Seagrass beds are highly productive shallow-water marine ecosystems and serve as a habitat for numerous invertebratesa nd fishes, some of which constitute important fisheries species.S eagrassb eds play an important role in physical coastal dynamics, since they stabilise soft bottoms, while their leaf canopy attenuates strong wave action and currents, thereby, reducing coastal erosion. However, seagrass habitats are undergoing a decline worldwide, hence, there is an urgent need for data on the influence of bed fragmentation and changes in plant architecture on the diversity of associated biota. In the Mediterranean Sea, the endemic seagrass Posidonia oceanica forms extensive beds that have a high structural complexity and support a high biodiversity, making them one of the most important marine habitat types in the region. The present study investigated the influence of P. oceanica bed structure (seagrassb ed morphology and within-bed architecture) on the diversity of associatedm acrofauna.P ilot surveys were made to map the occurrence and spatial distribution of different types of P. oceanica beds present off the northeastern coast of the Maltese Islands, to enable selection of four appropriate study locations. A quantitative sampler, which samples the total macroinvertebrate assemblages of P. oceanica beds, was designed and tested to determine the appropriate area and number of sampling units required to achieve given levels of accuracy and precision. Using data gathered from the pilot survey, studies were made to examine for differences in P. oceanica plant architecture, and in the species richness and abundance of motile macroinvertebrates between continuous (non-fragmented) and reticulate (fragmented)b eds of the seagrassa, crosst hree different spatial levels: (a) large (severalk m), (b) medium (circa 100 in) and small (tens of metres). A separate study examined differences in the species richness and abundance of associated motile macro invertebrates between living matte (bearing living shoots) and dead matte (without living shoots) of P. oceanica. Results of univariate and multivariate analyses indicated that there was no consistent pattern of significant differences in plant architecture,a nd in the speciesr ichnessa nd abundanceo f associatedm otile macroinvertebratesb etween the two P. oceanica bed types. Significant differences in macroinvertebrate species richness and abundance were detected between P. oceanica beds at the large spatial level that were significantly related to seagrasse piphytesa nd mean sedimentg rain size. Significant differences in macroinvertebrate species richness and abundance were detected between P. oceanica beds at the medium spatial level that were significantly related to shoot biomass. Significant differences in the assemblage composition of macroinvertebrates associated with P. oceanica beds were detected by multivariate analysis at the large and small spatial levels, with the best explanation for the multivariate assemblage structure being a combination of environmental variables that included epiphytes and sediment grain size. The results of the study comparing the motile macroinvertebrates of living and dead matte showed that dead matte supported significantly higher species richness and abundance of associated motile macro invertebrates, and had a significantly different macroinvertebrate assemblage structure, compared to living matte. Furthermore, a greater number of species occurred exclusively in dead matte than in living matte. It was concluded that natural fragmentation of continuous to reticulate P. oceanica beds does not lead to changes in plant architectural differences, or to a reduction in diversity of associated macroinvertebrates, nor does complete loss of the foliar canopy necessarily lead to a reduced diversity of macrofauna. Local factors, including nutrient levels in the water colunm (which enhanced epiphytic growth on the seagrass), sediment grain size and the hydrodynamic regime, were more important in influencing the diversity of macrofauna associated with P. oceanica beds, than bed type, leading to significant spatial variation in the associatedm acroi nvertebrate assemblagesT. he study concluded that the relative importance of different factors in influencing the diversity of macro invertebrates associated with P. oceanica beds varied, depending on scale. The results are discussed in the light of the available published data on fragmentation of seagrass habitat and on current Euro-Mediterranean proposals for the conservation of different bed types of P. oceanica. Attempts to classify different bed types of P. oceanica in categories according to value should be based on adequate ecological data. Different bed types of P. oceanica all form part of a dynamic heterogeneousla ndscapew here transformation from one bed type to another may occur, but which does not necessarily lead to loss of biodiversity. Hence, equal effort should be directed at conserving and managing different bed types of P. oceanica.

No description available.

Habitat structural complexity (HSC) is a key component of natural ecosystems. It helps to describe the morphological characteristics of the three-dimensional space that floral and faunal communities exist within. The physical structure can have a profound influence on the associated diversity and functioning of the ecosystem. This thesis is a macroecological study of Caribbean coral reef HSC and how changes in it might influence associated fish communities. An assessment of contemporary reef HSC in 15 countries and its relationship with spatial, environmental and anthropogenic variables, identified the degree and frequency of physical disturbance as significant drivers. There was also a strong link between live coral and HSC indicating that, despite region-wide declines in coral cover, there has not been a decoupling of complexity and coral, and that healthy coral populations facilitate the persistence of structurally complex coral habitats. To examine the relationships between different measures of HSC and their relevance to the fish community, a multidimensional model of HSC was developed. Through this approach it was possible to combine a number of variables which individually indicate different components of reef HSC, and derive compound ‘refuge’ and ‘complexity’ variables which are expected to be of greater relevance to the fish community, and potentially of greater use to understanding macroecological relationships on coral reefs. This approach was used to identify relationships between the Caribbean reef fish functional community structure and the derived refuge variable. Sites with low refuge levels were dominated by generalist fish species and had low levels of functional diversity. Additionally, the biomass of a number of fish functional groups and the size ii structure of groups was found to be related to the refuge characteristics of coral reefs. The examination of fish communities at night revealed a continued relationship with HSC over the entire diel cycle for some functional groups while others varied in their association with reef structure. Nocturnally active groups decoupled from HSC as they moved over flatter areas of reef to feed, while diurnally active groups moved into the reef to shelter. Observations of behaviour and feature preferences identified potential drivers behind the numerical relationships between the fish community and HSC. Some species have strong behavioural associations with HSC or features of reef structure and may suffer as result of habitat complexity declines. However, small non-fisheries target species such as wrasse and damsels, with less direct physical ties to the habitat structure are those most likely to persist after habitat degradation. The thesis highlights the state of and threats to Caribbean HSC and the pivotal role it plays in determining fish community structure.

Planktonic foraminifera are unicellular zooplankton, whose calcium carbonate ‘shells’, wide geographic distributions and very large population sizes combine to give them perhaps the best fossil record over the last 66 Ma of any group. Site-level assemblage diversity can be estimated comparably in the Recent and through geological time. In this thesis I model the environmental correlates of coretop (Recent) planktonic foraminiferal diversity (Chapter 2), with the aim of applying the model to the Eocene. Present-day diversity patterns are shaped by a richer combination of factors than suggested by previous work. I assess the potential of several non-biological biases to distort diversity patterns (Chapter 3). Functional and evolutionary diversity are less prone to bias than are species richness and evenness, while water depth has little impact on diversity in sites deeper than 500m. Asexuality has been suggested as an adaptation in low diversity environments. I used NanoCT scans of proloculi to test whether Neogloboquadrina pachyderma, the dominant species in polar waters, contains an asexual morph (Chapter 4), finding no support for this hypothesis. Having dealt with potential sources of bias, I use models from Chapter 2 to predict diversity in another time period, the Eocene (Chapter 5), based on current understanding of Eocene environments. The latitudinal gradient of species richness developed through the Eocene in both planktonic foraminifera and coccolithophores. Predicted and observed diversity fit well in the late – but not the early – Eocene. My analyses support two explanations for the early-Eocene mismatch. First, early Eocene climate model estimates of environment differ from the proxy records (which fit the fossil data better). Second, the intercorrelations among facets of diversity have changed through time (Chapter 6). Despite our limited understanding of some aspects of their biology, planktonic foraminifera have much to offer as a model system for macroevolution.

An investigation of a copper contaminated mine drainage stream, the Darley Brook, revealed that the water quality was relatively constant over time, with an exceptionally high copper concentration (mean 0.89 mg lˉ¹ ) and low pH (mean 5.2) at its source. Virtually all of the copper was in a 'soluble' form with upto 29% as the free cupric ion. Overall mean leachable copper concentrations of 2139 µg gˉ¹ and 101 µg gˉ¹ were recorded in the uppermost reaches of the Darley Brook and a control stream respectively, with little temporal variation. Biotic communities in the uppermost sampling stations were of reduced diversity. Excessive growth and high copper concentrations (maximum plant tissue concentration of 3664 µg gˉ¹) were recorded in Jungermannia atrovirens and Microspora sp. in the riffles and the pool-dwelling Juncus bulbosus. In these tolerant plants copper concentrations were higher in the summer than in the preceding winter. Macroinvertebrates were reduced to chironomid larvae, Coleoptera and the net-spinning caddis Plectrocnemia conspersa. This species was univoltine and larvae were more aggregated and occurred at higher densities (upto a mean of 126 mˉ²) in the contaminated stream than in the control stream. Each larval instar of P. conspersa from the Darley Brook contained significantly more copper than the same instar in the control stream. There was an exponential decrease in larval copper concentration with increasing weight for both populations, and maximum concentrations were always recorded in the summer. An histological and ultrastructural investigation of P. conspersa larvae revealed the presence of copper containing granules in the cells of the malpighian tubules and in the subcuticular region. These may be primarily pigment granules, which serve to immobilise excess copper within the larva. The metal tolerance of P. conspersa was confirmed by transfer experiments and discussed in relation to the occurrence of this species in metal rich waters.

Research into speciation and t~e processes governing divergence within species has received considerable attel1tion from ecologists and evolutionary biologists in recent years. However, few empirical studies have focused on marine systems, despite the potential evolutionary consequences of the unique conditions in this environment. The coral reef fish genus Hypoplectrus represents an excellent system with which to study marine divergence and comprises 10-14 distinct colour morphotypes. This thesis focuses on the Hypolectrus colour polymorphism and considers how these morphotypes differ genetically, morphometrically and ecologically. Furthermore, the distribution of morphotypes is resolved and the potential implications of these geographical patterns considered. All of these studies include data on multiple colour morphotypes from widely dispersed geographical locations. AFLP molecular marker analysis shows morphotypes to have consistent low level genetic isolation among each other. Some of the markers tested showed significantly higher levels of isolation than expected, suggesting they may be linked to genes under selection. Stable isotope analysis shows little evidence to suggest ecological divergence between morphotypes and they do not appear to possess distinct feeding niches. Morphotypes show quite variable dietary patterns across their range, suggesting against ecological adaptation involving diet. Geometric morphometric results show small but significant differences between morphotypes, which may be due to genetic isolation rather than physical adaptation. The diversity of morphotypes across the Hypoplectrus distribution shows interesting patterns, which suggest morphotypes do not co-occur randomly and that the number of morphotypes present on a reef is positively correlated with general reef fish species richness. Following these results, we now have an improved understanding of the Hypoplectrus colour polymorphism and there are clear directions for future research into the origin of this phenotypic radiation.

In the last century, seagrass ecosystems have suffered a worldwide decline ascribed to multiple environmental stressors, among which the reduction of light available for the photosynthesis and the increase in temperature represent the strongest constraints for their growth and survival. Despite conservation, this decline is at present still continuing. in order to understand the genetic adaptive response to light and temperature in the seagrass Posidonia ocxanica, two different strategies have been pursued: a genome scan approach along a latitudinal and a bathymetric gradient and a differential gene expression analysis along the bathymetric gradient, where light and temperature were the two main selective factors. For the genome scan approach 6 populations (Delimara - Malta, Lacco Ameno - Island of Ischia, Italy, Marettimo Island- Italy, Me10ria - Italy, Piombino - Italy and Stareso - Corsica, France) were sampled along the bathymetric gradient at two different depths (-5m and -25m). The same populations were used for the latitudinal gradient analysis by grouping them on the basis of their geographic location (Southern group: Delimara. Lacco Ameno and Marettimo; Northern group: Meloria, Piombino and Stareso). No genes under selection were identified in the genome scan along the bathymetric gradient. Three putative genes under selection were identified in the genome scan along the latitudinal gradient and were involved in the photosynthesis and in the translation process. For assessing differential gene expression, a transcriptome sequencing of plants sampled at two different depths and different times of the day in the Stareso meadow was performed by RNAseq technology. The analysis highlighted the capability of plants living in shallow waters to cope with environmental stresses imposed by high light and high temperature. Transcriptome data generated from this study increased the resources available in P. oceanica and will be very useful for further investigations of the adaptation of in this plant

Continental shelves are characterised by an active benthic system which is closely coupled to the pelagic: hence, the distinction between new and recycled production is not clear since the short and long-term processes are effectively coincident in space. In shelf ecosystems up to 40% of phytoplankton production can settle to the seabed and a large proportion of organic matter remineralisation occurs in the benthic compartment. However the benthic response of phytoplankton sedimentation is not unique and the receiving sediment type is a key factor determining the organic matter remineralisation rates. The aim of this project was to determine the relative importance of benthic regeneration processes at the Stonehaven monitoring site, Scotland, UK. Single and multi-beam acoustic and ground truthing surveys were carried out in the study area and the seabed landscape was described by sediment classes ranging from muddy sand to boulders and rock. General additive models were developed to link full coverage acoustic properties with sediment chemical concentrations developing full coverage maps of organic matter sources identifying the factors that can influence them. Direct measurements of sediment-water fluxes were carried out at monthly intervals throughout the year at a selected number of sampling stations identified by the acoustic analysis. The results revealed that water temperature, turbidity, sediment chlorophyll-a content and permeability were the main drivers of sediment biogeochemistry. Sediment nutrient regeneration rates contributed to the water annual budget of 25% and 89% for nitrogen and silicate respectively. Contrarily the sediment was a net sink for phosphate suggesting the importance of external supplying sources for this element. Finally, using a combined statistical modelling, we up-scaled results from a limited number of stations developing full coverage biogeochemical maps in the study area identifying 'hot spot' of sediment activity in a depositional area characterised by fine-grained sediment.

The occurrence of regime shifts in large marine ecosystems is unproven and controversial. Despite this, a number of marine regime shifts are reported to have occurred during the last century. These shifts have been largely attributed to overfishing and climate change, and have had important consequences for marine ecosystem structure and function. Understanding the relationships between ecosystem pressures and responses is a critical step towards an effective ecosystem approach to management. Ecosystem models are a useful way of investigating how ecosystems respond to changes in pressures at the component and system level. This thesis aims to (i) define a set of quantitative thresholds for defining large-scale changes in marine ecosystem components in historical data and model simulations, (ii) characterise changes in the Irish Sea marine ecosystem that were observed between 1973 and 2002, and (iii) investigate whether 3 UK Ecosim models are capable of simulating the observed ecosystem level changes in response to trends in fishing and climate. System-level changes are measured using 12 model-based indicators of ecosystem state. Marine ecological regime shifts are defined as changes in marine community structure that occur over a 1-7 year period with an amplitude of at least 46% of the largest difference in indicator value. They persist on a decadal to multi-decadal scale and are associated with changes in biological communities spanning over at least 3 trophic levels. The criteria are used to diagnose an ecological regime shift in the Irish Sea in the late 1980s. Despite the inherent uncertainties associated with ecosystem models, all 3 models show that they are able to provide insights into the ecosystem responses to fishing and climate pressures, and that other factors are likely to have been important in explaining observed changes in the UK's marine ecosystem state. This suggests that they will be useful tools for "end-to-end" ecosystem analysis, which will be important in understanding the relationships between ecosystem pressures and responses.

Ecosystem scale, adaptive management of the marine environment, emphasised by the Convention on Biological Diversity, requires a clear understanding of the spatial and temporal dynamics of the system in question. Classical species-based biogeography alone is not sufficient to describe those dynamics at the appropriate scale for whole ecosystem management. A novel, multidisciplinary, complex systems approach has been developed for characterising the biogeographic distribution of benthic fauna in a temperate shelf sea system through application of remote sensing to the principles of benthic-pelagic coupling. A six year time series of satellite remote sensing data (AVHRR SST and SeaWiFS Chl-α, LwN(555, 670)) was analysed using multivariate statistical techniques to identify the emergent patterns (temporal and spatial) of water column physical structure and associated patterns of productivity in North-West European Shelf waters. Three persistent biogeographic regions were identified from horizontal patterns in .sea surface properties. Comparison with the results of an epibenthic field survey, have shown gradients of epibenthic megafaunal distribution to correspond closely to the biogeographic regions identified by remote sensing. The application of this technique to marine monitoring programmes and ecosystem management is discussed.