Detection of Freeze and Thaw States using Sentinel SAR Measurements and Ground Observations; A Case Study in Alaska

The objective of this study is to evaluate the potential of Sentinel-1 Synthetic Aperture Radar "SAR" data (C-band) for monitoring the states of the earth surface in terms of freeze and thaw (FT) cycles. FT cycles especially in high-latitude regions have a crucial role in many applications such as agriculture, biogeochemical transitions, hydrology and ecosystem studies. Satellite-based passive microwave observations successfully have been utilized for the freeze and thaw detections in the past. However, their coarse resolution (about 25 km) makes them less helpful for many regional and local-scale applications. However, alternatively active microwave backscatter measurements from radar sensors can provide much higher spatial resolution (about 10 m). Here, we examine C-band backscatter data from Sentinel-1A Synthetic Aperture Radar (SAR) from April 2014 to June 2018 to detect high-resolution freeze/thaw states in Alaska where has shown signs of thawing permafrost. Under cold winter conditions, much of soil water freezes, which leads to a significant decrease of the soil dielectric constant. A decrease in the radar signal of several dB, therefore, could serve as a proxy for high resolution FT states detection. The contrasts between frozen and thawed states are used to define FT states after performing several required corrections such as radiometric corrections, calibrations, and earth flattening terrain corrections. Similar studies using other sensors have tried to use average backscattering values between the winter and summer time as a threshold for defining the soil state which may not be accurate for all surfaces and soil conditions. As a more novel approach, we define the threshold values using in situ observations of soil temperature, air temperature, snow depth, and soil moisture which consider the soil conditions and land cover type. The results of this method were compared against independent ground measurements from SNOw TELemetry (SNOTEL) observations. Varying values of backscattering, when the soil temperature was around 0 ⁰C, were observed in the processed data which further analysis suggested that they might be due to the effect of snow cover and melting snow in transition season.