Constraining the three-dimensional orbits of galaxies under ram pressure stripping with convolutional neural networks

Research output: Contribution to journalArticlepeer-review

7 Citations (Scopus)


Ram pressure stripping (RPS) of gas from disk galaxies has long been considered to play vital roles in galaxy evolution within groups and clusters. For a given density of intracluster medium (ICM) and a given velocity of a disk galaxy, RPS can be controlled by two angles (θ and ϕ) that define the angular relationship between the direction vector of the galaxy's three-dimensional (3D) motion within its host cluster and the galaxy's spin vector. We here propose a new method in which convolutional neutral networks (CNNs) are used to constrain θ and ϕ of disk galaxies under RPS. We first train a CNN by using ̃105 synthesized images of gaseous distributions of the galaxies from numerous RPS models with different θ and ϕ. We then apply the trained CNN to a new test RPS model to predict θ and ϕ. The similarity between the correct and predicted θ and ϕ is measured by cosine similarity (cos Θ) with cos Θ = 1 being perfectly accurate prediction. We show that the average cos Θ among test models is ≈0.95 (≈18° deviation), which means that θ and ϕ can be constrained by applying the CNN to the gaseous distributions. This result suggests that if the ICM is in hydrostatic equilibrium (thus not moving), the 3D orbit of a disk galaxy within its host cluster can be constrained by the spatial distribution of the gas being stripped by RPS. We discuss how this new method can be applied to H I studies of galaxies by ongoing and future large H I surveys such as the WALLABY and the SKA projects.
Original languageEnglish
Pages (from-to)1924-1937
Number of pages14
JournalMonthly Notices of the Royal Astronomical Society
Issue number2
Early online date17 Aug 2018
Publication statusPublished - 15 Feb 2019


Dive into the research topics of 'Constraining the three-dimensional orbits of galaxies under ram pressure stripping with convolutional neural networks'. Together they form a unique fingerprint.

Cite this