A testrun on merger trees part V

—    —

  previous

next  

Summary: In this post we compare the redshift evolution of halo properties using descendant information from a halo finder directly and from a tree finder run on top. We found that there is little difference between the two approaches and that surprisingly less variation in e.g. the fraction of shared particle IDs between two snapshots was found (compare Fig. 3 and Fig. 4). However, if a tree builder is not used then the full redshift history of the particular halo might not be traced correctly (see Fig. 2).

This is the continuation of the post from 2021-02-26 were we showed results on the median variation of halo properties when following halos on the their main progenitor branch. Thereby we used only descendant information provided directly by the halo finder Rockstar (RS). In this post we compare our findings using this simplified approach with results when running the tree finder Consistent-Trees (CT).

Full merger tree and main branch only of most massive halo with top node TreeID79509. The halo was found by Rockstar and the merger trees build by Consitent-Trees (CT).

Projection on X-Z axis of main branch of most massive halo. The red circles represent halos on the main branch if only Rockstar's (RS) descendant IDs were used to build the main branch and the colored dots the same information when Consistent-Trees (CT) was used to correct connections.

The variation in the shared fraction of particle IDs between two direct progenitor halos found by Rockstar (RS) and connected by Consistent-Trees (CT). The fraction of particle IDs of a halo at a certain snapshot SN(z) in comparison to the fraction found in the direct progenitor at the subsequent SN(z+1)) is given in red and the fraction of particle IDs of the direct progenitor at certain snapshot SN(z+1) in comparison to the halo at SN(z) in blue for the main branch halos with TreeID79509.

Same as in Fig.4 but here the halos were connected only by the descendant ID descID provided directly by Rockstar. Interestingly our results show less variations in the fraction of shared particle IDs and halo mass for the most massive halo found in our simulation treeID3 (Note that this halo is not the same as in Fig. 3).

Possible reason of uncertainties might be:

• the resolution of the simulation (only 256^3 particles in a box of 50h-1Mpc)
• too large time steps between simulation
• parameter using for Consitent-Trees were not particularly chosen but the default setting used

  A testrun on merger trees part IV

Parameterizing the mass growth history  

—    —