Infrastructure Engineering - Theses
Permanent URI for this collection
Search Results
1 - 3 of 3
-
ItemThe behaviour of salinity and density stratified flows in the Wimmera River, Australia( 1994)A quantitative understanding of the behaviour of salinity and density stratification in the Wimmera River is developed using a combination of field, laboratory and numerical modelling techniques. The Wimmera River, which is located in north-western Victoria, Australia, is a saline stream with a seasonal and highly variable flow regime. Large salt fluxes enter the Wimmera River as a result of surface water inflows from the upper catchment and groundwater inflows in the upper and lower reaches of the stream. During periods of low and zero flow, a series of long deep pools exist along the river, particularly downstream of Horsham. Inflows of saline groundwater accumulate in scour depressions within these large pools and a series of density stratified or saline pools results. Small flows of saline and fresh water down the river can also lead to density stratification. Larger flow events lead to destruction of the density stratification. A model of flow and salinity in a 200 km reach of the Wimmera River is developed using the MIKE 11 model (DHI, 1992a; 1992b). MIKE 11 solves the St Venant equations for gradually varied, unsteady flow and the Advection-Dispersion equation for solute transport. A time-series data base of discharges and salinities for all surface water and groundwater inflows to the river is developed. This was an important step in the model development due to the existence of a significant number of ungauged tributaries and the importance of groundwater as a source of salt. Stream channels are specified in MIKE 11 by defining a channel network and specifying a series of cross-sections along each channel. The channel morphology of the Wimmera River is studied and a methodology for characterising channel variability is developed. It is shown that the Wimmera River channel can be divided into two statistically different channel types which are characterised by a typical length-scale of several kilometres. Using the above analysis as a basis, a stochastic model of stream channel cross-sections is developed for the Wimmera River and used to infill the existing cross-sectional data. The hydraulic implications of along-channel cross-sectional variation are investigated numerically. A one-dimensional model of the Wimmera River is calibrated and tested. This model is applicable to in-bank flows and their associated salinities. The model adequately simulates the routing of water and salt down the Wimmera River. Variations in salinity associated with flow events and the seasonal variation of salinity are reproduced. Field and laboratory investigations of density stratified pools are described. Density stratified pools form as a result of saline groundwater inflows when the stream discharge is less than 200 - 300 Ml/d. The rate at which the stratification develops is quantified for four field sites. Saline water is mixed from density stratified pools during flow events. The mechanism responsible for most of the mixing involves a thin outflow of saline water up the downstream slope of the scour depression. Turbulent entrainment is also responsible for some mixing. During the autumn, convection associated with surface cooling can also mix some density stratified pools. A model of individual density stratified pools, known as Salipool, is developed and tested. Salipool is applied to four density stratified pools in the Wimmera River. A generalised calibration of the mixing relationship incorporated in Salipool is suggested. This generalisation is based on bend sharpness. It is hypothesised that bends have a significant impact on mixing of density stratified pools due to their effect on the vertical velocity profile and the direction of nearbed currents. Salipool is used as a basis for modifying MIKE 11 to incorporate the effect of density stratification.
-
ItemEstimating groundwater recharge using an integrated surface and groundwater model( 1990)Since European settlement, the Riverine Plain of south-eastern Australia has undergone extreme land use changes through the clearing of natural vegetation and the introduction of intensive irrigation. This has led to increased rates of groundwater recharge with consequent land salinisation and waterlogging. For effective evaluation of management options of salinity and high water-tables, reliable estimates of the spatial and temporal distribution of recharge rates throughout the region are required. This thesis describes the development of a methodology to estimate regional recharge rates using an integrated surface and groundwater modelling approach. The integrated model was calibrated against streamflow data and groundwater potentiometric head data with recharge estimated as an output from the calibrated model. The main input data are rainfall, irrigation rates and climatic data required to calculate potential evapotranspiration using Morton's procedure. The applicability of the modelling approach was tested in both the irrigated and non-irrigated areas within the Campaspe River Basin in north-central Victoria. The modelling aspects of this study can be divided into three phases: (1) surface modelling, where the surface model, HYDROLOG, was calibrated against streamflow records to investigate the use of streamflow data to optimise model parameters in recharge studies; (2) modelling of the groundwater flow of the Campaspe Valley using a quasi three-dimensional finite element numerical model, AQUIFEM-N; and (3) estimation of groundwater recharge using the integrated surface and groundwater model. The results indicated that there are limitations in using alone the surface model or the groundwater model to estimate recharge. The simulations improved when the surface and groundwater models were coupled. The integrated modelling approach utilises the essential features of both models in the estimation of recharge. The surface model accounts for the movement of moisture through the soil and can therefore simulate adequately the temporal variations in the recharge rates. The use of groundwater data improves the optimisation of the parameter values because the groundwater data provide a larger data set to calibrate the model. It also enables the validation of the estimated recharge rates as recharge is directly related to the ground water potentiometric levels. The integrated modelling approach can estimate satisfactorily the spatial and temporal distribution of recharge rates resulting from rainfall and irrigation water, and is essential in any predictive groundwater recharge study. The recharge rates predicted by the model compare favourably with other published studies. This research was initiated by the Department of Industry, Technology and Resources of Victoria and later commissioned by the Rural Water Commission of Victoria as part of their salinity control and management studies. The integrated surface and groundwater model is now used by the Commission to study different management practices in the formulation of preferred options for salinity control. The Commission is also using the model for the evaluation of resource development through better management of the conjunctive use of surface and ground water in the Campaspe Valley.
-
ItemThe evaluation of surge flow in border irrigation (with special reference to cracking soils)( 1993)Surge flow was identified as a potential method of improving farm water management to both reduce water use and minimise accessions to water table, and therefore to help address the salinity and water-logging problems facing the irrigated areas of the Murray-Darling Basin. Surge flow has been intensively researched in furrow irrigation in the USA, but border strips (which are the dominant application method in south-eastern Australia) and cracking soils have received minimal attention. This study investigated surge response in border irrigation with particular reference to cracking clay soils. Two years' field work resulted in a substantial data set covering four soil types and a range of management options. Surge flow reduces water use on cracking clays by 15-40% compared with continuous irrigation under similar conditions and given the same level of management. A negative response was observed on homogenous fine sandy loams, with up to 30% more water applied in surge flow. The positive response was obtained over a range of cycle ratios from 0.2-0.5 and with off-times from 30-300 minutes. Measured soil moisture distributions in the second season showed that application uniformity was better using long offtimes of around 300 minutes. Calculations of the water balance indicate that a greater proportion of applied water is retained in the root zone in surge flow and accessions are a small proportion of those in conventional irrigation. The strong field response was not supported by infiltration test data which recorded greater water intake in surge flow than in continuous tests. The recirculating infiltrometer over-estimated cumulative infiltration for both surge and continuous application, in both seasons. It was later observed that water advances in subsurface cracks in the off-time, up to 70 m ahead of the surface stream. Additional analysis indicates that the surge response is largely due to a mechanism connected with this phenomenon, which helps reconcile the disagreement between field and infiltration test data. Surges overlapped in every test and this phenomenon cannot be simulated by existing hydraulic models of surge flow and substantial revisions are needed before meaningful calibration can be undertaken. The field work provided a clear statement of the difficulty of obtaining infiltration data on cracking soils, and highlights the need for effective real-time estimation of infiltration parameters to overcome severe temporal and spatial variability in infiltration conditions. An inverse solution of the Zero Inertia model was developed, using a constrained Simplex optimisation algorithm to determine infiltration and roughness parameters from advance and depth profile data plus a compound objective function of the two. Global parameters, with reasonably effective predictive performance, were obtained using a compound objective function with continuous flow data sets and, more tentatively, for surge flow. This work provides practical possibilities for automation of border irrigation, using as little as two sensors to determine one roughness and one infiltration parameter from advance data, providing the Kostiakov exponent can be reliably classified. A minimum offour sensors are needed to identify two Simple Kostiakov infiltration parameters and one roughness value. Longer term, farm-scale trials are now required.