A New Halo Finding Method for N-body Simulations

We have developed a new halo finding method, Physically Self-Bound (PSB) group finding algorithm, which can efficiently identify halos located even at crowded regions. This method combines two physical criteria such as the tidal radius of a halo and the total energy of each particle to find member particles. No subtle dependence of halo mass functions on various parameters of the PSB method has been found. Two hierarchical meshes are used to increase the speed and the power of halo identification in the parallel computing environments. First, a coarse mesh with cell size equal to the mean particle separation lmean is used to obtain the density field over the whole simulation box. Mesh cells having density contrast higher than a local cutoff threshold δLOC are extracted and linked together for those adjacent to each other. This produces local-cell groups. We analyze the group of particles located at each local-cell group region separately. This treatment makes the halo finding method easily implemented on the parallel computing environments since each computational rank takes the halo identification job in each particle group independently. Second, a finer mesh is used to obtain density field within each local-cell group and to identify halos. We set the cell size of the refined mesh to be twice the gravitational force softening length ǫ. The density peaks in the fine mesh are the halo candidates. Based on the fine mesh, we split particles into many groups located in the density shells with different density levels. If a density shell contains only one density peak, its particles are assigned to the density peak. But in the case of a density shell surrounding at least two density peaks, we use both the tidal radii of halo candidates enclosed by the shell and the total energy criterion to find physically bound particles with respect to each halo. We have tested the PSB using a binary halo model against other popular halo finding methods, such as the Friend-of-Friend (FoF), Spherical Overdensity (SO), DENMAX, and HOP. Similar to DENMAX and HOP, the PSB method can efficiently identify small halos embedded in a large halo, while the FoF and the SO do not resolve such small halos. We apply our new halo finding method to a 1Giga particle simulation of the ΛCDM model and compare the resulting mass function with those of previous studies. The abundance of physically self-bound halos is larger at the low mass scale and smaller at the high mass scale than proposed by the Jenkins et al. (2001) who used the FoF and SO methods. Subject headings: Cosmology: N-body simulation: halo : halo-finding methods: Numerical