Amazon River and floodplain hydrodynamics

The annual monomodal flood pulse of the Amazon River is a key driver for this globally important hydrological system. Understanding the behaviour and characteristics of this flood wave and its influence on the dynamics of river and floodplain interaction is important to many studies attempting to quantify processes dependent upon it, including estimates of carbon fluxes from the wetlands and sediment movement to the ocean. The main aim of the research presented in this thesis was to investigate the hydrodynamic behaviour of the Amazon River and its floodplain, in order to better understand its components and processes. The key scientific question underpinning this aim is: How do the floodplain storage and fluxes affect the passage of the Amazon flood wave and in turn how does the passage of the flood wave control the dynamics on the floodplain? Research was carried out using a synthesis of hydraulic characterisation, numerical modelling, spatial analysis of remote sensing data and field surveys. In combination with information from published floodplain studies, these results were then used to inform the development of a conceptual hydrodynamics framework for the Amazon floodplain. Hydraulic characterisation of the Amazon flood wave was undertaken showing that it is subcritical and diffusive in nature with dominating backwater effects. Experiments with the main channel using hydraulic models showed that main channel water levels were relatively insensitive (0.5 m error on a 12 m flood wave amplitude) to the geometric representation of the channel and that simplified bathymetry is adequate for calibrated predictions, provided the mean cross sectional area can be reasonably well approximated. Hydraulic model results compare well with observed data despite explicit exclusion of the floodplain, indicating that the storage volume of the floodplain has a minimal affect on the passage of the Amazon flood wave for this reach. The spatial analysis of Landsat TM images shows a total of 1,762 floodplain channels in the study area with a mean width of 47 m. Shuttle radar topography mission data has difficulty in resolving many of these floodplain channels due to 96% of the channels having a width less than the SRTM spatial resolution of 90 m. Comparing floodplain channel widths with their frequency reveals a power law relationship, showing patterns of structure that are self-similar over many orders of magnitude. Analysis of the floodplain channel network connectivity showed that the complex floodplain can be divided into floodplain hydrologic units (FHU) and that each unit type has different geomorphic characteristics resulting from a different mixture of water inputs, with each unit isolated from other units for much of the flood cycle by sediment barriers. A field survey of the floodplain channels was conducted in order to measure and characterise their morphology. The findings of this survey together with numerical connectivity experiments, demonstrate that floodplain channels could be playing a much more significant role in the floodplain hydrodynamics then previously acknowledged. Floodplain channels were grouped into three types by depth: (i) for channels carrying only river flood water, depth was strongly correlated with the flood wave’s mean annual range of 11.4 m; (ii) for channels which carried river flood water and local runoff, the mean depth was substantially deeper at 15.9 m; and, (iii) main river island channels were deeper again, with a mean depth of 17.7 m. Finally, a coherent, evidence based, conceptual floodplain hydrodynamics framework was developed to inform a broad range of future Amazon wetland research. The framework represents an extension to the macro level understanding available in the scientific literature and provides an important link between the macro and detailed level study scale.