A fine-scale marine mammal movement model for assessing long-term aggregate noise exposure

Understanding the impacts of anthropogenic sound on marine mammals is important for effective mitigation and management. Sound impacts can cause behavioral changes that lead to displacement from preferred habitat and can have negative influence on vital rates. Here, we develop a movement model to better understand and simulate how whales respond to anthropogenic sound over ecologically meaningful space and time scales. The stochastic model is based on a sequential Monte Carlo sampler (a particle filter). The movement model takes account of vertical dive information and is influenced by the underwater soundscape and the historical whale distribution in the region. In the absence of noise disturbance, the simulator is shown to recover the historical whale distribution in the region. When noise disturbance is incorporated, the whale's behavioral response is determined through a dose-response function dependent on the received level of sound. The aggregate impact is assessed by considering both the duration of foraging loss and the spatial shift to alternate (and potentially less favorable) habitat. Persistence of the behavioral response in time is treated through a 'disruption' parameter. We apply the approach to a population of fin whales whose distribution overlaps naval sonar testing activities beside the Southern California range complex. The simulation shows the consequences of one year of naval sonar disturbance are a function of: i) how loud the sound source is, ii) where the disturbed whales are relative to preferred (high density) habitat, and iii) how long a whale takes before returning to a pre-disturbance state. The movement simulator developed here is a generic movement modeling tool that can be adapted for different species, different regions, and any acoustic disturbances with known impacts on animal populations.