No description available.

The dry valley systems in the Kalahari of southern Africa are traditionally considered to have developed as a result of past fluvial activity. However, it has also been suggested that erosion by groundwater processes (sapping and deep-weathering) had an important role in development. This thesis aims to establish the relative role of each of these process areas in mekgacha evolution using a combined geological and geomorphological approach. The study area is restricted to the valley systems of Botswana, eastern Namibia and the Northern Cape Province of South Africa, which can be subdivided into exorcic and endoreic systems directed towards the Orange River and the continental interior, respectively. Field studies, analyses of remotely-sensed imagery and a consideration of network orientation identify evidence for the role of both fluvial and groundwater processes in valley development. However, whilst both groups of processes have operated, the importance of each is suggested to have varied both spatially and temporally. Fluvial processes are indicated by sequences of sediments, relict channels and terrace levels, and appear to have operated most recently. Sapping processes are implied in the formation of certain valley systems, primarily from morphological properties and the presence of relict spring lines. Deep-weathering processes are implicated from the close parallelism of many valleys with geological structures now buried beneath thicknesses of Kalahari Group sediments. Borehole records also indicate deep-weathering of bedrock beneath valleys developed above fracture zones, which is suggested to have operated over the longest timescales. Thin-section studies of duricrusts from valley flanks, together with duricrust profiles and records from lithological boreholes, indicate the role of groundwater in their formation. Results suggest an intrinsic link between duricrust formation and valley development. Geochemical and thin-section analyses of duricrusts further suggest that previous considerations of the palaeoenvironmental significance of Kalahari silcretes based upon TiO2 levels may not be wholly appropriate.

Techn. Hochsch., Diss., 2005--Aachen.

Hillslopes comprise the majority of unglaciated upland landscapes; they are the primary source for the production of sediment from bedrock, and the routing system by which sediment is delivered to the channel network. Yet the nature of hillslope response to changes in tectonic, climatic or base-level boundary conditions is poorly understood in terms of the spatial and temporal distribution of hillslope morphology. Here I exploit a previously published framework for exploring hillslope morphology in high relief landscapes (Roering et al., 2007), to address several critical questions: Does high resolution topography allow understanding of the processes and rates by which sediment is redistributed on hillslopes? If so, can hillslope morphology be used to map the spatial distribution of erosion rates and facilitate interpretation of the timing and magnitude of tectonic forcing, particularly in transient landscapes which are adjusting their erosion rates? And to what extent does variation in lithology influence hillslope evolution and morphology, and the ability to interpret process rates from hillslope form? In this thesis I sought to explain hillslope adjustment to changing boundary conditions through combining the predictions of analytical and numerical models with detailed analysis of real, high resolution topographic datasets (derived from LiDAR), focusing on two landscapes where the influence of tectonic forcing on base-level history is relatively well constrained, the Middle Fork Feather River in the northern Sierra Nevada, and the Dragon’s Back Pressure Ridge, on the Carrizo Plain, both in California. The Sierra Nevada of California is a west-tilted fault block composed primarily of granitoids formed during Mesozoic arc volcanism. The block underwent acceleration in uplift 5 - 3.5 Ma which is hypothesised to be caused be the drop-off of a dense root from the lower crust and replacement by hot asthenosphere, causing crustal buoyancy. A relict landscape has thus been uplifted and dissected by the major drainage routes crossing the range, which have eroded rapidly to form deep canyons. The fluvial network is characterised by breaks in slope (knickpoints) which migrate into the landscape to transmit the signal of increased erosion, setting baselevel conditions for adjacent hillslopes. Theoretical predictions for the morphology of hillslopes governed by a nonlinear sediment transport law, if the hillslopes have attained steady state (i.e. they are eroding in concert with base-level fall in adjacent valleys) reveal that the curvature of hilltops will be linearly proportional to erosion rates or rate of base-level fall. I present innovative techniques to extract hilltop networks and sample their adjacent hillslopes in order to test the utility of hilltop curvature for estimating erosion rates. This work is carried out in granitoid lithologies where the influence of bedrock heterogeneity is assumed no to be a first order control on hillslope morphology. Existing and new cosmogenic radionuclide analyses in the Feather River basin, California, suggest that erosion rates vary by over an order of magnitude from the remnant upland landscape to the incised river canyon. Hilltop curvature increases with erosion rates, allowing calibration of the hillslope sediment transport coefficient, which controls the relationship between hillslope gradient and sediment flux. This in turn allows the estimation of erosion rates throughout the landscape by mapping the spatial distribution of hilltop curvature. Additionally, despite the landscape containing gradient-limited hillslopes, hilltop curvature continues to increase with rising erosion rates, reflecting higher erosion rates than can be predicted by hillslope gradient. The distribution of hillslope morphology conforms well to predictions of a nonlinear sediment transport model, with measured values of hillslope relief varying with the product of hilltop curvature and hillslope length (proxy for erosion rate) in a manner similar to that predicted by Roering et al. (2007). Hilltop curvature can thus be used to estimate erosion rates in landscapes undergoing a transient adjustment to changing boundary conditions provided that the response timescale of hillslopes is short relative to channels. Having focused on a landscape with roughly uniform bedrock geology to isolate drivers of geomorphic change, I sought to evaluate whether these techniques could be extended across lithologic contacts and throughout the landscape. Underlying geology influences the efficacy of soil production and transport on hillslopes, and resistance to erosion by valley-forming processes. Here, quantitative analysis of LiDAR digital terrain models was performed to search for a topographic signature in two distinct lithologies in the Feather River catchment in northern California; granodiorite and deformed volcanics. The two sites, separated by <2 km and spanning similar elevations, are assumed to have similar climatic and denudation histories. Responding to increased erosion rates, transient hillslopes exhibit high gradient but low hilltop curvature in the metavolcanics relative to theoretical predictions for steady state hillslopes. However, hillslopes in the granodiorite have, for the most part, variation in hilltop curvature, hillslope length and hillslope relief similar to model predictions for steady state hillslopes. The curvature of hilltops adjacent to main stem channels implies that the coefficient of sediment transport is two times larger in the granodiorite (c. 8.8 m2 ka-1) than in the metavolcanics (c. 4.8 m2 ka-1). The data suggest that hillslopes get shorter as erosion rates increase due to the increased influence of debris flows in valley incision, suggesting that drainage density increases with erosion rate. The incision wave associated with more rapid erosion in the Feather River has propagated further into a basin developed on the metavolcanics and hence this substrate is less resistant to channel incision. I review an inventory of values for the transport coefficient for hillslope sediment transport but find that no clear patterns emerge with varying lithology. However in unconsolidated substrates, precipitation may play an important role in modulating sediment transport through variation in rain splash impact frequency and the frequency of wetting/drying, freeze/thaw, and expansion/contraction cycles. Finally I apply the same techniques to study hillslope morphology to a landscape where the tectonic history has a documented influence on landscape development. The Dragon’s Back pressure ridge, Carrizo Plain, CA, consists of a series of small catchments adjacent to the San Andreas fault, where previous detailed geologic mapping has allowed the spatial and temporal distribution of uplift to be constrained. This landscape offers a hitherto unique opportunity to study the temporal evolution of hillslope morphology via ergodic substitution. I show that the time evolution of a sensitive indicator of erosion rate, hilltop curvature, can be predicted using a nonlinear sediment flux law. Further to this, the temporal evolution of relief and hilltop curvature experiences hysteresis as the landscape grows and decays. Relative to steady-state predictions, hillslope morphologies exhibit higher than expected values for relief during active uplift or landscape growth, and lower than expected relief during landscape decay. Therefore landscapes growing due to fault activity can be distinguished from those with quiescent faults undergoing topographic decay.

Sediment erosion is a major challenge for run-of-river power plants, especially during flood periods. Due to the high content of hard minerals such as quartz and feldspar carried in the river, substantial damage is observed on the turbine components. Material is gradually removed, thus the efficiency of the turbine decreases and the operating time of the turbine reduces. Hydro power plants situated in areas with high sediment concentration suffer under hard conditions, where turbine components could be worn out after only a short period of three months. This short life expectation causes trouble for energy production since the replacement of new turbine parts is a time consuming and costly procedure. It is desirable to design a Francis runner which will withstand sediment erosion better than the traditional designs. The literature states that an expression for erosion is velocity to the power of three. By reducing the relative velocities in the runner by 10%, the erosion will decrease almost 30%. The objective is to improve the design of a Francis turbine which operates in rivers with high sediment concentration, by looking at the design parameters in order to reduce erosion wear. A Francis turbine design tool was developed to accomplish the parameter study. In the search for an optimized Francis runner, several design proposals were compared against a reference design by evaluating the turbine’s performance. The hydraulic flow conditions and the prediction of erosion on the turbine components are simulated by analyzing the models with a Computational Fluid Dynamic (CFD) tool. A Fluid Structure Interaction (FSI) analysis ensures that the structural integrity of the design is within a desired value. Results from this research show that it is feasible to design a runner with an extended lifetime, without affecting the main dimensions and hydraulic efficiency.