meeting 2025 07 09 n410 - JacobPilawa/TriaxSchwarzschild_wiki_6 GitHub Wiki
Context
Moving along with N410 now that we have N57 mostly sorted out. I'm continuing here with some additional spectra tests that we've run for N57 to verify that N410 behaves similarly.
Extending to h8
Note that up to now, I had been fitting out to h6 to check agreement with Irina. I seem to be able to reproduce her fiducial results, so extending to h8 now seems to make sense. Also note that the results below are using the symmetric binning scheme (again, so we can compare to Irina), but we have generally shifted to the asymmetric binning in the past, so we'll likely want to do that too at some point (maybe...?).
So these results are comparing our fiducial set of fits (symmetric binning out to h6) compared to some of the higher order results. These all use the Barth library ONLY.
Case
h6 vs h8
h6 vs h10
h6 vs h12
One-to-One
Differences
And here's a comparison of the additive polynomial distributions:
Case
h6 vs h8
h8 vs h10
h8 vs h12
And the fit qualities for these cases:
h6
h8
h10
h12
One interesting thing to note -- it looks like scan603.fits is once again being selected from the Barth subset as being extremely important to the fit. Here are the Barth templates and their weight for the fiducial fit
Summed
By Bin
Various Library Tests
One thing I wanted to test was the sensitivity of the kinematics to various template choices, just like we did for N57. I just want another example galaxy here that we can try to make sense of with the various template choices.
Barth vs. Full CaT
Barth vs. Trimmed 209
Barth vs. Trimmed 384 + 3
Barth vs. FGK
Superspectrum Testing
I then proceeded with the superspectrum testing that we did for N57 -- I started with the library of 384 + 3 missing Barth templates, fit the superspectrum of N410, extracted the non-zero weights, and trimmed away the strange looking templates that were picked out.
Takeaways:
In the initial superspectrum test, there are once again 11 templates that are given non-zero weight, with two of the templates being "weird looking", with one looking like a B star (scan374.fits) and the other looking like an M star (scan387.fits).
Initial Superspec Fit
Top Templates
And now comparisons of the fiducial Barth h8 fits vs. the fits using the 11/11 and 9/11 superspectra results:
Seems like I'm getting slightly different behavior here than we saw with N57;
Barth h8 vs. 11/11 Superspec
Barth h8 vs. 9/11 Superspec
Asymmetric Binning
Given how some of the results are turning out above, I wanted to turn toward the "asymmetrically binned" spectra to see if we are seeing the same behavior in the kinematics.
Extending to h8
Case
h6 vs h8
h6 vs h10
h6 vs h12
One-to-One
Differences
Superspectrum Testing
Similarly to above, I constructued the superspectrum from the asymmetrically binned spectra, fit the superspectrum with the 387 set of templates, and then compared the fiducial h8 Barth to the resulting kinematics:
Notably, the same exact set of templates get picked (11) as we had in the symmetrically binned case, including the same two "weird" spectra:
Initial Superspec Fit
Top Templates
And now comparisons of the fiducial Barth h8 fits vs. the fits using the 11/11 and 9/11 superspectra results:
Seems like I'm getting slightly different behavior here than we saw with N57;
