Development of Bayesian Network Decision Support Tools to Support River Rehabilitation Works in the Lower Snowy River

Abstract

ABSTRACT Community concern at the continual environmental decline of the Snowy River (New South Wales and Victoria, Australia) has resulted in a substantial investment in river rehabilitation works, referred to as the Snowy Rehabilitation Project. Much of the investment in the science component of the project went into developing physical models to better understand the behavior of sediments that have accumulated over time in the river channel. The outcomes of these models were intended to assist river managers in better controlling sediment impacts to restore instream ecological function. To synthesize the learnings from these studies for use by catchment managers, a simple decision support tool, referred to as the Snowy tool, was constructed. The Snowy tool was designed to link outcomes from the models with management activities (or interventions) to outcomes within the river channel. It took the form of a probabilistic model (Bayesian network) that incorporates data from a hydraulic model (HEC-RAS), combined with expert opinion, and riverine response models. This article overviews the Snowy tool, and stresses the importance of the use of Bayesian networks in adaptive management frameworks and in guiding investments in research and on-ground decision-making.