meeting 2024 05 02 n57 - JacobPilawa/TriaxSchwarzschild_wiki_5 GitHub Wiki

Context

Some plots and diagnositcs that hopefully will finalize N57!
First up, I switched from rho0 to M15 for our cornerplot, and here's how that looks:

Cornerplot
images/240430/240430_M15-1.png

First up, here are some additional plots I made to potentially end up in the appendix/for internal sanity checks. Note that the location on the M-sigma relation was done by eye/just for a general idea of where it lands.

I also tried to re-run the V(R, Theta) fitting routine and clean up the results from that. Here's two options, one which excludes the outer 4 Mitchell bins and one that includes them in the fit.
- Note that I pulled the major/minor photo axes from Goullaud+2018 with webplot digitizer.

Without Outer 4	With Outer 4
[images/240430/without_outer_bins.png]]](/JacobPilawa/TriaxSchwarzschild_wiki_5/wiki/[[images/240430/with_outer_bins.png)
[images/240430/without_outer_bins-1.png]]](/JacobPilawa/TriaxSchwarzschild_wiki_5/wiki/[[images/240430/with_outer_bins-1.png)

I've also plotted Matthew's CFHT fits on top of the MGE surface brightness fit for comparsion. Here's Matthew's full plot:

CFHT
images/240430/NGC_0057_profile.png

No Zoom	Zoomed
[images/240430/MGE_sersic_comp.png]]](/JacobPilawa/TriaxSchwarzschild_wiki_5/wiki/[[images/240430/MGE_sersic_comp_zoom.png)

I tried to run kinemetry on the GMOS/Mitchell data individually to (a) reproduce roughly what Irina had quoted in her papers, and (b) to see if we can make any more sense of the kinematic PA's we are getting from our V(R, Theta) plot.
I'm using the "pafit" Capellari python package to do this, which is quite straightforward. I think (?) this is what Irina must have used, or potentially the IDL version, because I can reproduce her results on BOX with the correct set of inputs.
- This code takes in a set of (x,y,V,dV) and generates a bisymmetric model map for a bunch of different PA's. The code compares the data to the bisymmetric map to find the angle which best minimizes the difference between the two maps.

I first started with some of the files on BOX (https://berkeley.app.box.com/folder/11161021148) to make sure I could reproduce what Irina had.
In MASSIVE X/XIV (https://arxiv.org/pdf/2001.11046), she says that there was no Mitchell detection for the kinematic axis, and her diagnostics plots on BOX are:
- Note: I am taking the "no detection" to mean an uncertainty of essentially 90 degrees

Plot 1	Plot2
[images/240430/NGC0057_kinemetry-1.png]]](/JacobPilawa/TriaxSchwarzschild_wiki_5/wiki/[[images/240430/NGC0057_kinemetry-2.png)

I re-ran the Python routine with her exact inputs and obtained:
- Kin PA: 151.0 +/- 89.8 (3*sigma error); Velocity Offset: -1.31
- Compared to her values of: 151.06+/-89.55; Velocity offset: -1.32
And the resulting plot generated from the routine is:

Irina Reproduction
images/240430/IRINA_MITCHELL.png

I went ahead and ran the routine on my GMOS data (which I fit with pPXF) and on my Mitchell data (unfolded data, also with three bins removed relative to what Irina had because of overlap with GMOS data)
When running on GMOS alone, I find:
- Kin PA: 109.5 +/- 89.8 (3*sigma error); Velocity Offset: 0.23
  - Note that my value agrees quite well with hers, but I need to look a bit into how the uncertainty is computed here/how to get the code to compute 1 sigma uncertainity for a more fair comparison. At the very least, the point estimates are quite similar to one another.
- Irina quotes: 100 ± 22 for the 1 sigma error

GMOS Alone
images/240430/GMOS_Kinemtry.png

And doing the same on Mitchell alone, I find:
- Kin PA: 146.5 +/- 89.8 (3*sigma error); Velocity Offset: 0.00
- Note that this is really similar to Irina's values on BOX, but are not quoted in the paper beacuse I think of the +/- 90 degree difference.

GMOS Alone
images/240430/MITCHELL_Kinemtry.png