Title: Effects of topographic feedback on erosion and deposition prediction Authors
Submitted to: Meeting Proceedings
Publication Type: Proceedings
Publication Acceptance Date: July 5, 2011
Publication Date: September 18, 2011
Citation: Dabney, S.M., Vieira, D.A.N., Yoder, D.C. 2011. Effects of topographic feedback on erosion and deposition prediction. Proceedings of the International Symposium on Erosion and Landscape Evolution. 18-21 Sept. 2011. Anchorage, AK. 2011. pp. 577-585. Interpretive Summary: Soil erosion devastates arable land and infrastructure and strains the balance between economic stability and viability. Recent studies indicate that ephemeral gully erosion (small channels eroded by concentrated runoff that can be filled by normal tillage) may be a significant form of erosion and source of sediment on cropland in the U.S. (averaging around 40% of the sediment delivered to the edge of the field in some documented studies). The current study was designed to isolate one aspect of the ephemeral gully erosion problem: channel widening. Experiments on channel widening rates, quantified under controlled experimental conditions, revealed that discharges of ~50 L min-1 in the lower sloped beds (1-4%) and ~30 L min-1 in the steeper slope beds (8%) were responsible for excessive widening. Lower flow rates produced small amounts of widening and larger flow rates merely flushed sediment, no significant widening took place. Both water surface and sediment discharge records show good agreement with expectations; i.e. increased flow rates bring about an increase in the occurrence of bank failures and the flow changes accordingly. Ephemeral gully channel widening is not currently well understood. This research program will provide much needed predictive equations for gully channel width.
Technical Abstract: Soil erosion and depositional processes result in changes in topographic and soil profile properties over time. In spite of this, few current soil erosion models account for these changes. To begin to address this deficiency, a distributed version of RUSLE2 has been developed and used in combination with the tillage erosion model TELEM (Vieira and Dabney, 2010). These models were applied to a 6.6 ha research watershed near Treynor, IA, where runoff and sediment yield were measured on a daily basis from 1975 through 2002. The watershed contained a grassed waterway and, beginning in 1991, ~1-m wide grass hedges were established with 15.4 m interval between the grass hedges. Beginning in 1996, no-till management replaced conventional tillage management in the watershed. Using a 3-m raster DEM, concentrated flow areas ending RUSLE2 hillslope profiles were defined where contributing areas were > 600 m2. Average monthly runoff and sediment delivery to concentrated flow areas for a representative RUSLE2 hillslope profile were compared with measured monthly runoff and sediment delivery at the watershed outlet. For conventional tillage management, monthly runoff patterns reasonably matched observations, but watershed sediment yield was considerably lower than that predicted by RUSLE2 sediment delivery. These results suggest that sediment was deposited in the grassed waterway in the footslope area of the watershed. The establishment of grass hedges reduced sediment yield similarly in the RUSLE2 hillslope estimate and the watershed sediment yield. Conversion to no-till had a much greater impact in reducing RUSLE2 sheet and rill erosion than it did on watershed sediment yield, suggesting that ephemeral gully erosion remained an important sediment source. Distributed patterns of soil erosion appeared reasonable and the impact of grass hedges on average soil loss was similar to effects on observed sediment yield. To assess the potential influence of topographic feedback on future erosion estimates, terrain elevation was modified by summing estimates of water and tillage erosion that would be expected from 50 years of conventional tillage. As expected, reduced slope gradients generally reduced future erosion. However, results highlighted limitations of the current approach. The use of a 3x3 kernel to determine local slope steepness resulted in excessive steepness estimates in rasters immediately downslope of aggrading grass hedges. Also, the lack of consideration of erosion or deposition or concentrated flow channels limited the ability of the modeling scheme to reinforce existing ephemeral gully channels. A more sophisticated and complete modeling system that will include an ephemeral gully model and should update terrain elevation, slope steepness, and flow vectors on at least an annual basis when assessing conservation impacts of land management systems that result in the development of numerous permanent slope breaks.