Bank erosion in regulated navigable rivers

Banks constitute important areas for the river ecology since they provide a multitude of favourable conditions for flora and fauna. The hydromorphological diversity typical of these transitional zones between water and land, and the associated processes of erosion and accretion, make riverbanks vital for many aquatic and riparian plants and animals. In recent decades, the increasing awareness of the ecological significance of rivers and water bodies resulted in the gradual implementation of extensive stream, river and floodplain restoration. In the EU, these practices are regulated by the Water Framework Directive. An important and largely applied re-naturalization measure in highly trained watercourses is the removal of bank protections to reactivate erosion processes and promote habitat diversity. In rivers used as waterways, ship waves can be an important cause of bank erosion and ecological disturbance. The sediment yield from bank erosion may alter navigable depths, the water quality, and flood conveyance, for which enhancing the hydromorphology is a challenge in multifunctional rivers. Due to pressing needs to improve riverine habitats, large-scale restoration works have been implemented based on conceptual schemes without a comprehensive knowledge of wave erosion processes or a precise estimate of long-term bank retreat. The Meuse River in the Netherlands constitutes a remarkable example of systematic rehabilitation, where bank protections have been removed along 100 km between 2008 and 2020.Given that ship-induced erosion is still poorly understood, the management of navigable rivers and the planning of restoration measures would benefit from a solid and deeper understanding of natural bank dynamics induced by ship waves, for both economic and ecological reasons. Moreover, more precise estimates of long-term bank retreat would help to optimize different functions and reduce conflicts of interest within the river system. Therefore, the main objective of this investigation is to understand and predict erosion processes and the morphological evolution of natural banks in regulated navigable rivers. The research goal is pursued through the thorough investigation of a river reach that presents a wide range of erosion rates after the removal of bank protections. This main case study consists of a 1.2-km straight reach in the Meuse River, near Oeffelt in the Netherlands, the left bank of which was re-naturalized in 2010 by extracting the riprap. The Meuse is a midsize river with a pluvial regime, which has been canalized and is regulated with a series of weirs to enable navigation. Here, field techniques and complementary laboratory tests are utilized including topographic surveys with UAV, wave measurements with ADV, soil coring, geotechnical tests, and RTK GPS profiling. Processing and analysis of data are carried out with MATLAB. Four research steps are conducted. First, a methodology to quickly survey the 3D bank topography along a midsize river reach is determined to measure bank erosion processes. Second, distinct patterns of bank erosion that appeared along the Meuse River after protection removal are investigated. The aim is to disentangle the causes of the size, location and asymmetry of large embayments before analysing erosion processes at single river sections. Third, bank erosion processes in regulated navigable rivers are characterized and conceptualized. Fourth, a tool to estimate long-term or final retreat of re-naturalized banks in regulated navigable rivers is developed. The results of the first research component show that structure from motion photogrammetry applied to photos taken from an UAV is a practical and accurate method to measure riverbank erosion. By distributing ground-control points sufficiently spaced from the bank into the floodplain, digital surface models are georeferenced with sufficient accuracy to compare bank profiles between successive surveys. The identification of ground-control points in photographs is facilitated by placing oblique plaques on the floodplain, reducing the need for another perspective along banks. A single UAV flight with an oblique perspective of the bank becomes then sufficient to capture its three-dimensional complexity. Eight overlaps among consecutive images is the minimum number not to reduce the precision potential of a single UAV flight. The proposed methodology is fast to deploy in the field and surveys reach-scale riverbanks in sufficient resolution and accuracy to quantify bank retreat and identify morphological features of the complete erosion cycle, which enables the characterization of bank erosion at the process scale. Second, the oblique orientation of heterogeneous sedimentary strata with respect to the canalized Meuse River alignment explains the formation and asymmetry of large embayments. Depositional layers of varying compositions, structured by scroll-bar formation during former river meandering, led to wide-ranging erosion rates within a relatively short reach, which formed distinct bankline patterns across diverse lithologies and above the controlled water level of the river. The frequent occurrence of this water level and the persistent ship wave attack shaped bank profiles of varying strengths with a mild sloping terrace. The presence of isolated trees on the floodplain only locally delay erosion rates. Bank retreat rates at single cross sections primarily depend on the lithology near the minimum regulated water stage. Third, the evolution of bank profiles revealed the active role of ship waves in erosion progression, even at well-developed terraces. Currents initially contribute to all phases of the erosion cycle, but they gradually exert less shear stresses on the upper bank as the terrace elongates. Their later role at intermediate stages of development is reduced to the destabilization of steep high banks through water level fluctuations, without capacity to transport slump blocks. The resistance to erosion of the bank lithology defines the terrace geometrical proportions and the pace of morphological evolution of bank profiles. For instance, at a given time after protection removal, less cohesive banks can be present at intermediate stages of development while more cohesive banks remain at early stages. The latter present shorter and shallower terraces whereas the opposite holds for the former. Vegetation temporarily protects the upper bank from failure and toe erosion, but its permanence is subject to terrace stability and effectiveness to dissipate waves. Biofilms are able to partially cover well-developed terraces, changing entrainment thresholds. Fourth, based on the above conceptual framework of bank profile evolution, a model was developed which captures the observed non-linear morphodynamics driven by ship waves in regulated settings. This new tool estimates long-term retreat by accounting for the main erosion drivers and essential mechanisms. Equilibrium bank profiles are reached once wave-induced shear stresses fall below the threshold for entrainment of cohesive soils. Unlike previous models of ship-induced erosion, the process-based approach enables to distinguish the contribution of each factor to erosion. Primary waves are found to exert the highest loads on the terrace, shaping long-term profiles and defining ultimate retreat. To apply the model, it is necessary to measure or estimate the largest primary wave and the soil cohesion at the controlled level, preferably in the range -1.00 m to +0.50 m with respect to it. The above findings are based on cohesive banks in a straight reach of a regulated river. The presence of gravel layers in the bank changes the morphological response to ship waves due to the armouring of lower strata. In such cases, the bank terrace can reach a transverse slope in dynamic equilibrium defined by grain size, as long as longitudinal currents do not transport the gravel to the lower bank. The lower non-cohesive layer of composite banks responds in a similar way, eventually reaching a dynamic equilibrium, after which a final retreat of the upper cohesive layer is possible. The position of banks in the river planform affects the magnitude and duration of the contribution of currents to upper bank erosion. Their direct impact, especially during high floods, can dominate bank retreat during long periods if the flow is persistently steered against the upper bank, as at outer bends. Unregulated rivers present higher shear stresses than those with controlled stages. Their sandy strata of composite banks are normally exposed to currents and waves, creating larger morphodynamics and more challenging conditions for vegetation growth. The new model to estimate final retreat of cohesive banks may be used to prepare a reach scale strategy that defines the most convenient approach for stretches with similar morphological behaviour and available space to develop. In this way, the eventual need to reduce or stop erosion at sections with future excess retreat is determined in advance. In order to make the most of re-naturalized banks in terms of their benefits for ecological processes and habitat diversity in navigable rivers, the advantages of shallow areas with less perturbated zones should be sought where possible. Two phases of interventions are recommended, a first phase where ship waves freely reach the bank for terrace creation, responding to local lithologies, and a second phase with lowered erosive loads, facilitated by slightly submerged pre-banks. The latter phase increases the possibilities for vegetation, and likely other living organisms, to develop. The knowledge and tools now available create new possibilities for improved management of re-naturalized banks in navigable rivers. The progress made helps to better understand the contribution of different drivers to bank erosion and to identify which factors control retreat at different bank types, stages of development, and settings. The new insights explain how to apply SfM-UAV to monitor bank erosion processes along river reaches, interpret bankline patterns, assess the role of isolated trees in bank retreat, and manage expectations regarding bank retreat and the role of vegetation to control erosion. The understanding of erosion processes in regulated navigable rivers and the possibility to estimate final erosion magnitudes open future opportunities to analyse the river system from a holistic perspective and to find creative ways to balance diverse river functions.Rivers, Ports, Waterways and Dredging Engineerin