Verification of Downscaling Method for Near-Surface Freeze/Thaw State Monitoring in Genhe Area of China

摘要

The high-resolution freeze/thaw (F/T) monitoring plays an important role in studying carbon-nitrogen cycle, soil erosion and climate change in Genhe, China. In this paper, high-resolution downscaled land surface temperature (LST) retrieved from AMSR2 [1] were obtained from previous study [1]. And then were used to downscale the passive microwave (PMW) brightness temperature (TB) from 0.25° to 0.01° through downscaling method of PMW TB [2]. Finally, the downscaled TB data and F/T discriminant function algorithms [3], [4] were adopted to discriminate the surface freeze/thaw status. A comparison between high-resolution F/T state and soil temperature measured at 0~5 cm over Genhe area turned out that the F/T discriminant function algorithm [3] has a total classification accuracy higher than and 70%, and the improved F/T discriminant function algorithm [4] has a total classification accuracy higher than and 60%. From the perspective of orbit, both algorithms had freezing distinguished accuracy high than 90% at ascending and descending orbits. At last, we discussed and analyzed the possible problems of F/T discriminant function algorithms and downscaling method of TB.