1- The global issue of water scarcity caused by climate change and human utilisation highlights the importance of an efficient assessment of water quality in freshwater systems. One of the challenges facing water management in environmental impact studies is the difficulty of inferring causality in complex systems. Traditional water assessment methods are inadequate because they are challenged to separate natural variation from the effect of human activities. 2- Knowing the causal structure of a complex ecosystem will enable managers to identify key anthropogenic, climate and flow drivers of water quality, and make informed decisions about interventions that improve water and environmental quality. In this study, I show how causal modelling can facilitate decision making for water treatment plant managers to improve their environmental management of this valuable resource. 3- Models built using causal modelling techniques, including structural equation modelling and the principles of Bayesian Networks, were utilised for management decision making purposes. The discharge load values were manipulated in the models to predict the effect of a potential intervention, e.g. treatment plant upgrade, on the values of the water quality variables in the creek. That is, water quality variables were predicted when an imaginary or counterfactual situation was imposed on the models. 4- This study showed that there would not be any observable effect of effluent on macroinvertebrate communities if the discharge loads of chlorophyll a, total organic carbon, total phosphorus, nitrate, and conductivity were reduced to 0.1 of observed values. The concentrations of environmental variables in the creek would return to their baseline levels when their corresponding discharge loads in the effluent were halved or divided by 10. 5- Based on the findings of this study, managers in the field of environmental impact studies can predict the response of a system in the presence of potential interventions under complex and uncertain conditions. The implementation of such techniques offers great promise in the wider field of environmental management where accounting for multiple factors structuring ecosystems in necessary to adequately represent causality.