THE BRIGHTEST SIDE OF ARCTAS: USING SUNGLINT AND POLARIZATION IN RETRIEVALS OF AEROSOL PROPERTIES

The Research Scanning Polarimeter (RSP), a prototype for the Aerosol Polarimetry Sensor on the upcoming Glory NASA mission, flew on board the NASA LaRC B200 aircraft together with the High Spectral Resolution Lidar (HSRL) during the summer deployment of the Arctic Research of the Composition of the Troposphere from Aircraft and Satellites (ARCTAS). While scouting for wildfire smoke plumes, flight transects over water bodies occasionally revealed the strong signal of sunglint. In the context of retrieval of aerosol microphysical properties, this surface signature provides a valuable cross-check on the robustness of aerosol retrievals over land whenever the assumption holds that the aerosol properties do not appreciably vary across the boundary between water and land. The Cox-Munk parameterization may not be appropriate for those inland and coastal waters where the conditions for specular reflection are determined by the bathymetry and the currents rather than the windspeed. The RSP-observed surface reflectance in the visible range can nonetheless be modeled using the 2.2um channel information to characterize the distribution of wave slopes of the water surface. In this paper we evaluate the use of sunglint radiances as a means for characterizing total extinction while off-glint measurements are used to characterize scattering. The various corrections and limitations of the sun-glint technique for small inland water bodies, for which adjacency effects can be important, are discussed. We compare the absorption estimates obtained using the sunglint technique with those obtained using only polarimetric measurements over nearby land surfaces. The two remote sensing methods are then evaluated using the measurements of total aerosol optical depth (AOD), aerosol size distribution and absorption provided by the AATS-14 and the HiGEAR teams on the NASA P3 aircraft, together with more spatially and temporally extensive comparisons with the HSRL AOD and intensive parameter (backscatter ratio, backscatter Angstrom exponent) estimates.