Impacts of DEM uncertainties on critical source areas identification for non-point source pollution control based on SWAT model

Abstract

The impacts of different digital elevation model (DEM) resolutions, sources and resampling techniques on nutrient simulations using the Soil and Water Assessment Tool (SWAT) model have not been well studied. The objective of this study was to evaluate the sensitivities of DEM resolutions (from 30m to 1000m), sources (ASTER GDEM2, SRTM and Topo-DEM) and resampling techniques (nearest neighbor, bilinear interpolation, cubic convolution and majority) to identification of non-point source (NPS) critical source area (CSA) based on nutrient loads using the SWAT model. The Xiangxi River, one of the main tributaries of Three Gorges Reservoir in China, was selected as the study area. The following findings were obtained: (1) Elevation and slope extracted from the DEMs were more sensitive to DEM resolution changes. Compared with the results of the 30m DEM, 1000m DEM underestimated the elevation and slope by 104m and 41.57°, respectively; (2) The numbers of subwatersheds and hydrologic response units (HRUs) were considerably influenced by DEM resolutions, but the total nitrogen (TN) and total phosphorus (TP) loads of each subwatershed showed higher correlations with different DEM sources; (3) DEM resolutions and sources had larger effects on CSAs identifications, while TN and TP CSAs showed different response to DEM uncertainties. TN CSAs were more sensitive to resolution changes, exhibiting six distribution patterns at all DEM resolutions. TP CSAs were sensitive to source and resampling technique changes, exhibiting three distribution patterns for DEM sources and two distribution patterns for DEM resampling techniques. DEM resolutions and sources are the two most sensitive SWAT model DEM parameters that must be considered when nutrient CSAs are identified.