Effects of DEM horizontal resolution and methods on calculating the slope length factor in gently rolling landscapes

The USLE is used world-wide to predict soil loss on the field scale from sheet and rill erosion. The slope length (L) factor is derived as its topographical factor. The accuracy of L factor determines the precision of soil loss estimation with USLE. Uncertainties on L factor are caused by DEM resolution and the choice of the processing algorithm. In the present study we made two comparisons to evaluate the effects of DEM horizontal resolution and processing algorithm on the accuracy of the L factor in gently sloped landscapes: one is between the grid cumulating method (GC) and the contributing area method (CA) using D8 flow-routing algorithm, the other is among single (08, Rho8) and multiple (FD8. FRho8 and DEMON) flow-routing algorithms for processing the contributing area method. In two comparisons, 5 m, 10 m, 25 m, 50 m and 100 m OEM of a 0.88 km(2) catchment in the lowland of Northern Germany were applied. The results indicate that L factor calculated with any of the six methods is sensitive to horizontal resolution, which strongly affects the accuracy. With decreasing resolution, correlations of L(CA_Rho8) and L(CA_D8), L(CA_FD8) and L(CA_D8), L(CA_FD8) and L(CA_Rho8) increase while those between DEMON and the other Flow-routing algorithms do not change significantly. With decreasing resolution, the difference between L(GC_D8) and L(CA_D8) is enlarged, while differences between any two flow algorithms using CA did not change significantly. The L factor variation between any two methods is larger on the upslope than the flat valley for the 5 m and 10 m OEM while terrain characteristics are not visible on the 25 m, 50 m and 100 m DEM. The L factor also depends on the computation method. L(GC_D8) is approximately half of L(CA). It is concluded that OEM horizontal resolution is very important for L factor calculation. The most suitable calculation method is L(GC_D8) for gently rolling landscapes. This study can be used for selecting a suitable method and OEM resolution for accurate calculation of L factor and soil loss in gently rolling landscapes. (C) 2011 Elsevier B.V. All rights reserved.