Full Polarimetric GNSS-R Assessment of the Freeze and Thaw States of the Terrestrial Cryosphere

The shift between frozen and thawed conditions on the Earth’s surface influences climate, hydrology, and ecology. A primary objective of the Soil Moisture Active Passive (SMAP) mission is to estimate the surface binary freeze/thaw (F/T) state for the northern hemisphere above 45° N latitude with a classification accuracy of 80% at 3-km spatial resolution and 2-day intervals. The objective was to be accomplished by the SMAP L-band radar measurements. In July 2015, the radar transmitter suffered an anomaly that prevented it from nominal operation. The mission partially met the goal exceeding the targeted 80% accuracy but at 36 km resolution. After the radar anomaly, the SMAP mission switched the radar receiver bandpass filter frequency to Global Positioning System (GPS) L2c, acting as a full-polarimetric Global Navigation Satellite System-Reflectometry (GNSS-R) receiver, known as SMAP-Reflectometer (SMAP-R). This work focuses on using SMAP-R signals to classify the F/T state over the Northern latitudes, within the limitations of the dataset. An algorithm based on the seasonal threshold approach that was originally envisioned for the SMAP radar is applied to the SMAP-R data (i.e., bistatic radar). Then the algorithm is evolved using a Random Forest algorithm to aid threshold selection from the discriminator built in the seasonal threshold approach. This algorithm is applied for years 2016 to 2022 over the Northern Hemisphere terrestrial cryosphere and shows an F/T classification accuracy agreement better than 97% with respect to the classification of the official SMAP F/T Radiometry product, proving the potential of polarimetric GNSS-R to derive F/T.