Flow threads in surface run‐off: implications for the assessment of flow properties and friction coefficients in soil erosion and hydraulics investigations

Soil surface microtopography produces non‐uniform surface run‐off, in which narrow threads of relatively deep and fast ow move within broader, shallower, slower‐moving regions. This kind of ow is probably widespread, given that microtopography is itself common. Methods used to record the properties of surface run‐off include grid‐ or transect‐based depth observations, with a single mean ow speed derived by calculation from V = Q/WD , and the use of dye timing to estimate velocity, with an effective mean depth calculated from D = Q/WV . Because these methods allow only single, ow‐eld mean values to be derived for V or D , neither is well suited to non‐uniform ows. The use of depth data to derive a ow‐eld mean V furthermore implicitly applies area weighting to the depth data; likewise, the use of dye speeds for V inherently overestimates mean V because dye dominantly follows the faster ow threads. The associated errors in derived parameters such as friction coefcients are not readily quantied and appear not to have been addressed previously. New eld experiments made on untilled soil surfaces in arid western NSW, Australia, explore these circumstances and the implications for deriving meaningful measures of ow properties, including friction coefcients. On surfaces deliberately chosen for their very subtle microtopography, average thread velocities are shown to be commonly 2·5 times greater than the ow‐eld mean, and locally 6–7 times greater. On the other hand, non‐thread ow speeds lie below the ow‐eld mean, on average reaching only 84 per cent of this value, and often considerably less. Flow‐eld means conceal the existence of regions of the ow eld whose properties are statistically distinct. Results conrm that a reliance on ow‐eld average depths yields estimates of friction coefcients that are biased toward the shallower, high‐roughness parts of the ow, while if dye speeds are relied upon the results are biased toward the deeper, smoother threads of ow. A new approach to the evaluation of friction coefcients in non‐uniform ows is advanced, involving the determination of separate coefcients for threads and non‐thread zones of the ow eld. In contrast, ow‐eld friction coefcients as they are customarily derived in run‐off plot experiments subsume these distinct coefcients in proportions that are generally unknown. The value of such coefcients is therefore questionable. Copyright © 2004 John Wiley & Sons, Ltd.