Assessing the positioning accuracy of channel extracted from DEM: a test of channel offset

River network extraction from DEMs is crucial for hydrological analysis. While the accuracy of watershed area and river length have been widely tested, spatial positioning accuracy lacks attention. This study develops the Test of Channel Offset (TCO) to assess the positioning errors of extracted river networks. Using Landsat TM interpreted data as reference, TCO was applied to evaluated the accuracy of networks extracted from ASTER DEM (30 m) and SRTM DEM (90 m) in the Hanjiang River Basin. The results indicate channel offsets of 0.07–0.4 km for SRTM and 0.09–1.32 km for ASTER. The SRTM DEM achieved higher accuracy than the ASTER DEM, demonstrating that accuracy depends not only on the DEM resolution but also on data source. Furthermore, the channel offset increases with higher channel levels within the river network hierarchy. We suggest that this is a cumulative effect of errors originating from the D8 algorithm. The spatial similarity values calculated by the line group similarity (LGS) model are 0.945 for ASTER DEM and 0.96 for SRTM DEM, respectively, confirming the reliability of the TCO results.