Hector�s dolphin (Cephalorhynchus hectori) is an endangered coastal species endemic to New Zealand. Their distribution, like other marine organisms, is intertwined with the dynamics of their local habitats, and at a larger scale, the coastal waters around New Zealand. The main purpose of this thesis was to identify specific habitat requirements of this rare dolphin. Hector�s dolphin distribution around the South Island was quantified along several temporal and spatial scales. Large-scale density analyses of abundance surveys found over half of the South Island�s current population occurred within only three main regions. Two of these strongholds are along the west coast and the third is located around Banks Peninsula on the east coast. Smaller-scale analyses at Banks Peninsula found the majority of the dolphin community was preferentially using core regions within the marine mammal sanctuary. Monthly surveys showed that in summer and autumn statistically more dolphins occurred within inshore regions ([less than or equal to]one kilometre), spread throughout the surveyed coastline. From May through winter, dolphin densities rapidly declined. Remaining dolphins were significantly clumped in more offshore waters of eastern regions. The lowest encounter rates occurred between August and September. Certain 'hotspots' consistently had higher dolphin densities throughout the study period while others were preferred seasonally. To address habitat preferences, surveys simultaneously collected oceanographic samples using a CTD profiler. In general, physical variables of the Peninsula�s eastern and southeastern waters varied less, despite being regularly exposed to upwellings and the varied presence of sub-tropical waters. Semi-sheltered bays and shallow inshore waters were highly variable and more susceptible to spatially discrete influences, such as localised river outflows and exchange events. Several hydrographic features were seasonally predictable due to their dependence on climate. The stratification and location of the two dominant water masses (neritic and sub-tropical) accounted for over half of the temporal and spatial variability observed in oceanographic data. Possible relationships between oceanographic features and aggregations of dolphins within Banks Peninsula were examined using global regression and a spatial technique known as geographical weighted regression (GWR). GWR models out-performed corresponding global models, despite differences in degrees of freedom and increased model complexity. GWR results found relationships varied over localised scales that were concealed by global methods. Monthly GWR models suggested the seasonal presence and strength of local oceanographic fronts influenced dolphin distribution. Dolphin aggregations coincided with the steepest gradients between water masses along eastern regions of the Peninsula, and strong exchange events along the edges of the study area. The continued survival of this endangered species is contingent on its protection. Long-term monitoring programmes are needed for the three main strongholds identified in this study. The occurrence of Hector�s dolphin 'hotspots' along frontal zones within Banks Peninsula also suggests alternative and increased protection strategies are needed for this sanctuary to be effective. In light of this thesis� findings and based on marine protection research, future sanctuaries need to consider why Hector�s dolphins are preferentially using particular regions and how their association with certain oceanographic features can help make informed decisions on more appropriate protected areas.

Hector�s dolphin (Cephalorhynchus hectori) is an endangered coastal species endemic to New Zealand. Their distribution, like other marine organisms, is intertwined with the dynamics of their local habitats, and at a larger scale, the coastal waters around New Zealand. The main purpose of this thesis was to identify specific habitat requirements of this rare dolphin. Hector�s dolphin distribution around the South Island was quantified along several temporal and spatial scales. Large-scale density analyses of abundance surveys found over half of the South Island�s current population occurred within only three main regions. Two of these strongholds are along the west coast and the third is located around Banks Peninsula on the east coast. Smaller-scale analyses at Banks Peninsula found the majority of the dolphin community was preferentially using core regions within the marine mammal sanctuary. Monthly surveys showed that in summer and autumn statistically more dolphins occurred within inshore regions ([less than or equal to]one kilometre), spread throughout the surveyed coastline. From May through winter, dolphin densities rapidly declined. Remaining dolphins were significantly clumped in more offshore waters of eastern regions. The lowest encounter rates occurred between August and September. Certain 'hotspots' consistently had higher dolphin densities throughout the study period while others were preferred seasonally. To address habitat preferences, surveys simultaneously collected oceanographic samples using a CTD profiler. In general, physical variables of the Peninsula�s eastern and southeastern waters varied less, despite being regularly exposed to upwellings and the varied presence of sub-tropical waters. Semi-sheltered bays and shallow inshore waters were highly variable and more susceptible to spatially discrete influences, such as localised river outflows and exchange events. Several hydrographic features were seasonally predictable due to their dependence on climate. The stratification and location of the two dominant water masses (neritic and sub-tropical) accounted for over half of the temporal and spatial variability observed in oceanographic data. Possible relationships between oceanographic features and aggregations of dolphins within Banks Peninsula were examined using global regression and a spatial technique known as geographical weighted regression (GWR). GWR models out-performed corresponding global models, despite differences in degrees of freedom and increased model complexity. GWR results found relationships varied over localised scales that were concealed by global methods. Monthly GWR models suggested the seasonal presence and strength of local oceanographic fronts influenced dolphin distribution. Dolphin aggregations coincided with the steepest gradients between water masses along eastern regions of the Peninsula, and strong exchange events along the edges of the study area. The continued survival of this endangered species is contingent on its protection. Long-term monitoring programmes are needed for the three main strongholds identified in this study. The occurrence of Hector�s dolphin 'hotspots' along frontal zones within Banks Peninsula also suggests alternative and increased protection strategies are needed for this sanctuary to be effective. In light of this thesis� findings and based on marine protection research, future sanctuaries need to consider why Hector�s dolphins are preferentially using particular regions and how their association with certain oceanographic features can help make informed decisions on more appropriate protected areas.