Abstract:This paper provides an adaptive waveform re-tracking algorithm based on the primary peak (AMWRT). First, the primary peak is extracted, and second, the primary peaks are retracked by offset center of gravity (OCOG) method and threshold method. Finally, the retracked results are compared to obtain the best observed water level. The AMWRT, OCOG and threshold methods are applied to the middle reaches of the Yangtze river for comparison. The results show that: 1) The long-term water level sequence observed by altimetry satellites is more consistent with the trends measured by the hydrological station; 2) When the AMWRT algorithm is compared with OCOG and threshold, the AMWRT method performs well, and the accuracy is improved by 30% compared to the best-performing classic threshold_0.5 method. The absolute mean error (MAE) and root mean square error (RMSE) are reduced to 0.55 m and 0.72 m, respectively.