Timing of detector-specific beam-matching convolution within the PPI mapmaking pipeline

Determine whether the detector-specific beam-matching convolution step in the PRIMAger Polarimetric Imager (PPI) mapmaking pipeline can be deferred until after destriping-based baseline estimation is completed, or whether this convolution must be performed during each iteration of the iterative (I, Q, U) and baseline solution to adequately mitigate intensity-to-polarization leakage from beam-shape mismatches.

Background

Beam-shape variations across detectors in the PRIMAger Polarimetric Imager (PPI) can convert total intensity structure into spurious Stokes Q and U signals. To mitigate this, the authors propose a science-pipeline step that smooths each detector’s data with a customized convolution kernel to better match beams before forming polarization maps.

The PPI mapmaking algorithm iteratively solves for Stokes (I, Q, U) given assumed baselines, and then solves for baselines given the current maps (destriping). The open question concerns when to apply the detector-specific beam-matching convolution: either once after destriping prior to the final Stokes solution, or within every iteration of the joint baseline and Stokes reconstruction. Resolving this requires improved beam knowledge and further simulation.