The southern rock lobster (SRL), Jasus edwardsii (Hutton, 1875) (Crustacea: Decapoda: Palinuridae), is one of Australia and New Zealand's most valuable fishery resources. A large scale and prolonged reduction in the recruitment of SRL during years 2000 ‚Äö- 2010 translated into significant stock declines across the Australian fisheries. The geographical range of SRL spans more than 5000 km from Western Australia to New South Wales along the southern coast of Australia including Tasmania and around the entire New Zealand. Connectivity between distant populations is achieved solely through larval dispersal with adult movement being limited (Booth, 1997; Linnane et al., 2015). The SRL has a pelagic larval duration (PLD) of up to 24 months, one of the longest known in the marine environment and the longest among all rock lobsters (Booth, 1994; Bradbury and Snelgrove, 2001). SRL larvae can be carried hundreds of kilometres offshore and away from their origin, potentially connecting spawning grounds and recruitment sites hundreds of kilometres apart. This potential for widespread dispersal combined with unpredictable inter-annual and spatial variability of egg production and recruitment, make biophysical modelling an ideal approach to examine population connectivity for this species. This thesis assesses the population connectivity of SRL throughout its geographical distribution by the means of a larval dispersal model built utilising the best available hydrodynamic models and evidence based species-specific biological parameters. Prior to setting up the larval dispersal model, I performed a validation of two hydrodynamic models available in the study area, by comparing the model predicted time series of seawater temperature and ocean currents to in situ mooring measurements (Chapter 2). I found that the accuracy of the hydrodynamic models varied with the parameter investigated, the depth of the measurement and the geographical region. The model predictions of water temperature were more accurate than the predictions of ocean current velocities. This study identified important inaccuracies in the hydrodynamic models' estimations of ocean parameters and on time scales relevant to larval dispersal studies. The largest errors in global ocean models are seen in coastal regions, where models have poor coverage due to their lower spatial resolution. Many global ocean models also do not explicitly resolve key hydrodynamic features in the coastal regions. I investigated the effect of nesting a highly-resolved coastal hydrodynamic model ‚Äö- ETAS ‚Äö- within the global BRAN model, on the passive dispersal of larvae released on the east coast of Tasmania (Chapter 3). I found significant differences in larval trajectories and dispersal metrics between the simulations using the nested ocean models and the simulations using only the global ocean model. Finally, an individual-based biophysical model was built for SRL larval dispersal (Chapter 4), the larval survival during dispersal and the probability of pueruli successfully settling in suitable habitat were estimated (Chapter 5). Larvae were released throughout the species geographic range during the egg hatching period over release dates spanning twenty years. In addition to dispersal metrics, I report the connectivity matrix between 16 fishery management zones across the dispersal domain. Dominated by the large west-flowing currents in the study region, the main larval transport was from west to east. The highest rates of survival to settlement were seen in larvae that metamorphosed early during the competency window or within few hundred of km from their release locations. South Australian and Victorian fisheries were important larval sources for most other fisheries east from them, while Victorian and Tasmanian fisheries received the largest proportion of successful pueruli from all other fisheries. The highest self-recruitment rates were observed in New Zealand, Northern South Australia and Tasmania 2 fisheries. This study offers valuable insight into the applications and limitations of hydrodynamic models in larval dispersal modelling and calls for the development of highly-resolved coastal ocean models. The SRL larval dispersal model developed in this work provides new information on this species' potential for dispersal and the predicted population connectivity can be used to inform new fishery policies and help improve management of the SRL stock.
Copyright 2018 the author Chapter 2 appears to be the equivalent of a post-print version of an article published as: Vasile, R., Hartmann, K., Hobday, A. J., Oliver, E., Tracey, S., 2018. Evaluation of hydrodynamic ocean models as a first step in larval dispersal modelling. Continental shelf research, 152, 38‚Äö-49