ARS | Publication request: DEM AGGREGATION FOR WATERSHED MODELING

Submitted to: Journal of the American Water Resources Association
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: May 18, 1999
Publication Date: N/A

Interpretive Summary: Digital Elevating Models (DEM) provide topographic information in digital form organized in a 30m x 30m grid. This is very useful for watershed modeling, but the level of detail is overwhelming for watersheds of even small size. An aggregation procedure is described that enlarges the grid size but ensures that watershed properties, such as flow paths, are maintained. The procedure is demonstrated using the ARS research watershed, Goodwater Creek, located in central Missouri. The DEM, which originally contained more than 80,000 cells, was reduced to approximately 5000 and to 3000 cells in two aggregation trials without loss of significant information. These results make the digital form of elevation maps more usable to water resource planners and users who must increasingly rely upon comprehensive, spatially-distributed computer models for evaluation of proposed changes.

Technical Abstract: The widely available USGS digital elevation model (DEM) data have a horizontal resolution of 30 m x 30 m. This high resolution topographic information is impractical for many large scale applications of grid-based hydrologic and water quality models because it requires an excessive amount of computational time and memory. In this study, a grid aggregation method dis developed to approximate flow direction and land slope for macro-scale grids from which topographic watershed characteristics can be parameterized as input to distributed models. The approximation method for flow direction and land slope of macro-scale grids accounts for spatial variations of topography and makes the flow direction closely follow the flow pattern suggested by the micro-scale DEM data. The aggregation method was applied to the DEM of Goodwater Creek, a flat watershed which is located in central Missouri. The drainage networks derived for different levels of aggregation and the corresponding changes on land slopes show that for thi application the landscape topography is not over-smoothed by aggregation and that the aggregated macro-scale grid of the Goodwater Creek watershed provides an adequate representation of the landscape topography for large scale applications of grid-based hydrologic and water quality models.