Incorporating ancillary data to refine anthropogenically modified overland flow paths

A study carried out in the Oldman River watershed in southern Alberta, Canada, to delineate potential surface water pollution sources revealed that the use of conventional raster-based digital elevation models (DEMs) and flow-direction algorithms resulted in the delineation of incorrect watershed boundaries. The inaccuracies resulted from the resolution of the available DEM, which is too coarse to represent the presence of anthropogenically modified terrain features, such as roads, ditches and canals, which can significantly alter overland flow paths. This observation prompted the development of the computer model RIDEM (Rural Infrastructure Digital Elevation Model). By incorporating ancillary data to refine anthropogenically modified overland flow paths into the grid-based watershed delineation process, the scale of runoff flow paths created with RIDEM is reduced below the scale defined by the representation of terrain with the DEM. RIDEM also resolves the inability of grid-based flow-direction algorithms to account for the split-flow patterns that occur in irrigated landscapes. The model outputs include a corrected flow-direction matrix, gross watershed boundaries, and road-induced dead drainages. The watershed areas derived with RIDEM differed by up to 49% from watersheds derived using a conventional DEM with a 20 m grid cell size. However, results of watershed boundaries could not be verified, because highly accurate watershed boundaries do not exist for our study area. Copyright © 2005 John Wiley & Sons, Ltd.