Do floods matter? Bridging the gap between fluvial morphodynamics and alluvial architecture

点击翻译按钮获取中文摘要

Most lowland rivers flow across large floodplain complexes composed of sediment deposited during floods. These floodplains, and the sedimentary deposits of which they are composed, are of considerable environmental significance. For example, floodplain deposits are regularly used to infer the nature and timing of past climate change, or to assess the impact of upstream landscape disturbance by human activity. Furthermore, the balance between processes of sediment deposition and reworking, due to river migration, determines the residence time of sediment in the floodplain. This is critically important for biogeochemical cycling (eg. of Carbon) and for the transport and fate of sediment associated nutrients and contaminants. Numerical models are important tools that are needed to predict the way that floodplains build up over time and recycle sediment, in order to provide quantitative understanding of floodplain functioning in the context of the diverse environmental applications outlined above. However, despite the need for such models, no model currently exists that is capable of representing the processes involved in the construction and evolution of floodplains over the timescales relevant to these applications (decades to millennia). The reason for this is that realistic models of floodplain evolution need to represent the complex behaviour of the floodwaters that control sediment transport and deposition. However, to do this it is necessary to solve the equations of fluid motion, which is time consuming (in computational terms). For this reason, existing models of floodplain construction neglect these hydraulic and hydrologic controls and, consequently, are unable to predict how floodplains evolve in a way that is physically realistic. This project aims to address this fundamental problem by developing & evaluating a new generation of hydraulically-driven approaches to modelling floodplain construction and evolution. The model developed here will be applicable over periods of up to 100,000 years, yet will have at its core a physically-based hydrodynamic model more usually restricted to applications involving individual floods. This will be achieved by using two approaches to reduce model run times: (1) Parallelising the code for implementation using High Performance Computing; and (2) Developing a series of novel methods of parameterising the effects of fine scale floodplain topography to allow the model to be implemented at reduced grid resolutions, thus substantially increasing model efficiency. The key strength of this modelling approach is that it will allow long-term floodplain evolution to be simulated using an approach underpinned by sound fluid dynamics principles. The model will be evaluated using field and remote sensing data collected from an extensive, natural floodplain system, that is unaffected by either human activity or the effects of Holocene sea level change. Model evaluation will be carried out over the past century and over the Holocene. Following this, the model will be used to conduct a series of numerical experiments designed to investigate the relationships between floodplain evolution, sedimentary deposits and environmental conditions (climate, flood regime, sediment supply, flood basin geometry, & tectonic setting). In combination, field evidence and model simulations will provide new quantitative insight into questions concerning floodplain functioning that have never been addressed due to the current lack of a physically-realistic hydraulically-driven model of long-term floodplain evolution.