Account for MIRI trace foldover <4.5 microns

[jaymedina]

MIRI spectra folds < ~4.5 microns, so we should make new simulations to account for this.

" The dispersion profile however folds over below 4.5 µm (where the prism throughput is very low), superimposing two parts of the spectrum onto each other. A dedicated filter is mounted over the slit to block radiation shortward of 4.5 µm to avoid this contamination in the slit. This effect is not mitigated for LRS in slitless mode, causing some spectral contamination at the shortest wavelengths. "
Source: https://jwst-docs.stsci.edu/mid-infrared-instrument/miri-observing-modes/miri-low-resolution-spectroscopy

[skendrew]

just to summarise here also my comments from the meeting notes in confluence:
this is not currently in Pandeia so there is no easy way to incorporate it is important and definitely will be measured in more detail once were in flight with data from MIRI commissioning. The spectral foldover is mitigated by the very steep drop in transmission of the prism below 5 µm (though there is a small bump in the transmission around 3.8 micron) - so yes, the wavelengths fold over but not a lot of flux is getting through. It's very hard to characterise as the turnover in dispersion means that there is a "pile-up" of flux so you need a very good model of the source.

This will basically be captured in our spectral response function, for which we already have separate slit and slitless versions. Right now we have only 1 wavelength calibration reference file, and I think that will be okay going forward as well (but with pixel values for shorter wavelengths than are currently populated). I don't think any significant changes in format will be required, the pipeline will just have to be able to deal with a pixel having multiple wavelength values in the wcs (which may not be trivial?).