Habitat use of the bottlenose dolphins (Tursiops truncatus) of Fiordland: Where, why and the implications for management

Understanding the distribution of a species gives important clues about its ecology, and can provide key information and guidance for conservation management. The bottlenose dolphins (Tursiops truncatus) of Fiordland, New Zealand, form three small subpopulations, two of which are mostly resident within separate fjord systems: Doubtful Sound and Dusky Sound. Within these fjords, the dolphins’ distribution and resulting habitat use varies, with high and low use areas, and seasonal variation evident. In this thesis I investigated the distribution patterns of the dolphins, to better understand what drives them and how this relates to the way the dolphins are managed. Specifically, I used Kernel Density Estimation (KDE) from sighting information collected between 2005 and 2018 in Doubtful Sound, and 2009 and 2018 in Dusky Sound, to identify patterns in habitat use over time. Drivers of habitat use were investigated using species distribution models (SDMs), in the form of generalised additive models. Abiotic predictor variables were modelled using long term occurrence data as the response variable. Biotic predictors were included in SDMs for 2018, a year in which I collected data on potential prey and dolphin distribution concurrently. Information on dolphin prey was collected using Baited Underwater Video (BUV). I found that although seasonal variation in habitat use was present, general distribution patterns were consistent through time. I was able to identify some useful abiotic predictors, but found that they were generally poor at explaining dolphin distribution (Doubtful Sound top model deviance explained = 9.45%; Dusky Sound top model deviance explained = 5.20%). In Doubtful Sound during 2018, including biotic predictors improved model performance (abiotic-only model deviance explained = 19.8%; biotic predictors included deviance explained = 39.1%) and suggested that the main driver of distribution for the dolphins was the abundance of potential prey. These results are significant because they indicate the importance of particular areas within the fjords for the dolphins, and show that they remain important through time. This confirms that both populations would be good candidates for spatial based management. The results of this thesis also suggest that prey distribution plays an important role in the distribution of the dolphins and management should therefore have a more ecosystem-based focus, protecting the fish species that are important for the dolphins