Flow Detachment by Concentrated Flow on Smooth and Irregular Beds

Historically, soil detachment by overland flow has mainly been studied using small samples with smooth surfaces. It may be questioned to what extent information from such experiments can be applied to predict flow detachment in actively eroding rills which are characterized by rough irregular bed surfaces. To evaluate how flow detachment on rough and smooth beds are related, two types of laboratory experiments were carried out: natural rill experiments, where a rill could freely develop; and small sample experiments, where small boxes with smooth surface were used. Soil conditions in both types of experiments were kept as constant as possible. Flow hydraulics and sediment load were recorded in detail during each experiment. Our experiments show that unit length shear force and shear stress are the most universal detachment predictor in rills (i.e., the relationship between these variables and soil detachment is independent of bed geometry). Other hydraulic variables can also be successfully related to flow detachment on a given bed type (smooth or rough), but they cannot accommodate for the effect of changing bed geometry. The fact that total shear stress controls flow detachment rather than grain shear stress or unit stream power implies that the coupling between sediment detachment and sediment transport is more complex than is presently assumed. A detachment prediction approach using hydraulic parameters based on unit length may allow a more accurate and simpler way to estimate sediment detachment in rills than a predictor, which needs an a priori calculation of flow width.