Authors: Alan Hsu (Harvard University), Jenna Samra (Smithsonian Astrophysical Observatory), Steven Tomczyk (Solar Scientific LLC), Daniel Durusky (Smithsonian Astrophysical Observatory), Jacob Hawthorne (Smithsonian Astrophysical Observatory), Vanessa Marquez (Smithsonian Astrophysical Observatory), Abigail Unger (Smithsonian Astrophysical Observatory)

We show preliminary results for a probabilistic calibration of the CORSAIR polarimeter. CORSAIR is a new balloon-borne coronagraph and spectropolarimeter that will provide the first uninterrupted observations of the global coronal magnetic field over one month-long solar rotation. We calibrate the polarimeter using a newly developed probabilistic calibration scheme that quantifies the calibration uncertainty explicitly. First, we place a calibration unit in front of the polarimeter to generate known Stokes inputs. Using an infrared source and detector, we project light through the system and record the observed intensities at different modulation states. We model our polarimeter using a probabilistic Mueller matrix framework and initialize prior distributions on free parameters centered on values using point-estimate methods employed by existing polarimeters. We then estimate the posterior distribution of the free parameters, effectively computing the calibration uncertainty in the form of a distribution shift on our initial estimates.