ABSTRACT
In recent years, various studies have appeared on (automated) processing of river morphodynamics from satellite images. A key challenge in such approaches is to use appropriate and consistent methods to quantify eroded and deposited areas and channel migration rates. This is particularly important in highly dynamic multi-threaded rivers, where apparent bank lines can move significantly with changing water levels. To assure a consistent treatment of bank lines we use a methodology where channels are identified by detecting the vegetation boundary. We apply this methodology to a 250 km reach of the Ayeyarwady river (Myanmar) and extract metrics for eroded and deposited areas and channel migration. These morphological metrics are compared to hydrologic conditions, and significant correlations were found. It shows that the proposed method may help to classify and better understand river dynamics rivers of large multithreaded rivers.
