An Algorithm for Freeze/Thaw State Detection Using GNSS-R Reflectivity Time Series

The detection of soil freeze-thaw (F/T) states is crucial for understanding surface hydrological processes, carbon cycle dynamics, and their climatic implications. Recent advancements in spaceborne Global Navigation Satellite System Reflectometry (GNSS-R) have demonstrated significant potential for soil state classification. Surface reflectivity, derived from GNSS-R measurements, is a key parameter for distinguishing between frozen and thawed soil conditions. This study implements an edge detection algorithm to analyze reflectivity time series obtained from Cyclone Global Navigation Satellite System (CYGNSS) observations, enabling the estimation of soil F/T transition onset dates. The algorithm’s performance was validated by comparing the results with ERA5_Land surface temperature data, showing a mean absolute deviation (MAD) of 20.2 days in seasonal transition date estimates and achieving an overall detection accuracy of 88.95%. These validation results indicate strong consistency between the predicted and reference data, showing the algorithm’s efficacy in determining soil-state transitions. This work contributes to the advancement of soil F/T mapping techniques and enhances our understanding of seasonal soil-state dynamics.