A coastal altimetry retracking strategy based on waveform classification and sub-waveform extraction

Satellite radar altimetry is successful in the open ocean. However, the application of altimetry is limited in coastal regions, in part because the complex waveforms returned from the coastal sea are contaminated by land reflections. We performed a systematic analysis in which we compared a large number of waveform and retracking algorithms. Based on this analysis, we developed a retracking strategy consisting of waveform classification and sub-waveform extraction (OceanCS). The altimetry waveforms are classified into three types (ocean-like, coastal-like and land-like), according to a set of indicators based on waveform shapes. The sub-waveform of coastal-like type that corresponds to the ocean return signal is extracted. The retracking processes are carried out for ocean-like and for sub-waveforms of coastal-like types by using the Ocean retracker. Thirty-one cycles (from March 2006 to February 2007, cycle 155 to cycle 188) of 20 Hz waveform data from Jason1 were reprocessed using this retracking strategy to obtain not only the improved sea surface height (SSH) but also the significant wave height (SWH) estimates. The sea-level measurements from a tide gauge station and SWH from a buoy were used to validate the new retracking strategy. Comparisons show that the OceanCS retracking strategy was more suitable than the other five algorithms tested (Ocean, OCOG, Ice-2, Beta5 and Threshold) and has uniform performance in both the open ocean and coastal regions. The waveform retracking algorithm proposed in this article is helpful for solving the problem of coastal altimeter waveform processing based on explicit physical description and without using a geoid model or other external data source. Furthermore, the OceanCS retracking strategy furnishes not only the SSH but also the SWH in the coastal region.