fermiPy / fermipy

Fermi-LAT Python Analysis Framework
http://fermipy.readthedocs.org/
BSD 3-Clause "New" or "Revised" License
51 stars 53 forks source link

SED fit error regions too small to be right #573

Open ranieremenezes opened 7 months ago

ranieremenezes commented 7 months ago

Hello,

The fit uncertainty region saved in the TARGET_NAME_sed.fits by the GTAnalysis function sed() is often unreasonable small, as shown in the figure below (it looks like a single solid line, but there are three lines in total, i.e. one line for the fit, and two delimiting the uncertainty region).

BLLac_SED_2months

This analysis was done for BL Lac with the following configuration:

emin : 100.0 emax : 200000.0 zmax : 90 evclass : 128 evtype : 3 ra: 330.68038041666665 dec: 42.277771944444446 tmin: 239557417 tmax: 245691781

The same issue frequently happens with other data sets and spectral models, even if the fit quality returned by GTAnalysis.fit() is 3 (as is the case for the plot shown above).