Tracking of striped marlin (kajikia audax) in the southwest Pacific Ocean : environmental influences on movement and behaviour

Abstract

The aim of this thesis was to investigate the movement and behaviour of striped marlin {Kajikia audax} within the southwest Pacific Ocean. Satellite tags were successfully deployed on 29 marlin over the years 2005-2008, including 25 being double tagged with both pop-up satellite archival tags and satellite-linked radio telemetry tags. Double tagging allowed assessment of the accuracy of light level geolocation models, and enabled error distributions to be parameterized into a recently developed state-space interpolation technique. Movement data was then used in a behavioural classification model which provided a framework for addressing specific questions about the effects of capture on post-release behaviour, foraging patterns, spatio-temporal distribution of the population, and environmental factors influencing these patterns. Behaviour classification modelling enabled objective estimates of when and where switches between area-restricted behaviours (a proxy for foraging behaviour) and transiting behaviours occurred. Behaviour estimates analyzed in conjunction with depth distributions of double tagged individuals revealed a tendency to forage closer to the ocean's surface, and to utilize more of the water column when in searching mode. In addition to identifying these behaviours, striped marlin tended to reverse directions upon reaching 20-21⁰S latitude, suggesting behavioural and environmental mechanisms may help explain the distribution of the population within the southwest Pacific Ocean. The probability of foraging was diminished during the first 16 days at liberty, indicating fitness and behaviour are impacted in the short/medium term by the capture and/or tagging process. The behavioural bias identified during the two week post-release period suggests the timing of migrations may be altered by human interaction, and has implications for the spatio-temporal structure of stock assessment models. Linear discriminant function analysis (LDFA) indicated directional reversals at 20-21⁰S latitude were driven by bathymetric features, sea surface height, wind speed and mixed layer depth. The most significant predictors of foraging behaviour were sea surface temperature, chlorophyll-a concentrations, mixed layer depth, sea surface height and wind speed. The East Australian Current is a dominant oceanographic feature in the southwest Pacific Ocean, and is a strong influence on striped marlin population distribution within the Tasman Sea and northern New Zealand.