The MIRI MRS and Imager exhibit a gradual, time-dependent reduction in count rates produced by fixed-brightness targets, primarily impacting long-wavelength data. The JWST calibration pipeline and exposure time calculator (ETC) correct for this reduction in count rate using a parametric model derived from regular monitoring observations.

The time baseline of these observations is now long enough to demonstrate that the exponential model previously implemented in the pipeline in 2023 is not the best representation of the late-time behavior of the data. For the MIRI MRS, observations in Channel 4C in 2025 experience a few percent greater loss than predicted by the previous model. The MIRI imager is less affected by the count rate loss than the MRS, but shows qualitatively similar declines relative to the previous exponential model.

Updated models of the count rate decrease for both the MIRI MRS (Figure 1) and Imager reproduce the long-term trends more effectively, allowing for better extrapolation of the predicted decrease in future observing cycles. These new models will be incorporated into upcoming Build 12.1 of the JWST calibration pipeline (expected to be released in November 2025) and Cycle 6 of the JWST ETC. The difference between the old and new models suggests that SNR estimates provided by the Cycle 5 ETC will be too high in Channel 4C by about 3-5% in the faint-source regime and about 13% in the bright-source regime.  Differences for all other channels are expected to be less significant and fall within the typical range of uncertainty.

The MIRI team will continue to monitor the count rate loss via regular observations and make further updates as necessary. More information on the latest models can be found on the MIRI Calibration Status pages on JDox. Observers are encouraged to contact the JWST Help Desk with any questions.