An adiabatic Galactic wind simulation
The first CGOLS simulation used a spherical injection model to pump mass and energy into the central 300 pc of the disk with rates that were set to test both a high mass-loading and low mass-loading state. The gas was not allowed to cool radiatively, so this is essentially a numerical test of the classic Chevlier & Clegg analytic outflow model, but with a disk.
A radiative galactic wind simulation
The second CGOLS simulation was identical to the first, but with radiative cooling allowed. This model numerically validated the analytic model of Thompson et al. 2016. At high mass-loading rates, a cooling radius develops in the hot wind within a few kpc of the disk.
Centrally clustered star formation
The third CGOLS simulation used the same tuned mass and energy input rates, but this time distributed the feedback amongst 8 “clusters” placed within the central kiloparsec of the disk. At early times, the mass input rate is set to be 40 percent of the total assumed star formation rate (SFR), and the result is a hot outflow that is mass-loaded enough to cool radiatively on large scales. At later times, the mass input rate is only 10 percent of the SFR, and the hot phase stays hot at all radii. These are the same rates as were used in the previous simulation with spherically symmetric feedback, but as the videos show, the results are quite different.
Centrally clustered star formation with realistic mass and energy input rates
This CGOLS simulation extends the earlier work by using stellar population synthesis models to generate time-dependent mass and energy input rates for a few massive star clusters centered within R = 1 kpc. Each cluster represents the collective input of thousands of supernovae, and is designed to be massive enough break out of the disk and drive a hot wind. As more gas is transferred from the disk into the outflow by the interaction with these clusters, a multiphase outflow develops.