search

spacetelescope / pysynphot

Synthetic Photometry.
http://pysynphot.readthedocs.io/en/latest/
BSD 3-Clause "New" or "Revised" License
23 stars 21 forks source link

Phoenix model spectra problematic #138

Open CBAH opened 4 years ago

CBAH commented 4 years ago

phoenix_peterplot_full

phoenix_peterplot_ha

full70008800 full70008800ha

phoenix_9000

phoenix_6200

I'm working with phoenix models, and ran a quick analysis to test their similarities to the ATLAS9 grid (Castelli & Kurucz 2004) also in Icat. At certain temperatures/metallicities there was a large discrepency, so I investigated those temperatures and it shows that something isn't right with the Phoenix grid.

Attached is a series of spectra from 7000- 8800K which shows that something isn't right with the grid at 7800 and 8000K. The two lines have much lower overall fluxes and sit virtually on top of each other and the Halpha line at 6600A is an emission line.

I talked to the author of the Phoenix grids and he made comparison plots from his side (also attached) which shows that this isn't present in the model grids.

The conditions (8000, 0.0, 4.0) are not the only conditions where there appear to be anomolous spectra, it's just a good example of it. There have been two issues raised with the Phoenix models #102 and #74, which were reported in 2018.

I'm aware that this package is no longer being actively developed, but as it stands a significant portion of the Phonix catalogue is unreliable. Thanks!

pllim commented 4 years ago

@CBAH , while the software reads the data, it is not responsible for curating it. Can you please contact the help desk at https://jwsthelp.stsci.edu ? Thanks.

MartianColonist commented 2 years ago

It appears that the reference PHOENIX models for synphot were updated in May, 2021. Perhaps this issue has now been fixed?

https://archive.stsci.edu/hlsp/reference-atlases

"Update 2021-05-27: New synphot reference files in the "synphot5" tar file and PHOENIX individual files. The models within the phoenixm00/ directory have been updated to include more data points, and fix other issues discovered with the files."

CBAH commented 2 years ago

phoenix_peterplot_updated_m00_Ha phoenix_peterplot_updated_met0 0_logg4 0 phoenix_updated_met0 5_logg4 0

Hi, Thank you for bringing me back to this because I actually thought all the issues had been resolved. The files with met = 0.0 have been updated and now reflect the models I received from the Husser grid when we contacted directly. Which is great news and has proved very helpful in my work!

However there are still issues with other metallicites, I've included a plot of one example at met=0.5 and log(g) = 4.0, where the T = 6200K line is much reduced, this is a repeat of an earlier plot, but remade using the current grid. I have also found examples at around T = 2900K.

I'm writing up results for a paper using these models and now must carefully examine to see if my results have been affected.

In any case I will email the helpdesk again in case they thought the update had fixed all the issues.

MartianColonist commented 2 years ago

Very good to know, thanks for the updated plots.

It's encouraging to see that the solar metallicity models have been fixed, and thanks in advance for reaching out to the helpdesk to ask about the other metallicities.

pllim commented 1 year ago

Hello, did you get a resolution from the help desk?

MartianColonist commented 1 year ago

As far as I'm aware, only the 0.0 metallicity PHOENIX models have been fixed. Other metallicities still have the issues described above.