meeting 2025 03 04 n315 - JacobPilawa/TriaxSchwarzschild_wiki_5 GitHub Wiki
Context
Just processing the raw kinematics we've been discussing over email into kinematic we can feed into the code. Here are some diagnositcs for these new kinematics.
Plots
As a quick reminder, we want to reminimize a set of kinematics using the Barth stars with bias set to 0.2. I'm using my verison of the kinematics (Emily and I perfectly agree in the fiducial fits) since it's a bit easier to put these in input-file form.
I've then point-symmetrized these kinematics as we have done in the past, and here's the resulting diagnostic plots:
Radial Moments
Input Maps
Cropped/Mitchell
And here are the kinematics post-symmetrization:
No Crop
Cropped
I've pulled the same plots from the paper (with the unfolded moments) for quick comparison:
Radial Moments
Maps
Strange Spectra
As we talked about briefly over email, three of the bins in both mine and Emily's reductions have sigma ~> 700 km/s, but this was not the case in my original reduction. Here are those three spectra:
Bin
Plot
148
238
247
I tried track down why these specific bins were behaving strange, and it looks like they might have a slightly different weight template weight distribution than the majority of other bins. Below I've got pie charts for each spectrum, showing the fraction of each template weight from the Barth library that contributes to this kinematic bin. For bins 148/238/247, there appears to be much more weight given to scan437.fits and scan574.fits which hardly contribute as much to other bins:
Pie Charts of Templates
And here are the templates themselves, though it's not obvious to me that there's anything different about the two "problematic" templates. I've also included a ratio of all the templates relative to the first template (scan399.fits):
Template Spectra
Ratio of Templates
Quick follow up:
It seems like the template issue is mostly resolved from the discussion with Emily. in short, it seems like the three problematic spectra have a bit of different weight distribution than the others. In particular one of the templates is being picked out by these bins which seems to have a much deeper absorption feature at ~8500 angstroms or so. As a result, the fits are placing tons of weight in this template causing the sigma to be inflated. We can discuss this in the meeting with the plots above + the refits from Emily's computer.
