The work performed in this project fell into three main areas; (i) the development and application of an experimental flume facility for the study of bedrock channel and knickpoint erosion processes, (ii) numerical modelling to identify and explore the hydraulic-geomorphic interactions that can lead to rapid change in river systems, and (iii) study of a natural landscape (New Zealand) that has recently undergone rapid erosion by knickpoint processes following a change in climate.
The experimental flume facility was developed at the start of the project and a successful laboratory campaign was completed that considered the relative effect of the following factors on knickpoint erosion in a controlled environment: bedrock strength, water discharge and sediment flux. Additional experiments considered the role of sediment flux in setting bedrock channel width, and the role of sediment flux and slope in driving the transition of a channel from fully alluvial to incised bedrock conditions. Data showing the evolution of the topography during the experiments was collected at two-minute intervals using a terrestrial laser scanner (TLS), and were analysed to calculate the rates of knickpoint retreat, the evolution of the river long profiles and the evolution of channel width and slope.
Results from the experimental part of the project have been published in two peer-reviewed journals. The first, in Scientific Reports, identified that the strength of the bedrock material is the key factor in setting the retreat rate of knickpoints. Importantly, these experiments also showed that the river discharge does not control the knickpoint retreat rate, which is contrary to existing theory and many common landscape evolution modelling studies. River channels adapt their width to the river discharge (wider channels under higher discharges) which maintains a constant shear stress acting on the bed and constant erosion rate at the knickpoints. In the second publication, published in Geology, the experiments identified the physical process that leads to the transition of river channels from a wide, fully-alluviated state, to narrow, deeply incised bedrock conditions. Critically, the results identified a multi-stage erosion process that included both a downstream and upstream incision wave component. The erosion into bedrock and the onset of the upstream incision wave is triggered under supercritical flow conditions, a geomorphic-hydraulic interaction that had never been previously identified in large-scale rivers.
The hydraulic-geomorphic interaction leading to bedrock incision and knickpoint development was identified using the output from a numerical hydrodynamic model. This allowed a detailed investigation of the previously impossible to recognise role of interactions between channel hydraulics and channel form in determining the evolution of the system. Further investigation using the numerical model has validated the dynamics of water flow in plunge pools at the base of waterfalls, and work is ongoing to implement an erosion law for the evolution and retreat of knickpoints into the morphodynamic part of the model.
The results from the Rangitikei river area of New Zealand support the experimental results from part 1 of the project. The region is a unique natural laboratory for the study of the interactions between sediment flux, bedrock river morphology and knickpoint dynamics, due to the uniform lithology of the region, with the exception of the headwater tributaries where some supply hard, coarse, gravel to the river channel. As a result, some rivers in the region have high sediment fluxes while others have no coarse sediment flux, allowing a spatial comparison of the impact of sediment on channel geometry and also the rate and style of knickpoint erosion. Rivers containing coarse sediment are systematically wider than those without coarse sediment. A paper exploring the role of sediment flux in setting channel width, combining data from experiments and the Rangitikei region is currently being prepared for publication. Results combining experimental data from part 1 and data from the Rangitikei region will also be published showing that sediment has an important role in setting the morphology and retreat rate of waterfalls. Waterfalls in channels containing sediment, either experimental or in a natural system, are characterised by a series of steps in a narrow incised canyon while those without sediment are characterised by a single vertical drop in a wider, open, canyon. These results provide insights into important factors controlling both the morphology of bedrock river channels, the complexities of waterfall/knickpoint erosion processes and the drivers of rapid landscape change over multiple spatial and temporal scales.
