Morphodynamics and driftwood dispersal in braided rivers

Driftwood is widely recognized as a relevant component of riverine systems due to its complex interactions with flow, sediment transport and vegetation dynamics. In-channel large wood has a relevant geomorphic and ecological role as it enhances morphological diversity and creates a variety of physical habitats that sustain high biodiversity. Its presence can also increase flood risk and therefore wood is often removed from streams especially in densely populated areas. Recent river restoration policies aim to maximise the environmental benefits of driftwood and minimise risks. The study of wood dynamics can provide useful information to define guidelines for sustainable wood management. Multi-thread systems represent a particularly interesting and challenging context for the investigation of wood dynamics because of their complex geometry, the presence of vegetated islands and the frequent, intense changes in channel pattern observed even for moderate discharge fluctuations. However, comparatively few studies focus on driftwood in large braided rivers and limited quantitative information is available on wood transport, deposition and remobilisation in these systems. The goals of the present work are: a) characterising the spatial organisation of wood deposits and identifying typical retention sites and styles; b) analysing the influence of flow regime, channel morphology, wood supply and log properties (size and shape) on dispersal patterns; c) investigating wood remobilisation induced by discharge fluctuations and bed reworking; and d) analysing long-term wood storage volume and budget. A combination of field-scale direct observations, remote sensing techniques and physical modelling was used to investigate wood and channel dynamics. Field-scale monitoring carried out on the Tagliamento River (Italy) allowed the observation of complex interactions and feedbacks between channel, vegetation and wood dynamics. Laboratory simulations – carried out in two large flumes at the University of Trento (Italy) and at the University of Hull (UK) – were employed to investigate individual wood dispersal mechanisms under controlled conditions and to explore the role of governing parameters. In large rivers, floods are the primary driver of wood recruitment through the erosion of vegetated banks and islands; field-scale observations showed that these localised wood inputs control wood storage at sub-reach scale because a large proportion of eroded trees is retained close to the input point in sparse, small jams. Physical modelling highlighted a complex relationships between flow stage and the longitudinal and vertical distribution of wood; high discharge increases the ability of the system to transfer wood, but at the same time generates complex inundation patterns where a larger number of sites are available for wood retention. No clear link between flow stage and the vertical distribution of wood is observed, probably because water surface elevation exhibits small changes with discharge in flat braided river cross-sections. Driftwood element properties also influence deposition patterns; log diameter controls travel distance as it governs flotation and therefore the likeliness of deposition. High element length and complex piece shape sustain the formation of large jams. The presence of a root bole is also associated to short travel distance and low relative elevation. At reach-scale, the spatial distribution of wood is the product of local inputs during major floods and reorganisation of deposits induced by minor events. Wood pattern exhibits a threshold behaviour with supply. High input rates determine very high spatial density and the formation of large, stable jams. Two processes govern wood reorganisation over different time scales, namely network inundation – inducing rapid changes in flow field – and bed reworking. In the first case, the persistence of deposits depends on the magnitude of discharge fluctuations as wood dispersed by small floods is easily removed by larger events. High relative elevation and large jam size enhance wood stability, while the presence of a root wad has a dual effect as it determines large accumulations at low elevation. Channel pattern reworking determines intense turnover of driftwood deposits regardless of supply rate, piece properties and jam size, save for very large accumulations. As a consequence, wood deposition occurs mostly on empty braidplain areas as opposed to pre-existing sites. These results suggests that wood (alone) has little direct effect on reach-scale bed geometry in a large braided river; however, deposited wood significantly influences local hydraulics and morphology, enhancing physical habitat diversity. Moreover, deposited wood favours the accumulation of fine sediment, nutrients and seeds and often exhibits vegetative regeneration. These processes lead to the transformation of instable driftwood pieces into vegetated islands, which in turn can trap more wood. Therefore, wood has a relevant, indirect effect on braided river morphodynamics through the establishment of vegetation, whose presence influences network complexity and evolution.