http://arxiv.org/abs/1812.06708
The implementation of ACACIA (\textbf{a}nother \textbf{c}atchy \textbf{a}cronym for \textbf{c}lump-based mergertree \textbf{i}dentification in r\textbf{a}mses), an algorithm to generate dark matter halo merger trees on the fly, into the AMR code RAMSES is presented.
The algorithm is fully parallel using MPI, works on the fly, and tracks dark matter substructures individually through particle IDs. Once a clump merges into another, it is still being tracked through the last identified most tightly bound particle of that clump, allowing to check at later snapshots whether the identified merging event truly was one, or whether a misidentification by the density field gradient based clump finding algorithm might have occurred, as well as to track orphan galaxies.
The influence on the merger trees of various definitions of substructure and the maximal number of particles tracked per clump have been tested.
Furthermore, with the known formation history of dark matter structure, galaxies can be introduced in a simulation containing only dark matter particles in a forward modelling approach through use of a parametrised stellar-mass-to-halo-mass (SMHM) relation.
The obtained predicted stellar mass functions of central galaxies from $z\sim 0 – 8$ and correlation functions at $z\sim0$ are compared to observations.
Considering that the mock galaxy catalogues were obtained using simulations without particularly high resolutions ($512^3$ particles), the results show satisfactory agreement with observational data.
We have however demonstrated that our new merger tree algorithm and the generation of the corresponding mock galaxy catalogue can be performed on the fly, using the RAMSES built-in clump finder.
M. Ivkovic and R. Teyssier
Tue, 18 Dec 18
85/91
Comments: 18 pages, 9 figures
You must be logged in to post a comment.