| Presentation Title: | A Probabilistic Calibration Procedure for the CORSAIR Polarimeter |
| Author(s): | Alan Hsu, Jenna Samra, Steven Tomczyk |
Abstract:
The Coronal Spectropolarimeter for Airborne Infrared Research (CORSAIR) is a balloon-based coronagraph that will make spectropolarimetric observations of the 3.9 micron Si IX emission line. Its 2025 commissioning flight will be followed by a long-duration Antarctic science mission in 2028. CORSAIR is a pathfinder for making routine, continuous, high-sensitivity measurements of the global coronal magnetic field from high altitudes. CORSAIR uses a rotating-waveplate polarimeter to measure the full Stokes polarization state of the emission line, which includes the Stokes Q, U signals that encode the plane-of-sky direction of the magnetic field, and the Stokes V signal that encodes the line-of-sight magnetic field strength. To accurately measure the faint polarization signals, however, requires a comprehensive understanding of the systematic and random errors in our polarimeter. We present a novel, robust calibration routine that computes the posterior distribution of systematic random variables in our polarimeter, which subsequently allows us to compute downstream distributions of the demodulation matrix and thus Stokes Q, U, V from observations. In our setup, we use calibration optics to generate representative Stokes inputs that are fed into the polarimeter, which are subsequently modulated by a rotating waveplate. We build upon the calibration procedure used in the Upgraded Coronal Multi-channel Polarimeter (UCoMP) by extending the theoretical model to account for more systematic errors, and utilize the calibration data to compute posterior distributions instead of point-estimates of these errors. This allows us to perform Bayesian inference on the Stokes signals and propagate uncertainties in an end-to-end manner.