When the target is dispersed, it is challenging to reconstruct precisely where the target was positioned on the detector, either from blind pointing or from target acquisition. Verification images have proven very useful as a pointing verification check and as a diagnostic tool for assessing potential calibration issues. Verification images are taken after the science target has been placed at the initial pointing position for the science observation sequence, but before the filter wheel is moved to the P750L prism setting. This capability has been given extra flexibility as of APT version 2023.1.1 to make this feature maximally useful. For both slit and slitless observations, this feature is optional, but highly recommended. 

A verification image can be taken with any of the MIRI imager filters, and with custom groups/integration settings that can differ from those selected for the TA exposure. Verification images can be taken regardless of whether a TA sequence is part of the observation. 

For observations of point sources in the LRS slit, the ALONG SLIT NOD dither pattern should be specified. In this configuration, the nods will be subtracted from each other pair-wise for background subtraction purposes, before being recombined into a single spectral image in the final stage of the calibration pipeline. 

For extended sources in the LRS slit, observers are strongly encouraged to obtain dedicated background observations by defining a separate background target.  The coordinates should point to a suitable region nearby, preferably within 20", and the background observations should use the same number of groups per integration and read mode as the science observations. 

The science and background observations should be linked as a non-interruptible sequence in APT; for the science target, the corresponding background target must be specified (see Specifying APT Background Targets). This method will obtain 2 images of the science target, one in each nod position, and 2 corresponding background images. When a user assigns a background to a science target, that creates a formal  association between them. By doing this, the pipeline will automatically subtract the background exposure from the target exposure. The background exposures are co-added into a single background image, which is subtracted from each nod separately. This method is preferred, because it mitigates for bad pixels on the array (whether permanent or due to a recent cosmic ray hit).

A MAPPING dither pattern with either 2 spatial positions or 2 spectral positions could also be used, but this method is strongly discouraged, because the pipeline will have no means of identifying the background position and will therefore not subtract it. In addition, this approach does not mitigate for bad pixels. If the observer wishes to pursue this option, the extended science target will be observed in the center of one of the 2 slit positions. The background position will rotate with the position angle of the slit (at roughly 1° per day) and can only be constrained by restricting the observing time.

For slitless TSO observations, the background can be measured from the portions of the subarray not covered by the science spectrum. However, due to the extensive detector-level scattering for bright targets, it may be difficult to get a "clean" measurement of the background, and the background subtraction step may end up removing some portion of the flux from the science target. To avoid this, it is advisable to take a dedicated background exposure immediately following the science exposure. This should have the same N_groups as the science exposure, with a small number of integrations (e.g., 10), and the observation should be linked as a non-interruptible sequence. Further details are provided on the MIRI TSO Recommended Strategies page.