Shore Zone Classification from ICESat-2 Data over Saint Lawrence Island
Huan Xie,Yuan Sun,Xiaoshuai Liu,Qi Xu,Yalei Guo,Shijie Liu,Xiong Xu,Sicong Liu,Xiaohua Tong
DOI: https://doi.org/10.1080/01490419.2021.1898498
2021-01-01
Marine Geodesy
Abstract:The shore zone is the most active zone in the atmosphere, hydrosphere, biosphere and lithosphere of nature, and has the environmental characteristics of both ocean and land. The ICESat-2 satellite provides height measurements of shore zone using a photon-counting LiDAR. The purpose of this study is to explore the application potential of ICESat-2 satellite data in shore zone classification. Saint Lawrence Island, Alaska, was chosen as the study area. Firstly, in this study, the upper and lower boundaries of the shore zone of the study area were extracted based on Google Earth images. The slope and width between the two boundaries were then calculated according to the formula. Secondly, six statistical indicators (standard deviation, relative standard deviation, average absolute deviation, relative average deviation, absolute median error and quartile deviation) related to the substrate and sediment classification that could reflect the characteristics of the shore zone profile were extracted, and the statistical indicators were used as input parameters of the softmax regression model for classification. Finally, the accuracy of the shore zone classification was validated using the ShoreZone classification system. The results show that, among the 246 shore zone sections in the study area, 86% (212) has been correctly classified. The results therefore indicate that ICESat-2 data can be used to support the characterization of shore zone morphology.