Stellar models do not extend to MIRI wavelengths

[liztino]

Hello, I am using BEAST to generate ASTs for JWST MIRI imaging observations.

While MIRI filters are implemented in BEAST, the stellar models in the physics grids do not extend to MIRI wavelengths. During the create_physicsmodel step, this results in NaNs in the wavelength field of the output SEDs for MIRI bands. From inspecting the spectra files, it appears the stellar models only extend to 55444 Angstroms, which is insufficient to cover the full wavelength range of the MIRI filters.

Extended the model grids to MIRI wavelengths is therefore critical, since multi-filter ASTs are not easily generated without the BEAST.

[karllark]

Related to issue #180

[karllark]

All stellar models would need to be updated. Focus on BOSZ (LTE) and TLUSTY (nLTE). Would probably want to update Aringer as well - or at least make sure it extends to 30 micron.

[liztino]

I took a closer look at the kurucz2004.grid.fits and tlusty.lowres.grid.fits files in the .beast directory to determine whether the issue originates from the grids themselves or from sampling within create_physicsmodel. The files are a bit difficult to parse, but I am assuming that the first HDU contains the model spectra and the last row corresponds to the wavelength in Angstroms.

Assuming this interpretation is correct, I plotted the spectrum for the first model in both the TLUSTY and Kurucz2004 grids (see attachments). The sampling looks good up to about 10 microns, but then falls off with only 8 samples extrapolated to about 100 micron. This means that while the grids do extend further than the 55444 Angstroms suggested from the spectrum generated by create_physicsmodel, they are not well sampled enough to cover most of the MIRI filters (except perhaps F560W and F770W).

Therefore, I agree, @karllark, that ingesting the new models is likely the way to go.

[karllark]

Thanks Liz. Yes the x-axis would be angstroms.