Quelques Aspects Du Transport Des Alluvions Grossieres Et Du Façonnement Des Lits Fluviaux

Click to increase image sizeClick to decrease image size

Abstract

The Centre de Géographie Appliquéeh as used a new approach in the field of fluvial dynamics: a combination of field observations, including accurate mapping (scales of 1/10,000 & 1/5,000), and sedimentological studies. The following conclusions have been reached:

1)	Pebbles suffer only a very slow transportation and the “traffic” is low (in t/km/year). Thus, abrasion and chemical weathering during transportation are very important and explain both the rounding and the petrographic spectras of the material. When pebble discharge is not too high, pebbles disappear in the river bed without any fan aggradation, by weathering. But pebbles, as a consequence of their slow movement, are an obstacle to the flow of water and play a determinant part in the bed morphology. Normally, their vanishing along main streams, principally in piedmont regions, originates the following sequence of bed types: braided channels, alluvial meanders, stable channels.
2)	Braided channels and meanders are typical of rivers with a relatively high bed load, either pebbles or sand. Various rivers present alternatively both bed types, geographically and chronologically. For instance, the Adour river, where bed evolution has been reconstructed over 200 years, shows an alternance of both types with time in certain stretches and an alternance of stretches with meanders and other with braided channels. It seems that both granulometry of the bank material and irregularity of discharge are involved. Bank sapping occurs during discharge increase before flooding. A quick increase of discharge gives more violent sapping and originates more abundant bed load and braided channels Meanders develop when these conditions are less intense and when the material yielded by sapping is nearly equivalent to the competence. Under such conditions, its transport can occur only under short distance.
3)	Granulometrie curves of sands, drawn on semilogaritmic paper, belong to 3 basic types: logarithmic (transport without selection, the percentages of the various sizes being governed by a probability law), sigmoidal (selection in a narrow gap between greater sizes equal to competence and smaller ones for which transportation remains possible) and parabolic(exponential law for deposition, as the result of a positive retroaction). Logarithmic curves are characteristic of mass transport, sigmoidal of free accumulations as the result of changes in potential energy of the stream, and parabolic, of impeded transportation (base-level accumulations). Competence seems to be the main hydraulic factor, and not the total alluvial charge, as frequently admitted.
4)	Competence has been underestimated for the explanation of fluvial dynamics, because it cannot be studied experimentally with flumes, where the similitude laws can be respected only when using artificial materials. In fact, as proved by soil mechanics practice since half a century, water and fine material can be mixed in any proportion. Properties of the mixture don't change abruptly and the limits of Atterberg are conventional. When above plasticity limit, the mixture remains in liquid condition whatever the mud content, and is able to flow. The concept of charge-limit, so frequent among geomorphologists, must be discarded, every time we have a fine fraction under transportation. This is the reason for which we can observe a complete transitional series of flow types from torrential clear water to mud flow. In such a series, when suspended matter content increases, velocity of flow diminishes but density increases and, as a consequence, comptetence. Mudflows, with a low total discharge, are able to transport huge blocks and have a tremendous competence, somewhat similar to that of glaciers. Their slow flowage broadens torrential beds where they occur.

It seems that the gradual changes of parameters in the flow characteristics with an increase in mud content of water can be studied only through an analysis of viscosity. Viscosity classes must be defined in a conventional scale, as it was made by Atterberg with plasticity and liquidity limits for soil mechanics. Such a classification is, a necessary basis in order to define limits between clear water dynamics, mud torrents, torrential lavas, mud flows etc…