High Resolution Freeze and Thaw States Detection Using Combination of Sentinel 1A SAR and Passive Microwave Measurements

The freeze and thaw (FT) cycles in high-latitude regions have great impact on many biogeochemical transitions, hydrology and ecosystem especially in wetland areas. Passive and active microwave remote sensing data from satellite observations have been deployed in the past to define the status of the surface in terms of freeze and thaw. While many progresses have been made in this field, the limitations attached to such observations have hindered our ability to fully predict the change of surface state in the scale that is appropriate for the aforementioned applications. The transition between freeze and thaw states may occur frequently (even within a day) especially during shifts from cold to warm seasons and vice versa. Passive microwave sensors have different acquisition times, and data fusion of these sensors may provide a complete diurnal variation estimate of FT states. However, the coarse spatial resolution of these measurements may undermine their applicability. However, active microwave backscatter measurements from sensors such as Sentinel 1A and the Advanced Land Observing Satellite Phased Array L-Band SAR (ALOS PALSAR) can deliver high resolution information about wetlands and FT status. In this project, Synthetic Aperture Radar (SAR) c-band backscatter data from Sentinel 1 from April 2014 to June 2017 are deployed to detect high resolution freeze/thaw states and wetland areas. The contrasts between frozen and thawed seasons are used to define FT states after performing required radiometric corrections and calibrations. A method based on phase changes in polarized images is developed for different land cover types to maximize the accuracy of the detections. The aggregated (up-scaled) estimates from active measurements are compared to passive microwave-based FT product. The results of this method reveal that the estimates are relatively in good agreement with SNOw TELemetry (SNOTEL) ground measurements. Finally, a downscaling method is tried to link passive emissivity-based FT product to high resolution active FT estimates to increase the temporal frequency of the high-resolution Sentinel data. The results of this study contribute to better understanding sources of positive carbon and methane (CH4) feedback to the atmosphere.