Analysis and Geovisualisation of Hector’s Dolphin Abundance and Distribution Patterns in Space and Time

Abstract

While Hector’s dolphin (Cephalorhynchus hectorii) has been the topic of many research projects within the first Marine Mammal Sanctuary in New Zealand, few long-term analytical abundance and distribution projects in other population strongholds have been conducted. The primary purpose of this thesis project was to test quantitative observations that suggested that this unprotected population of Hector’s dolphin at Te Waewae Bay, on the south coast of the South Island, New Zealand, may be in decline and utilises non-continuous portions of the coastline. Seasonal patterns of distribution and density were extracted from a rich data set collected over 24 consecutive months that provided fine-scale data of encounters with dolphins along four preplanned transects that followed the concave nature of the bay. Monthly data were binned into seasons producing eight seasons of data over the two years. Survey results revealed that Hector’s dolphin in warmer seasons were found in greater densities closer to shore and that in the cooler seasons the range extended outward and across more offshore areas. Individual seasons did not have as strong a pattern as the complete two year data set that indicated hotspots of higher densities of dolphins in the vicinity of freshwater inputs into Te Waewae Bay. To explore individual spatio-temporal movement patterns and how the individual patterns relate to group spatio-temporal patterns, 58 individual Hector’s dolphin movements were extracted from geo-tagged photographic data and then analysed. Visual analysis of movement patterns of individual dolphins were found to vary dramatically, having distribution patterns that exhibited a high degree of site fidelity. Most notable were the twenty one dolphins that remained in relatively small areas on either the east (ten dolphins) or west (eleven dolphins) halves of the bay. This evidence of strong site fidelity may suggest partitioning along as yet unidentified social or environmental parameters. Abundance estimates were calculated from mark-recapture photographic identifications. Calculations using Pollock’s Robust Design were limited to seasonal estimates of the total population of Hector’s dolphins, which ranged from the low in winter 2005 of 380 (CV=13%; 95% CI, 300-500) to the high in summer 2005/2006 of 580 (CV=9%; 95% CI, 480-700). The estimates from these eight seasons correspond to the numbers of dolphins that utilise the bay as their primary homerange and indicate that the population is not yet in a critical decline. However, caution is urged in interpretation because two years of field data is insufficient to calculate robust survival or reproduction rates for such a long lived species. To examine whether statistically quantifiable relationships exist between environmental variables and dolphin distribution patterns, both global (ordinary least squares; OLS) and local regression (geographically weighted regression; GWR) modelling techniques were applied. The local model was a spatially explicit model. The GWR model outperformed the OLS model, revealing statistically significant hotspots directly related to the amount of rain falling four days prior to the surveys being conducted as well as to distance from the main source of freshwater in the bay. The outcomes from this thesis offer a robust baseline of information regarding the population of Hector’s dolphin in Te Waewae Bay, such that management will have the capacity to monitor long term changes in abundance and distribution. This thesis findings have suggested that freshwater input may play a crucial role in Hector’s dolphin distribution in Te Waewae Bay, which when added to previous research results indicating the importance of oceanic frontal zones, water clarity, and depth, suggests that the picture of habitat requirements for Hector’s dolphin is becoming less obscure.