Providing high quality and safe drinking water is challenging for suppliers, who must follow strict national and international standards in the treatment processes required for drinking water. Dissolved organic matter (DOM) is one of the most important indicators for water treatment as it designates the ideal pH as well as the amounts and types of chemicals used in water treatment processes. Although previous research and innovations have helped to optimise the physical and chemical processes utilised in drinking water treatment plants and maximised the removal of pathogenic substances and pollutants, sometimes carcinogenic substances recognised as disinfection by-products are produced during the water treatment processes due to a combination of complex raw water characteristics, inefficient treatment and the type of chemicals added (Bekbolet et al., 2005). Water containing disinfection by-products has a harmful effect on consumers’ health. It has been found that in over 600 different disinfection by-products, the formation of each disinfection by-product was potentially related to the DOM present in the water (Richardson et al., 2007). Consequently, the monitoring and prediction of concentrations of DOM are of paramount importance in drinking water reservoirs. In recent times, probes measuring fluorescent dissolved organic matter (fDOM) have been developed and installed in several Australian reservoirs, providing high-frequency data on the vertical distribution of DOM in lakes such as the Tingalpa Reservoir in South-East Queensland. However, the probe measurement is limited to a vertical direction and cannot directly observe DOM transport within the whole lake. In this regard, prior studies have shown hydrodynamic modelling offers a basis for simulating the transport of nutrients and pollutants in response to meteorological forcing functions. Thus, the objective of this research is to (i) develop a three-dimensional (3D) hydrodynamic and sediment transport model, (ii) develop a data-driven model using high-frequency fDOM data, and (iii) to integrate them to predict a 3D fDOM for the Tingalpa Reservoir in South-East Queensland. A better understanding of the fate of DOM, especially with regard to extreme weather events, will benefit water utilities by enabling more proactive and agile DOM removal operations. To this end, firstly, an investigation into the water circulation and mixing processes occurring in the shallow, subtropical Tingalpa Reservoir in Australia was conducted. Bathymetrical, meteorological, chemical and physical data collected from field measurements, laboratory analysis of water sampling and an in-situ vertical profiling system (VPS) were analysed. Based upon the high-frequency VPS dataset, a one-dimensional model was developed to calculate the vertical velocity and diffusion coefficient. The results demonstrate that a persistently high air temperature and stable reservoir water depth lead to prolonged thermal stratification. Analysis of the collected water quality data indicates that heavy rainfalls have a significant impact on water quality when the dam level is low. The peak value of dissolved organic carbon (DOC) concentrations occurred in the wet season, while the specific ultraviolet (UV) absorbance value decreased when solar radiation increased from spring to summer. The one-dimensional model provides a comprehensive approach for understanding and modelling the water mixing processes in similar lakes with high-frequency data from VPS or other monitoring systems. Secondly, the effect of winds and storms was studied through a 3D numerical investigation based on the hydrodynamic and sediment transport model in the Tingalpa Reservoir. Data-driven models were also established to generate the inflow conditions. Based upon the simulation results, sediment transport is driven by storm events, during which sediment delivery to the reservoir is dominated by allochthonous flux. In this regard, the sediment accumulation during storm events in 2015 was estimated. The effects of the wind upon the lake mixing processes and sediment transport were also investigated. Thirdly, the historical database derived from water sampling and the VPS was collected and analysed. An innovative, coupled data-driven and process-based model was developed and assessed to simulate the transport processes of fDOM. These models proved to be able to predict fDOM in both calm and storm conditions. Given the scenario analysis of the modelling results, it was concluded that fDOM concentrations increase along with water depth during storm events, and the area close to the riverine zone had the sharpest increase in fDOM concentrations in the whole reservoir during such events. The simulated results indicated that simulated fDOM can be regarded as a proxy for DOC concentrations. This innovative model enables operators to obtain a more thorough understanding of the DOM cycle in a reservoir and to receive more guidance in efficiently removing the DOM. It is well-known that analysing and understanding the cycle of DOM is a difficult task and, to date, past research has been limited to the study of the biological and chemical processes of DOM in lakes and reservoirs. However, hydrodynamic processes are the basic force that transports the water, sediment and pollutants into lakes and reservoirs. This research considers the effects of hydrodynamic processes on the cycle of DOM. The 3D fDOM prediction model built in this research extends the understanding of the DOM cycle into whole reservoir systems and enhances the interpretation of spatially and temporally incomplete field measurements. Moreover, the fDOM model costs less and can help the water managers or operators to make good decisions on DOM removal. Finally, despite being limited to fDOM predictions, there is the potential for the application of such a modelling approach to forecast other parameters of possible concern, such as nitrogen or phosphorus. This model concept can be used in other lakes or reservoirs with the same DOM issues.