<p>Monitoring marine ecosystems is essential for the conservation and management of marine biodiversity as it is central to the development of sustainable management practices and for assessing the effectiveness of the increasing number of marine reserves (MR) globally. Monitoring data are often collected in MRs to assess the state of natural marine systems in the absence of anthropogenic disturbance or to assess recovery of previously impacted species. In recent years, MR designation has attempted to move away from ad hoc approaches to MR establishment and towards using existing species distribution and abundance data to define protected areas. Given the logistics and cost of collecting biological data in the marine environment, effective methods are required to successfully demonstrate changes associated with MRs and to identify the spatial distribution of organisms and habitats for the planning of further MRs. The aim of this thesis was to identify effective protocols for the monitoring of fish and invertebrate species inside MRs in New Zealand, and to develop and apply methodologies to identify spatial distribution patterns relevant to marine spatial planning. Using baseline data of fish and invertebrate species abundances for the Taputeranga MR I performed prospective power analyses to identify the most cost-effective monitoring approach for subsequent monitoring. Based on before-after-control-impact (BACI) tests the power to conclude statistically that abundances were higher at MR sites was low for even large simulated changes in abundance (two-fold or four-fold increases) for most species. Due to differences in baseline abundance and spatio-temporal variance terms, power varied considerably among species, highlighting the difficulty of monitoring all species to the same degree, whilst also remaining cost-effective. Furthermore, the results highlight the need for temporally replicated survey designs as “one-off” surveys had much lower power than those that were temporally replicated. Longer term monitoring effectiveness was analysed using three long-term datasets from MRs in the South Island of New Zealand. I analysed the power of alternate underwater visual census (UVC) monitoring configurations to conclude statistically that there were increasing/decreasing trends in abundance, as well as the precision and accuracy of trend estimates. Overall even the highest replication designs considered had low power (< 80%) to conclude there was a non-zero trend even when simulated data represented trends equivalent to the population doubling or halving over ten years. The most cost-effective monitoring design varied among species and MRs, further highlighting that monitoring choices need to be location- and species-specific. A general finding, however, was that increasing the number of sites was almost always more beneficial than increasing the number of transects per site. Based on these results, I recommend that monitoring design planning focuses more specifically on assessments of precision and accuracy of estimated parameters, with less focus on power, as this places greater emphasis on interpreting monitoring data in terms of potential biological significance rather than testing for statistical significance. Monitoring can never achieve complete coverage of large areas therefore methods for extrapolating or predicting species or habitats to un-surveyed locations are necessary for evaluating large-scale spatial distributions. To address this I used modelling techniques to identify the spatial variation in species and habitats along the Wellington south coast, with a particular focus on elucidating the potential and realised effects of wave exposure. A wave simulation model (SWAN) was used to identify the spatial variation in wave exposure relevant to intertidal and subtidal communities. In particular the spatial variation in wave forces was compared to the distribution of two subtidal macroalgal species, Macrocystis pyrifera and Ecklonia radiata, taking into consideration the biomechanical thresholds of damage for these plants. Despite considerable wave forces during winter storms, healthy E. radiata is unlikely to be damaged, whilst larger (>15 m stipe length) M. pyrifera plants are likely to be damaged in certain locations dependent on local sheltering effects. Furthermore, the distribution of M. pyrifera from aerial imagery coincided with areas that were predicted to have lower wave forces, suggesting that the distribution of M. pyrifera may be related to wave exposure. I subsequently constructed species distribution models revealing the relationship between intertidal species distributions and environmental factors, as a predictive baseline of the current distributions of species. The abundances of Chamaesipho barnacle species were found to be best described by wave exposure, with increased cover correlated with increasing wave exposure, while contrasting patterns were observed for C. brunnea and C. columna with respect to distance from the harbour entrance, suggesting differential larval supply or differential responses to changing water column characteristics. Macroalgal assemblage composition was explained predominantly by wave exposure, with a rich macroalgal assemblage at the less exposed locations, and more exposed locations exhibiting a community consisting of coralline algal species and the large brown alga Durvillaea antarctica. The predictive models were then used to predict species distributions for a section of coastline demonstrating how this form of modelling can be used to maximise the potential of monitoring data. Finally, a literature keyword search along with methodological developments and results from previous chapters are used in the final chapter to develop a framework for the collection of data from the planning phase all the way through to long-term monitoring of MRs.</p>
Covid Patient Health Monitoring Using IoT During Quarantine
