published_or_final_version / Civil Engineering / Doctoral / Doctor of Philosophy

Submitted in fulfillment of the requirements of the Degree of Master of Technology: Civil Engineering, Durban University of Technology, 2014. / Streamflow modelling remains crucial to decision-making especially when it concerns planning and management of water resources systems in water-stressed regions. This study proposes a suitable method for streamflow modelling irrespective of the limited availability of historical datasets. Two data-driven modelling techniques were applied comparatively so as to achieve this aim. Genetic programming (GP), an evolutionary algorithm approach and a differential evolution (DE)-trained artificial neural network (ANN) were used for streamflow prediction in the upper Mkomazi River, South Africa. Historical records of streamflow and meteorological variables for a 19-year period (1994- 2012) were used for model development and also in the selection of predictor variables into the input vector space of the models. In both approaches, individual monthly predictive models were developed for each month of the year using a 1-year lead time. Two case studies were considered in development of the ANN models. Case study 1 involved the use of correlation analysis in selecting input variables as employed during GP model development, while the DE algorithm was used for training and optimizing the model parameters. However in case study 2, genetic programming was incorporated as a screening tool for determining the dimensionality of the ANN models, while the learning process was further fine-tuned by subjecting the DE algorithm to sensitivity analysis. Altogether, the performance of the three sets of predictive models were evaluated comparatively using three statistical measures namely, Mean Absolute Percent Error (MAPE), Root Mean-Squared Error (RMSE) and coefficient of determination (R2). Results showed better predictive performance by the GP models both during the training and validation phases when compared with the ANNs. Although the ANN models developed in case study 1 gave satisfactory results during the training phase, they were unable to extensively replicate those results during the validation phase. It was found that results from case study 1 were considerably influenced by the problems of overfitting and memorization, which are typical of ANNs when subjected to small amount of datasets. However, results from case study 2 showed great improvement across the three evaluation criteria, as the overfitting and memorization problems were significantly minimized, thus leading to improved accuracy in the predictions of the ANN models. It was concluded that the conjunctive use of the two evolutionary computation methods (GP and DE) can be used to improve the performance of artificial neural networks models, especially when availability of datasets is limited. In addition, the GP models can be deployed as predictive tools for the purpose of planning and management of water resources within the Mkomazi region and KwaZulu-Natal province as a whole.

The purpose of this study is to develop a method for predicting monthly flow duration curves for ungaged basins that are suitable for estimating average annual flow, and installed capacity and average annual energy generation at potential sites for hydropower development. The procedures were tested by developing monthly rainfall duration curves for five sample watersheds and then developing flow duration curves from the rainfall data. The methods were evaluated by comparing the predicted monthly flow duration curves to daily and monthly flow duration curves based on field data from the selected sites because a plant's potential energy output can be computed directly from a flow duration curve. The methods tested fit duration curves based on field data reasonably well and are suitable for preliminary evaluation of hydropower developments in ungaged basins.

A finite element numerical model is developed for the modeling of coupled surface-water flow and ground-water flow. The mathematical treatment of subsurface flows follows the confined aquifer theory or the classical Dupuit approximation for unconfined aquifers whereas surface-water flows are treated with the kinematic wave approximation for open channel flow. A detailed discussion of the standard approaches to represent the coupling term is provided. In this work, a mathematical expression similar to Ohm's law is used to simulate the interacting term between the two major hydrological components. Contrary to the standard approach, the coupling term is incorporated through a boundary flux integral that arises naturally in the weak form of the governing equations rather than through a source term. It is found that in some cases, a branch cut needs to be introduced along the internal boundary representing the stream in order to define a simply connected domain, which is an essential requirement in the derivation of the weak form of the ground-water flow equation. The fast time scale characteristic of surface-water flows and the slow time scale characteristic of ground-water flows are clearly established, leading to the definition of three dimensionless parameters, namely, a Peclet number that inherits the disparity between both time scales, a flow number that relates the pumping rate and the streamflow, and a Biot number that relates the conductance at the river-aquifer interface to the aquifer conductance. The model, implemented in the Bill Williams River Basin, reproduces the observed streamflow patterns and the ground-water flow patterns. Fairly good results are obtained using multiple time steps in the simulation process.

A finite element numerical model is developed for the modeling of coupled surface-water flow and ground-water flow. The mathematical treatment of subsurface flows follows the confined aquifer theory or the classical Dupuit approximation for unconfined aquifers whereas surface-water flows are treated with the kinematic wave approximation for open channel flow. A detailed discussion of the standard approaches to represent the coupling term is provided. In this work, a mathematical expression similar to Ohm's law is used to simulate the interacting term between the two major hydrological components. Contrary to the standard approach, the coupling term is incorporated through a boundary flux integral that arises naturally in the weak form of the governing equations rather than through a source term. It is found that in some cases, a branch cut needs to be introduced along the internal boundary representing the stream in order to define a simply connected domain, which is an essential requirement in the derivation of the weak form of the ground-water flow equation. The fast time scale characteristic of surface-water flows and the slow time scale characteristic of ground-water flows are clearly established, leading to the definition of three dimensionless parameters, namely, a Peclet number that inherits the disparity between both time scales, a flow number that relates the pumping rate and the streamflow, and a Biot number that relates the conductance at the river-aquifer interface to the aquifer conductance. The model, implemented in the Bill Williams River Basin, reproduces the observed streamflow patterns and the ground-water flow patterns. Fairly good results are obtained using multiple time steps in the simulation process.

Several different approaches are applied in low flow frequency analysis. Each method's theory and application is explained. The methods are (1) physically based recession model dealing with time series, (2) log-Pearson type III and mixed log-Pearson type III using annual minimum series, (3) Double Bounded pdf using annual minimum series, (4) Partial Duration Series applying truncated and censored flows. Each method has a computer program for application. One day low flow analysis was applied to 15 stations, 10 perennial streams and 5 intermittent streams. The physically based method uses the exponential baseflow recession with duration, initial recession flow, and recharge due to incoming storm as random variables, and shows promise as an alternative to black box methods, and is appealing because it contains the effect of drought length. Log-Pearson is modified to handle zero flows by adding a point mass probability for zero flows. Another approach to zero flows is the Double Bounded probability density function which also includes a point mass probability for zero flows. Maximum likelihood estimation is used to estimate distribution parameters. Partial Duration Series is applied due to drawbacks of using only one low flow per year in annual minimum series. Two approaches were used in Partial Duration Series (i) truncation, and (ii) censorship which represent different low flow populations. The parameters are estimated by maximum likelihood estimation. / M.S.

Natural stream systems contain a variety of flow geometries which contain flow separation, turbulent shear layers, and recirculation zones. This work focuses on streams with dead zones. Characterized by slower flow and recirculation, dead zones are naturally occurring cutouts in stream banks. These dead zones play an important role in stream nutrient retention and solute transport. Previous experimental work has focused on idealized dead zone geometries studied in laboratory flumes. This work explores the capabilities of computational fluid dynamics (CFD) to investigate the scaling relationships between flow parameters of idealized geometries and the time scales of transport. The stream geometry can be split into two main regions, the main stream flow and the dead zone. Geometric parameters of the dead zone as well as the bulk stream velocity were varied to determine a scaling relationship for the transport time scales. These flow geometries are simulated using the RANS turbulence model with the standard k-ω closure. The standard first order dead zone model is expanded to a two region model to accommodate the multiple time scales observed in the simulation results. While this model currently has limited predictive capability, it provides physical insight into the functional dependence of the dead zone time scales. LES is used to evaluate the performance of the Reynolds Averaged Navier-Stokes (RANS) turbulence model and to describe the anisotropic turbulence characteristics. The differences between the time averaged flow field for Large Eddy Simulation (LES) and RANS was determined to have a significant impact on passive scalar transport. / Graduation date: 2012

This study presents a new Lagrangian model for predicting dynamics and transport in rivers and shallow water flows. A hydrostatic model is developed for the prediction of rivers and floodplain flow and lateral interactions between them. The model is extended to the Boussinesq weakly non-linear, non-hydrostatic model for the simulation of solitary waves and undular bores. A model for advection-diffusion transport of tracers in open channel flow is also presented. The simulation results are compared against an analytical solution and published laboratory data, field data and theoretical results. It is demonstrated that the Lagrangian moving grid eliminates numerical diffusion and oscillations; the model is dynamically adaptive, providing higher resolution under the wave by compressing the parcels (grid). It also allows flow over dry beds and moving boundaries to be handled efficiently. The hydrostatic model results have shown that the model accurately simulates wave propagation and non-linear steepening until wave breaking. The model is successfully applied to simulate flow and lateral interactions in a compound channel and flood wave movement in a natural river. The non-hydrostatic model has successfully reproduced the general features of solitary waves such as the balance between non-linearity and wave dispersion and non-linear interactions of two solitary waves by phase-shift. Also, the model successfully reproduced undular bores (high frequency short waves) from a long wave and the predicted maximum height of the leading wave agreed very well with the published results. It is shown that the simple second order accurate Lagrangian scheme efficiently simulates dispersive waves without any numerical diffusion. Lagrangian modeling of advection-diffusion transport of Gaussian tracer distributions, top hat tracer distributions and steep fronts (step function) in steady, uniform flow has provided exact results and has shown that the scheme allows the use of a large time step without any numerical diffusion and oscillations, including for the advection of steep fronts. The scheme can handle large Courant numbers (results are presented for Cr = 0 to 20) and the entire range of grid Peclet numbers from zero to infinity. The model is successfully applied to tracer transport due to flow induced by simple waves, solitary waves and undular bores

Many hydrologic basins in the southwest have seen their perennial streamflows turn to ephemeral, their riparian communities disappear or be jeopardized, and their aquifers suffer from severe overdrafts. Under -management of ground -water exploitation and of conjunctive use of surface and ground waters are the main reasons for these events.