Flood variability recorded by crevasse-splay sedimentation of large river floodplains

Wednesday, 16 December 2015
Poster Hall (Moscone South)
Christopher R Hackney1, Stephen E Darby2, Rolf E Aalto3, Daniel R Parsons4, Alexander Clayton1, Arved Schwendel5, Julian Leyland1, Andrew Paul Nicholas6 and Jim Best7, (1)University of Southampton, Geography and Environment, Southampton, United Kingdom, (2)University of Southampton, Southampton, SO14, United Kingdom, (3)University of Exeter, Geography, Exeter, EX4, United Kingdom, (4)University of Hull, Hull, United Kingdom, (5)University of Exeter, Geography, Exeter, United Kingdom, (6)University of Exeter, Exeter, United Kingdom, (7)University of Illinois at Urbana Champaign, Geography, Mechanical Science and Engineering and Ven Te Chow Hydrosystems Laboratory, Urbana, IL, United States
During rapid rise flood events, crevasse-splay complexes are a key conduit through which sediment and water are passed from the main channel onto the floodplain, particularly for large rivers. These crevasse-splay systems may therefore be key loci of floodplain sedimentation, as well as conditioning the location of avulsions. As such they may preserve climatic signals within the depositional record of large alluvial rivers. Despite recent advances in our capabilities to model the development and evolution of these systems, our understanding of the passage, storage and reworking of water and sediment across them remains relatively poor. A key limitation concerns the point that, since floodplain topography is a first-order control on the hydrodynamics of crevasse-splays, publicly available topographic data sets (e.g. SRTM, ASTER) are currently unable to resolve key processes at the necessary spatial resolution.

Here we employ Structure-from-Motion (SfM) on low-level aerial photography to obtain high-resolution (3m grid cell) georectified topographic data (horizontal error = 0.02 m; vertical error = 0.2 m) for a representative crevasse-splay complex (27 km2) located along the Mekong River, Cambodia. We use the coupled hydrodynamic and morphodynamic model, Delft-3D to simulate sedimentation patterns for a series of idealised and observed rapid-rise flood events. We model floodplain deposition and erosion and validate simulated spatial and temporal variations against observed patterns of sedimentation determined through analysis of 210Pb geochronology of a network of floodplain cores and in-situ, post-event surface scrapings.