A linear reservoir cell model is presented which is proposed as a good candidate for real time flood forecasting applications. The model is designed to be computationally efficient since it should be able to run on a P.C and must operate online in real time. The model parameters and forecasts can be easily updated in order to allow for a more accurate forecast based on real time observations of streamflow and rainfall. The final model, once calibrated, should be able to operate effectively without requiring highly skilled and knowledgeable operators. Thus it is hoped to provide a tool which can be incorporated into an early warning system for mitigation of flood damage, giving water resources managers the extra lead-time to implement any contingency plans which may be neccssary to ensure the safety of people and prevent damage to property. The use of linear models for describing hydrological systems is not new, however the model presented in this thesis departs from previous implementations. A particular departure is the novel method used in the conversion of observed to effective rainlfall. The physical processes that result in the rainfall to runoff conversion are non-linear in nature. Most of the significant non-linearity results from rainfall losses, which occur largely due to evaporation and human extraction. The remaining rainfall is converted to runoff. These losses are particularly significant in the South African climate and in some regions may be as much as 70-90 % of the total observed rainfall. Loss parameters are an integral part of the model formulation and allow for losses to be dealt with directly. Thus, input to the model is observed rainfall and not the "effective" rainfall normally associated with conceptual catchment models. The model is formulated in Finite Difference form similar to an Auto Regressive Moving Average (ARMA) model; it is this formulation which provides the required computational efficiency. The ARMA equation is a discretely coincident form of the State-Space equations that govern the response of an arrangement of linear reservoirs. This results in a functional relationship between the reservoir response constants and the ARMA coefficients, which guarantees stationarity of the ARMA model. / Thesis (M.Sc.Eng.)-University of Natal, Durban, 2001.

Studies are currently underway to determine the suitability of Yucca Mountain in Nevada as the nation's first high-level nuclear waste repository. Values of net infiltration are required to determine pre-waste emplacement groundwater travel times and the performance of the repository as a waste containment system. The objective of this study was to develop a numerical model to perform water balance calculations and predict rates of net infiltration for the site. The model included processes of precipitation, runoff, evapotranspiration, infiltration, and redistribution of water within a soil profile. The watershed was divided into 477 grid cells 75.7 x 75.7 m. The elevation, slope, aspect, and hydrologic properties were assumed to be constant within a grid cell but varied from one cell to the next Water balance calculations were performed for each cell using a one-dimensional form of Richards equation. The solution was obtained using the finite difference method with Newton-Raphson iteration. The model was calibrated using water content data obtained from neutron-moisture meter measurements in boreholes located in Pagany Wash Watershed Measurements were made in channel and terrace alluvium and in tuffs. Computer simulations reproduced water content data for a major precipitation event that occurred in 1984. Simulations verified the importance of antecedent soil water content in controlling the occurrence of runoff. Sensitivity analysis indicated that the soil and alluvium grain-size distributions, which are used to calculate unsaturated hydraulic conductivity, can greatly affect predicted rates of water movement / Graduation date: 1991

The object of current efforts at investigating catchment response is to derive a physically based stochastic model of the watershed. Recent studies have, however, indicated that a limiting factor in deriving such models is the dependence of hydrologic response on initial soil moisture. The dependence affects the distributions and moments of the hydrological processes being investigated. A stochastic model of soil moisture dynamics is developed in the form of a pair of stochastic differential equations (SDE's) of the Ito type. The sources of stochasticity are linked to the random inputs of rainfall and evapotranspiration (ET). One of the SDE's describes the "surplus" case, in which sufficient infiltration always occurs to allow for moisture depletion by the processes of drainage through and ET out of the root zone. The other SDE represents the "deficit" case, in which lack of adequate moisture leads only to an ET-controlled depletion process. Sample functions and moments of moisture evolution are obtained from the SDE's. From the general model of soil moisture, a specific model of initial soil moisture (the moisture at the beginning of a rainstorm event) is developed and its moments are derived. Furthermore, the probability distribution of initial moisture is postulated to permit the assessment of how initial moisture affects the estimation of hydrologic response. The moisture dynamics model reveals that the stochastic properties of moisture ae sensitive to initial conditions in the watershed only for less permeable soils under the "surplus" state but are practically insensitive to such conditions for more permeable soils. The stochastic properties are also less sensitive to initial conditions for all soil types whenever under the "deficit" state. These results suggest that hydrologic processes, such as precipitation excess and infiltration, depend on initial moisture only in regions where the soils are generally less permeable and where the climate tends to sustain a "wet" environment, whereas in arid or semi-arid regions, such processes would not depend on initial moisture. These conclusions imply that, in arid regions, an effective value of initial moisture such as the mean can be used to estimate the properties of the hydrologic processes, whereas in "wet" environments, more accurate values of the properties must be "weighted" based on the probability distribution of initial soil moisture.

[Truncated abstract] The balance equations of mass and momentum, defined at the scale of what has been defined as a Representative Elementary Watershed (REW) has been proposed by Reggiani et al. (1998, 1999). While it has been acknowledged that the REW approach and the associated balance equations can be the basis for the development of a new generation of distributed physically based hydrological models, four building blocks have been identified as necessary to transform the REW approach into, at the very least least, a workable modelling framework beyond the theoretical achievements. These are: 1) the development of reasonable closure relations for the mass exchange fluxes within and between various REW sub-regions that effectively parameterize the effects of sub-REW heterogeneity of climatic and landscape properties, 2) the design of numerical algorithms capable of generating numerical solutions of the REW-scale balance equations composed of a set of coupled ordinary differential and algebraic equations for the number of REWs constituting a study catchment and the sub-regions within the REWs, 3) applications of the resulting numerical model to real catchments to assess its performance in the prediction of any specified hydrological variables, and 4) the assessment of the model reliability through estimation of model predictive uncertainty and parameter uncertainty. This thesis is aimed at making substantial progress in developing each of these building blocks. Chapter 1 presents the background and motivation for the thesis, while Chapter 2 summarizes its main contributions. Chapter 3 presents a description of the closure problem that the REW approach faces, and presents and implements various approaches to develop closure relations needed for the completeness of balance equations of the REW approach. ... In addition, Chapter 4 also shows an initial application of CREW to a small catchment, Susannah Brook in the south-west of Western Australia. Chapter 5 presents the application of CREW to two meso-scale catchments in Australia, namely Collie and Howard Springs, located in contrasting climates. Chapter 6 presents results of the estimation of predictive uncertainty and parameter sensitivity through the application of CREW to two catchments in Australia, namely Susannah Brook and Howard Springs, by using the Generalized Likelihood Uncertainty Estimation (GLUE) methodology. Finally, Chapter 7 presents recommendations for future work for the further advancement of the REW approach. Through these exercises this PhD thesis has successfully transformed the REW-scale coupled balance equations derived by Reggiani et al. (1998, 1999) into a new, well tested numerical model blueprint for the development and implementation of distributed, physically based models applicable at the catchment, or REW scale.

The aim of this dissertation was to develop a multi-criteria decision support system (MCDSS) to allow a specified manager to select with confidence one or many of these BMPs for a particular site. The principal design approach was a review of South African and international literature pertaining to stormwater management techniques, in particular BMPs. This information was formulated into a primary matrix using a rank-and-weighting method. The scores were then checked against the literature to ensure that they were reasonable, culminating in the initial MCDSS. The MCDSS was then provided with seven scenarios, described in the literature, and the output reviewed. Although, the MCDSS would select appropriately when given few criteria for selection when these were increased, inappropriate outcomes resulted. Consequently, weighting factors were assigned to each criterion. The MCDSS was further tested using all the selection criteria and the output deemed satisfactory. The MCDSS was then tested in a case study of the Town Bush stream catchment at eleven sites along the river network and the results were adequate. Taking into consideration the economic aspects of BMP implementation a need also arose for the sites to be allocated to certain authorities depending upon ownership or responsibility. The sites were prioritised depending on potential threat to property and lastly by the hydrological nature of the stream at each site. A stormwater plan for the study area was also proposed. Although the MCDSS was functioning adequately it was not without its limitations. Limitations included the use of drainage areas as a surrogate measure for peak discharge thus, not allowing the user to design a series of BMPs or treatment chain. A second limitation was that initially the BMPs were designed as offline systems where stormwater is managed before entering the channel but in this study they were used as inline systems. Hence the ultimate selection was biased towards those BMPs able to deal with large drainage areas. Recommendations for further improvement include the development of a surrogate measure for drainage area thus allowing the user to design a treatment chain of BMPs; testing the MCDSS in more diverse circumstances; developing a more comprehensive set of selection criteria; and developing a clearer priority-setting model as the one used was rather simplistic. In conclusion the MCDSS provides the user with a useful tool where the selection and implementation of BMPs no longer has to take place in an ad hoc manner. / Thesis (M.Sc.)-University of Natal, Pietermaritzburg, 2001