Дисертації з теми "Tidal basins Mathematical models"

A two-dimensional hydrodynamic model is used to simulate the flow field and the concentration distribution of a conservative tracer in shallow basins. A series of numerical test are performed to evaluate different numerical schemes and problems which arise for the use of the Second Moment Method (SMM) in diffusion dominated flows are reported. The results of the basin simulations are compared with experimental data. The model predicts the location and the size of the dead zones, bypassing, recirculation, and local concentrations within the basin. The positioning of the inlet and outlet, and the presence of baffles are important parameters for the location and size of dead zones. The model gives results which are in agreement with the experimental data. The results show that the hydrodynamic model is quite powerful in terms of predicting correctly the residence time distribution for ponds of various dimensions and shapes.

Semi-distributed hydrological models are often used for streamflow forecasting, hydrological climate change impact assessments, and other applications. In such models, basins are broken up into hydrologic response units (HRUs), which are assumed to have a relatively homogenous response to precipitation. HRUs are delineated in a variety of ways, and the procedure used may impact model performance. HRU delineation procedures have been researched, but it is still not clear how important these subdivision schemes are or which delineation methods are most effective. To start addressing this knowledge gap, this project investigated whether or not HRU size has a significant effect on streamflow simulation at the mouth of a watershed. To test this, 30 gaged, relatively unimpaired western U.S. basins were each modeled with 6 HRU sets of different sizes using the Precipitation Runoff Modeling System (PRMS). To isolate size as a variable, HRUs were delineated using stream catchments. For each basin, streams were defined with 6 different threshold levels, producing HRUs of differing sizes. Nineteen model parameters were derived for each HRU using nationally consistent GIS datasets, and all other model parameters were left at default values. Climate inputs were derived from a national 4-km2 gridded daily climate dataset. After calibration, 4 goodness-of-fit metrics were calculated for daily streamflow for each HRU set. Uncalibrated model performance was generally poor for a variety of reasons, but comparison of the models was still informative. Results for the 30 basins across the 6 HRU size classes showed that HRU size did not significantly impact model performance across all basins. However, in basins that had less total precipitation and higher elevation, sensitivity of model performance to HRU subdivision levels was slightly greater, though not significantly so. Findings indicate that, in most basins, little subdivision may be required for good model performance, allowing for desirable simplicity and fewer degrees of freedom without sacrificing runoff simulation accuracy.

In this thesis, the phenomenon of flow-topography interaction is considered in the context of two dynamically distinct case studies. In the first study, tidally-driven upwelling is investigated usingfield data collected in Hydrographers Passage (20????S), a narrow, navigable channel in the dense outer reef matrix of the southern Great Barrier Reef, Australia. In the second study, island wake formations at Cato Island (155????32????E, 23????15????S) in the deep, Western Coral Sea are examined using a combination of field data and numerical experiments. The result of the Hydrographers Passage study are of considerable scientific interest since they apply to numerous smaller non-navigable reef-edge passages dotted throughout the southern Great Barrier Reef. Strong, semi-diurnal flood tides flowing through a gap in a distal patch reef system at the shelf break generate strong upwelling, providing a pulsed, semi-diurnal input of nutrients to the reefs offshore of the passage. If stable in the long term, this mechanism could have profound evolutionary implications for large reefal areas in the southern Great Barrier Reef. In the second study, two sets of field observations at Cato Island coincided with conditions of strong (~0.7m s-1), vertically sheared incident currents and weaker (~0.3m s-1), more variable incident flows. The combination of dynamically distinct flow regimes and a tall, steep-sided island penetrating oligotrophic surface waters provides a unique opportunity to investigate the impact of island wakes on hydrographic structure and biological enhancement. Field data indicate that flow disturbances downstream of Cato Island are likely to generate biological enhancement during conditions of eddy shedding and non-shedding wakes. A primitive equation numerical model configured on the basis of field observations faithfully reproduces the key features of both data sets; mechanisms responsible for producing these key features are proposed. Previous numerical studies of island wakes have concentrated primarily on eddy shedding flows. In this thesis, the sub-critical (non-shedding) flow scenario is also considered. It is demonstrated that particle retention in island wakes has a ????hair trigger???? characteristic controlled by incident flow speed. This observation leads to a new proposal to explain the long-standing recruitment problem of biological oceanography.

A numerical model has been developed for the study of stratified tidal circulation in Indian Arm - a representative inlet on the southern coast of British Columbia. Equations for horizontal velocity, salt conservation, continuity, density (calculated as a linear function of salinity), and the hydrostatic approximation govern the dynamics. All equations have been laterally integrated under the assumption of negligible cross-inlet variability. The model is time dependent and includes nonlinear advective terms, horizontal and vertical turbulent diffusion of salt and momentum, and variations in width and depth. Provisions for surface wind stress and a flux of fresh water are also included, although neither was utilized in this study. An explicit finite difference scheme centred in both time and space was used to solve for the horizontal and vertical velocity components, salinity, and surface elevation on a staggered rectangular grid. A backward Euler scheme was used to suppress the computational mode. Tests using a semi-implicit scheme to solve the finite difference equations over realistic topography led to numerical instabilities at modest values of the time step - in spite of the unconditional stability criteria - suggesting that linear stability analysis may give misleading results for strongly nonlinear systems. Surface elevations calculated from tidal harmonic analysis and salinity timeseries derived from linearly interpolated CTD casts were prescribed at the open boundary. Initial and boundary conditions based on observations in Burrard Inlet and Indian Arm during the winter of 1974-75 were used to study the inlet's response to tidal forcing and to simulate the deep-water renewal that occurred during this period. Coefficients for the horizontal turbulent diffusion of momentum and salt were set equal to 10⁶ cm² s⁻¹. Reducing this value by a factor of two was found to have little impact on the solution. A further reduction to 10³ cm² s⁻¹ led to numerical instabilities under conditions of dense water inflow. The side friction term in the momentum balance was tuned to match calculated and observed dissipation rates in Burrard Inlet; leading to good agreement between the observed and calculated barotropic tide. Contour plots of tidal amplitudes and phases for model currents and salinities revealed a standing wave pattern for the K₁ and M₂ internal tides in Indian Arm; thus allowing for the possibility of resonance. A comparison of model results with vertical amplitude and phase profiles from harmonic analysis of Cyclesonde current meter timeseries at two locations in Indian Arm was consistent with this result. A least-squares fit was made of the vertical modal structure in the model to the complex tidal amplitudes. This led to calculations of the kinetic energy contained in each of the modes along the model inlet for the M₂ and K₁ constituents. Most of the energy was found to be contained in the barotropic and first baroclinic modes, with the latter dominating in the deep basin, and the former dominating near the sill. Second mode energy was significant for the K₁ constituent at some locations in Indian Arm. There are clear indications in the model of barotropic tidal energy being radiated into the inlet basin via the internal tide. Simulations of the influx of dense water into Indian Arm yielded exchange rates that are consistent with observed values and suggest the possibility of fine-tuning the model coefficients to allow prediction of future overturning events.
Science, Faculty of
Earth, Ocean and Atmospheric Sciences, Department of
Graduate

The world is undergoing rapid changes and the future is uncertain. The changes are related to modification of the landscape due to human activities, such as large and small scale irrigation, afforestation and changes to the climate system. Understanding and predicting hydrologic change is one of the challenges facing hydrologists today. Part of this understanding can be developed from observed data, however, there often too few observations and those that are available are frequently affected by uncertainties. Hydrological models have become essential tools for understanding historical variations of catchment hydrology and for predicting future possible trends. However, most developing countries are faced with poor spatial distributions of rainfall and evaporation stations that provide the data used to force models, as well as stream flow gauging stations to provide the data for establishing models and for evaluating their success. Hydrological models are faced with a number of challenges which include poor input data (data quality and poorly quantified human activities on observed stream flow data), uncertainties associated with model complexity and structure, the methods used to quantify model parameters, together with the difficulties of understanding hydrological processes at the catchment or subbasin. Within hydrological modelling, there is currently a trend of dealing with equifinality through the evaluation of parameter identifiability and the quantification of uncertainty bands associated with the predictions of the model. Hydrological models should not only focus on reproducing the past behaviour of a basin, but also on evaluating the representativeness of the surface and subsurface model components and their ability to simulate reality for the correct reasons. Part of this modelling process therefore involves quantifying and including all the possible sources of uncertainty. Uncertainty analysis has become the standard approach to most hydrological modelling studies, but has yet to be effectively used in practical water resources assessment. This study applied a hydrological modelling approach for understanding the hydrology of a large Tanzanian drainage basin, the Great Ruaha River that has many areas that are ungauged and where the available data (climate, stream flow and existing water use) are subject to varying degrees of uncertainty. The Great Ruaha River (GRR) is an upstream tributary of the Rufiji River Basin within Tanzania and covers an area of 86 000 km2. The basin is drained by four main tributaries; the Upper Great Ruaha, the Kisigo, the Little Ruaha and the Lukosi. The majority of the runoff is generated from the Chunya escarpment, the Kipengere ranges and the Poroto Mountains. The runoff generated feeds the alluvial and seasonally flooded Usangu plains (including the Ihefu perennial swamp). The majority of the irrigation water use in the basin is located where headwater sub‐basins drain towards the Usangu plains. The overall objective was to establish uncertain but behavioural hydrological models that could be useful for future water resources assessments that are likely to include issues of land use change, changes in patterns of abstraction and water use, as well the possibility of change in future climates.

Rates of snowmelt distributed across Emerald Lake watershed, an alpine basin located in the Sierra Nevada, California, were estimated for water year 1987 using a point snowmelt model applied to regions that were classified based on distributed snow water equivalence and net solar radiation (NSR). A 5-m resolution digital elevation model (DEM) and a 5-m classified digital terrain model of snow water equivalence (SWE) were resampled to coarser resolutions (25-m, 30-m, 50-m, and 100-m) using the nearest neighbor approach. These images were used to define other snowmelt physical parameters and the initial state of the snowpack before melting. Topographic parameters calculated at 50-m and 100-m resolution exhibited significant differences in their histogram distribution as compared to the 5-m DEM. The most important were variations in slope, aspect, sky view factor, and terrain configuration factor, which influenced radiation calculations and the definition of distributed parameters for snowmelt calculations. Elevations, however, did not change significantly from one resolution to the other. The distribution of topographic parameters modeled at 25-m and 30-m, remained almost unchanged. Four, seven and ten classes of snow water equivalence and net solar radiation were combined using a band interleave process to determine the maximum number of combined classes. The point snowmelt model was then applied to these areas, which shared similar SWE and NSR characteristics, to obtain hourly melt rates. Modeled snowmelt rates were compared to the total daily discharge observed at the outlet of Emerald Lake watershed. There was good agreement for resolutions S-, 25-, 30-, and 50-m but not for the 100-m OEM, as modeled net solar radiation was too high and water was released from the basin too early. Model performance using three tests (Nash-Sutcliffe criteria, sum of squares of the deviations and the sum of the absolute differences between observed discharge and computed melting) showed that the 30-m resolution OEM with combined classes of 7 SWE and 7 NSR provided the best snowmelt performance for this distributed approach. Finally, fractional snow cover area at one month intervals were estimated, showing that this approach offers the potential to model spatially distributed snow covered area in alpine regions.

La flexión de la Litosfera es un proceso observado en distintos contextos tectónicos y su estudio tiene especial relevancia para la comprensión de los procesos de formación de cuencas de antepaís y fosas oceánicas. Dicha flexión es consecuencia de la aplicación de cargas superficiales (p. e., apilamiento de mantos de corrimiento) y/o cargas intemas de origen incierto (p.e., deformación intema de la litosfera). Por tanto, la evolución de la deflexión y la consiguiente geometría del relleno sedimentario están relacionados con la evolución de la carga. El objetivo del trabajo que se presenta en esta Memoria es el desarrollo de modelos numéricos de flexión litosférica y su aplicación a la formación de fosas oceánicas y cuencas de antepaís. En la primera parte de la Memoria se describen los modelos conceptuales que relacionan los procesos de carga con la deflexión del basamento y la geometría sedimentaria resultante. Esta relación está basicamente determinada por el comportamiento reológico de la Litosfera y los procesos superficiales de erosión/sedimentación. En la segunda parte, se desarrollan los algoritmos de cálculo de la deflexión que están basados en el modelo de placa delgada y permiten incorporar distintas reologías: elástica, viscoelástica y elastoplástica. Una de las aportaciones de este trabajo es la integración en el modelo de flexión de la cinemática en 2D y 3D del proceso de carga, permitiendo una simulación más rigurosa del emplazamiento de láminas cabalgantes. El tratamiento de los procesos de erosión y sedimentación es distinto en los modelos 2D y 3D. En el primer caso se han considerado los modelos simples de tasa constante y transporte difusivo, mientras que en el segundo caso se ha incorporado el transporte advectivo a través de la red de drenaje. Ambos modelos consideran el efecto de las variaciones eustáticas del nivel del mar y de la paleobatimetría. La versatilidad con que se han diseñado los algoritmos de cálculo ha permitido su aplicación a dos contextos geodinámicos distintos. En primer lugar, el modelo de flexión de placa con diferentes reologías se ha aplicado a las fosas oceánicas de Tonga y Kermadec, permitiendo acotar la estructura de la litosfera en la Placa Pacífica y la magnitud de las fuerzas que actúan sobre la misma. En segundo lugar, el mismo modelo de flexión, conjuntamente con las aproximaciones adoptadas para la formación de la carga y el transporte de sedimentos en 2D y 3D, se ha aplicado a la cuenca de antepaís del Guadalquivir, permitiendo relacionar el movimiento de las Zonas Béticas Externas con la geometría de las unidades sedimentarias. El modelo de placa elástica homogenea no permite reproducir el conjunto de observables (deflexión, arquitectura de la cuenca, etc.) y predice una distribución de esfuerzos en el interior de la placa que es incompatible con el comportarniento reológico de la litosfera. Por el contrario, el modelo de placa viscoelástica permite reproducir geometrías complejas del relleno sedimentario, mientras que el modelo de placa elastoplástica no homogénea predice, además, una distribución mas realista de los esfuerzos generados en el interior de la placa De la aplicación de los modelos a las fosas del Tonga y Kermadec se deduce que es necesario imponer una fuerza horizontal extensiva para ajustar correctamente la deflexión del basamento. Ello implica que el mecanismo preponderante en la subducción de la Placa Pacífica es la contribución gravitatoria del slab que subduce (slab pull), y no el arrastre del manto astenosférico. De acuerdo a los parámetros reológicos escogidos, los resultados obtenidos favorecen el modelo termico de Stein & Stein (1992) frente al de Parsons & Sclater (1977). No obstante, la validez de este resultado debe tomarse con precaución dado el alto grado de incertidumbre en Ia deterrninación de los parámetros reológicos. Del estudio realizado en la Cuenca del Guadalquivir se deduce que la influencia de la carga asociada a las Zonas Béticas Internas es despreciable en el proceso de formación de la Cuenca. La subsidencia de la Cuenca precisa de la concurrencia de la carga topográfica, la carga asociada a la paleobatimetría, la contribución de la deformación cortical y la presencia de una carga de origen subcortical que, de acuerdo con las anomalías gravimétricas y geoidales, debe localizarse en la base de la litosfera. Por otro lado, el retroceso que se observa en los depósitos del Tortoniense y el avance subsiguiente de los depósitos del Mesiniense se han reproducido satisfactoriamente considerando el modelo de placa elastoplástica no homogénea y el modelo de placa viscoelástica. Por tanto, el retroceso de la unidad Tortoniense puede atribuirse a un efecto combinado de variaciones del nivel del mar, relajación viscosa de los esfuerzos y estratificación de la resistencia litosférica. Por último, el modelo elastoplástico permite, además, explicar la existencia de fallas extensionales en el basamento coetáneas al proceso de carga.
Flexure of the lithosphere is a widely observed process in different tectonic settings. Its study is particularly relevant to understand the formation of foreland basins and oceanic trenches. The lithosphere bends in response to external loads (e.g., thrusts) and/or hidden Loads (e.g., lithospheric deformation). Thereby, the evolution of the deflection and the basin infill geometry are directly related to loading processes. The main goal of this work is to develop numerical models of lithospheric flexure and their application to the study of oceanic trenches and foreland basin formation. The first part of this Thesis deals with the conceptual models that link loading processes to basement deflection and basin infill geometry. This relationship is mainly controlled by the rheological behavior of the litosphere and by surface processes such as erosion and sedimentation. The second part includes the development of thin plate model based algorithms with different rheologies: elastic, viscoelastic, and elastoplastic. A relevant contribution of this work is the integration in the flexural model of a 2D and 3D fully kinematic approach which realistically simulates thrust emplacement. Erosion and sedimentation processes are treated distinctly in the 2D and 3D approaches. The 2D approach considers either a constant erosion/sedimentation rate or a diffusive model, whereas the 3D can also incorporate advection through a drainage network. Both approaches include eustatic sea level variations and paleobathymetry. Elastic and homogeneous elastoplastic plate models do not account for the observed deflection and basin architecture. Moreover, these models predict an unrealistic lithospheric stress distribution. In contrast, the viscoelastic plate model permits to reproduce complex infill geometries, whereas the non-homogeneous elastoplastic plate model predicts, in addition, a more realistic stress distribution. The application of the 2D algorithms to the Tonga and Kennadec trenches has allowed to constrain the lithospheric structure of the Pacific Plate and the magnitude of the forces acting on the subducting slab. The most outstanding result is the necessity to impose a horizontal extensional force to fit the deflection of the basement. This implies that the driving mechanism is mainly governed by slab pull rather than mantle drag. Results also favour the half plate cooling model proposed by Stein & Stein (1992) instead of that proposed by Parsons & Sclater (1977), although the uncertainties on the rheological parameters do not allow for firm conclusions. The application of the 2D and 3D algorithms to Southern Spain reveals that the Internal Betics play a negligible role in the formation of the Guadalquivir foreland basin. The basement deflection can only be explained considering the topography and paleobathymetry loads, the Load associated with crustal deformation, and a hidden or extra load seated at subcrustal levels. The analysis of gravity and geoid anomalies suggests that the hidden load is likely related to lithospheric thickening beneath the Betics. The complex stratigraphy of the Guadalquivir basin, particularly during Tortonian and Messinian times, is successfully reproduced by the viscoelastic and the non-homogeneous elastoplastic plate models suggesting that the basin infill resulted from the combination of eustatic sea-level variations, viscous stress relaxation, and layered lithospheric strength. Furthermore, the non-homogeneous elastoplastic plate model accounts for the presence of basement extensional faults that developed during thrusting.

Ocean tides worldwide have exhibited secular changes in the past century, simultaneous with a global secular rise in mean sea level (MSL). The combination of these two factors contributes to higher water levels, and may increase threats to coastal regions and populations over the next century. Equally as important as these long-term changes are the short-term fluctuations in sea levels and tidal properties. These fluctuations may interact to yield locally extreme water level events, especially when combined with storm surge. This study, presented in three parts, examines the relationships between tidal anomalies and MSL anomalies on yearly and monthly timescales, with a goal of diagnosing dynamical factors that may influence the long-term evolution of tides in the Pacific Ocean. Correlations between yearly averaged properties are denoted tidal anomaly trends (TATs), and will be used to explore interannual behavior. Correlations of monthly averaged properties are denoted seasonal tidal anomaly trends (STATs), and are used to examine seasonal behavior. Four tidal constituents are analyzed: the two largest semidiurnal (twice daily) constituents, M2 and S2, and the two largest diurnal (once daily) constituents, K1 and O1.

Part I surveys TATs and STATs at 153 Pacific Ocean tide gauges, and discusses regional patterns within the entire Pacific Ocean. TATs with statistically significant relations between MSL and amplitudes (A-TATs) are seen at 89% of all gauges; 92 gauges for M2, 66 for S2, 82 for K1, and 59 for O1. TATs with statistically significant relations between tidal phase (the relative timing of high water of the tide) and MSL (P-TATs) are observed at 55 gauges for M2, 47 for S2, 42 for K1, and 61 for O1. Significant seasonal variations (STATs) are observed at about a third of all gauges, with the largest concentration in Southeast Asia. The effect of combined A-TATs was also considered. At selected stations, observed tidal sensitivity with MSL was extrapolated forward in time to the predicted sea level in 2100. Results suggest that stations with large positive combined A-TATs produce total water levels that are greater than those predicted by an increase in MSL alone, increasing the chances of high-water events.

Part II examines the mechanisms behind the yearly (TAT) variability in the Western Tropical Pacific Ocean. Significant amplitude TATs are found at more than half of 26 gauges for each of the two strongest tidal constituents, K1 (diurnal) and M2 (semidiurnal). For the lesser constituents analyzed (O1 and S2), significant trends are observed at ten gauges.

Part III analyzes the seasonal behavior of tides (STATs) at twenty tide gauges in the Southeast Asian waters, which exhibit variation by 10 – 30% of mean tidal amplitudes. A barotropic ocean tide model that considers the seasonal effects of MSL, stratification, and geostrophic and Ekman velocity is used to explain the observed seasonal variability in tides due to variations in monsoon-influenced climate forcing, with successful results at about half of all gauges. The observed changes in tides are best explained by the influence of non-tidal velocities (geostrophic and Ekman), though the effect of changing stratification is also an important secondary causative mechanism.

From the results of these surveys and investigations, it is concluded that short-term fluctuations in MSL and tidal properties at multiple time scales may be as important in determining the state of future water levels as the long-term trends. Global explanations for the observed tidal behavior have not been found in this study; however, significant regional explanations are found at the yearly time scale in the Solomon Sea, and at the seasonal time scale in Southeast Asia. It is likely that tidal sensitivity to annual and seasonal variations in MSL at other locations also are driven by locally specific processes, rather than factors with basin-wide coherence. (Abstract shortened by ProQuest.)

L'objectif de ce travail est de contribuer a la comprehension du fonctionnement hydrologique d'un bassin versant mediterraneen, notamment a l'occasion de ses crues. Il est compose d'un premier volet experimental destine a l'observation du comportement des eaux souterraines. Ce volet inclut des mesures de pression de l'eau dans le sol et de niveaux des nappes (fugitives ou perennes) sous un versant (ligne de plus grande pente perpendiculaire a la riviere), et une analyse approfondie des sols du versant: granulometries, conductivites hydrauliques in situ mesurees a l'aide de differents appareillages, mesures de la retention en eau des sols etc. Les proprietes hydrodynamiques des sols en place sont egalement decrites. Utilisant cette description hydrodynamique, on tente ensuite de modeliser le comportement du versant lorsqu'il est soumis a la pluie. La modelisation cherche a reproduire les infiltrations et les mouvements de nappes observes. Pour ce faire on developpe un modele d'infiltration et de montee de nappe dans une zone inclinee d'un sol multicouche. Pour resoudre l'equation de richards differentes methodes ont ete developpees. La meilleure, dans notre travail, s'avere etre celle qui utilise les differences finies avec un procede iteratif des directions alternees. Les resultats de la modelisation confirment qu'en cas de pluies importantes, une nappe peut s'elever relativement rapidement a proximite de la riviere et l'alimenter. Toutefois, les forts debits ne peuvent s'expliquer que par la saturation de l'ensemble du profil du sol qui engendre la montee de ces nappes et par les ecoulements de surface que ces aires saturees entrainent

Tide-topography coupling is important for understanding surface-tide energy loss, the intermittency of internal tides, and the cascade of internal-tide energy from large to small scales. Although tide-topography coupling has been observed and modeled for 50 years, the identification of surface and internal tides over arbitrary topography has not been standardized. Here, we begin by examining five surface/internal-tide decompositions and find that only one is (i) consistent with the normal-mode description of tides over a flat bottom, (ii) produces a physically meaningful depth-structure of internal-tide energy flux, and (iii) results in an established expression for internal-tide generation. Next, we examine the expression for internal-tide generation and identify how it is influenced by remotely-generated shoaling internal tides. We show that internal-tide generation is subject to both resonance and intermittency, and can not always be predicted from isolated regional models. Lastly, we quantify internal-tide generation and scattering on the Oregon Continental slope. First, we derive a previously unpublished expression for inter-modal energy conversion. Then we evaluate it using observations and numerical simulations. We find that the surface tide generates internal tides, which propagate offshore; while at the same time, low-mode internal tides shoal on the slope, scatter, and drive turbulent mixing. These results suggest that internal tides are unlikely to survive reflection from continental slopes, and that continental margins play an important role in deep-ocean tidal-energy dissipation.
Graduation date: 2011

Bibliography : leaves 214-246.
xiii, 246 leaves : ill., maps ; 30 cm.
Title page, contents and abstract only. The complete thesis in print form is available from the University Library.
Thesis (Ph.D.)--University of Adelaide, Dept. of Geology and Geophysics, 1992

The efficacy of sedimentation ponds as a means of sediment and heavy metal remediation is investigated, with particular regard to the physical and chemical conditions that may lead to remobilisation of metals from the sediments
Thesis (M.Eng.Sc.) -- University of Adelaide, Dept. of Civil and Environmental Engineering, 2001