Constructing merger trees of density peaks using phase-space watershed segmentation algorithm

ABSTRACT Structure identification in cosmological simulations plays an important role in analysing simulation outputs. The definition of these structures directly impacts the inferred properties derived from these simulations. This paper proposes a more straightforward definition and model of structure by focusing on density peaks rather than haloes and clumps. It introduces a new watershed algorithm that uses phase-space analysis to identify structures, especially in complex environments where traditional methods may struggle due to spatially overlapping structures. Additionally, a merger tree code is introduced to track density peaks across time-steps, making use of the boosted potential for identifying the most bound particles for each peak.