Trade-offs between biodiversity conservation and maintaining fisheries yield from Australian marine environments : approaches using the Atlantis ecosystem modelling framework

There is often a conflict between conservationists and the users of natural resources. This is just as much the case in marine management as it is on land. In order to practise ecosystem based fisheries management (EBFM), as mandated by law, we must address this conflict and find a system of management that protects both the ecosystems and those that exploit them for food production. This can only be done by increased understanding of the systems, both through empirical research and by developing tools, such as models that allow us to investigate trade-offs and alternative management strategies. Ecosystem models are becoming a popular tool for management strategy evaluation (MSE), as well as to explore ecosystem dynamics in the marine environment. This thesis examines aspects of the utility of the Atlantis ecosystem modelling framework, by: ‚Äö testing its ability to capture ecosystem dynamics in southern Australian waters, especially in and around marine protected areas, which are data rich areas; ‚Äö assessing the effects of alternative parameterisations and spatial structures within the model; and ‚Äö using it to assess possible impacts of various fishing management regimes. Preliminary examinations in the first chapter of the thesis describe the results of simulations using a relatively simple model domain. The main findings of these examinations indicate that mesopelagic fish have the potential to form a critical link in south Australian ecosystems, possibly acting as a keystone species in the system. The ecosystem model is then further developed in subsequent chapters, with a new modelling approach introduced for managing the spatial scale in models. A telescoping technique is explained that allows a large spatial domain to be modelled with a fine scale spatial structure where warranted and a coarse scale used to represent areas with fewer data available for parameterisation. This approach is an innovative extension of previous spatial resolution concepts that extends well beyond the range of scales previously considered. The implementation of this approach means the modelling framework is appropriate for exploring finely resolved spatial questions (e.g. around spatial management). This new capability is then used to test the results of the preliminary findings of the simple model, with additional model structures and parameterisations also compared to examine how the model structure may influence results. This examination reveals that spatial structure, as well as the way that trophic connectivity is parameterised, plays an important role in determining the importance of different species within the modelled ecosystem. The telescoping approach is also used to explore the further management issue of the effectiveness of no-take zones for providing increased yield outside of the protected area. This examination indicates that although no-take zones may help preserve biodiversity within the protected area, they are not sufficient to ensure sustainable fishing within the wider ecosystem context. This thesis describes the benefits and appropriate utilisation of large, end-to-end ecosystem models, while also highlighting the limitations that must be acknowledged and taken into account. End-to-end ecosystem models cannot and should not be used as absolute predictors of ecosystem dynamics under perturbation. However, they are useful for informing on qualitative impacts that may arise from various ecosystem processes and uses. Furthermore, the work presented here demonstrates that the parameter set chosen for a model will have significant impact on model output, highlighting the importance of empirical research to provide high quality data to furnish ecosystem models.