Abstract:Global Navigation Satellite System-interferometric reflectometry (GNSS-IR) has become an important means to monitor soil moisture content (SMC). It works based on the assumption that the detrended signal-to-noise ratio (DSNR) sequence has a cosine waveform. However, when the signal reflection ground is undulating, its pattern may deviate from the assumed waveform, eventually influencing the SMC retrieval. To resolve this limitation, we proposed an arc editing method whose basic idea is to edit the DSNR sequence and keep only the DSNR data whose interference pattern is a typical cosine waveform. We tested the arc editing method in comparison with the conventional method in SMC retrieval using 4-year DSNR data from 4 simulated satellite arcs and 3-year DSNR data from 3 GNSS stations deployed in undulating terrains. The simulation results show that the new method eliminated the multi-dominant frequencies appeared in the DSNR sequences and improved the dominant frequency power, resulting in better SMC retrieval accuracy than the conventional method. When comparing the retrieval results to in situ probe-measured SMC data, the SMC retrieved by the new method has an average correlation coefficient of 0.758 and an average root mean square error (RMSE) of 0.045 cm3 cm−3, which are 48.1% and 27.0% better than those from the conventional method (0.512 cm3 cm−3 and 0.062 cm3 cm−3). These results suggest that the new method can improve the retrieval accuracy of SMC in undulating terrains, which helps in applying the GNSS-IR to much wider regions.

Editing arcs to improve the capacity of GNSS-IR for soil moisture retrieval in undulating terrains