Spectral model vs points mismatch for HESS J1731-347, 2011A&A...531A..81H

[cdeil]

I noticed that for HESS J1731-347, 2011A&A...531A..81H there's a small mismatch between the spectral model and points we have in gamma-cat:
https://github.com/cdeil/gamma-cat-status/blob/master/sed_png/source_000103.png
(the spectral model is ~ 10% too high or the points ~ 10% too low; e.g. see the first points from the left in the residuals in Fig 4 in the paper and in this plot)

Resolving this and getting both correct is high priority, because this is an "extern" source in HGPS, i.e. we copy the spectral model and points over into the HGPS catalog from gamma-cat.

The current spectral parameters are here, they match the ones given in Table 1 of the paper:
https://github.com/gammapy/gamma-cat/blob/master/input/data/2011/2011A%2526A...531A..81H/tev-000103.yaml

The current spectral points are
https://github.com/gammapy/gamma-cat/blob/master/input/data/2011/2011A%2526A...531A..81H/tev-000103-sed.ecsv
again, as far as I can see they match what's given here:
https://www.mpi-hd.mpg.de/hfm/HESS/pages/publications/auxiliary/AA531_A81.html

@facero as corresponding author: could you please have a look? What's wrong?

[facero]

Just to be sure I understand, you mean that by taking the spectral parameters from the paper in Table 1 and overlaying the data points from the original paper they don't match by 10% is that right ?

[cdeil]

@facero - Yes. See this plot: https://github.com/cdeil/gamma-cat-status/blob/master/sed_png/source_000103.png

[cdeil]

Daniel Gotschall (@gottschall) just commented by email that they also noticed this issue concerning the HESS J1731 flux in Tübingen. He was wondering if it could be a typo in the paper, specifically for a norm of 3.67e-12 instead of the stated 4.67e-12 the spectral model would appear compatible with the published spectral points.

@facero - We'll circulate the HGPS catalog and paper again in HESS this Friday. If you do have time to have a look into this issue (e.g. compare to your old analysis files or HESS-internal notes on wiki page or Confluence or slides to figure that mismatch out, that would be very helpful for HGPS!

[facero]

I did some archeology in my J1731 directory and found some of the old ROOT scripts generated by the Model++ framework. The best-fit parameter seems to support the typo hypothesis.
By inspecting the ROOT script I found this
TText text = pt->AddText("* SpectrumPowerLaw **");
text = pt->AddText("#Phi (243.594 GeV) = (5.951e+01 #pm 4.887e+00) 10^{-12} cm^{-2}.s^{-1}.TeV^{-1}");
text = pt->AddText("#Phi (810.395 GeV) = (3.669e+00 #pm 1.747e-01) 10^{-12} cm^{-2}.s^{-1}.TeV^{-1}");
text = pt->AddText("Norm: (3.669 #pm 0.175) 10^{-8} m^{-2}.s^{-1}.TeV^{-1}");
text = pt->AddText("Index: (2.318 #pm 0.056) ");
text = pt->AddText("Decorrelation energy = 0.810 TeV");
text = pt->AddText("I(>244 GeV) = (1.100e+00 #pm 5.900e-02).10^{-11} cm^{-2}.s^{-1}");
text = pt->AddText("I(>810 GeV) = (2.256e-01 #pm 1.436e-02).10^{-11} cm^{-2}.s^{-1}");
text = pt->AddText("F_{E}(>244 GeV) = (1.110e+00 #pm 5.956e-02).10^{-11} TeV cm^{-2}.s^{-1}");
text = pt->AddText("F_{E}(>810 GeV) = (7.577e-01 #pm 4.823e-02).10^{-11} TeV cm^{-2}.s^{-1}");

This is probably not the final analysis but the numbers are pretty close to the published value.
There is indeed a typo in Table 1. Thank you for spotting it after many years ;-)