| Abstract: | Friction coefficients in overland flows are customarily estimated from mean flow properties (depth, velocity, slope) that subsume spatial variations in flow arising from two major causes: microtopography and obstacles. This paper uses laboratory experiments in shallow flumes to examine the extent of non‐uniformity in flow conditions associated with each cause. Randomly placed emergent obstacles in a flume with a shallow axial channel generally yielded higher hydraulic roughness than the same pattern of obstacles on a planar flume, as well as greater variation in roughness as the obstacle locations were altered. In both flumes, hydraulic roughness fell with increasing Reynolds number for 10% obstacle cover, showed a flattening trend at 20% cover, and exhibited a convex‐downward trend at 30% obstacle cover. These results indicate the progressive onset of flow controls at narrow gaps in the obstacle field. In such flows, the use of mean flow properties conceals the existence of two main subdivisions of flow: flow filaments and backwater flows. In the experiments, flow filaments involved velocities more than twice the overall mean, whereas backwater flows were much slower than the mean. The existence of fast‐moving flow filaments may be significant in understanding soil transport in surface runoff, and backwater depths may modify splash detachment. Similarly, friction coefficients that fail to reflect these important non‐uniform flow components may not be optimal for hydraulic calculations or in erosion models. It is concluded that new approaches to observing and processing flow data may be required, in order to avoid the loss of important flow detail that is entailed in assuming uniform flow conditions. Copyright © 2003 John Wiley & Sons, Ltd. |
