Snow depth estimation from GNSS SNR data using variational mode decomposition.

In recent years, Global Navigation Satellite System-Interferometric Reflectometry (GNSS-IR), a new remote sensing technique, has been widely used to monitor surface signature parameters. In the classical GNSS-IR technology, poor signal separation will seriously affect the accuracy of the inversion results. In order to better separate the signal-to-noise ratio trend item, the variational mode decomposition (VMD) algorithm is introduced. We use the GNSS data of P351 station in 2013–2014 and AB33 station in 2017 in the Earthscope Plate Boundary Observatory network to carry out snow depth inversion experiments. The measured snow depths provided by the Snowpack Telemetry network were used for the validation of the inversion accuracy. The feasibility and superiority of the VMD algorithm in GNSS-IR snow depth inversion experiments were verified by analyzing the experimental results. The root-mean-square error (RMSE) and correlation coefficient of the inversion results of P351 station in 2013–2014 were 13.41 cm and 0.99, respectively, which improved the inversion accuracy by about 54%. Moreover, the number of inversion points during the experimental period increased from 19,997 to about 26,958, which is an increase of about 35%. Similarly, the RMSE and correlation coefficient of the inversion results of AB33 station in 2017 reached 8.55 cm and 0.97. Compared with the traditional algorithm, the accuracy and the number of inversion points increased by about 15% and 22%, respectively. [ABSTRACT FROM AUTHOR]