Informations et ressources scientifiques
sur le développement des zones arides et semi-arides

Accueil du site → Doctorat → États-Unis → 1972 → THE DYNAMIC STRUCTURE OF EPHEMERAL STREAMS

University of Arizona (1972)


Renard, Kenneth G.


Auteur : Renard, Kenneth G.

Université de soutenance : University of Arizona

Grade : Doctor of Philosophy (PhD) 1972

The hydraulic features of ephemeral streams are dynamic in response to the variable streamflow available for sediment transport. For a given water and sediment discharge, the hydraulic stream features of width, depth, velocity, slope, and roughness result from mutual adjustments. Although the problem of quantifying the stream behavior can be approximated using laboratory and field developed hydraulic relations, variations in the temporal and spatial stream roughness and sediment sizes available for transport add an additional element of uncertainty because of the difficulties of measuring these factors in the prototype. The dynamic stream behavior is discussed qualitatively with consideration of the profile shape, bed slope, channel width, flow depth, sediment discharge and channel controls. Transmission losses (that water infiltrating to the alluvium) in ephemeral streams result in a reduced volume of water to transport sediment. The losses tend to be offset by additional runoff from tributaries when the runoff-producing thunderstorms, which are of limited areal extent, are confined to the lower portion of the drainage basin. There are two tendencies apparent in an ephemeral stream ; to be concave down because of loss of discharge by infiltration through the normally dry channel alluvium, and to be concave up because there is more flow downstream than upstream due to tributary inflow. These two tendencies seem to be in balance on Walnut Gulch resulting in a remarkably constant slope for the main channel from one end to the other. The principal channel reach on the Walnut Gulch Experimental Watershed was modeled incorporating a geomorphic parameter approach based on the Horton-Strahler stream order numbering concept for tributary intersections. The intersections by stream order are described by a geometric probablity function with the probability and the number of intersections related to the rate of drainage area increase per length of stream channel. The channel slope and width and the watershed drainage area for the tributaries were found to be related to stream order. The log-normal probability distribution was used to model the particle size distribution of the alluvial beds. Although minor deviations from the theoretical distributions were sometimes observed at the extremes, the distributions can be specified using the mean and standard deviation. These values were then used in the Laursen relation to describe the percentages in various sizes. Prediction equations developed using multiple linear regression were used to predict the mean and standard deviations from the channel width in the tributaries. For the main channel, however, the regression equations indicated an increase in the mean grain size at the downstream end of the channel which did not agree with the sampling data at the ends of the reach. To quantify the inputs and output of water and sediment, a stochastic model of runoff based on the Diskin-Lane model was used with the Manning open channel flow relation and the Laursen sediment transport relation. The composite model was then tested against Walnut Gulch data for the 36,200 foot channel reach between Flume 6 and Flume 1 and was found to produce synthetic sediment volumes and peak discharges which agree quite well with the prototype data. The importance of the bed material size distribution was demonstrated using several sizes at the upstream end of the channel reach. For example, reducing the mean of the logarithmic transformations of the bed sediment from 1.85 to 1.00 mm changed the sediment balance for the reach from about 7 acre-feet per year of erosion to about the same amount of deposition. This rate represents about one-half of a foot over the channel area for a 10 year period. Because it would involve the combined effects of bank scour and bed scour or deposition, the depth change might be expected to be less. In reality, however, the bed composition probably reacts and changes in response to the flow sequences and the availability of sediment from tributaries. Because of complex conditions across a typical channel section, changes in bed form and hydraulic roughness temporally and spatially are important when predicting the channel behavior

Mots clés : Bed load — Measurement. ; Sediment transport — Mathematical models. ; Ephemeral streams — Mathematical models.


Version intégrale (5,35 Mb)

Page publiée le 27 avril 2016, mise à jour le 23 mars 2018