Formation of riffle—pool sequences: field evidence for an autogenetic process

Abstract

Riffle—pool sequences are the characteristic reach-scale bedforms of mixed- and gravel-bedded rivers. While most research has been directed to the quantitative description of these sequences and to their maintenance, comparatively little attention has been given to the processes responsible for their formation. In this paper, field results from quasi-continuous velocity records obtained with an array of electro-magnetic current meters are analysed with respect to a model of macro-scale structure in the turbulent velocity field of natural channels. This model has been suggested as the most plausible basis for riffle—pool formation. The results demonstrate the occurrence of a variety of coherent flow structures, some of which scale on channel dimensions as predicted by the model, but which also suggest that some revision of the association between flow structure and riffle—pool formation is necessary. It is suggested that riffle—pool units are initiated with the generation of roller eddies upstream and downstream from a major flow obstacle corresponding to engineering criteria for scour around bridge piers. The obstacle persists long enough to fix the flow pattern and hence account for significant modification of channel form, but is ultimately removed as part of the process of bed modification via the extension of scour. Three distinct stages in the process are involved: local scour of a single pool creates deposition downstream, which then generates the next-downstream flow irregularity. A riffle—pool sequence is thus created autogenetically as the summation of a sequence of irregularities, each unit of which is formed and maintained only by flow dynamics operating at the local scale. The process is essentially both deterministic and statistical in character. It is further suggested that spatial differences in the near-bed turbulence field arising from incipient riffle—pool topography themselves create differences in surface sediment entrainment which enhance and maintain the sequence in a form process feedback mechanism. Similar reasoning might possibly be extended to the entire range of bedform sizes.