Articles | Volume 1, issue 3
30 Sep 1997
30 Sep 1997

Bank erosion events and processes in the Upper Severn basin

D. M. Lawler, J. Couperthwaite, L. J. Bull, and N. M. Harris

Abstract. This paper examines river bank retreat rates, individual erosion events, and the processes that drive them in the Upper Severn basin, mid-Wales, UK. Traditional erosion pin networks were used to deliver information on patterns of downstream change in erosion rates. In addition, the novel automatic Photo-Electronic Erosion Pin (PEEP) monitoring system was deployed to generate near-continuous data on the temporal distribution of bank erosion and accretion: this allowed focus on the magnitude and timing of individual erosional and depositional events in relation to specific flow episodes. Erosion dynamics data from throughout the Upper Severn basin are combined with detailed information on bank material properties and spatial change in channel hydraulics derived from direct field survey, to assess the relationships between flow properties and bank erosion rates.
Results show that bank erosion rates generally increase downstream, but relate more strongly to discharge than to reach-mean shear stress, which peaks near the basin head. Downstream changes in erosion mechanisms and boundary materials, across the upland/lowland transition (especially the degree of development of composite bank material profiles), are especially significant. Examples of sequences of bank erosion events show how the PEEP system can (a) quantify the impact of individual, rather than aggregated, forcing events, (b) reveal the full complexity of bank response to given driving agents, including delayed erosion events, and (c) establish hypotheses of process-control in bank erosion systems. These findings have important implications for the way in which bank erosion problems are researched and managed. The complex responses demonstrated have special significance for the way in which bank processes and channel-margin sediment injections should be handled in river dynamics models.