Estimation of significant wave height using Sentinel-3 data

Abstract

Coastal area is one of the most important area for us. More than 600 million people (around 10% of the word’s population) live in coastal areas that are less than 10m above sea level. Nearly 2.4 billion people (about 40% of the world’s population) live within 100km of the coast. Therefore, monitoring of coastal waters is extremely important. Due to the limitation of the number and location, the tide gauge stations around the world cannot provide a sufficient amount of in-situ data. Therefore, satellite altimetry plays an increasingly important role, especially when the SAR altimeter is put into use. However, due to the complexity of the coastal water surfaces, the performance of the satellite altimeter over the coastal area is far worse than over ocean. This thesis is dedicated to developing a method to determine one of the essential characters of the water surfaces - the significant wave height (SWH), using the Sentinel-3 data in the coastal area. The three primary steps of the method are extracting the thermal noise and the leading edge, fitting this part of waveform and determining the relationship between the new retracker and the physical model. In the first step, an algorithm is developed to avoid the interferences of the noise on the trailing edge. Therefore, the peak of the leading edge could be determined more accurately. The condition for the start point of the leading edge of the PLRM waveforms is Dwf > 0.01, inherited from ALES, whereas a more appropriate threshold for the SAR waveforms has been found as Dwf > 0.03. In the second step, the limitation of the Gauss-Markov model for the waveform adjustment has been discussed. Thus, the Levenberg-Marquardt method has been chosen to adjust the waveform. In the third step, the relationship between the raising time and the beta4 has been found. Then, we could estimate the SWH directly from the raising time which makes it possible to estimate the SWH from some complicated waveforms in coastal areas. We have employed the developed methodology to determine the significant wave height in the coastal area near the Cuxhaven. The quality of the results has been proved by comparing with the in-situ data from the Elbe measuring station provided by the Federal Maritime and Hydrographic Agency of Germany. The validation showed that the proposed method can determine reliable SWH from approximately 1km offshore, which is an improvement of earlier results.