Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
6345614 | Remote Sensing of Environment | 2016 | 10 Pages |
â¢A 3D topography reconstruction method was proposed to measure surface roughness.â¢The physical meaning of the SERD was reinterpreted in the Lop Nur lake basin area.â¢Calculated the distributions of different salt crust types in the certain lake area.â¢The fluctuation processes of an ancient lake in Lop Nur were interpreted theoretically.
Lop Nur once was a huge lake located in northwestern China. Its environmental evolution is significant for understanding historical global climate change. At present, there is no surface water in Lop Nur Lake Basin, and on SAR images it looks like an “Ear”. The objective of this article is to recognize different types of salt crust by surface roughness parameters and interpret their environmental meanings in Lop Nur. Surface roughness parameter estimation, by means of microwave remote sensing technology, is a topic which is intensively studied yet not solved satisfactorily. In this article, the potential of using Polarimetric Synthetic Aperture Radar (PolSAR) acquisitions for the estimation of surface roughness parameter in Lop Nur Lake Basin is investigated. A new parameter Sl is proposed to characterize surface roughness and to discriminate salt crust types. Through the analysis of scattering mechanisms, the physical meaning of the PolSAR parameter Single Bounce Eigenvalue Relative Difference (SERD) is reinterpreted and it is used to retrieve parameterSl. A good linear correlation between the measured surface roughness parameter Sl and SERD with an R2 value of 0.72 was obtained. Furthermore, the distributions of different salt crusts in the certain lake area were analyzed. Based on hypotheses of the evolution of surface salt crust and the drying-up cycle of lakes, the fluctuation processes of an ancient lake in Lop Nur were discussed theoretically. These can help to have a better understanding of environmental changes in arid and semi-arid regions.