Modelling and understanding knickpoint dynamics in homogenous substrates.

In fluvial environments, knickpoints are key geomorphological landscapes exerting a disproportionate control on landscape evolution. Their influence extends both along the fluvial network and into adjacent hillslopes through coupled channel–hillslope processes. Despite this, previous studies focus on individual elements of knickpoint retreat (retreat rate, height, occurrence rate) with little focus on how they come together to wholly govern knickpoint retreat and the impact this has on wider channel evolution. We present a geometric model to illustrate the ways in which a channel can respond to changing forcing conditions – by changing knickpoint retreat rate, knickpoint spacing or knickpoint height. To evaluate our model, we conducted 22 analogue experiments in the Bedrock River Experimental Incision Tank at the Université de Rennes. To emulate bedrock, we use a silica paste of 45 μm silica, glass beads, and 18% water, with the ratio of silica to glass beads controlling substrate strength. Sediment was fed into the channel (±2% tolerance) via an infinite screw feeder, and base-level fall was simulated by a constant speed motor lowering a movable outlet gate. Conditions were constant throughout each of the experiments with sediment flux ranging from 0 g min^-1 to 30 g min^-1, base-level fall rate from 2.5 cm hr^-1 to 5.0 cm hr^-1 and silica to beads ratio from 2.5:1 (weakest) to 4:1 (strongest). A terrestrial laser scanner with a green water-penetrating laser scanned the bathymetry of the channel every 2 minutes, and the 2mm resolution digital elevation models (gridded point cloud data) used as input topography for the FLOODOS hydrodynamic model. Using Z-score normalised multi-variate regression we find knickpoint spacing is set primarily by base-level fall rate where base-level fall rate has 2.41x the impact of sediment flux and 3.61x the impact of bedrock strength. This results from the rate the channel can go through the cycle of initiating knickpoints. We find knickpoint retreat rate is set almost exclusively by bedrock strength explained by the impact of excess shear stress, with bedrock strength having 2.15x the impact of base-level fall rate and 3.61x the impact of sediment flux. Finally, knickpoint height is found to be set by base-level fall rate with base-level fall rate having 3.92x the impact of sediment flux and 4.59x the impact of bedrock strength. Next, the study looks at the impact of knickpoint dynamics on reach scale width and slope and find that these factors are governed by knickpoint morphology, with implications for channel-hillslope interactions. We find that vertical, steep knickpoints impact channel width and slope on a local scale, compared to long, elongated steepened reaches where knickpoint impacts extend beyond the local scale, reducing both overall channel width and slope. Overall, this study enhances the of understanding holistic knickpoint dynamics by assessing the interplay between multiple factors. This is important due to the broader implications for hillslope processes and landscape evolution resulting from knickpoint migration.