The erosion of headwater wetlands in the Sand River catchment, in the lowveld of north-eastern South Africa has led to a focus on their rehabilitation, both for livelihood security for those that use them for subsistence agriculture, as well as for provision of streamflow regulation services for the Sand River itself. One such wetland, the Craigieburn-Manalana itself undergoing severe erosion was subject to technical rehabilitation using concrete weirs and gabion dams to stabilize the erosion gullies during 2007. Through a series of papers the research discussed in this thesis examined the response of the wetland?s hydrodynamics to the implementation of these measures. Through the installation of a network of hydrometric apparatus the research has shown that the wetlands hydrology is largely controlled by the presence of both horizontal and vertical clay aquicludes within a hydraulically conductive sandy matrix. The sequence of these aquicludes had allowed for artesian phreatic surface phenomena identified in a relatively hydrologically intact region of the wetland. The gully erosion had initiated hydraulic drawdown of the wetland?s water table leading to the desiccation of the system. The construction of a buttress weir within the erosion gully had restored the wetlands hydrodynamics to that typical of conditions upstream of a clay-plug. The research also explored the role that clay plays in terms of controlling the wetland?s hydro-geomorphic setting through geophysical analysis. A conceptual model was then derived that states that these wetlands are held in place by clay-plugs that form through clay illuviation from the hillslopes at regions of valley confinement. This has important implications for the connectivity of wetland process domains. The research also determined the inputs of surface and subsurface flows to the wetland and it was found through detailed examination of soil moisture responses and variably saturated soil physics modelling using the HYDRUS model, that the wetland is hydrologically connected to its contributing hillslope by threshold induced preferential flow pathways, via macropores, that only respond after specific antecedent soil moisture conditions are met. In addition, the thesis describes novel approaches to use information provided by soil scientists for the development of catchment hydrological models. It was shown that the use of this hydropedology information improved the low flow response function of the catchment model, ACRU. This development has important implications for up-scaling of catchment process domains, or hydrological response units by being able to generalize on hillslope hydrological responses based on configuration of their soil type elements. The research also undertook to examine the role that the wetlands play in catchment processes. It was found through water budgeting, supported by hydrological time-series, stable isotope analysis and the quantification of vegetation water use within the wetland and contributing catchment, that these wetlands do not augment baseflows during the dry season. Furthermore, it is only early on during the wet season that these systems may attenuate peak flows, thereafter they act as conduits for high storm flows. Similarities emanated from this research with previous hydrological studies of headwater wetland systems in southern Africa and these are discussed.

Thesis (Ph.D.)-University of KwaZulu-Natal, Pietermaritzburg, 2011.