Rill erosion initiation and effect of DEM resolution on erosion prediction

Rill erosion dominates the erosion process in many parts of the world. Quantifying conditions for rill initiation and determining the effect of slope steepness and rainfall intensity on rill erosion can be useful for a better understanding of soil erosion processes and estimating soil loss. An indoor facility with a large sloping plot (8 m x 3 m) was used to investigate rill initiation using simulated rainfall with five different slopes and three rainfall intensities. The study found that the distance from start of runoff to rill initiation depended on both overland flow rate and slope steepness and decreased as slope and rainfall intensity increased. The impact of slope on the distance from start of runoff to rill initiation was relatively more important than that of rainfall intensity. The use of a GIS (Geographic Information System) has greatly enhanced the capability of predicting soil loss in watershed modeling. Selecting the appropriate DEM (Digital Elevation Model) resolution is important for estimating topographic values and subsequently estimating soil erosion and sediment delivery rates. Field slope length and steepness data were compared to slope length and steepness estimates from 10-m and 30-m DEMs in ArcGIS. Predicted erosion rates from the RUSLE2 model for observed and DEM topographic data in ArcGIS were also compared. Multiple statistical analyses resulted in the quantified rank of error for slope steepness and slope length estimates ascending order as compared to the observed data in general as: DEM resolution (10-m, 30-m); data extraction (3DA, identification); for calculation scenarios, slope steepness estimation (ArcGIS neighborhood, 3DA followed by curve fitting derivative); slope length estimation (flow length, flow accumulation) in ArcGIS. For soil loss in the NWRR, the predictions of RUSLE2 from 10-m DEM resolution underestimated 13%, and those from 30-m DEM resolution underestimated 23%, RUSLE2 estimates from observed topographic data. For this steep and varying topographic area, the impact of slope steepness on soil loss was more important than that of slope length. The results can be applied to hillslope and watershed models using slope steepness or flow path length elements from a GIS in the NWRR or other similar area.