The week of June 19, 2023 saw multiple major updates to the MIRI Medium Resolution Spectrometer (MRS) pipeline and the associated reference files based on calibration observations taken throughout Cycle 1. Here is information about these changes and the work that is still in progress.

The new pipeline version 1.11.0 (see release notes) implements the correction for the MRS long-wavelength count rate loss described in a previous JWST Observer article​​​​. Using a model of the rate of loss at a given wavelength as a function of time, the pipeline will now properly return spectral data in calibrated physical units regardless of the date of observation (see Figure 1). This correction will continue to be monitored throughout future observing cycles and improved as better models are developed.

This new pipeline version also implements three additional improvements for MRS data processing:

1. The outlier detection routine has been significantly revised to better identify evolving warm pixels that represent the bulk of the outliers in MRS data cubes. 
2. The 1d spectral extraction code has been modified to add an auto-centroiding feature for point sources, rather than performing it at the nominal RA/DEC coordinates of the target which could often produce incorrect extractions if the telescope attitude information was inaccurate. This auto-centroiding can be enabled by setting the optional flag spec3.extract_1d.ifu_autocen=True.
3. A new one-dimensional residual fringe correction has been added to the 1d spectral extraction code; this routine (P. Kavanagh et al.  in preparation) looks for periodic amplitude variations of known frequency and removes them from the extracted spectra. If desired, this correction can be enabled in the pipeline by setting the optional flag spec3.extract_1d.ifu_rfcorr=True.

These updates will be available in MAST data products starting in Fall 2023, but are available now for anyone re-reducing their own data (see, e.g., an MRS example pipeline notebook).

At the same time, many of the MIRI MRS reference files have also been updated to incorporate Cycle 1 revisions to the instrument flatfield, photometric response vectors, and default point source extraction widths and aperture corrections. 

A few additional photometric calibration issues remain and will continue to be worked by the MRS team. Calibration artifacts, for instance, are still present at around 24 microns (due to disagreement among the standard star observations) and at 12 microns (due to the 12.2 microns spectral leak). This spectral leak feature is the result of a few-percent leak of second-order light from 6 microns into the 12 micron bandpass (Wells et al 2015). In the previous photometric calibration, this leak was incorporated into the response curve derived from the single A star HD 163466. As a result, the leak would be removed from spectra of similarly blue sources and produce a dip-like artifact in observations of source with a redder spectral shape. In the new calibration, the standard star spectra have been corrected for the leak prior to deriving the photometric calibration vectors, so the effect will now be most pronounced for observations of blue sources (where it will manifest as excess flux around 12.2 microns) and minimal for observations of very red sources for which there is not significant 6 micron light to contaminate the 12 micron regime (see Figure 2). A future update to the pipeline will incorporate a correction for this leak for individual science observations.

If you have any questions about the MIRI MRS pipeline and calibration, please let us know via the JWST Help Desk.