Effect of Surface Characteristics on Sentinel-1 SAR Observations for Detecting Freeze and Thaw States in Boreal Regions

More than a third of the global land surfaces undergo freeze and thaw (FT) cycles that have wide range of implications in hydrology, carbon cycle, climate, ecological, and environmental studies. This is in particular more pronounced in boreal regions with significant existence of wetland surfaces. Passive and active microwave observations have shown to be helpful in detecting FT and wetland surfaces. In particular, Sentinel-1 A/B SAR observations have the advantages of having high spatial resolution and being free of atmospheric effect. Various models and analysis have been proposed to define the threshold needed to detect FT for different land-cover types, which generally rely on the contrasts between summer and winter back-scatterings values. While defining the state of freeze and thaw may not be very challenging for summer and winter times, defining the transition times and onset between these states still requires more investigations. Here we conduct a thorough analysis of SAR data from Sentinel-1 to define the sensitivities of observations to several physical conditions such as moisture and water content, vegetation type and elevation. A wealth of ground based observations such as temperature, land-cover type, vegetation and snow depth are used to characterize the opportunities and limitation of SAR data when is used for FT and wetland detections. The results reveal that a combination of elevation, water presence, and vegetation conditions should be considered for accurately detect FT. The behavior SAR signal changes could be misleading especially in transition times and previous simple contrasts between summer and winter may not be sufficient. Different patterns were seen when transitions from freeze to thaw happen in low land and high land areas even with similar surface conditions such as land cover type. This is the first study that has comprehensively investigated four years of SAR data from Sentinel-1 in cold regions based on a rigorous set of in situ data.