Among efforts to develop sustainable approaches towards the intensive rearing of finfish within open marine waters, is the development of integrated aquaculture techniques. Integrated Multi-Trophic Aquaculture (IMTA), has been promoted as a way to reduce unwanted environmental impacts associated with the intensive production of marine finfish within net-pens. The principle aim of this concept, is the bioremediation of nutrient discharges from fish aquaculture. This is to be achieved by integrating fish cultivation with the growing of species from lower trophic levels, which use the nutrient discharges as a food source. Many studies have been performed that investigate the ability of various species of macroalgae to remove dissolved nutrient discharges, and the ability bivalves to remove solid-bound nutrients, presented as either fish faeces, or an enhanced production of phytoplankton that may be promoted by nutrients emitted by fish-farms. IMTA has also been suggested as a means to improve overall productivity per unit of feed applied to fish, through the conversion of nutrient emissions into additional biomass, such as the tissues of macroalgae or bivalves. Within the research community which focuses upon the environmental impacts of aquaculture, there is a growing awareness that sustainable solutions to aquaculture production cannot be realised through a focus restricted to the growing-phase, and to a limited set of environmental impacts which may this activity may produce. This is because changes to a specific production phase often promote changes at phases located elsewhere along a products value chain. Life-Cycle Assessment (LCA), is a method employed for modelling the environmental impacts that may potentially be generated across the value chain of a product. It is particularly useful for identifying instances of environmental impact shifting; a term used to describe situations where efforts to reduce the contribution of a specific production phase towards one or more environmental impacts, has the effect of either displacing this contribution to another phase, or increases the contribution of production towards different environmental impacts. Despite its apparent suitability, LCA has not previously been fully applied to the environmental impact modelling of open-water IMTA systems. The work presented in the following thesis advances this research front, by using LCA to explore the potential for environmental problem shifting occurring as a consequence of replacing intensive monoculture production, with IMTA. Comprehensive datasets have been acquired from the Chilean aquaculture industry, describing the production of aquafeed and Salmo salar, as well as for the production of the Phaeophytic macroalga, Macrocystis pyrifera, and the molluscan bivalve, Mytilus chilensis. Using LCA methodology, the production of salmon feed, and the production of S.salar, M.pyrifera and M.chilensis, have been assessed for their capacity to contribute towards a variety of global-scale, environmental impacts. IMTA consisting of either S.salar and M.pyrifera, S.salar and M.chilensis, or all three of these species, and combined at ratios required for a bioremediation efficiency of 100 %, 50%, or 20 % of either nitrogen or phosphorous emission from fish, is compared to the monoculture production of S.salar. The comparison is based upon a standardised functional unit, with each species produced through IMTA, being modelled as part of the reference flow required to fulfil the functional unit. Monoculture is compared to IMTA upon the basis of nutritional function, by using a functional unit of mass-adjusted protein content, and mass-adjusted economic value. The use of economic value is based upon the ‘best-case’ assumption, that it serves as a proxy for the total nutritional function that each product offers to human society. The LCAs presented in this study have produced a number of results. Salmon ingredients derived from agricultural crops and animals account for the majority (between 71 % to 98 %) of contributions towards the impacts of compound salmon feed. In general, agricultural crops ingredients contribute more to these impacts than do agricultural animal ingredients, and account for between 31 % and 87 % of the contributions from all ingredients and inputs. In contrast, the combined supply of fish meal and fish oil from capture fisheries is responsible for between 0.13 % and 11 % of all impacts. Vegetable oil accounts for the vast majority of contributions from ingredients derived from agricultural crops. Vegetable oil is modelled as a 50 : 50 blend of sunflower oil and rapeseed, oil, but sunflower oil accounts not only for most of the contributions from vegetable oil, it is responsible for over 50 % of the contributions that all agricultural crop based ingredients contributes towards some impact categories. Replacing sunflower oil with rapeseed oil reduces the contributions of salmon feed by between 6 % and 24 % across 10 out of the 11 impact categories. When compared upon the basis of equal weight, the contributions of fish oil are between 18 % and 99 % lower than those from rapeseed oil. The production of feed is responsible for the majority of contributions (between 32 % and 86 % ) to all impacts of salmon grow-out production. The production of salmon-smolts accounts for between 3 % and 18 %. The majority (64 %) of contributions towards the eutrophication potential of salmon production are from nutrient emissions, which are the result of fish metabolism, whilst nutrients released through the production of feed, the majority of these being from the agricultural production of crop and animals, account for 32 %. Feed production is also a major contributor to the impacts of land-based smolt production, but these contributions (between 12 % and 37 % across all impact categories) are of a lower magnitude than those from the supply of feed to the grow-out phase. Inputs of salt, and inputs of both electricity produced in a diesel power generator and obtained from the national electricity network, are also notable contributors (between 5 % and 67 %, 4 % and 29 %, and 2 % 47 %, respectively) towards the impacts of smolt-production. The main contributors towards the potential impacts of kelp grow-out production (excluding eutrophication potential) are the supply of infrastructure (between 14 % and 89 %), operation of a diesel-powered motorboat for maintenance purposes (between 1 % and 89 %), and the supply ‘of seeded cartridges’ (between 9 and 49 %). The major contributors from the production of ‘seeded cartriges’ in a land-based facility are the supply of electricity from the national electricity network, the supply of fresh water, and the treatment of waste water. The impact potentials of producing seed in this facility might be reduced if the scale of operation is increased. Removal of nitrogen and phosphorous upon the harvesting of kelp is calculated based upon kelp tissue contents of these nutrients. The harvesting of 200 tonnes ha / yr-1, results in a eutrophic potential with a negative value (-376.51 kg of phosphate equivalents). The removal of such a quantity of nutrients might be beneficial if the local marine environment is at risk of hypernutrification, but when no such problem is present, the potential for undesirable consequences of nutrient sequestration should be considered. The major contributor towards the impacts of mussels is the provision of infrastructure (between 25 % and 99.5 %, excluding eutrophication potential). Infrastructure is also responsible for the majority of contributions from mussel seed production. The provision of cotton mesh bags, which are used to aid attatchment of seed to drop-ropes in the grow-out phase, account for between 37% and 99 % of the contributions from the infrastructure from the grow-out phase. This result suggest that either the impacts of mussel production can be reduced by using an alternative material with lower environmental impact potentials, or the inventory data describing the producing of cottonmesh bags requires some improvement. The outcomes of the LCAs of the different IMTA scenarios, are interesting. The results show that choice of species, and the ratios of their combination as required for the different efficiencies of bioremediation, can have a significant effect upon the comparison between IMTA and monoculture. / The study demonstrates a potential for environmental problem shifting as being a consequence of IMTA, especially when the functional unit is mass-adjusted economic value. As bioremediation efficiency increases, contributions towards eutrophication decrease. However, this reduction is achieved at the cost of increasing the contributions of IMTA towards those impact categories, such as ‘ozone layer depletion,’ for which it has a greater contribution than does monoculture. In general, it cannot be concluded from these results that open-water IMTA represents a more sustainable alternative to the monoculture production of Atlantic salmon. The sustainability of IMTA is shown to be dependent upon a variety of trade-offs, between individual environmental impacts, and between these impacts and the nutritional function that the system is capable of providing.

Macroalgal beds provide the ecological foundations for most shallow reef ecosystems in temperate environments. With distinctive canopies primarily of brown laminarian algae (northern hemisphere), or laminarian or fucalean algae (southern hemisphere), in many areas these habitats are at risk from human activity. Overexploitation, pollution, and other effects of coastal activities have resulted in significant habitat loss in coastal ecosystems, and human-induced climate change is now seen as a major threat to ecosystem health in marine systems. Understanding the impact of climate change is particularly important for habitat-forming ecosystem engineers like kelps, as these species form the basis of hierarchically organised communities and play a fundamental role in determining community structure and ecological processes. South eastern Australia has experienced increases in marine temperatures at nearly four times the global average, and there is now evidence that, in some locations, macroalgae communities are retreating in a manner consistent with ocean warming. Successful management of marine systems requires understanding ecosystem processes, particularly the patterns and magnitude of production. Macroalgal communities often show relatively low resistance to disturbance, yet rapid recovery once disturbances are removed, hence they are generally highly dynamic in response to environmental perturbations. As a result, macroalgae are likely to play an increasingly important role in buffering the short term/dynamic effects of climate change on temperate reef communities.Knowledge of the productivity of seaweed-dominated temperate reef systems is largely a synthesis from studies conducted over small spatial scales utilising a variety of methods that generally measure different characteristics of both individual seaweeds and collectively. As a result of the diversity of measurement methods, estimates of gross primary productivity (GPP), production potential, and macroalgal biomass for temperate reefs are numerous and variable. This can lead to challenges for ecologists attempting to amalgamate research findings to facilitate long-term, broad-scale perspectives or compare short-term research between spatially separated communities. However, to date there has been relatively little research to compare measurement approaches and quantify differences in productivity estimates across the different techniques.The present research provides a unique investigation into some of the techniques and methodology involved in measuring primary productivity in marine systems, particularly kelp forests, using the macroalgae Ecklonia radiata, Phyllospora comosa and Macrocystis pyrifera as study species. The work is based on both field and laboratory exploration of productivity measurements and associated parameters. In situ measurements of primary productivity (diel oxygen modelling, benthic oxygen exchange chambers) or PSII electron transport (PAM fluorometry) are compared, and the possibility of using acoustics as a means of quantifying oxygen production at large scales is explored, as has already been applied in seagrass beds. This thesis also provides an in depth investigation of the effect of variability in sampling methodology with regards to interpretation of PAM fluorometry-derived parameters. Chapter 2 investigates the acoustic properties of Ecklonia radiata. The density, sound speed and resulting adabiatic compressibility of E. radiata tissue were investigated in the laboratory. Four methods were developed and trialled to determine the intrinsic sound speed of Ecklonia radiata tissue based on measurement of the time of flight of an ultrasonic pulse, while compressibility was calculated from density measurements. The results show that Ecklonia radiata sound speed and density are higher, and compressibility lower, than that of seawater. Properties varied according to size and tissue type and the variation likely reflected differences in cell type, packing and structure as well as the concentrations of alginates and other carbohydrates. These are important considerations for acoustic propagation and the results provide valuable inputs for future acoustic work. Chapter 3 focuses on the acoustic modelling of different scenarios of primary production in a shallow water rocky reef environment of Fortescue Bay (Canoe Bay), Tasmania, where E. radiata dominates the canopy. In February 2012, the environment was continuously probed by acoustic signal transmission and monitored by a comprehensive set of oceanographic sensors with the aim to assess the potential for acoustics to quantify excess oxygen production in bubble form. Ray-theory acoustic modelling results indicate that ecologically-significant void fractions of oxygen in the canopy layer from production would be clearly seen in diel variation of propagation features such as the energy decay rate of the medium impulse response. The model can then be used to invert empirical data for retrieving void fraction. However, comparative analysis of part of FORTES 12 data and model suggests that no large excess of bubbles was produced by photosynthesis under the present environmental conditions, in contrast to earlier observations made in seagrasses. As a result, the use of acoustics as a means of measuring primary productivity in kelp could not be further explored during the course of this research.Chapter 4 provides a unique comparison of the estimates of photosynthetic O2 production rates in an Ecklonia radiata dominated community using three different measurement methods: diel oxygen GPP models, benthic oxygen exchange chambers, and electron transport rate from PAM fluorometry which is usually interpreted as a measure of production potential. All three methods were run concurrently in situ in Fortescue Bay, Tasmania. The first diel oxygen model was fitted to in situ measures of dissolved oxygen (DO) in the environment and demonstrated a good fit, however, a consequence of this approach is that large variation in oxygen production was predicted at low PAR levels. A second model was created which utilised an explicit relationship between DO production and PAR, but it didn’t represent DO at the surface as well as the first model. Importantly, the two models indicate similar daily production rates of the seaweed bed (all species combined) that are ~ 2 times that predicted for the kelp alone based on incubations in the benthic chambers and scaling for the average size of adult kelp sporophytes and their population density. Oxygen evolution from incubation of sporophytes in benthic chambers and PAM fluorometry derived electron transport rates showed similar patterns, but the results indicate that the latter method may overestimate potential photosynthesis. The results suggest that diel oxygen modelling, benthic oxygen exchange chambers and PAM fluorescence can all provide good indications of productivity in shallow water marine environments. However, care must be taken in interpretation of results as each method differs in the type of productivity estimates it provides. As a direct measure of total seaweed production per unit area of reef, estimates from models based on empirical measures of environmental DO have much to recommend them.Chapter 5 details a final analysis investigating the effects of diurnal, seasonal and latitudinal variability in ambient light on PAM-derived parameters, as well as possible effects associated with depth, within- and between-alga variation in PSII performance, and latitudinal effects unrelated to the light climate. This research was based on field measurements undertaken in Tasmania, Western Australia and New South Wales, Australia in both summer and winter during 2012 and 2013, focussing on Ecklonia radiata, Macrocystis pyrifera and Phyllospora comosa. Photosynthetic characteristics of all species were highly dependent on the time of day, depth, latitude/region, season, and part of the thallus from which measurements were taken. Patterns dependent on time-of-day, depth and thalli placement varied with season and/or geographic region, and the nature of these patterns varied between species. It is clear from this work that efforts to standardise approaches to taking measurements of seaweeds using PAM fluorometry will be essential if measurements are to be compared meaningfully across studies.The key findings of this thesis are: (1) a first determination of the acoustic properties of E. radiata tissue which enable the development of scattering models to interpret scientific echosounder data collected in kelp beds; (2) a Gaussian beam/finite element beam code (Bellhop) with detailed environmental input and a huge number of beams can predict the acoustic character of a shallow water rocky reef and bubble layers with low-frequency effective sound speed; (3) the model allows prediction of the acoustic energy decay rates due to various scenarios of ecologically-relevant photosynthetic O2 production rates; (4) day vs night acoustic measurement and model data comparisons challenge void fraction predictions made from well established measurements and methods; (5) diel oxygen modelling, benthic oxygen exchange chambers and PAM fluorescence can all provide good indications of productivity, however, understanding the limitations of each method is essential when interpreting the results as the measurements they provide are not directly comparable; and (6) applying a consistent sampling methodology is a key consideration when planning research utilising PAM fluorometry as diurnal, seasonal, and latitudinal variability, as well as effects associated with depth and within- and between-alga variation in PSII performance will have significant impact on PAM-derived parameters. The results of this work give valuable insight into the advantages and disadvantages involved with several main techniques currently utilised to measure production of macroalgal/seagrass beds, and the challenges faced by ecologists attempting to interpret results and compare research between methods and across studies. Last but not least, this study provides important and relevant information on the potential use of acoustics as a future means of determining productivity of benthic habitat on large scales in marine environments. The work presented herein will assist in both development and interpretation of future study of productivity in marine systems. / Doctorat en Sciences de l'ingénieur et technologie / info:eu-repo/semantics/nonPublished