Global navigation satellite system interferometric reflectometry (GNSS-IR) is an effective technique for fully automated, all-weather, cost-effective tide level monitoring using signal-to-noise ratio (SNR) data. However, environmental noise and non-sea surface reflections often interfere with SNR signals, reducing retrieval accuracy and limiting satellite arc utilization rate. To address these limitations, this study proposes a joint denoising algorithm combining wavelet decomposition (WD) and empirical mode decomposition (EMD). By integrating the adaptive decomposition capabilities of EMD with the time–frequency localization strength of WD, the proposed algorithm effectively isolates noise components in the SNR data and extracts cleaner sea surface reflection signals. A case study using data from station SC02 at Friday Harbor evaluates the algorithm on short-term (10-day) and long-term (1-year) SNR time series. For 10 days, the root mean square error (RMSE) between retrieved and observed tide levels was reduced from 0.11 m (raw) and 0.11 m (EMD only) to 0.10 m, while arc utilization increased from 82% and 90% to 93%. For 1 year, the RMSE decreased from 0.12m and 0.10 m to 0.09 m, and arc utilization improved from 86% and 92% to 95%. These results demonstrate the WD–EMD algorithm improves retrieval accuracy and efficiency in GNSS-IR tide level monitoring.
Wang et al. (Tue,) studied this question.