# Climbing Halo Merger Trees with TreeFrog

## TL;DR

TreeFrog is a high-performance, parallel halo merger tree builder that improves the accuracy of halo tracking across snapshots, reduces pathologies, and aligns well with established merger rate models.

## Contribution

The paper introduces TreeFrog, a novel parallel code for constructing halo merger trees with improved matching algorithms and pathology corrections, enhancing accuracy over previous methods.

## Key findings

- Reduces stranded halos by searching multiple snapshots.
- Improves primary-secondary progenitor distinction with a new merit function.
- Aligns measured merger rates with established models.

## Abstract

We present TreeFrog, a massively parallel halo merger tree builder that is capable comparing different halo catalogues and producing halo merger trees. The code is written in c++11, use the MPI and OpenMP API's for parallelisation, and includes python tools to read/manipulate the data products produced. The code correlates binding energy sorted particle ID lists between halo catalogues, determining optimal descendant/progenitor matches using multiple snapshots, a merit function that maximises the number of shared particles using pseudo-radial moments, and a scheme for correcting halo merger tree pathologies. Focusing on VELOCIraptor catalogues for this work, we demonstrate how searching multiple snapshots spanning a dynamical time significantly reduces the number of stranded halos, those lacking a descendant or a progenitor, critically correcting poorly resolved halos. We present a new merit function that improves the distinction between primary and secondary progenitors, reducing tree pathologies. We find FOF accretion rates and merger rates show similar mass ratio dependence. The model merger rates from Poole et al, (2017) agree with the measured net growth of halos through mergers.