Extracting a Connected River Network from DEM by Incorporating Surface River Occurrence Data and Sentinel-2 Imagery in the Danjiangkou Reservoir Area

Accurate extraction of river network from the Digital Elevation Model (DEM) is a significant content in application distributed hydrological model. However, study based on DEM has some limitations, such as location offset, inaccurate parallel channel and short circuit meandering channels. In this study, we proposed new enhancement method for NASADEM V001 Danjiangkou Reservoir area. We used Surface Water Occurrence (SWO) Sentinel-2 data to describe vertical limit differences between morphological units complement actual flow path information by stream burning method. The extracted were evaluated three different geographical regions. Compared with centerline, error was significantly reduced. average offset distances 68.38, 36.99, 21.59 m test areas. V001, areas reduced 7.26, 40.29, 42.35%, respectively. To better estimate accuracy, also calculated compared accuracy MERIT Hrdro HydroSHEDS. experimental results demonstrated that can effectively improve meet needs simulation.