Quantifying and modeling sediment yields from interrill erosion under armouring

Abstract

Abstract Surface soil properties can change as a result of soil disturbances, erosion, or deposition. When soils contain rock, surface soil properties can also change over time as a result of the process of soil armouring, which is the selective removal of finer particles by erosion, leaving an armoured layer of coarser particles that may reduce further soil loss. Rapid armouring is typically reported in steep and bare slopes on mine sites, construction sites, road embankments, and also rangelands. Changes in surface soil properties over time induced by armouring are not accounted for in current erosion models such as WEPP or RUSLE2 because little is known about rates of armouring over time as a function of rainfall intensity, rock content, slopes, and other factors. In this paper we simulate soil armouring induced by interrill erosion in two sets of experiments and propose ways to account for the process in WEPP and RUSLE2 without modifying the science behind the models. The first set of experiments was conducted to demonstrate and quantify the effect of armouring on sediment yields under varying rainfall intensities. Rainfall with intensities ranging from 22 to 80\u202fmm h−1 was simulated on 0.56\u202fm2 plots at slopes of 18 degrees (32.5%) using topsoil with high rock content from a mine restoration site. Results showed a clear relationship between rainfall intensity and armouring. There was an over 75% reduction in total soil loss under 22\u202fmm h−1 rainfall between freshly applied soils and highly armoured soils at the same slope. A second set of experiments was conducted to understand the relationships between soil rock content, rate of surface rock cover change, slope change and sediment yields. Sediment yields and surface rock cover were quantified for non-cohesive soils consisting of glass beads with a diameter range of 45 to 90\u202fμm and 0, 20 and 40% rock content in 0.22 m2 plots at a 15° (26.8%) slope under 80\u202fmm h−1 simulated rainfall. Linear relationships were observed between cumulative sediment yields and rock cover, and exponential relationships between total runoff and rock cover. The armouring process was modeled at the experimental scale with WEPP and RUSLE2 by iteratively altering the rock cover and the slope over time. WEPP event based simulations at this scale resulted in reasonable predictions of sediment yields throughout the armouring process, and RUSLE2 required modification of soil erodibility values to account for high rock content and changes in runoff over time. Automation of this process, however, would require modifications of the models to expose rock cover and change slope over time. A method for doing this is discussed for interrill processes, but interactions between rill erosion and armouring need further study. The effect of long-slope high-energy rill flows on armouring also needs further investigation through field scale experiments, as this study did not examine the extent to which rock is uncovered and/or moved on long slopes differently from the residues already modeled in WEPP and RUSLE2.