Modeling polarized solar radiation from a snow surface for correction of polarization-induced error in satellite data

• Published in: Journal of quantitative spectroscopy & radiative transfer 2019-01, Vol.222-223, p.154-169
• Main Authors: Sun, Wenbo, Wielicki, Bruce A., Baize, Rosemary R., Lukashin, Constantine, Hu, Yongxiang, Zubko, Evgenij, Videen, Gorden, Kim, Sungsoo S., Choi, Young-Jun
• Format: Article
• Language: English
• Subjects: Measurement error; Polarization of light; Polarization sensitivity; Radiative transfer model; Snow reflection; Solar radiation
• ISSN: 0022-4073; 1879-1352
• DOI: 10.1016/j.jqsrt.2018.10.011

•Reflected solar radiation is strongly polarized by some snow surfaces in NIR.•We model polarized reflected solar radiation from snow surfaces.•We provide comparisons with PARASOL data. Snow surfaces can be composed of ice crystals, melt-form snow or melt-freeze snow. Solar radiation scattered by such surfaces is polarized and the amount of polarization depends on the surface composition. This can be a source of measurement errors in satellite data if a non-polarimetric radiometric sensor is sensitive to the polarization state of light. To obtain highly accurate spectral solar radiation data from the Earth-atmosphere system for the space-borne inter-calibration studies as proposed in NASA's Climate Absolute Radiance and Refractivity Observatory (CLARREO) mission and the CLARREO Pathfinder (CPF) mission, the spectral polarization state of the reflected solar light at the top of atmosphere (TOA) must be known with sufficient accuracy. The degree of polarization (DOP) and the angle of linear polarization (AOLP) of the light at the TOA as functions of incident and viewing geometry and scene type construct the Polarization Distribution Models (PDMs) for correction of polarization-induced error of satellite data [3]. In this work, algorithms for modeling the spectral polarization state of reflected sunlight from various types of snow, including particulate snow, melt-form snow, and melt-freeze snow crusts are developed. A particulate snow surface model based on a mixture of spherical and hexagonal-column particle shapes is used. A rough-surface ice/water facet model is introduced to approximate the melt-form snow and melt-freeze snow crusts. The area fraction of facets for typical (mostly particulate) snow, and for melt-form and melt-freeze snow crusts, is estimated by comparing the model results with the PARASOL satellite data and with ground measurements conducted in Hokkaido, Japan, and in the northwest Greenland ice sheet. Our numerical results demonstrate that the model can provide a reliable approach for making the spectral PDMs for wavelengths between 320 and 2300 nm for satellite inter-calibration applications as proposed in the CLARREO and the CLARREO CPF missions for all kinds of snow surfaces.