Soil moisture monitoring using GNSS interference signal: proposing a signal reconstruction method

Soil moisture monitoring using Global Navigation Satellite System (GNSS) interference signal has gained wide interests in recent years. It utilizes the variation pattern of the routinely measured Signal-to-Noise Ratio (SNR) that contains the interference information between the direct and soil reflected signal. Previous studies have shown that the amplitude of the de-trended SNR data is not a good indicator for soil moisture variation. The possible cause found in this study is that the de-trending operation only reduces the additive direct signal component in the SNR data, while the multiplicative direct signal component in the amplitude of SNR data is unaffected. Therefore, a method is proposed to reduce the contamination of the direct signal component on the amplitude through signal reconstruction and then normalization. Experiment data are collected and processed to calculate the normalized amplitude based on the reconstruction method. The results show that the overall correlation coefficient of the normalized amplitude with in-situ soil moisture reaches to 0.6966 under highly rough surface condition, while it is only 0.4314 for the amplitude obtained from the conventional method. A quadratic model is used to retrieve soil moisture from the normalized amplitude, the retrieval error is less than 0.085 cm³ cm⁻³.