A Bayesian Probability Network Approach To Predictive Modeling in Support of EffectiveManagement of Underwater Noise in Marine Mammal Habitat
Marine mammals have evolved to send and receive underwater sound as their primary means for communicating, finding food, and sensing their environment. Introducing human-generated and other unnatural sources of noise into their environment can cause potentially serious consequences, particularly for those species at conservation risk. Past work on predicting the behavioural response to excessive noise has ignored the variability in the sound field and in marine mammal behaviour. An important research gap is to incorporate these uncertainties in the predictions of marine mammal response to proposed industrial development projects. We propose to develop a probability-based network model that will quantify the uncertainties of marine mammal behaviour in response to various intensities and frequencies of industrial noise. This model will be both science and data driven and will provide a framework for assessing the likelihood of various behavioural responses, thereby acting as a management and decision support tool.