Reducing Europa #1250 NIRSpec IFU data

[YoffeG]

Hey guys,

I'm attempting to reduce NIRSpec IFU data of Europa taken for observation ID 1250. I'm using the newest .pmap and it seems like everything is in order.
For every band, I concentrated the _uncal.fits files in the stage0 folder, and it seems to have combined the two dithers in the final reduced cube.
But compared to e.g., Trumbo+23, the resulting spectra seem considerably more noisy. I'm wondering if there's any tips you could direct me to concerning the reduction process? My initial guess is that, of course, the issue is on my end.

Many thanks in advance,

Gidi

[ortk95]

Hi!

Trumbo+2023 used a custom data reduction pipeline, which may explain some of the discrepancy you're seeing. The supplementary material for Trumbo+2023 talks about their data processing, and in particular some of the techniques used to correct 1/f noise.

I know a 1/f noise correction has since been incorporated into the main JWST pipeline as the nsclean step - https://jwst-pipeline.readthedocs.io/en/stable/jwst/nsclean/main.html. You can enable this step using our pipeline code by running something like:

python3 nirspec_pipeline.py /path/to/data --kwargs '{"stage2": {"steps": {"nsclean": {"skip": false}}}}'

The documentation implies this nsclean step is being superseded by the clean_flicker_noise step, so you may want to experiment with using this instead (e.g. with --kwargs '{"stage1": {"steps": {"clean_flicker_noise": {"skip": false}}}}').

A few other things that may be useful to try:

• The pipeline should produce individual dithers (e.g. stage3/d1/...), as well as dither combined data files (stage3/combined/...). In some situations, the JWST pipeline can have trouble with processing data for moving targets (although I think this has improved in the last year or so), so it may be more useful to work with the individual dithers rather than the combined files.
• The SNR obviously improves the more spectra you average. I think the Trumbo paper averaged over multiple spaxels to produce their plotted spectra, so the noise in these spectra will be a bit lower than the noise in an individual spaxel.

Hope this helps, and let us know if you have any more questions!

Oliver