Abstract
To address the degradation of ocean current retrieval accuracy in shore-based coherent C-band radar caused by non-wave components (e.g., broken waves) and low signal-to- noise ratio (SNR) under complex sea states, a novel retrieval framework based on Spectral Feature Reconstruction (SFR) is proposed. By utilizing Spectral Proper Orthogonal Decomposition (SPOD) combined with a spatiotemporal joint constraint mechanism, the dominant wave spectrum is adaptively extracted from radar echoes contaminated by nonlinear interferences. Subsequently, two current retrieval methods are presented: 1) the SFR-based Comprehensive Weighted Least Squares (SFR-CWLS) method, which enables rapid and robust current estimation by constructing a comprehensive weight matrix incorporating spectral energy density and spatiotemporal joint membership; and 2) the SFR-based Spectrum Matching (SFR-SM) method, which introduces forward spectral modeling to precisely simulate spectral asymmetry and energy broadening effects induced by the finite spatiotemporal aperture and observational geometry, thereby achieving high-precision ocean current retrieval via global spectral matching. Validation using field data collected at Dongshan Island during the typhoon, along with High-Frequency Surface Wave Radar (HFSWR) and buoy measurements, demonstrates that both methods are in good agreement with HFSWR observations. Notably, the SFR-SM method exhibits superior accuracy and adaptability. Moreover, correcting the dispersion filtering with the retrieved currents reduces the root-mean-square error (RMSE) of significant wave height (Hs) retrieval from 0.74 m to 0.24 m. This pronounced “wave-current coupling” further validates the effectiveness of the proposed framework in complex observational environments.
Funding Information
  • Research Fund of Key Laboratory of Ocean Observation Technology, Ministry of Natural Resources (2024klootA06)
  • National Key Research and Development Program of China (2025YFE0219300)

This publication has 36 references indexed in Scilit: