Speaker
Description
The evolution of galaxies is dependent on their present-day cosmic environment; whether the galaxies are isolated, or live in dense regions such as galaxy clusters. However, their evolution also depends on the environments they have experienced in the past. Backsplash galaxies are a key example of this -- galaxies that have previously passed through the centre of a galaxy cluster, but now reside in the cluster outskirts. These galaxies cannot easily be distinguished from those infalling for the first time, and so it is difficult to know whether to attribute galaxy properties in the cluster outskirts to the cluster itself, or to pre-processing en-route to the cluster.
Using The300 Project, a suite of hydrodynamical simulations of 324 galaxy clusters, we compare the properties of backsplash galaxies to those approaching a cluster for the first time. We develop a machine learning model, trained on these simulations, which is able to distinguish between backsplash and infalling galaxies based only on their present-day properties, with an accuracy of up to 90%. Crucially, this model only uses observationally measurable galaxy properties, such as their line-of-sight velocities and stellar masses, meaning it can be easily applied to real observations of galaxy cluster members to build pure samples of backsplash and infalling galaxies. This tool can therefore be used to disentangle different environmental effects, by better constraining the environmental histories of cluster member galaxies.
