Quantifying sediment loads from streambank erosion and potential load reductions from streambank stabilization using process-based modeling

Unstable streambanks contribute a significant sediment load to surface waters in some watersheds. Streambank stabilization techniques are available to increase stability of streambanks or reduce erodibility, thereby reducing sediment loads. Process-based models can be used to evaluate the stability of stream channels and predict sediment yields with and without potential stabilization to determine the effectiveness of stabilization. Two fluvial erosion models are commonly used with in process-based models to simulate the erosion rate of soils: the excess shear stress equation and the Wilson model. Both models include two soil parameters which may be highly variable within a stream system. In order to simulate stabilization practices in process-based models, each practice must be appropriately parameterized. The objectives of this research were to investigate the variability of fluvial erodibility parameters within a watershed and resulting implications for erosion prediction, parameterize stabilization practices for simulation in process-based models, and determine stabilization effectiveness for stream-scale sediment reduction. Jet erosion tests were completed along two streams in both the Illinois River watershed and Fort Cobb Reservoir watershed to determine erodibility parameters. Erodibility parameters were incorporated into a process-based model, CONCEPTS, to simulate bank retreat. Erodibility parameters varied by two to five and one to two orders of magnitude in the Illinois River and Fort Cobb Reservoir watersheds, respectively. Less variation was observed in lateral retreat prediction from CONCEPTS simulations than input erodibility parameters. Two stabilization practices were selected for simulation, riprap and vegetation. Each practice was simulated using two parameters, median particle size, d50 and riprap height, h for riprap and added root cohesion, Cr and shear stress adjustment factor, v for vegetation. An uncertainty analysis showed sediment reduction and retreat predictions were not sensitive to d50 or Cr, but were highly sensitive to h and v. Finally, a framework was developed to evaluate streambank stabilization practices for sediment reduction using process-based models by accounting for public and landowner perception, costs and effectiveness. The methodology was applied using the CONCEPTS model setup for Fivemile Creek in the Fort Cobb Reservoir Watershed. Vegetation with 2:1 bank slopes was the most cost-effective stabilization practice for this stream