An iterative coastal altimetry retracking strategy based on fuzzy expert system for improving sea surface height estimates

This paper improves the accuracy of altimeter-derived sea level anomalies (SLAs) near coast through an iterative waveform retracking system. The principle of this system is twofold. First is to reprocess the altimeter waveforms using the optimal retracker, which is searched base on the analysis from a fuzzy expert system. Second is to minimize the relative offset in the retrieved SLAs when switching from one retracker to another, using a neural network. The system reprocesses 20-Hz waveforms from Jason-2/OSTM in the Great Barrier Reef, Australia. When compare the retrieved SLAs with tide gauge data from Townsville and Bundaberg stations, results show the SLAs from this study generally outperform SLAs from MLE4 and Ice retrackers. It yields higher correlations (≥0.8) and smaller root mean square errors (≤16.6 cm) than those of MLE4 (≤0.78 and ≤19 cm) and Ice (≤0.78 and ≤18.7 cm) retrackers.