Retracking Sentinel-3A SAR waveforms to monitor the water level of a small inland water body (Case study: Doroudzan Dam Reservoir, Shiraz, Iran)

Document Type : Research


1 Student, School of Surveying and Geospatial Engineering, College of Engineering, University of Tehran, Tehran, Iran

2 Professor, School of Surveying and Geospatial Engineering, College of Engineering, University of Tehran, Tehran, Iran

3 Assistant Professor, Department of Geomatics Engineering, South Tehran Branch, Islamic Azad University, Tehran, Iran

4 Ph.D. Student, Department of Geomatics, College of Geodesy & Geomatics Engineering, K.N. Toosi University of Technology, Tehran, Iran


In inland water bodies, the water level obtained from the Level-2 data of the altimetry missions is not correct. Therefore, to correct the water level measured in these areas, it is necessary to retracking the return waveforms. In this study, data from level-2 and level-1 SRAL altimeter of Sentinel-3A mission, measured in SAR mode, in the period from March 2016 to November 2019 to monitor the water level of Doroudzan Dam, has been used. The threshold retracking algorithm with different thresholds has also been used to retrack the waveforms in the level one data. The results showed that the OCOG retracker in L-2 data with an RMSE value of 38.23 cm and a correlation of 99.23% with in situ gauge data compared to other retrackers in L-2 data from Doroudzan dam has higher accuracy in estimating the time series of the water level. The Ocean retracker also has results close to those of the OCOG retracker, indicating that these two retrackers perform well in restoring water levels. After obtaining the water level time series from the retrackers in the L-2 data and selecting the optimal level two retracker, the return waveforms from the L-1 data were first retracked using the threshold algorithm. Then the time series of the water level for different thresholds were obtained and compared with in situ gauge data, which showed that the threshold of 60% with a value of RMSE 37.73 cm and a correlation of 99.30% improved %1.3 accuracies and increase of %0.07 correlation with in situ gauge data has been optimized for the time series of water level obtained from L-2 retracker. Also, the results showed that, especially in the period from 2017 to 2018, the difference in water levels resulting from the retracking of the return waveforms with the optimal threshold algorithm (60%) with in situ gauge data less than the optimal L-2 retracker (OCOG). The average water level of Doroudzan Dam from the threshold of 60% was analyzed. Results showed the highest growth in water level with 4.09 m from March 6 to April 2, 2019, which corresponds to usually rainy months. The most significant decrease in the water level with 2.80 meters occurred from April 29, 2019, to May 26, 2019, which are usually low rainfall months. The results also showed during the study period a slight increase in the water level of Doroudzan Dam. Due to the hard, challenging shape, and topography of Doroudzan Dam and its confused waveforms, therefore, in the above study area, it is not possible to expect high accuracy from both the retrackers in the L-2 data and the results of the waveform retracking. Therefore, the proximity of RMSE results and correlation goes back to the shape and topography of the Doroudzan Dam reservoir. The results of this study show the high suitability of the Sentinel-3 mission in monitoring the water level from inland water bodies, which is still a challenging area for satellite altimetry to monitor. Indeed, for a better understanding of the performance of this mission, more samples need to be analyzed.


Main Subjects