Why the same cut for TNG50, TNG100, TNG300 give similar catalog sizes?
Hangci Du
20 Jan '22
Here I apply the same cuts on TNG50, TNG100 and TNG300 (directly click the link to the webpage of Subhalo Search), and I always make sure the stellar mass of the resulted subhalo is large.
What confused me is that the resulted catalog size is similar: 9 for TNG50, 7 for TNG100, 10 for TNG300, but as I simply thought, there should be ~70 for TNG100, ~2000 for TNG300. Can anyone help to explain this?
Best.
Dylan Nelson
20 Jan '22
Dear Hangci,
If you made a simpler selection, you would see what you expect. For instance, if you keep only the total mass criterion, then we have 97, 912, and 20114 subhalos in TNG50-1, TNG100-1, and TNG300-1, respectively. These ratios correspond roughly to the volume ratios of the different boxes.
But when you add cuts on stellar mass, or SFR, and especially both, you are selecting against galaxies properties which are not entirely converged with resolution. I suspect that in one of the simulations your cuts select a portion of the SFMS (star forming main sequence), or similar? But this will shift slightly with resolution, e.g. stellar masses are about 1.4x higher in TNG100-1 versus TNG300-1 (see discussion in Pillepich+18 Appendix). I would suggest to check the M*-SFR plane in each of the three runs (e.g. with the Catalog Plotting tool). You could try shifting your search bounds on these parameters along with the general shift with numerical resolution between the runs. This would certainly result in more comparable galaxy samples between the three simulations.
Here I apply the same cuts on TNG50, TNG100 and TNG300 (directly click the link to the webpage of Subhalo Search), and I always make sure the stellar mass of the resulted subhalo is large.
What confused me is that the resulted catalog size is similar: 9 for TNG50, 7 for TNG100, 10 for TNG300, but as I simply thought, there should be ~70 for TNG100, ~2000 for TNG300. Can anyone help to explain this?
Best.
Dear Hangci,
If you made a simpler selection, you would see what you expect. For instance, if you keep only the total mass criterion, then we have 97, 912, and 20114 subhalos in TNG50-1, TNG100-1, and TNG300-1, respectively. These ratios correspond roughly to the volume ratios of the different boxes.
But when you add cuts on stellar mass, or SFR, and especially both, you are selecting against galaxies properties which are not entirely converged with resolution. I suspect that in one of the simulations your cuts select a portion of the SFMS (star forming main sequence), or similar? But this will shift slightly with resolution, e.g. stellar masses are about 1.4x higher in TNG100-1 versus TNG300-1 (see discussion in Pillepich+18 Appendix). I would suggest to check the M*-SFR plane in each of the three runs (e.g. with the Catalog Plotting tool). You could try shifting your search bounds on these parameters along with the general shift with numerical resolution between the runs. This would certainly result in more comparable galaxy samples between the three simulations.