Summary: In this post we show the median variations of main branch halo properties connected by the descendant information without running a tree builder on top. Our results and their estimated uncertainties are reasonable given the fact that we do not used a tree builder to improve the connection between halos.
This is the continuation of the post from 2021-02-22. Thereby we use only descendant information as the ‘DescID’ provided directly by the halo finder Rockstar (RS). The halos between snapshots are connected by the descendant ID provided by Rockstar as a standard output. When we need to decide whether a halo is serves a main progenitor or not in case on halo has various progenitors, then we choose the most massive (MM) one.
We show in Fig. 1 the median values of variation of shared fraction of particle IDs within \(10^{th}\) and \(90^{th}\) percentiles and in Fig. 2 the variaton of the virial halo mass \(M_{vir}\) \(h^{-1}M_{\odot}\), virial radius \(R_{vir}\) \(h^{-1}\)Mpc, and X/Y/Z-positions in \(h^{-1}\)Mpc. In Fig. 3 and Fig. 4 show the same results but use the median absolute deviation (MAD) as an error estimator which is the median of the absolute deviations and a robust measurement for uncertainties when dealing with outliers.