structure. By performing a large suite (60 simulations) of nested-box N-body simulations with careful convergence and error controls, we determine the mass function and its evolution with excellent statistical and systematic errors, reaching a few percent over most of the considered redshift and mass range. Our results are consistent with a 'universal' form for the mass function, and are in a good agreement with the Warren et al. analytic fit. Next, we. study the· structure of halos and ratification of different halo mass definitions. This analysis is important for connecting structure formation theory with observations, and also impacts the widely used approaches of assigning visible galaxies to dark matter halos - the halo occupancy distribution models. We find that the vast majority of halos (80-85%) appear as isolated objects, allowing for an accurate mapping between the two main mass definitions (friends-of-friends and spherical overdensity). Based on results from Monte Carfo realizations of .ideal Navarro-Frenk-White halos and N-body simulations we provide a mass mapping formula. Furthermore, investigation of non-isolated, bridged halos, reveals that the fraction of these halos and their satellite mass distribution is cosmology dependent, and can be expressed in a cosmology universal form. Third, we turn to the spatial distribution of. halos, which serves as a 'biased' mass tracer. While this bias is scale dependent, at large distances it asymptotes to a constant value. We show that commonly used, heuristic approach to relatingthe mass function to the bias (peak-background split) clearly fails at the accuracy we are interested in (less or equal to 10%). Using our large set of simulations we provide universal formula for halo bias as a function· of mass. This formula fit well not only our data, put the current state of the art simulation data (Millenium simulation). Finally, we present the results of a comparison between 10 different cosmology codes. These include virtually all major codes used today, and more importantly, they completely cover the range of numerical algorithms used in cosmological N-body simulations. For the mass function, the matter power spectrum, arid halo profiles - the most important statistics for this thesis - codes agree at less then 10% over wide dynamic ranges. This robustness gives us additional confidence in our numerical results.

Type of Resource

text

Permalink

http://hdl.handle.net/2142/79031

Owning Collections