I am a science fiction writer, looking for a method for constructing a realistic galactic map. I would like to know if it's possible to get the 3D density distribution of gas within a given Subhalo, such that (say, using volumetric fog) a 3D image of that galaxy could be constructed. Is it possible to compute this from data in the Illustris database?
The 3D gas density within galaxies/subhalos is, as you suspect, a direct output of the simulation. The form this data is in is somewhat complex - it could be considered a scattered/irregular point set. More accurately, it should (could) be interpreted as a Voronoi tessellation with the gas density and gradient (first derivative) defined within each control volume.
Making an image from this data is exactly what we've done in most of the movies (which show often gas temperature, or sometimes gas density). But with two differences to what you want: they are more zoomed out, showing many galaxies at once, and we typically neglect any fog/attenuation effects.
You could directly download, for a given subhalo, the gas Coordinates and Density fields, and with just these attempt some renders.
Let me know if I can help more.
Thanks for the reply.
I managed to identify those fields and am using them to place stars in my artificial galaxy.
I noticed that the gas density and coordinates data appears present only for the final snapshot of the simulation Illustris-1. I tried to grab the data for other snapshots (in the sub halo I'm using for my galaxy - 340843) and found it contained only the dark matter concentration. Is this intentional?
It is true that some subhalos will contain only e.g. dark matter.
But, subhalo ID 340843 at snapshots other than 135 has nothing to do with subhalo ID 340843 at snapshot 135 (z=0). To figure out what subhalo this guy originated from, you will need to follow the merger tree.
For example, looking at the sublink_progenitor field in the online API, the immediate prog (at snapshot 134) is 338003, which does indeed have gas, stars, etc.
I see. I'll have a look. So far the data has what I'm looking for. Thanks so much for your help and setting the simulation up.