Lake Ice Surface Type Correlation with GNSS Surface Reflectivity

We present estimates and uncertainties using Global Navigation Satellite System reflectometry (GNSS-R) of surface reflectivity (SR) at L-band for a set of basic lake surface types - snow-covered ice and mixed ice and water - based on GNSS signals reflected from the surface with a dedicated downward-pointing GNSS antenna. The surface reflectivity, assumed to be from the first Fresnel zone of the L1 signal, is compared to camera images of the surface mapped to their location on the surface using lidar range measurements. The mean red value (MRV) of camera pixels within the Fresnel zone is used as a truth reference to distinguish surface ice from surface water and, separately, to distinguish bare surface ice from snow-covered ice. For the mixed ice and water surface, the surface reflectivity of three Global Positioning System satellites' L1 signals are examined from data collected on 14 Feb 2020 at Lake Michigan in Chicago. Two of the three satellites show moderate positive correlation, which agrees with prior work, and the third shows a correlation coefficient of 0.2. For comparing snow-covered ice to bare lake ice, data from 19 Feb 2019 is analyzed. While we anticipate that the snow having higher albedo than the bare lake ice, but also being rougher, should lead to a high MRV and low GNSS-R SR for snow, and thus negative SR-MRV correlation, our two SR versus MRV data sets show opposite correlations. GNSS-R snow versus ice sensing requires further investigation.