Producing Synthetic Spectra From Simulations of Star Formation

Loke Ohlin (presented by Simon Glover)

Wednesday, Dec. 7th, 15:10CET

Linking simulations to their counterparts in observational surveys is a key aspect of validating these simulations and improving their predictive power. Such comparisons requires post-processing of the simulation to emulate observable quantities. In the context of studies of the interstellar medium, this involves reproducing the emission from chemical tracers in the gas, requiring post-processing of the simulations using radiative transfer models. We have developed GASSPY, a GPU-Accelerated Spectral Simulator in PYthon, as a tool to generate synthetic observation maps from hydrodynamical simulations. We group cells with similar physical conditions, and run an advanced chemical network for each group to produce emission and absorption spectra. These models are used as sources in a GPU accelerated radiative transfer step, where we can track thousands of spectral energy bins at a time. This allows for fast generation of maps with per pixel spectra, including line and continuum emission.