Calibration of Airborne Multiangle SpectroPolarimetric Imager (AirMSPI) polarization measurements, and in-flight comparisons with the Research Scanning Polarimeter (RSP) and the Spectropolarimeter for Planetary EXploration (SPEX)

The Airborne Multiangle SpectroPolarimetric Imager (AirMSPI) is a remote sensing instrument for the characterization of atmospheric aerosols and clouds. AirMSPI, flying onboard the NASA ER-2 aircraft at 20 km altitude, participates in field campaigns since 2013, including ORACLES (2016). The pushbroom camera is mounted on a programmable, motorized gimbal for multi-angle observations at 10x10 m2 resolution. Eight spectral bands within 355-935 nm are recorded, 3 of which also measure linear polarization. Photoelastic modulators (PEMs) encode the polarized and total intensities in each polarimetric pixel as the amplitude and offset of a modulated intensity pattern, such that the ratio of the two is insensitive to pixel-to-pixel differences. This technique, developed to enable the high-accuracy imaging polarimetry required for aerosol species discrimination, will also be applied in the Multi-Angle Imager for Aerosols (MAIA) satellite instrument. We present the calibration and accuracy validation of AirMSPI polarization measurements. The main calibration, describing the instrument's response to any degree (DoLP) and angle of linear polarization, is performed in the lab using a recently updated, carefully designed and characterized polarization state generator (PSG-2). Validation measurements using an independent polarimeter show agreement in DoLP to within 0.001 for several DoLPs across the 0-1 range. The PEMs' retardances and phases, which are different and not necessarily stable in flight, are extracted from measurements of the on-board validator, a partially polarized light source located inside the instrument housing, which is viewed before and after each target. Although this calibration does not rely on the validator's DoLP, and the validator was not designed for DoLP calibration, the frequent measurements of its DoLP provide an upper limit for AirMSPI's in-flight polarimetric stability, which is 0.001. A correction for the actual PEM retardances and phases in the science data is finally extracted from the data itself. The aforementioned calibration techniques were recently implemented in the data processing pipeline. We present in-flight polarimetric comparisons against the RSP and SPEX instruments, and comparisons of aerosol retrievals against collocated AERONET stations.