Doctor of Philosophy
Department of Civil and Mining Engineering
Musavi-Jahromi, S. H., Water quality modelling of river systems, Doctor of Philosophy thesis, Department of Civil and Mining Engineering, University of Wollongong, 1996. http://ro.uow.edu.au/theses/1282
Environmental pollution, including river pollution, has been keeping pace with industrial development. It has become one of the most important world - wide problem, and is one of the key factors by which economic development and the conservation of natural resources can be constrained.
Rivers play important role not only for humans, but also for the aquatic organisms both flora and fauna. They play a significant role in local community activities. They are used for water supply, irrigation, recreation, sand mining, and the disposal of industrial and domestic wastewater. These uses of river systems raise concern about river management and water quality modelling is a very important river management tool for water industries.
Water quality models can assess the water quality impact of various management options for the river systems under industrial and domestic wastewater loading. They can be applied to develop sewerage strategies, to estimate the volumes of fresh water for flushing effluents, determination of the location of new sewage treatment plants (STPs) and to determine critical location of water quality sampling stations in a given river system.
In this thesis, a one-dimensional (1-D) numerical unsteady water quality model called 'Model for Australian River Systems' (MARS) has been developed. The flow continuity, momentum and advection-dispersion equations are the fundamental equations of MARS. MARS includes a hydrodynamic module (MARS-HD) , a transport module (MARS-AD) and a water quality module (MARS-WQ).
The MARS-HD module is based on the flow continuity and momentum equations. It is developed using the Method Of Characteristics (MOC) . The prediction of flow velocity and depth using MARS-HD module has shown excellent agreement with published experimental laboratory data. Furthermore, it was able to reproduce field data of the Nepean River, Australia, with an excellent accuracy.
The MARS-AD module is established using a method known as 'Quadratic Upstream Interpolation with Convective Kinematics and Estimated Streaming Terms' (QUICKEST). The MA S-A D module was examined against analytical solutions. Good agreement between the prediction by the MARS-A D module and analytical solution has been obtained.
The MARS-W Q module has been developed based on QUICKEST numerical algorithm. The Monod Model of population dynamics which is similar to the Michaelis-Menton relationship for enzyme reactions has been widely used in the development of the MARS-WQ module. The MARS-WQ module can predict concentration of dissolved oxygen (DO), biochemical oxygen demand (BOD), ammonia-nitrogen (NH3-N), nitratenitrite nitrogen (NOx-N), nonfilterable phosphorus (NFP), filterable phosphorus (FP) and total phosphorus (TP) in a given river system. It has shown high accuracy in the modelling of water quality variables in calibration and verification processes in the Nepean River.
MA S is involved with 53 water quality parameters. The sensitivity analysis of MARS has shown that it is very sensitive to parameters like maximum growth rate of algae (GMAX), respiration rate (kr), hydrolysis rate of NFP (k7), and half saturation concentration for uptake of inorganic phosphorus by algae (HP). However, parameters such as the benthic flux of ammonia (BNH), Manning's roughness coefficient (Cm), dispersion coefficient (DK), half-saturation concentration for uptake of inorganic nitrogen (HN), half-saturation concentration for nitrification (HCN), BOD decay rate (K1), settling rate of nonfilterable phosphorus (k8), and benthic flux of nonfilterable phosphorus (BNFP) can also affect the prediction but to a lesser degree.
MARS has been applied to the Nepean River for decision making. Two sewage treatment plant (STP) discharges (Camden STP and proposed Picton STP) and one source of freshwater (Cordeaux D a m ) discharges were modelled. Picton STP will be constructed in two stages with an outflow of 3.3 ML/day in year 2003 and 4.8 ML/day in year 2012. Therefore, a total of 24 practical management scenarios have been assumed for Upper Nepean River. Based on calibration results for worst situations in winter 1991 and summer 1992 low flow conditions, water quality variables have been predicted for the year 2003 and 2012.
It was found out that the water quality of the Nepean River will be deteriorated further with an increase in nutrient concentration. Total nitrogen concentration will be increased to above ANZECC (1992) guidelines along the river in year 2003 and 2012. In the year 2012, total phosphorus concentration will not meet guideline value between the West Camden STP and about 10 k m downstream. To satisfy guideline values, 20 ML/d fresh water is required for dilution and flushing nitrogen along the river. However, a 10 ML/d discharge of freshwater drops the T P concentration below guideline value. Freshwater can be provided from Cordeaux Dam in the upstream.
Accidental discharges from the West Camden STP including the raw sewage, treatment malfunction in the STP and the stormwater overflow from the STP have been investigated using MARS . This study has shown that the influence of raw sewage discharge on water quality of the Nepean River is several times more than the others. It causes significant increase in BOD and nutrient concentrations downstream of the West Camden STP and DO level less than 1.0 mg/L.
Finally, MARS has been compared with an existing river water quality model called MIKE 11 (version 3.1, 1993). MARS has shown better accuracy than MIKE 11 in the prediction of flow hydrographs at four Sydney Water gauging stations on the Nepean River. It has predicted DO level along the Nepean River more accurate than MIKE 11. MARS has simulated NH3-N concentration with average and root mean square errors more closer to zero than MIKE 11. The predicted NOx-N concentration using MARS is also in better agreement with measured values than MIKE 11.