Dissertationen zum Thema „Marine habitats mapping“

Virtually all marine conservation planning and management models in place or proposed have in common the need for improved scientific rigour in identifying and characterising the marine habitats encompassed. An emerging central theme in the last few years has been the concept of representativeness, or representative systems of Marine Protected Areas (MPAs). The habitat classification and mapping needed to incorporate considerations of representativeness into MPA planning must logically be carried out at the same scale at which management occurs. Management of highly protected areas occurs almost exclusively at local scales or finer, independent of the reservation model or philosophy employed. Moreton Bay, on Australia’s east coast, was selected for studies at the local scale to map and classify macrobenthic habitats. In a site scale (1 km) trial for the major habitat classification study, remote underwater videography was used to map and characterise an unusual assemblage of epibenthic invertebrates on soft sediments. The assemblage included congregations of the comatulid crinoid Zygometra cf. Z. microdiscus (Bell) at densities up to 0.88 individuals.m-2, comparable to those found in coral reef habitats. There was no correlation between the distribution of this species and commonly used abiotic surrogates depth (6 – 18 m), sediment composition and residual current. This site scale trial is the first quantitative assessment of crinoid density and distribution in shallow water soft-sediment environments. The high densities found are significant in terms of the generally accepted picture of shallow-water crinoids as essentially reefal fauna. The findings highlight the conservation benefits of an inclusive approach to marine habitat survey and mapping. Assemblages such as the one described, although they may be of scientific and ecological significance, would have been overlooked by common approaches to marine conservation planning which emphasise highly productive or aesthetically appealing habitats. Most habitat mapping studies rely solely or in part on abiotic surrogates for patterns of biodiversity. The utility of abiotic variables in predicting biological distributions at the local scale (10 km) was tested. Habitat classifications of the same set of 41 sites based on 6 abiotic variables and abundances of 89 taxa and bioturbation indicators were compared using correlation, regression and ordination analyses. The concepts of false homogeneity and false heterogeneity were defined to describe types of errors associated with using abiotic surrogates to construct habitat maps. The best prediction by abiotic surrogates explained less than 30% of the pattern of biological similarity. Errors of false homogeneity were between 20 and 62%, depending on the methods of estimation. Predictive capability of abiotic surrogates at the taxon level was poor, with only 6% of taxon / surrogate correlations significant. These results have implications for the widespread use of abiotic surrogates in marine habitat mapping to plan for, or assess, representation in Marine Protected Areas. Abiotic factors did not discriminate sufficiently between different soft bottom communities to be a reliable basis for mapping. Habitat mapping for the design of Marine Protected Areas is critically affected by the scale of the source information. The relationship between biological similarity of macrobenthos and the distance between sites was investigated at both site and local scales, and for separate biotic groups. There was a significant negative correlation between similarity and distance, in that sites further apart were less similar than sites close together. The relationship, although significant, was quite weak at the site scale. Rank correlograms showed that similarity was high at scales of 10 km or less, and declined markedly with increasing distance. There was evidence of patchiness in the distributions of some biotic groups, especially seagrass and anthozoans, at scales less than 16 km. In other biotic groups there was an essentially monotonic decline in similarity with distance. The spatial agglomeration approach to habitat mapping was valid in the study area. Site spacing of less than 10 km was necessary to capture important components of biological similarity. Site spacing of less than 2.5 km did not appear to be warranted. Macrobenthic habitat types were classified and mapped at 78 sites spaced 5 km apart. The area mapped was about 2,400 km2 and extended from estuarine shallow subtidal waters to offshore areas to the 50 m isobath. Nine habitat types were recognised, with only one on hard substrate. The habitat mapping characterised several habitat types not previously described in the area and located deepwater algal and soft coral reefs not previously reported. Seagrass beds were encountered in several locations where their occurrence was either unknown or had not previously been quantified. The representation of the derived habitat types within an existing marine protected area was assessed. Only two habitat types were represented in highly protected zones, with less than 3% of each included The study represents the most spatially comprehensive survey of epibenthos undertaken in Moreton Bay, with over 40,000 m2 surveyed. Derived habitat maps provide a robust basis for inclusion of representative examples of all habitat types in marine protected area planning in and adjacent to Moreton Bay. The utility of video data to conduct a low-cost habitat survey over a comparatively large area was also demonstrated. The method used has potentially wide application for the survey and design of marine protected areas.

Virtually all marine conservation planning and management models in place or proposed have in common the need for improved scientific rigour in identifying and characterising the marine habitats encompassed. An emerging central theme in the last few years has been the concept of representativeness, or representative systems of Marine Protected Areas (MPAs). The habitat classification and mapping needed to incorporate considerations of representativeness into MPA planning must logically be carried out at the same scale at which management occurs. Management of highly protected areas occurs almost exclusively at local scales or finer, independent of the reservation model or philosophy employed. Moreton Bay, on Australia’s east coast, was selected for studies at the local scale to map and classify macrobenthic habitats. In a site scale (1 km) trial for the major habitat classification study, remote underwater videography was used to map and characterise an unusual assemblage of epibenthic invertebrates on soft sediments. The assemblage included congregations of the comatulid crinoid Zygometra cf. Z. microdiscus (Bell) at densities up to 0.88 individuals.m-2, comparable to those found in coral reef habitats. There was no correlation between the distribution of this species and commonly used abiotic surrogates depth (6 – 18 m), sediment composition and residual current. This site scale trial is the first quantitative assessment of crinoid density and distribution in shallow water soft-sediment environments. The high densities found are significant in terms of the generally accepted picture of shallow-water crinoids as essentially reefal fauna. The findings highlight the conservation benefits of an inclusive approach to marine habitat survey and mapping. Assemblages such as the one described, although they may be of scientific and ecological significance, would have been overlooked by common approaches to marine conservation planning which emphasise highly productive or aesthetically appealing habitats. Most habitat mapping studies rely solely or in part on abiotic surrogates for patterns of biodiversity. The utility of abiotic variables in predicting biological distributions at the local scale (10 km) was tested. Habitat classifications of the same set of 41 sites based on 6 abiotic variables and abundances of 89 taxa and bioturbation indicators were compared using correlation, regression and ordination analyses. The concepts of false homogeneity and false heterogeneity were defined to describe types of errors associated with using abiotic surrogates to construct habitat maps. The best prediction by abiotic surrogates explained less than 30% of the pattern of biological similarity. Errors of false homogeneity were between 20 and 62%, depending on the methods of estimation. Predictive capability of abiotic surrogates at the taxon level was poor, with only 6% of taxon / surrogate correlations significant. These results have implications for the widespread use of abiotic surrogates in marine habitat mapping to plan for, or assess, representation in Marine Protected Areas. Abiotic factors did not discriminate sufficiently between different soft bottom communities to be a reliable basis for mapping. Habitat mapping for the design of Marine Protected Areas is critically affected by the scale of the source information. The relationship between biological similarity of macrobenthos and the distance between sites was investigated at both site and local scales, and for separate biotic groups. There was a significant negative correlation between similarity and distance, in that sites further apart were less similar than sites close together. The relationship, although significant, was quite weak at the site scale. Rank correlograms showed that similarity was high at scales of 10 km or less, and declined markedly with increasing distance. There was evidence of patchiness in the distributions of some biotic groups, especially seagrass and anthozoans, at scales less than 16 km. In other biotic groups there was an essentially monotonic decline in similarity with distance. The spatial agglomeration approach to habitat mapping was valid in the study area. Site spacing of less than 10 km was necessary to capture important components of biological similarity. Site spacing of less than 2.5 km did not appear to be warranted. Macrobenthic habitat types were classified and mapped at 78 sites spaced 5 km apart. The area mapped was about 2,400 km2 and extended from estuarine shallow subtidal waters to offshore areas to the 50 m isobath. Nine habitat types were recognised, with only one on hard substrate. The habitat mapping characterised several habitat types not previously described in the area and located deepwater algal and soft coral reefs not previously reported. Seagrass beds were encountered in several locations where their occurrence was either unknown or had not previously been quantified. The representation of the derived habitat types within an existing marine protected area was assessed. Only two habitat types were represented in highly protected zones, with less than 3% of each included The study represents the most spatially comprehensive survey of epibenthos undertaken in Moreton Bay, with over 40,000 m2 surveyed. Derived habitat maps provide a robust basis for inclusion of representative examples of all habitat types in marine protected area planning in and adjacent to Moreton Bay. The utility of video data to conduct a low-cost habitat survey over a comparatively large area was also demonstrated. The method used has potentially wide application for the survey and design of marine protected areas.
Thesis (PhD Doctorate)
Doctor of Philosophy (PhD)
School of Environmental and Applied Science
Full Text

Marine planners need to know about ecosystems, such as Priority Marine Habitats (PMHs) in order to manage and conserve them effectively. The overarching theme of this thesis is to contribute to this knowledge through the development of “marine planning tools”. The primary focus is on the PMH, Modiolus modiolus beds, although other PMHs and Marine Protected Areas (MPAs) were also considered. Four key studies were designed and conducted, i) Species Distribution Modelling (SDM) of M. modiolus in UK waters; ii) SDM of PMHs in Europe; iii) assessment of MPA management effort; and iv) the genetic connectivity of M. modiolus beds Overall, the research provided information and knowledge to contribute to implementation of a truly ecosystem-based approach to management and effective PMH management. It is now known: i) where Modiolus modiolus beds occur; ii) where they have the potential to occur, now and in the future; iii) that there is the potential for them to be lost/ hindered or lack-viability if ocean temperatures increase; iv) that they may become more important to conservation at northern latitudes in the future; v) that European nations will have to work towards integrated marine conservation policies and protection when considering all PMHs; vi) that some MPAs may require more effort to manage than others and that it may be possible to predict which ones they will be; vii) that cumulative human impacts may not be the driving force for management effort; and viii) that some M. modiolus beds in the UK are potentially connected. The data and discussion points generated within this thesis will enable effective PMH management through the selection of appropriate management strategies.

There is an increasing need for more detailed knowledge about the spatial distribution and structure of shallow water benthic habitats for marine conservation and planning. This, linked with improvements in hyperspectral image sensors provides an increased opportunity to develop new techniques to better utilise these data in marine mapping projects. The oligotrophic, optically-shallow waters surrounding Rottnest Island, Western Australia, provide a unique opportunity to develop and apply these new mapping techniques. The three flight lines of HyMap hyperspectral data flown for the Rottnest Island Reserve (RIR) in April 2004 were corrected for atmospheric effects, sunglint and the influence of the water column using the Modular Inversion and Processing System. A digital bathymetry model was created for the RIR using existing soundings data and used to create a range of topographic variables (e.g. slope) and other spatially relevant environmental variables (e.g. exposure to waves) that could be used to improve the ecological description of the benthic habitats identified in the hyperspectral imagery. A hierarchical habitat classification scheme was developed for Rottnest Island based on the dominant habitat components, such as Ecklonia radiata or Posidonia sinuosa. A library of 296 spectral signatures at HyMap spectral resolution (~15 nm) was created from >6000 in situ measurements of the dominant habitat components and subjected to spectral separation analysis at all levels of the habitat classification scheme. A separation analysis technique was developed using a multivariate statistical optimisation approach that utilised a genetic algorithm in concert with a range of spectral metrics to determine the optimum set of image bands to achieve maximum separation at each classification level using the entire spectral library. These results determined that many of the dominant habitat components could be separated spectrally as pure spectra, although there were almost always some overlapping samples from most classes at each split in the scheme. This led to the development of a classification algorithm that accounted for these overlaps. This algorithm was tested using mixture analysis, which attempted to identify 10 000 synthetically mixed signatures, with a known dominant component, on each run. The algorithm was applied directly to the water-corrected bottom reflectance data to classify the benthic habitats. At the broadest scale, bio-substrate regions were separated from bare substrates in the image with an overall accuracy of 95% and, at the finest scale, bare substrates, Posidonia, Amphibolis, Ecklonia radiata, Sargassum species, algal turf and coral were separated with an accuracy of 70%. The application of these habitat maps to a number of marine planning and management scenarios, such as marine conservation and the placement of boat moorings at dive sites was demonstrated.

[Truncated abstract] Spatially explicit marine habitat data is required for effective resource planning and management across large areas, although mapped boundaries typically lack rigour in explaining what factors influence habitat distributions. Accurate, quantitative methods are needed. In this thesis I aimed to assess the utility of ecological models to determine what factors limit the spatial extent of marine habitats. I assessed what types of modeling methods were able to produce the most accurate predictions and what influenced model results. To achieve this, initially a broad scale marine habitat survey was undertaken in the Recherche Archipelago, on the south coast of Western Australia using video and sidescan sonar. Broad and more detailed functional habitats types were mapped for 1054km2 of the Archipelago. Broad habitats included high and low profile reefs, sand, seagrass and extensive rhodolith beds, although considerable variation could be identified from video within these broad types. Different densities of seagrass were identified and reefs were dominated by macroalgae, filter feeder communities, or a combination of both. Geophysical characteristics (depth, substrate, relief) and dominant benthic biota were recorded and then modelled using decision trees and a combination of generalised additive models (GAMs) and generalised linear models (GLMs) to determine the factors influencing broad and functional habitat variation. Models were developed for the entire Archipelago (n=2769) and a subset of data in Esperance Bay (n=797), which included exposure to wave conditions (mean maximum wave height and mean maximum shear stress) calculated from oceanographic models. Additional distance variables from the mainland and islands were also derived and used as model inputs for both datasets. Model performance varied across habitats, with no one method better than the other in terms of overall model accuracy for each habitat type, although prevalent classes (>20%) such as high profile reefs with macroalgae and dense seagrass were the most reliable (Area Under the Curve >0.7). ... This highlighted not only issues of data prevalence, but also how ecological models can be used to test the reliability of classification schemes. Care should be taken when mapping predicted habitat occurrence with broad habitat models. It should not be assumed that all habitats within the type will be defined spatially, as this may result in the distribution of distinctive and unique habitats such as filterfeeders being underestimated or not identified at all. More data is needed to improve prediction of these habitats. Despite the limitations identified, the results provide direction for future field sampling to ensure appropriate variables are sampled and classification schemes are carefully designed to improve descriptions of habitat distributions. Reliable habitat models that make ecological sense will assist future assessments of biodiversity within habitats as well as provide improved data on the probability of habitat occurrence. This data and the methods developed will be a valuable resource for reserve selection models that prioritise sites for management and planning of marine protected areas.

Dans un contexte de changement climatique et d’érosion de la biodiversité marine, la surveillance écologique des habitats marins les plus sensibles est primordiale et nécessite des méthodes opérationnelles de suivi permettant aux décideurs et gestionnaires d’établir des mesures de conservation pertinentes et d’évaluer leur efficacité. TEMPO et RECOR sont deux réseaux de surveillance centrés sur les herbiers de posidonie et les récifs coralligènes, les deux habitats les plus riches et sensibles de Méditerranée. L’objectif de cette thèse est de répondre aux besoins de la surveillance des habitats marins par le développement de méthodes d’évaluation de leur état de santé, basées sur deux techniques d’analyses d’images clés : les réseaux de neurones convolutifs et la photogrammétrie. Les résultats montrent que les réseaux de neurones convolutifs sont capables de reconnaître les principales espèces des assemblages coralligènes sur des photos sous-marines issues de RECOR, avec une précision semblable à celle d’un expert taxonomiste. Par ailleurs, nous avons montré que la photogrammétrie permettait de reproduire en 3D un habitat marin avec une grande précision, suffisante pour un suivi de la structure de l’habitat et de la distribution d’espèces à fine échelle. À partir de ces reconstructions, nous avons mis au point une méthode de cartographie automatique des herbiers de posidonie, permettant de réaliser un suivi temporel de la qualité écologique de cet habitat sensible. Enfin, nous avons caractérisé la structure 3D des récifs coralligènes à partir de leurs reconstructions photogrammétriques et étudié les liens avec la structuration des assemblages qui les composent. Ce travail de thèse a permis de développer des méthodes opérationnelles, aujourd’hui intégrées aux réseaux de surveillance TEMPO et RECOR, et ouvre la voie à de futures recherches, notamment la caractérisation de l’activité biologique des récifs coralligènes grâce au couplage entre photogrammétrie, réseaux de neurones et acoustique sous-marine
In a context of climate change and the erosion of marine biodiversity, ecological monitoring of the most sensitive marine habitats is of paramount importance. In particular, there is a need for operational methods that enable decision-makers and managers to establish relevant conservation measures and to evaluate their effectiveness. TEMPO and RECOR are two monitoring networks focusing on Posidonia meadows and coralligenous reefs, the two richest and most sensitive habitats in the Mediterranean. The objective of this thesis is to meet the needs of effective monitoring of marine habitats by developing methods for assessing their health, based on two key image analysis methods: convolutional neural networks and photogrammetry. The results show that convolutional neural networks are capable of recognizing the main species of coralligenous assemblages in underwater photographs from RECOR, with a precision similar to that of an expert taxonomist. Furthermore, we have shown that photogrammetry can reproduce a marine habitat in three dimensions with a high degree of accuracy, sufficient for monitoring habitat structure and species distribution at a fine scale. Based on these reconstructions, we have developed a method for automatic mapping of Posidonia meadows, enabling temporal monitoring of the ecological quality of this sensitive habitat. Finally, we characterized the three-dimensional structure of coralligenous reefs based on their photogrammetric reconstructions and studied the links with the structuring of the assemblages that make them up. This PhD work has led to the development of operational methods that are now integrated into the TEMPO and RECOR monitoring networks. Results of this work paves the way for future research, in particular concerning characterization of the biological activity of coralligenous reefs thanks to the coupling of photogrammetry, neural networks and underwater acoustics

Le sondeur multifaisceaux (SMF) est l'une des technologies d'acoustique sous-marine les plus avancées pour l'étude des fonds et de la colonne d'eau. Il requiert une réelle expertise pour son déploiement sur le terrain ainsi que pour l'élaboration de cartographies à partir des différentes données acquises. Ces traitements sont souvent chronophages en raison de la quantité de données acquises et demandent à être automatisés pour alléger le travail à l'hydrographe. C'est ce sur quoi portent les travaux réalisés durant cette thèse. Après des rappels sur des notions d'acoustique sous-marine, le fonctionnement du SMF est décrit et les types de données manipulées tout au long des traitements sont présentés. Le manuscrit s'articule ensuite autour de deux thématiques ˸ la cartographie bathymétrique et la cartographie biocénotique. Les développements sont intégrés dans les logiciels de l'entreprise Seaviews pour laquelle les travaux sont réalisés. Ils répondent à des besoins particuliers de l'entreprise.En ce qui concerne la cartographie bathymétrique, la donnée bathymétrique doit être préalablement triée pour écarter les sondes aberrantes et éviter qu'elles ne pénalisent la précision topographique. Ce tri d'innombrables sondes est une tâche que réalisent les hydrographes, assistés aujourd'hui d'outils numériques. Nous proposerons une méthode statistique rapide pour trier les sondes tout en réalisant une carte de profondeurs marines. Ce qui amène à se demander si les images de la colonne d'eau acquises également par le sondeur ne seraient pas exploitables pour déduire une bathymétrie exempte d'aberration. Nous testerons cette hypothèse à l'aide de l'apprentissage profond (deep learning) et en particulier par des réseaux de neurones convolutifs qui ont permis des progrès considérables en vision par ordinateur. La cartographie des habitats marins (les biocénoses) est un travail de classification de la nature des fonds à partir des données acoustiques du SMF en concordance avec les espèces vivant sur les lieux. La société Seaviews a développé une méthode de préparation des données SMF pour l'analyse des habitats. Nous nous orientons vers des méthodes de classification des habitats, à partir de ces données, par des techniques d'apprentissage automatique (machine learning). Plusieurs méthodes sont mises en place et testées, puis une zone d'étude est choisie pour évaluer et comparer les résultats des différentes approches
Among underwater acoustic technologies, multibeam echo sounder (MBES) is one of the most advanced tool to study and map the underwater floors and the above water column. Its deployment on-site requires expertise so as the whole data processing to map the information. These processing are very time-consuming due to the massive quantity of recorded data and thus needs to be automatised to shorten and alleviate the hydrographer's task. This PhD research works focus on the automatisation of the current activities in Seaviews society.After some reminders on the underwater acoustic sciences, the MBES operating is described as well the produced data that will be manipulated throughout the developments. This document presents two thematics˸ bathymetric (depths) and marine habitats mapping. The developments are integrated into the Seaviews' software in the aim to be used by all the employees.About seafloor depths mapping, the bathymetric sounding has to be sorted to avoid that the outlier errors distort the results. Sorting the uncountable measures is cumbersome but necessary, although the hydrographers are today happily computed-assisted. We propose a fast statistical method to exclude the outliers while mapping the information. This leads to wonder if the water column imagery would be workable to deduce the bathymetry without failure. We will test this hypothesis with some technics of deep learning, especially with convolutional neural networks.The marine habitats mapping is a seabed nature classification according to the local life. Seaviews has worked on a way to prepare MBES data and habitats analysis. Concerning the method of classification itself, we move towards machine learning technics. Several methods are implemented and assessed, and then an area is chosen to evaluate and compare the results

Bathymetrical mapping performed using multibeam sonar systems is widely used in marine science and for habitat mapping. The incoherent part of the multibeam data, the backscatter data, is less commonly used. Automatic classification of processed backscatter has a correlates well with three sediment classes, defined as fine-(clay-silt), medium- (sand) and coarse- (gravel–till) grained substrates. This relation is used directly as a theme in a modified habitat classification scheme, while a more detailed substrate classification is incorporated as another theme. This theme requires a manual interpretation and comprehensive knowledge of the substrate. This can partly be obtained by a newly developed technique using the backscatter strength plotted against the grazing angle. These plots make it possible to determine the critical angle and thereby calculate the compressional acoustic speed in seabed sediments. Marching a theoretical modeled backscatter curve to the measured backscatter strength at lower grazing angles provides estimates of four additional geoacoustic parameters.

This work, framed within the Marine Strategy Framework Directive, is focused on Sardinian southern continental margins marine habitat mapping. Aim of this thesis is to produce predictive marine habitat mapping, starting from a detailed geomorphologic study integrated with biological and oceanographic data coming both from original direct surveys than from bibliographic data. Underwater remotely operated vehicles (ROV), coupled with multibeam echo-sounder (MBES), enabled to perform interpretative hypothesis validation, controlled sampling and detailed observation of specific mesophotic habitats with noninvasive protocols, which are particularly relevant for habitats of conservation interest. Marine habitat mapping represents the best estimation of the distribution of habitats in a place and at a particular time, this goal have been focused on target biocoenosis A4.26 – “Mediterranean coralligenous communities moderately exposed to hydrodynamic action” and A4.713 – “Caves and overhangs with Corallium rubrum”, as defined by EUNIS classification (European Nature Information System), subsequently reunite under the name of Coralligenous Biocoenosis for the mapping of which has been reached a high level of confidence. Principal drivers for seabed habitat distributions include the type of seabed substrate, depth, light availability and the energy of water movements. For the very first time on Sardinian southern margin a multidisciplinary approach has been used to study the relationship between biotic and abiotic components of marine habitats and how seabed morphologic features influences different benthic biocoenosis development styles, geomorphologic characteristics of settlement for substrates could be important factors structuring benthic biodiversity, by affecting sediment accumulation rates, bottom currents and, ultimately, the rates of food supply. These biocoenosis hosts relevant commercially relevant species, as the blue and the giant red shrimp which represent the most important demersal resources for trawling fleet of Sardinia.

Global biodiversity is in decline, with the marine environment experiencing significant and increasing anthropogenic pressures. As a response, very large (105–106 km2) marine protected areas (MPAs) have become the dominant form of environmental protection in the marine environment. At present, however, paucity in scientific sampling makes prioritising which regions of the ocean to protect, especially over such large spatial scales, particularly problematic. One such very large MPA, covering an area of over 1 million Km2, is located at the sub-Antarctic South Georgia and South Sandwich Islands (SGSSI). This study uses the SGSSI MPA as a model system to assess the application of benthic habitat mapping as an evidence-based framework for the spatial prioritisation of marine conservation. This study presents an interdisciplinary methodology to marine landscape mapping, as a top-down, objective statistical approach to hierarchically partition and map the benthic environment into physical habitats types. Ordination analysis demonstrates a statistically significant relationship between environmentally-derived landscape mapping clusters and the composition of benthic species data from the region, thus attributing ecological relevance to the marine landscape map. Furthermore, this study adopts a bottom-up approach to habitat mapping, using an ensemble of habitat suitability models. Potential distributions are modelled for a range of benthic faunal attributes relevant to marine management, based on taxonomic classification, functional traits and vulnerability to disturbance. These modelled distributions are used to describe, for the first time, the bio-physical characteristics of SGSSI’s benthic environment. Synthesising both top-down and bottom-up approaches to habitat mapping, this study assesses the physical landscape clusters and modelled distribution results in relation to the spatial protection currently enforced at SGSSI. This synthesis addresses, (i) whether marine spatial planning in the region is representative in terms of the habitats and fauna it protects; and (ii) whether this interdisciplinary methodology at SGSSI can inform on MPA design and designation more universally, in what is an increasingly exploited, yet still poorly understood marine environment.

With an increase in demand on deep-sea resources comes a need for appropriate and effective management of this ecosystem. The establishment of a representative network of deep-sea Marine Protected Areas offers one tool with which to address the conservation needs of the deep sea. While a number of deep-sea habitats have been identified as vulnerable to anthropogenic activities (e.g. cold-water coral reefs and sponge aggregations), poor knowledge of the distribution of these habitats hinders conservation efforts and network planning, and thus we need habitat maps. With improvements in acoustic data resolution acquired from the deep sea, and the ability to cover large areas rapidly, the use of acoustic techniques in mapping biological habitats is growing. Multibeam bathymetry and its derived terrain variables can potentially provide important information that can aid in the delineation and characterisation of biological communities. A necessary prelude to mapping is therefore the definition of biological assemblages for use as mapping units. Two megahabitat features (seamount and submarine canyons) were sampled using acoustic and ground-truthing to characterise and map the distribution of benthic assemblages. Species were identified as distinct morpho-types and catalogued, and still images quantitatively analysed. Standard multivariate community analysis was undertaken to define distinct faunal assemblage that may act as mapping units. Those clusters identified by the SIMPROF routine were taken against a set of criteria to reject/accept as robust assemblages that may be used as mapping units. Twenty two benthic assemblages or biotopes were defined from multivariate analysis of quantitative species data, 11 from the SW Approaches and 11 from Anton Dohrn Seamount, and a further one from video observations (SW Approaches). Taken against current definitions, 11 of these were considered as Vulnerable Marine Ecosystems (VME). Diversity was measured to compliment the comprehensive description of biotopes. The use of multivariate diversity indices proved better for comparing diversity of biotopes as it captures a more than one aspect of diversity of the community. Two biotopes were common to both megahabitat features, cold-water coral reef habitats, and those from Anton Dohrn Seamount were more diverse than from the SW Approaches. Modelling techniques were employed to test the relationship between biotopes and environmental and geophysical parameters, which may be used as surrogates to map VME. Generalised Additive Models of Vulnerable Marine Ecosystems revealed multibeam bathymetry and its derived parameters to be significant surrogate for mapping the distribution of some assemblages, particularly those that appear to be influenced by current regime; whilst not so well for those whose distribution is not so strongly current driven e.g. soft sediment communities. In terms of deep-sea mapping, the use of multibeam can prove a useful mapping tool if the resolution of the data is at an appropriate scale that will identify meso-scale geomorphological features, such as cliff-top mounds, that may act as proxies for occurrence of biotopes, but this relationship is still unclear. Surrogates were used to map VME across the seamount and submarine canyons, and full coverage maps were produced for all biotopes occurring on these megahabitat features.

Autonomous vehicles are often deployed to perform exploration and monitoring missions in unseen environments. In such applications, there is often a compromise between the information richness and the acquisition cost of different sensor modalities. Visual data is usually very information-rich, but requires in-situ acquisition with the robot. In contrast, remotely sensed data has a larger range and footprint, and may be available prior to a mission. In order to effectively and efficiently explore and monitor the environment, it is critical to make use of all of the sensory information available to the robot. One important application is the use of an Autonomous Underwater Vehicle (AUV) to survey the ocean floor. AUVs can take high resolution in-situ photographs of the sea floor, which can be used to classify different regions into various habitat classes that summarise the observed physical and biological properties. This is known as benthic habitat mapping. However, since AUVs can only image a tiny fraction of the ocean floor, habitat mapping is usually performed with remotely sensed bathymetry (ocean depth) data, obtained from shipborne multibeam sonar. With the recent surge in unsupervised feature learning and deep learning techniques, a number of previous techniques have investigated the concept of multimodal learning: capturing the relationship between different sensor modalities in order to perform classification and other inference tasks. This thesis proposes related techniques for visual and remotely sensed data, applied to the task of autonomous exploration and monitoring with an AUV. Doing so enables more accurate classification of the benthic environment, and also assists autonomous survey planning. The first contribution of this thesis is to apply unsupervised feature learning techniques to marine data. The proposed techniques are used to extract features from image and bathymetric data separately, and the performance is compared to that with more traditionally used features for each sensor modality. The second contribution is the development of a multimodal learning architecture that captures the relationship between the two modalities. The model is robust to missing modalities, which means it can extract better features for large-scale benthic habitat mapping, where only bathymetry is available. The model is used to perform classification with various combinations of modalities, demonstrating that multimodal learning provides a large performance improvement over the baseline case. The third contribution is an extension of the standard learning architecture using a gated feature learning model, which enables the model to better capture the ‘one-to-many’ relationship between visual and bathymetric data. This opens up further inference capabilities, with the ability to predict visual features from bathymetric data, which allows image-based queries. Such queries are useful for AUV survey planning, especially when supervised labels are unavailable. The final contribution is the novel derivation of a number of information-theoretic measures to aid survey planning. The proposed measures predict the utility of unobserved areas, in terms of the amount of expected additional visual information. As such, they are able to produce utility maps over a large region that can be used by the AUV to determine the most informative locations from a set of candidate missions. The models proposed in this thesis are validated through extensive experiments on real marine data. Furthermore, the introduced techniques have applications in various other areas within robotics. As such, this thesis concludes with a discussion on the broader implications of these contributions, and the future research directions that arise as a result of this work.

A growing number of scientists are investigating applications of landscape ecology principles to marine studies, yet few coral reef scientists have examined spatial patterns across entire reefscapes with a holistic ecosystem-based view. This study was an effort to better understand reefscape ecology by quantitatively assessing spatial structures and habitat arrangements using remote sensing and geographic information systems (GIS). Quantifying recurring patterns in reef systems has implications for improving the efficiency of mapping efforts and lowering costs associated with collecting field data and acquiring satellite imagery. If a representative example of a reef is mapped with high accuracy, the data derived from habitat configurations could be extrapolated over a larger region to aid management decisions and focus conservation efforts. The aim of this project was to measure repeating spatial patterns at multiple scales (10s m2 to 10s km2) and to explain the environmental mechanisms which have formed the observed patterns. Because power laws have been recognized in size-frequency distributions of reef habitat patches, this study further investigated whether the property exists for expansive reefs with diverse geologic histories. Intra- and inter-reef patch relationships were studied at three sites: Andavadoaka (Madagascar), Vieques (Puerto Rico), and Saipan (Commonwealth of the Northern Mariana Islands). In situ ecological information, including benthic species composition and abundance, as well as substrate type, was collected with georeferenced video transects. LiDAR (Light Detection and Ranging) surveys were assembled into digital elevation models (DEMs), while vessel-based acoustic surveys were utilized to empirically tune bathymetry models where LiDAR data were unavailable. A GIS for each site was compiled by overlying groundtruth data, classifications, DEMs, and satellite images. Benthic cover classes were then digitized and analyzed based on a suite of metrics (e.g. patch complexity, principle axes ratio, and neighborhood transitions). Results from metric analyses were extremely comparable between sites suggesting that spatial prediction of habitat arrangements is very plausible. Further implications discussed include developing an automated habitat mapping technique and improving conservation planning and delimitation of marine protected areas.

L’utilizzo del Multibeam Echo sounder (MBES) in ambienti di transizione poco profondi, con condizioni ambientali complesse come la laguna di Venezia, è ancora in fase di studio e i dati biologici e sedimentologici inerenti ai canali della laguna di Venezia sono attualmente scarsi e datati in letteratura. Questo studio ha lo scopo di mappare gli habitat e gli oggetti antropici di un canale della laguna di Venezia in un intervallo di profondità tra 0.3 e 20 m (Canale San Felice) analizzando i dati batimetrici e di riflettività (backscatter) acquisiti da ISMAR-Venezia nell’ambito del progetto RITMARE. A tale scopo il fondale del canale San Felice (Venezia) è stato caratterizzato dal punto di vista geomorfologico, sedimentologico e biologico; descrivendo anche l’eventuale presenza di oggetti antropici. L’ecoscandaglio utilizzato è il Kongsberg EM2040 Dual-Compact Multibeam in grado di emettere 800 beam (400 per trasduttore) ad una frequenza massima di 400kHZ e ci ha consentito di ricavare ottimi risultati, nonostante le particolari caratteristiche degli ambienti lagunari. I dati acquisiti sono stati processati tramite il software CARIS Hydrographic information processing system (Hips) & Sips, attraverso cui è possibile applicare le correzioni di marea e velocità del suono e migliorare la qualità dei dati grezzi ricavati da MBES. I dati sono stati quindi convertiti in ESRI Grid, formato compatibile con il software ArcGIS 10.2.1 (2013) che abbiamo impiegato per le interpretazioni e per la produzione delle mappe. Tecniche di ground-truthing, basate su riprese video e prelievi di sedimento (benna Van Veen 7l), sono state utilizzate per validare il backscatter, dimostrandosi molto efficaci e soddisfacenti per poter descrivere i fondali dal punto di vista biologico e del substrato e quindi degli habitat del canale lagunare. Tutte le informazioni raccolte durante questo studio sono state organizzate all’interno di un geodatabase, realizzato per i dati relativi alla laguna di Venezia.

Il benthic habitat mapping è un processo che integra diverse discipline e dataset. Le mappe sono utili per la gestione e la tutela degli ecosistemi; tra questi, le lagune costiere sono obiettivo di protezione da parte di numerose direttive europee, come la Dir. Habitat e la Water Framework Directive. Le lagune sono sistemi altamente produttivi e forniscono habitat unici ed indispensabili per alcune specie animali e vegetali; inoltre, forniscono numerosi servizi ecosistemici. Nel Mediterraneo la più importante ed estesa è la Laguna di Venezia. Gli obiettivi di questa tesi sono mappare e classificare le forme fisiografiche e morfologiche, i tipi di substrato, la componente biologica e gli oggetti antropici dei canali circondanti le isole di Burano e Torcello, con particolare attenzione alla componente macroalgale, ottenendo alla fine una mappa degli habitat bentonici. Tramite MultiBeam EchoSounder ad alta risoluzione sono stati acquisiti dati batimetrici e di backscatter, analizzati con ArcGIS ed integrati con campionamenti mediante dropframe. Dall’analisi dei dati è risultato che nell’area sono presenti, in ordine decrescente di estensione, le forme fisiografiche pool, bar e point bar e le forme morfologiche 3d-dunes, 2d-dunes, furrow, scour hole, obstacle mark e scour. Le forme elencate sono coerenti con il tipo di substrato, che è risultato prevalentemente sabbioso-fangoso, con presenza di patches di tipo hard. Il substrato risulta spesso bioturbato da molluschi, crostacei e, soprattutto, policheti. Sono anche presenti associazioni di spugne e macroalghe, le quali sono state classificate secondo il gruppo morfo-funzionale, che tiene conto della struttura e non della specie. Questo studio ha dimostrato che i MBES ad alta risoluzione possono essere utilizzati in ambienti poco profondi per ottenere una mappa continua degli habitat bentonici, in tempi relativamente brevi e con grande dettaglio. Inoltre, ha fornito una base per un futuro studio comparativo.

La presente tesi di dottorato prende in considerazione tecniche di rilevamento acustiche (Remote sensing, RS) e ottiche per lo sviluppo di modelli di distribuzione e mappatura (Distribution and mapping modelling, DMS) come strumenti di gestione per valutare lo stato degli habitat bentonici e per supportare decisioni relative alla pianificazione dello spazio marittimo (Marine Spatial Planning). Modelli di distribuzione di habitat e specie sono stati analizzati congiuntamente ad analisi spaziali e analisi di immagine da video subacquei, e messi in relazione alle caratteristiche geomorfologiche del substrato, raccolte tramite RS e validate tramite immersioni di controllo. L’impiego del Sistema Informativo Geografico (GIS) ha permesso di disegnare spazialmente la distribuzione e l’estensione degli habitat così come la distribuzione e l’intensità delle pressioni antropiche, creando quindi un dataset utile a sostenere adeguate scelte gestionali. L’accuratezza dei modelli è stata testata e confrontata. I risultati hanno permesso di definire un quadro metodologico che potrebbe essere facilmente recepito a fini gestionali. Sono infatti state sviluppate delle linee guida da distribuire ai gestori di Aree Marine Protette e ai manager interessati alla gestione della fascia costiera, utili ad inviduare le aree prioritarie in termini di conservazione, sulla base di modelli predittivi che potrebbero suggerire anche eventuali azioni di recupero ambientale. I risultati ottenuti dal presente lavoro mettono quindi in relazione complessità geomorfologica e habitats, permettendo di sviluppare piani di gestione che prendono in considerazione la distribuzione e l’intensità degli impatti antropici.
The present doctoral degree thesis is based on the implementation of remote sensing (RS) methods habitat mapping and distribution modelling (DMs) techniques as management tools to assess the status of benthic habitats and to support Marine Spatial Planning (MSP) decisions. A Geographic Information System (GIS) was used to model the spatial boundaries of the physical and biological datasets, as well to assess the proximity of anthropogenic pressures. Through spatial examination, image analysis and underwater video, the biological patterns of habitats/species were related to the variation in geomorphology based on geophysical substrate properties gathered through RS techniques in combination with optical data, collected during the ground truthing sampling. In addition, DMs and classification approaches were applied and their accuracy tested. Finally, a methodological framework was suggested as guideline to inform and provide recommendations to managers and policymakers about how to accurately locate and best protect benthic habitats and its resources, how to evidence possible different sensitivities between habitats in relation to geomorphology, create or redefine different zones or levels of protections at Marine Protected Areas and how to forecast future changes due to global warming and/or anthropogenic activities. General results demonstrate that the produced maps provide information about where the habitats/species could be present and how they are related to the geomorphological context and/or the anthropogenic pressures. Results emphasize the role of critical expert evaluation of spatial predictions before they are used to guide policy. We conclude that RS and DMs could be very useful tools for understanding the distribution of species–habitat associations and to help resources managers make informed and ecologically relevant decisions.

Il mare Adriatico è una delle regioni del bacino Mediterraneo più impattate poiché soggetto a multipli fattori di stress, come cambiamenti climatici e una lunga storia di intenso sfruttamento delle risorse. Allo stesso tempo ospita una grande varietà di endemismi, aree di riproduzione, nursery e foraggiamento. Il drammatico declino delle risorse target e non del mare Adriatico richiede un urgente sviluppo di nuove ed idonee misure di gestione e conservazione degli ecosistemi marini. Il cambiamento delle comunità macrobentoniche di fondo mobile può innescare un’alterazione delle reti trofiche, della qualità delle acque, del riciclo dei nutrienti. Pertanto, un’efficiente gestione degli ecosistemi marini non può prescindere dal recupero degli habitat bentonici. I risultati di questo lavoro di dottorato hanno permesso di evidenziare i principali cambiamenti avvenuti sulle comunità bentoniche dei fondali del largo del mare Adriatico (centro e nord) nel corso di circa 60 anni (1934 – 1998). Si è osservato un declino degli organismi dell’epifauna e delle specie macrobentoniche più fragili con Spugne ed Echinodermi che si sono ridotti fino ad un 90-70%. Tuttavia, lo sviluppo di modelli di preferenza degli habitat dei pennatulacei ha confermato che i fondi mobili del largo Adriatico sono habitat ideali per le specie Funiculina quadrangularis e Pennatula phosphorea. Virgularia mirabilis, invece, predilige le zone sabbiose-fangose del nord e della costa occidentale. Una dettagliata descrizione morfologica di P. phosphorea e Pteroeides spinosum è stata condotta per cercare di fornire uno strumento che possa agevolare il riconoscimento di specie ancora oggi spesso confuse a causa di descrizioni finora poco dettagliate. La tesi è stata disegnata al fine di fornire utili elementi scientifici a supporto del progetto “Adriatic Marine Ecosystem Recovery” (AMER), avente lo scopo di avviare processi utili al recupero degli ormai sovrasfruttati e degradati habitat marini del mare Adriatico.
The Adriatic Sea is one of the most depleted regions of the Mediterranean Sea due to a cumulative impact of multiple stressors, such as climate change and a long history of intense exploitation. At the same time it hosts a variety of endemism, of vulnerable marine ecosystems and essential fish habitats. Severe decline of target and by-catch species call for urgent conservation measures. The ecological importance of soft bottoms macrobenthic community have been increase and their changes can lead to an alteration of food-webs, reduction of water quality, or nutrient cycling. An efficient ecosystem-based management focusing on the recovery of key benthic habitats is needed to promote recovery of stocks and to develop a sustainable exploitation of marine resources. The results of this PhD work gives a picture of the main changes occurred on macrobenthic soft bottoms communities of the north and central Adriatic Sea during 60 years (1934 – 1998). Overall change are the decline of epifauna organisms and of the most fragile macrobenthos species with sponges and echinoderms declined by 90-70% since the 1900s. However, the study of the spatial and bathymetric distribution of Adriatic sea-pens and the development of suitability habitats models have confirmed that the soft bottoms of the central offshore Adriatic Sea are favourable bottoms for Funiculina quadrangularis and Pennatula phosphorea. Virgularia mirabilis prefers the northern basin and the sandy-muddy occidental coastlines. A detailed morphological description of P. phosphorea and Pteroeides spinosum was also provided as a useful tool to sea-pens classification, usually misidentify because a lack of clear available taxonomical descriptions. The thesis has been designed in order to provide useful scientific elements to support the ‘Adriatic Marine Ecosystem Recover’ (AMER) project, which main aim is provide processes to promote recovery of depleted and overexploited fish stocks and habitats of the Adriatic Sea.

Le lagune costiere sono considerate tra i più produttivi ecosistemi al mondo, un nodo per lo scambio di merci e lo sviluppo industriale e sono zone rilevanti a livello turistico. Costituiscono inoltre un importante sito di elevata biodiversità. A causa dell'influenza dell'uomo sono andate incontro a un declino in tutto il mondo ed essendo ambienti delicati e in continuo cambiamento necessitano di un costante monitoraggio. A riguardo è fondamentale la mappatura degli habitat bentonici (Benthic Habitat Mapping) che permette una descrizione qualitativa e quantitativa degli habitat sommersi, sia sotto un profilo biotico che abiotico, considerando diverse scale spaziali e differenti intervalli temporali. Le innovative tecniche di remote sensing, come il multi-beam ecosounder (MBES), hanno aperto nuove frontiere nel campo dell'habitat mapping e solo di recente sono state applicate ad ambienti poco profondi come i canali della laguna di Venezia fino ad oggi poco esplorati. In questa tesi l'analisi, l'interpretazione e l'elaborazione dei dati di batimetria e backscatter acustico, congiuntamente alla realizzazione di video del fondali e all'analisi dei campioni fotografici estratti da essi, hanno permesso di mappare i canali che si estendono dalla bocca di porto di Lido fino all’isola di Murano (area a nord-est di Venezia), ponendo attenzione alla componente macroalgale. I risultati ottenuti confermano la natura prevalentemente sabbiosa-fangosa del substrato, in cui sono più o meno visibili tracce superficiali di infauna, in relazione alla quantità di detrito conchigliare presente, ed organismi bentonici animali e/o vegetali. Sono stati rinvenuti infatti organismi quali spugne, anemoni e ofiure e raggruppamenti più o meno isolati di macroalghe corticate o filamentose. Tra gli oggetti antropici mappati i più abbondanti sono le bricole, che delimitano i canali navigabili dell'area.

Sandy marine coastal habitats provide ecosystem services such as food supply, food web structure, and habitat provision. However, these habitats are subject to various anthropogenic disturbances, including coastal urban development and the construction of coastal defence structures. These habitats are very dynamic and difficult to investigate and monitor with conventional techniques. In this context, the TAO project, Technologies for Coastal Monitoring (Formally: Tecnologie per il monitorAggio cOstiero), focus on the implementation of a technological platform, helpful in developing monitoring plans with low environmental impact and low cost, for the investigation of littoral and supra / sub-littoral areas. This thesis, which is part of the TAO project, aims to provide useful data for evaluating the validity of these new monitoring technologies by mapping both the data obtained from the surveys carried out using acoustic instruments (Single-Beam Echosounder and Side-Scan Sonar installed aboard the autonomous OpenSWAP vehicle), and the data on the characteristics of sediments and benthic assemblages collected through direct sampling in scuba diving at Lido di Dante (northern Adriatic Sea) study site. A comparative integration of the physical-biological data and the geophysical data, recorded continuously by the acoustic instruments, was carried out. The data collected were interpolated to the entire study area and represented through cartographic software. The different variables were analyzed and correlated with each other with innovative spatial correlation techniques through linear models. This allowed highlighting possible anomalies in the characteristics of the sediments and in the distribution of the benthic assemblages that can be related both to natural sedimentological and ecological processes and to alterations induced by the presence of coastal defence structures. These integrated approaches provide innovative tools for integrated coastal zone management.

Anthropogenic changes to the landscape, storm events and sea level rise are contributing to the erosion of beaches leading to an increase of the sediment load in near shore marine environments. Palm Beach, Florida is host to unique near shore hardbottom habitats. These areas are distinct from the vast expanses of surrounding sediments and play and important role of habitat and shelter for many different species. In this study, remotely sensed images from 2000-2015 were used to look at the movement of sediment and how it contributes to exposure rates of near shore hardbottom habitats in Palm Beach, Florida and how these factors affect the benthic community. GIS was used to determine areas of hardbottom with high exposure (exposed in >60% of aerial images), medium exposure (40-60%), and low exposure ( I strived to determine if one can detect a successional relationship of benthic communities in a dynamic environment with annual mapping. I also examined if areas with higher exposure rates have more complex successive communities than those with lower exposure rates, and what implications this has on near shore benthic communities. In situ surveys conducted at 117 sites determined the community structure (corals, octocorals, macroalgae, and hydroids). This study confirmed that periodic mapping was successful in identifying hardbottom burial and exposure, which fluctuate both spatially and temporally. This periodic mapping along with manual delineation did identify hardbottom burials and exposures that fluctuate between years and relate to benthic community differences. The near shore hardbottom coral reef communities aligned with the observed exposure categories with the greater coral species richness and octocoral morphologies found at sites classified as highly exposed. Statistical analyses showed differences in communities shallower and deeper than three meters’ depth. Increasing the frequency of imagery captures and in situ observation would further increase our comprehension of the metrics of hardbottom exposures in reference to community structure.

Remote sensing has transformed coral reef science in the last decades. It is now possible to audit reef health and geomorphology at global scale and very high spatial resolution. This work utilizes the vast and no-cost archive of imagery housed within Google Earth (GE). GE was used to interrogate the morphometric properties (size, shape, complexity, etc) of the components that combine to yield a reef environment (reef structures, sediment sheets, and so forth). The data populated a morphometric database for reefs globally that are partitioned by their architecture. The database has been investigated using information theory. The aim of the work is to identify whether motifs of carbonate platforms within these reef types are predictable on the basis of environmental parameters. The results are relevant to assess depositional patterns to develop rules for predicting how facies are distributed in modern systems and ancient reservoirs.

Several studies have found separation of biogeographic provinces on the West Florida Shelf (WFS), but the location of this separation differs depending on different organisms with faunal boundaries proposed at Apalachicola, Cedar Key, Anclote Key. Tampa Bay, Charlotte Harbor, Cape Romano, or Cape Sable. Biogeographic boundaries can be gradual over a given space and are often species-specific. Analyses of marine benthic mapping and community characterization of Florida’s West-central coast shallow water (depth) hardbottom habitats indicate a major shift in the benthos across Tampa Bay. Quantitative benthic surveys of 29 sites yielded a total of 4,079 individuals of nine stony coral species and 1,918 soft coral colonies. Populations were dominated by four species of corals: Siderastrea radians, Oculina robusta, Solenastrea hyades, and Cladacora arbuscula. Most corals were less than 10 cm in diameter. Cluster analyses of coral density and major functional group percent cover showed distinct differences in hard and soft coral densities and species demographics from south to north with clear spatial patterns between regions. These benthic hardbottom coral communities change over a relatively small spatial scale (10’s of km), indicating a biogeographical province or ecosystem region boundary in marine benthic communities at, or very near, the mouth of Tampa Bay. Broader studies are needed to identify the shifts in benthic community biogeography along the West Florida Shelf.

L’imagerie hyperspectrale aéroportée constitue un candidat potentiel pour la cartographie et le suivi des récifs coralliens à large échelle et à forte résolution spatiale. Dans cette thèse, nous présentons tout d’abord les traitements à appliquer au signal hyperspectral afin d’en extraire des informations sur la nature du fond marin, la bathymétrie et les propriétés optiques de l’eau, et nous les mettons en perspectives devant deux principaux verrous : (i) le faible rapport signal sur bruit du signal mesuré, (ii) le nombre et la variabilité des interactions onde-matière ayant lieu entre l’entrée des rayons lumineux dans l’atmosphère et leur mesure par le capteur hyperspectral. En considérant ces limitations, nous examinons les performances d’une des méthodes existantes de traitement de la colonne d’eau : l’inversion de modèle semi-analytique par optimisation. Nous évaluons d’abord la robustesse des estimations des types de fond et de la bathymétrie pour six paramétrisations du schéma d’inversion. Les résultats sur des images hyperspectrales acquises à l’île de La Réunion en 2009 montrent que la paramétrisation de l’inversion joue un rôle crucial sur la qualité des estimations et que la paramétrisation la plus communément utilisée ne présente pas toujours les meilleures performances. Nous évaluons par la suite l’importance de la précision de la paramétrisation du modèle semi-analytique direct à travers une analyse de sensibilité réalisée sur des données hyperspectrales synthétiques puis sur des données réelles que nous avons acquises en 2015 à La Réunion. L’analyse est en outre effectuée pour toutes les paramétrisations de l’inversion précédemment étudiées. Cette étude montre qu’en milieu récifal l’importance de la précision de la paramétrisation du modèle direct est faible devant le choix de la paramétrisation de l’inversion. Nous montrons en outre qu’il n’est pas possible d’identifier le(s) paramètre(s) du modèle direct le(s) plus influent(s) car cela dépend des concentrations relatives de chacun des éléments optiquement actifs de l’eau
Airborne hyperspectral imaging is a potential candidate for mapping and monitoring coral reefs at large scale and with high spatial resolution. In this thesis, we first present the processing steps to be applied to hyperspectral signals for extracting information about seabed types, bathymetry and water optical properties, and we discuss their efficiency with respect to two main confounding factors: (i) low signal to noise ratio of the measured signals, and (ii) large number and variability of physical interactions occurring between the entrance of sunlight into the atmosphere and its measurement by the hyperspectral sensor. Considering these limitations, we examine the performance of an already existing water column processing method: semi-analytical model inversion by optimization. We first evaluate the robustness of seabed type and bathymetry estimation for six different inversion setups. The results on hyperspectral images acquired over Réunion Island reefs in 2009 show that the choice of the inversion setup plays an important role on the quality of the estimations and that the most widely used inversion setup does not always produce the best results. We then evaluate the importance of the accuracy of the parameterization of the direct semi-analytical model. This is done through a sensitivity analysis performed on both simulated and real hyperspectral data acquired in Réunion Island in 2015. The analysis is performed for each inversion setup previously studied. This study shows that in coral reef context the accuracy of the parameterization of the direct model is less important than the choice of the inversion setup. We also demonstrate that it is not possible to identify the most influent parameters of the direct model because it depends on the relative concentration of each optically active constituent

Seagrasses are core components of the nearshore environment and there is sustained interest in developing mapping and monitoring techniques of their extent and condition for management purposes. An identified gap in mapping methods is the capacity to monitor at landscape scales, that is, areas that are larger than the 1 m2 quadrat and smaller than those covered by broad area mapping (approximately 5km2 or greater). Monitoring at the landscape scale is required to investigate the dynamic patterning and patchiness present in seagrass beds, as well as providing inputs and validation for predictive modelling. However, the acquisition and use of remote sensing images for these purposes provides many challenges to the practitioner. The primary aim of this thesis is to develop effective optical remote sensing techniques for mapping and monitoring seagrass habitats in shallow temperate marine waters, over depth ranges of approximately 0-10 m and spatial scales of hundreds of square metres. Image capture is often compromised because of environmental conditions, such as sun glitter, water clarity, cloudiness and wind. Small format digital aerial photography was selected as the remote sensing platform for its flexibility and responsiveness regarding deployment when environmental conditions are favourable and its low cost, rapid access to imagery. To address the problem of sun glitter, a simplified algorithm was developed that allows the precise prediction of the extent of sun glitter on vertical, downward-looking imagery with the readily available inputs of sun elevation angle, wind speed and sensor field of view (FOV). Subsurface illumination was also investigated via the modelling of reflection and refraction at the water surface. These improvements and investigations enable more efficient and accurate image capture. Problems are also typically encountered during image interpretation, in part due to the characteristics of the seagrass habitats, including the common occurrence of uncertain boundaries and the high variability of vegetation density. Limitations on the detectability of the maximum depth limit (MDL) of seagrass were examined, with the discovery that if imagery is captured when water clarity is higher than the annual average, the limiting factor is the contrast between the seagrass and the surrounding substrate or submerged aquatic vegetation (SAV). A simple and inexpensive measurement of water clarity, Secchi depth (Zsd), was found to be suitable when applying this monitoring method. These findings have substantially increased the feasibility of monitoring seagrass condition and extent via the MDL, as well as the water quality parameter of average annual water clarity (Kz). A major challenge for image interpretation is presented by the high attenuation of light in water, which often means that spectral methods of image analysis, such as image classification, produce poor results. In response, an improved depth correction approach was developed that uses digital bathymetry (DEM) to assist in removing the spectral attenuation of light by the water column. The method lifted the accuracy of mapping seagrass epiphyte abundance (i.e. the amount of associated algae including epiphytic and drift algae present, related to biomass) by an average 25% to an overall average accuracy of 75%, though it made no difference to the accuracy of SAV density mapping (Note: SAV density relates to the proximity and length of the SAV blades such that high density SAV obscures the substrate and creates high levels of shadowing while lower densities have less shadowing and allow the substrate to be observed.). The improved depth correction method also enabled, for the first time from aerial photography, the production of a spatially explicit map of epiphytic biomass in the form of a continuous prediction surface with values ranging from 4 to 58 g dried weight m-2. In response to the shortcomings of the existing field observation measurements of seagrass density and cover for image interpretation purposes, a new measurement was created, called SAV structural density or SSD, which is designed to improve thematic coherence between aerial photography and field observations, such as downward-looking benthic videography or dive quadrats. This new measurement enabled the consistent discrimination of high and low density SAV with average overall accuracies of 77%, which supports the assessment of seagrass condition, particularly when complemented by the new maps of epiphyte abundance. This thesis presents methods that improve the quality of remote sensing of shallow marine habitats and provides a more reliable basis for further investigation of habitat change detection via spatial metrics and predictive modelling at landscape scales.

Seagrasses are core components of the nearshore environment and there is sustained interest in developing mapping and monitoring techniques of their extent and condition for management purposes. An identified gap in mapping methods is the capacity to monitor at landscape scales, that is, areas that are larger than the 1 m2 quadrat and smaller than those covered by broad area mapping (approximately 5km2 or greater). Monitoring at the landscape scale is required to investigate the dynamic patterning and patchiness present in seagrass beds, as well as providing inputs and validation for predictive modelling. However, the acquisition and use of remote sensing images for these purposes provides many challenges to the practitioner. The primary aim of this thesis is to develop effective optical remote sensing techniques for mapping and monitoring seagrass habitats in shallow temperate marine waters, over depth ranges of approximately 0-10 m and spatial scales of hundreds of square metres. Image capture is often compromised because of environmental conditions, such as sun glitter, water clarity, cloudiness and wind. Small format digital aerial photography was selected as the remote sensing platform for its flexibility and responsiveness regarding deployment when environmental conditions are favourable and its low cost, rapid access to imagery. To address the problem of sun glitter, a simplified algorithm was developed that allows the precise prediction of the extent of sun glitter on vertical, downward-looking imagery with the readily available inputs of sun elevation angle, wind speed and sensor field of view (FOV). Subsurface illumination was also investigated via the modelling of reflection and refraction at the water surface. These improvements and investigations enable more efficient and accurate image capture. Problems are also typically encountered during image interpretation, in part due to the characteristics of the seagrass habitats, including the common occurrence of uncertain boundaries and the high variability of vegetation density. Limitations on the detectability of the maximum depth limit (MDL) of seagrass were examined, with the discovery that if imagery is captured when water clarity is higher than the annual average, the limiting factor is the contrast between the seagrass and the surrounding substrate or submerged aquatic vegetation (SAV). A simple and inexpensive measurement of water clarity, Secchi depth (Zsd), was found to be suitable when applying this monitoring method. These findings have substantially increased the feasibility of monitoring seagrass condition and extent via the MDL, as well as the water quality parameter of average annual water clarity (Kz). A major challenge for image interpretation is presented by the high attenuation of light in water, which often means that spectral methods of image analysis, such as image classification, produce poor results. In response, an improved depth correction approach was developed that uses digital bathymetry (DEM) to assist in removing the spectral attenuation of light by the water column. The method lifted the accuracy of mapping seagrass epiphyte abundance (i.e. the amount of associated algae including epiphytic and drift algae present, related to biomass) by an average 25% to an overall average accuracy of 75%, though it made no difference to the accuracy of SAV density mapping (Note: SAV density relates to the proximity and length of the SAV blades such that high density SAV obscures the substrate and creates high levels of shadowing while lower densities have less shadowing and allow the substrate to be observed.). The improved depth correction method also enabled, for the first time from aerial photography, the production of a spatially explicit map of epiphytic biomass in the form of a continuous prediction surface with values ranging from 4 to 58 g dried weight m-2. In response to the shortcomings of the existing field observation measurements of seagrass density and cover for image interpretation purposes, a new measurement was created, called SAV structural density or SSD, which is designed to improve thematic coherence between aerial photography and field observations, such as downward-looking benthic videography or dive quadrats. This new measurement enabled the consistent discrimination of high and low density SAV with average overall accuracies of 77%, which supports the assessment of seagrass condition, particularly when complemented by the new maps of epiphyte abundance. This thesis presents methods that improve the quality of remote sensing of shallow marine habitats and provides a more reliable basis for further investigation of habitat change detection via spatial metrics and predictive modelling at landscape scales.

Tese no âmbito do Doutoramento em Biociências, ramo de especialização em Ecologia, apresentada ao Departamento de Ciências da Vida da Faculdade de Ciências e Tecnologia da Universidade de Coimbra.
Ecosystem services (ES) are defined as the final outcomes of the internal structures, processes, and functions derived from ecosystems that contribute to human well-being. It is a growing and valuable concept that supports the achievement of Sustainable Development Goals, facilitating connections between people and nature. Nevertheless, although ES have gained representation in science and policy agendas, a gap remains between ES research and the information required to support management decisions. To fulfil this gap, this thesis contributes to highlight the pathways through which the ES concept can be made usable to decision-makers, adding information and presenting methodological approaches organized in two operationalisation facets: A) Access to reliable datasets (i.e. availability of site-specific spatial datasets), and B) support to management actions (i.e. effects of pressures on the delivery of ES). This work is focused on coastal and marine ecosystems and uses the Atlantic coastal region adjacent to the Mondego River as a case study. This region was selected due to community interest to develop environmental management proposals and the willing of local decision-makers to implement an ecosystem-based management to promote the sustainable development of the area. The accomplishment of the abovementioned objective was driven by three main research questions: 1) Are current open source spatial databases suitable to informed decision-making related to coastal and marine ecosystem services in Europe?; 2) Can gap analyses based on site-specific databases contribute to the development of ES research and knowledge transfer?; and lastly, 3) How to explore the role of ES supply and demand in spatial explicit habitat risk assessments? The chapters cover the necessity to gather, systematize and organize information regarding ecosystem services status, trends and gaps (Chapters I, II and III), and the application of conceptual approaches (Chapter III) as well as methodological approaches (Chapters III and IV) to identify the relationships between the ecological characteristics of habitats, the pressures acting on the system, and the role of management actions and decision-making to support the operationalisation of the ecosystem services approach. Chapter I introduces the ES concept, describes the current state of ES studies, placing emphasis on coastal and marine ecosystems, and presents the needs for information to guide decision-making in terms of current conceptual, methodological, policy and applicability challenges. This chapter also describes the main characteristics of the Atlantic coastal region adjacent to the Mondego River, used as a case study to test our proposals. Chapter II covers facets A and B and addresses research question 1. It presents a set of useful databases to map European coastal and marine ES with a deep analysis linked to the Common International Classification of Ecosystem Services (CICES) framework, as well as the Driver-Pressure-Status-Impact and Response (DPSIR) management framework and their utility to support the Marine Strategy Framework Directive. 581 open source databases were analysed, but only 193 allowed further work to map ecosystem services. Most of these were related to regulation services and had the ability to report the status of ecosystems and the Marine Strategy Framework Directive (MSFD) descriptors linked with hydrogeological conditions (D7), Eutrophication (D5), and Biodiversity (D1). The results highlight the open source spatial data limitations to be integrated into in-depth study case analyses focused on mapping ecosystem services in Europe, as well as the need to improve the metadata behind the datasets for a better interpretation of the information. Chapter III is linked to facets A and B and addresses research question 2. It presents a multicriteria decision-making approach to identify ES research hotspots at the local level, based on a literature review of coastal and marine ES. This methodology identified 231 potential and current CICES ecosystem services, linked to 21 EUNIS (European Nature Information System) habitats in the study area. The results highlight the local importance of cultural services and the need to deepen ES research associated to saltworks’ habitats. In addition, the results emphasize the value and necessity to establish priority areas for management actions, since it is not possible to manage large areas at the same time with enough efficiency to obtain good results in all of them. Chapter IV focuses on the operationalisation of facet B and addresses research question 3. This chapter establishes an explicit spatial relationship between habitat risk and the vulnerability of ecosystems in the presence of ES supply, using a modified habitat risk assessment (HRA) model from the InVEST tool. The results indicate that the modified HRA model appears to be more in accordance with the social-environmental realm than a non-modified HRA model, suggesting its usefulness for decision-making processes. For the study area, intertidal habitats, specially seagrasses, were identified to be at risk due to sea level rise and Nitrogen enrichment, however the results also indicate that these habitats are sensitive to management actions, and thus may be positively influenced by actions that tend to reduce their exposure to risk. The findings could be a useful addition to the political management documents that recommend taking precautionary actions to protect potentially vulnerable areas. Chapter V offers a discussion about the need to manage information and create consensus in operative definitions, promote the legitimization of shared objectives and to prioritize actions to avoid the risk of service loss, taking into account the dynamic and context dependence of ecosystem services, as well as a friendly communication strategy to ensure the participation of all social actors in the accomplishment of the objectives defined. The study also suggests ways to introduce ES in decision making through understandable and replicable steps, especially in local scales with scarce information. In summary, this study focuses on the importance of information to describe the spatial heterogeneity that makes it possible to recognize the potential of ecosystems to meet the human demand for services and to improve the understanding and management of ES across scales. Likewise, it stresses the importance of the existence of conceptual and methodological frameworks to provide structure to the real-world, highlighting important avenues for the assessment of ecosystem services and ways to communicate results from the scientific community to policymakers and practitioners.
Os serviços dos ecossistemas (SE) podem-se definir como o resultado final das estruturas internas, processos e funções derivados dos ecossistemas, que contribuem para o bem-estar humano. Este é um conceito em ascensão e importante para atingir os Objetivos do Desenvolvimento Sustentável, nomeadamente por facilitar a conexão entre as pessoas e a natureza. Embora os SE tenham ganhado protagonismo nas agendas científica e política, ainda há uma lacuna entre a investigação em SE e a informação necessária para apoiar decisões de gestão. Neste contexto, este trabalho contribui para o processo através do qual o conceito de SE se torna útil para os decisores políticos, por meio da apresentação da informação e propostas metodológicas que têm sido organizadas em duas facetas do processo de operacionalização: A) Acessibilidade a bases de dados confiáveis (i.e., na disponibilidade de dados espaciais), e B) Apoio à gestão (i.e., nos efeitos das pressões no fornecimento de SE). O trabalho foca-se nos ecossistemas marinho-costeiros e utiliza a Região da Costa Atlântica adjacente ao Rio Mondego como caso de estudo. Esta área tem sido escolhida com base no interesse da comunidade para desenvolver propostas de gestão ambiental e pela disponibilidade dos decisores locais para implementar uma gestão na base dos ecossistemas para promover o desenvolvimento sustentável da área. Para atingir o objetivo acima mencionado o trabalho foi orientado para três questões científicas: 1) Serão as atuais bases de dados espaciais em formato aberto adequadas para apoiar decisões informadas sobre os SE marinho-costeiros na Europa? 2) Poderá a análise de lacunas em bases de dados locais contribuir para o desenvolvimento da investigação em SE e, assim, apoiar a transferência de conhecimento? 3) Como integrar a análise da oferta e da procura em SE numa análise de risco de habitats de forma espacialmente explícita? Os capítulos cobrem a necessidade de reunir, sistematizar e organizar informação relacionada sobre o status dos serviços dos ecossistemas, as tendências e as lacunas ali existentes (Capítulos I, II e III), e a aplicação de abordagens conceptuais (capítulo III) bem como as propostas metodológicas (Capítulo III e IV) a fim de identificar as relações entre as características ecológicas dos habitats, as pressões que atuam sobre o sistema e o papel das ações de gestão e decisões tomadas para apoiar a abordagem dos serviços dos ecossistemas. O Capítulo I apresenta o conceito dos SE, descreve o estado da arte do tema, com ênfase nos ecossistemas marinho-costeiros, e apresenta as necessidades de informação em SE para orientar a tomada de decisões, considerando desafios conceituais, metodológicos, políticos e de aplicação. Este capítulo também descreve as principais características da Região da Costa Atlântica adjacente ao Rio Mondego, usada como estudo de caso para aplicar nossas propostas. O Capítulo II integra as facetas de operacionalização A e B e responde à questão 1). Apresenta um conjunto de bases de dados úteis para mapear os SE marinho-costeiros na Europa e faz uma análise profunda vinculada à estrutura da Classificação Internacional dos SE (CICES) e à estrutura de gestão DPSIR, assim como à sua utilidade para apoiar a Diretiva-Quadro Estratégia Marinha. 581 bases de dados de código aberto foram analisadas, mas apenas 193 tiveram potencial para ser envolvidas em trabalhos de mapeamento de SE na Europa. A maioria das bases de dados está relacionada com os serviços de regulação e tem capacidade para informar sobre o status dos ecossistemas e sobre os descritores relacionados com as condições hidrogeológicas (D7), de eutrofização (D5) e a biodiversidade (D1) da Diretiva Quadro da Estratégia Marinha (MSFD). Assim, os resultados deste capítulo sublinham as limitações e os desafios das bases de dados de acesso aberto para se integrar em estudos mais profundos sobre mapeamento dos SE na Europa e identificam as lacunas e oportunidades que estas oferecem para a gestão sustentável dos ecossistemas marinho-costeiros. O Capítulo III integra as facetas A e B e responde à questão 2). Com base em revisão de literatura, apresenta uma abordagem para a tomada de decisões multicritério capaz de identificar hotspots de investigação em SE ao nível local. Os resultados permitiram identificar, na área em estudo, 231 SE atuais e potenciais, em conformidade com a CICES, vinculados a 21 habitats EUNIS, com destaque para a importância local dos serviços culturais e a necessidade de uma maior investigação associada aos habitats Salinas. Os resultados deste capítulo sublinham o valor e a necessidade de estabelecer áreas prioritárias para as ações de gestão com base na ideia que não é possível implementar ações em grandes áreas, ao mesmo tempo e com a mesma eficiência, para obter bons resultados em todas elas. O Capítulo IV foca-se na faceta B e está orientado para a questão 3). Este capítulo estabelece uma relação espacialmente explícita entre o risco a que um habitat está sujeito e a vulnerabilidade dos seus ecossistemas num contexto de usufruto de SE, utilizando um modelo modificado da Análise de Risco de Habitats (HRA) da ferramenta InVEST. Os resultados mostraram que o modelo modificado apresenta uma maior concordância com a realidade, por comparação com um modelo não-modificado, sugerindo que terá utilidade em processos de tomada de decisão. Para a área de estudo, os habitats interditais, nomeadamente as pastagens marinhas, foram identificadas em risco devido ao incremento no nível do mar e ao enriquecimento do nitrogénio. Os resultados também indicaram que esses habitats são sensíveis as ações de manejo para reduzir sua exposição ao risco. Esta proposta fornece um complemento útil para ser adicionado aos documentos de gestão que recomendam a adoção de ações preventivas para proteger areas potencialmente vulneráveis. O Capítulo V discute a necessidade de apresentar e usar a informação para criar consenso nas definições operativas, promover a legitimação de objetivos partilhados e priorizar ações para evitar o risco de perder SE, considerando a dinâmica e os contextos dos SE e uma estratégia de comunicação amigável que assegure a participação de todos os atores socias na concretização dos objetivos definidos. Na discussão, esta tese sugere formas para introduzir os SE nas decisões ao longo de passos compreensíveis e replicáveis, nomeadamente em escalas locais com escassa informação. Em resumo, este estudo focou sua atenção na importância da informação para descrever a heterogeneidade espacial que torna possível reconhecer o potencial dos ecossistemas para satisfazer a demanda humana de serviços e para melhorar a compreensão e gestão dos SE através de diferentes escalas. Da mesma forma, enfatiza a importância da existência de estruturas conceituais e metodológicas para fornecer estrutura ao mundo real, destacando importantes formas de avaliar os SE e maneiras de comunicar os resultados da comunidade científica aos formuladores de políticas e profissionais.

Tese de Doutoramento, Ciências do Mar (Biologia Marinha)
As áreas marinhas protegidas (AMPs) desempenham um papel chave na preservação da biodiversidade e dos habitats, e na gestão sustentável dos recursos naturais. Um desenho representativo, conectado, resiliente e adequado das redes de AMPs requere uma boa compreensão da distribuição das espécies e da sua preferência de habitats. No entanto, um conhecimento detalhado está geralmente concentrado em apenas alguns locais de amostragem e espécies. A presente tese está direcionada para os peixes marinhos em habitats costeiros até os 40 m. A abordagem que foi desenvolvida foca-se no desenho de redes de AMPs planeadas para otimizar benefícios ecológicos. Para esta finalidade, previsões preditivas de um conjunto de variáveis dos peixes foram mapeadas para identificar áreas prioritárias para a conservação de múltiplos espécies e objetivos. A dissertação está organizada em sete capítulos. O capítulo 1 dá uma introdução geral ao estado do conhecimento da ciência das AMPs, modelos de distribuição de espécies (MDEs) e ao ecossistema marinho da área de estudo, que integra duas ilhas vizinhas do arquipélago dos Açores (Nordeste Atlântico). Uma visão geral da rede existente de AMPs e das pescas regionais está descrita. O capítulo conclui com a motivação e objetivos desta dissertação. O capítulo 2 apresenta o método de amostragem para aquisição dos dados de ictiofauna, a seleção das espécies-alvos, e os dados ambientais que foram usados na dissertação. Os principais métodos de MDEs utilizados para a produção de mapas preditivos são explicados em detalhe. Este capítulo é a base metodológica para os seguintes três capítulos. Abundância ou presença-ausência de peixes de recife com diferentes ecologias tróficas são modelados e mapeados no capítulo 3. Os resultados mostram que o ambiente determinou o padrão espacial das espécies estudadas. Por exemplo, a abundância foi sempresuperior na interface entre os principais tipos de habitat: rocha e sediment. As áreas com a presença potencial de espécies individuais foram espalhadas na área de estudo mas mais pequenos para múltiplas espécies de um determinado nível trófico. Estes hotspots de multi-espécies são uma potencial contribuição para o 'efeito de reserva' minimizando a área necessária para a conservação. O capítulo 4 apresenta modelos espaciais para a biomassa desovante e a fecundidade potencial (o número de oócitos das fêmeas matures) de peixes de recife selecionados. As duas medidas mostraram uma distribuição espacial heterogénea por espécie e influenciada pela batimetria, forças oceanográficas e distância à fronteira do habitat. Mapas de fecundidade potencial refinaram os padrões espaciais da biomassa desovante para espécies com um sex-ratio altamente enviesado para os machos. Hotspots reprodutivos de várias espécies são potenciais "habitats fonte" aumentando a produção e exportação de larvas para áreas adjacentes. Como tal, vêm potencialmente apoiar o "efeito de recrutamento" das AMPs. Consequentemente, a sua total proteção deve ser promovida. Diferentes índices de biodiversidade e de vulnerabilidade intrínseca para a pesca são analisados no capítulo 5. A proteção de locais de alta biodiversidade é frequentemente exigida em conservação marinha. No entanto, os resultados mostraram que somente os padrões de biodiversidade podem não representar bem as áreas de maior interesse e necessidade para a conservação. A integração da vulnerabilidade intrínseca para a pesca no planeamento espacial resultou numa identificação mais precisa de sítios prioritários. A combinação de ambos os parâmetros é proposta como uma nova abordagem para apoiar o planeamento espacial marítimo que serve a gestão pesqueira e os objetivos de conservação. O capítulo 6 é uma análise abrangente que combina os mapas preditivos que foram produzidos nos três capítulos anteriores com outras características do habitat e sócio-econômicos. Cenários alternativos da rede de reserva foram produzidos com o software 'Marxan', considerando diferentes alvos de conservação e objetivos. Os resultados demonstraram que a estatística da rede (ex. tamanho, 'rácio da-borda-à-área', e percentagem da linha de costa protegida) e o posicionamento da reserva foram influenciados, principalmente, pelos diferentes níveis de proteção. As diferenças foram menos pronunciadas entre soluções que se focaram na pesca ou na conservação da biodiversidade. As soluções criadas pelo Marxan correspondem parcialmente à atual rede de AMPs. Estas mostram alternativas para a localização e tamanho das áreas protegidas, que podem ser usadas em processos de gestão adaptativa. O capítulo 7 combina a discussão geral dos resultados da dissertação, impactos e possíveis trabalhos futuros. Os resultados evidenciam que as AMPs não podem beneficiar igualmente todas as espécies, portanto, é fundamental incluir informação da ecologia espacial de multi-espécies no seu desenho. Análises de representatividade mostraram que todos os hotspots de multi-espécies são quantitativamente bem integrados na rede existente de AMPs. Porém dada a elevada importância biológica/ecológica e a pequena extensão destes hotspots, os futuros processos de gestão devem promover a proteção de toda a área de hotspots para assegurar o seu funcionamento ecológico. Devido à sua clareza a aplicação de mapas preditivos deve ser uma ferramenta prioritária para a gestão do espaço marítimo, especialmente em situações de escassez de dados, desde que rigorosos critérios de validação sejam aplicados. O enquadramento apresentado é simples, direto e eficiente na identificação de habitats com potencialmente alta abundância, fecundidade, diversidade e vulnerabilidade para a pesca. Propõe-se a integração desta abordagem promissora como um primeiro passo de um múltiplo processo para a identificação de sítios prioritários para a conservação que servem vários objetivos.
ABSTRACT: Marine protected areas (MPAs) can play a key role in preserving biodiversity and habitats and in managing the sustainable use of natural resources, including fish stocks. Designing representative, connected, resilient, and adequate MPA networks requires a good understanding of the species’ distribution and habitat preferences. Yet, detailed knowledge is often reduced to a few sampled locations and species. This thesis focuses on marine reef fishes in coastal habitats down to the 40 m isobath. A framework is developed to design MPA networks that optimise ecological benefits. For this purpose, predictive distributions of a set of fish variables are spatialised to identify potential sites of priority for conservation that can serve multiple species and objectives. The thesis is organised in seven chapters. Chapter 1 gives an overall introduction to the state of the art on the science of MPAs, species distribution models (SDMs) and the marine ecosystem of the study area, two neighbouring islands of the Azores archipelago (Northeast Atlantic). It includes an overview of the existing MPA network and regional fisheries. The motivation and objectives of the thesis are outlined. Chapter 2 presents the sampling method to acquire fish data, the selection of individual study species, and the environmental data that are used in the thesis. The main methods and background knowledge for the statistical models used to describe the species-environment relationships and to produce predictive maps are explained in detail, setting the basis for the following three chapters. The occurrence and abundance of reef fishes with different trophic ecologies are modelled and spatialised in chapter 3. Results showed that the environment shapes the spatial distribution patterns of the reef fishes. For instance, the abundance was typically highest at the interface between rock and sediment, highlighting the importance of this main ecotone for subtidal fish assemblages. Individual species were predicted to occur in large parts of the study area but these areas were much smaller if multiple species from the same trophic guild were considered. These multi-species abundance hotspots can be a major potential contribution to the 'reserve effect' of MPAs while minimising the area needed for protection. Chapter 4 presents spatially explicit models for the spawning biomass and the potential fecundity (number of oocytes of mature females) of selected reef fishes. The spatial distribution of both measures was heterogeneous, species-specific and influenced by bathymetry, oceanographic forces and the distance to the habitat edge. Maps of the potential fecundity further refined the spatial patterns of the spawning biomass for species with sex-ratios highly skewed towards males. Multi-species reproductive hotspots were identified and are potential 'source habitats' of increased larvae production and export to adjacent areas. As such, they potentially support the 'recruitment effect' of MPAs and their complete protection should be promoted. Chapter 5 evaluates different indices of biodiversity and vulnerability to fishing of fish assemblages. The protection of high-biodiversity sites is often demanded in marine conservation. Yet, results showed that biodiversity patterns alone may not represent well the areas of higher need for conservation. Integrating the intrinsic vulnerability to fishing in spatial planning resulted in a more precise identification of priority sites. The combination of both parameters is proposed as a novel approach to support marine spatial planning that serves fisheries management and conservation objectives. Chapter 6 is one comprehensive analysis that combines the predictive maps produced in the previous three chapters with additional habitat and socio-economic characteristics. Alternative scenarios for a reserve network are produced with the systematic conservation software 'Marxan' considering different conservation targets and objectives. Results demonstrated that the network statistics (e.g. size, edge-to-area ratio, and percentage of protected coastline) and reserve localisation were mainly influenced by the targeted level of protection. In contrast, differences were less pronounced between solutions that focused either on fisheries aspects or the protection of biodiversity within a given conservation target. The solutions provided by Marxan overlapped only partially with the existing MPA network. They provide potential alternatives for the location and size of protected areas that can be used in adaptive management processes. Chapter 7 combines a general discussion of the thesis results, impacts and possible future work. Results highlight that MPAs may not equally benefit all species, thus it is critical to include information of multi-species spatial ecology in their design. Analyses of representativeness showed that all multi-species hotspots are quantitatively well integrated in the existing MPA network. However, given the high biological/ecological significance and the rather small extent of these hotspots, future adaptive management processes should, possibly, promote the protection of the entire area of the hotspots to ensure their ecological functionality. Because of its clarity the application of predictive species distribution maps should be a principle tool for marine spatial management, especially in data scarce situations, provided that rigorous validation criteria are applied. The presented framework is simple, straightforward and efficient in identifying habitats with potentially high fish abundance, fecundity, biodiversity or vulnerability to fishing. It is proposed to integrate this promising approach as a first step of a manifold process for the identification of priority sites for conservation that serve multiple purposes.

Techniques for seabed habitat mapping have developed considerably over the past 10 years, principally through advances in acoustic technologies and tools for improved spatial analysis. The increasing need for information on the distribution and structure of seabed habitats in coastal waters (< 50 m deep) has developed in Australia due to increasing pressures from development and exploitation. A clear understanding of the uncertainties in creating spatial information from marine data is required within seabed mapping projects, particularly those using acoustic methods that vary in coverage and resolution. This thesis investigates methods to explore spatial uncertainty in the techniques used to characterise shallow water seabed habitats using local and regional scale case studies ranging from interpolation of sediment cores to classifying digital elevation models generated from multi-beam acoustic data. Uncertainties are investigated in a multidiscipliary approach to habitat mapping. Broad-scale and fine-scale mapping of marine seabed habitats can provide considerable information on patterns of physical seafloor structuring and this is now achievable given recent technological advances in echosounders and backscatter analysis, digital underwater video, differential GPS and Geographic Information Systems (GIS). The uncertainties in classifying single beam acoustic data are examined by comparing data visually classified into habitat classes in real time compared to those defined through post-processing in the laboratory. The influence of a range of transect spacings on qualitative knowledge-based interpolation of shallow rocky reef acoustic data is assessed. Parameters of physical reef characteristics from single beam acoustic data are investigated which aid in improving kriging interpolation techniques. A fuzzy classification method is applied to sediment core data collected over 100s of kms in order to map the distribution of unconsolidated sediments. This technique is tested on Australia's SeaScapes data. The SeaScapes layer was recreated with classes derived from an unsupervised fuzzy clustering algorithm. A sensitivity analysis was performed which explores the optimal number of clusters and clearly shows where classes overlap, highlighting confusion and therefore potential classification errors in the data. Conditional simulation was utilised to explore uncertainty in the interpolated data layers and animations produced which provided an advanced way of visualising the surfaces. Image segmentation techniques are applied at various values of splitting and merging thresholds to identify objects in sidescan sonar imagery. Grey Level Co-occurrence Matrices (GLCM) texture measures are integrated with segmentation to identify homogeneous texture regions in an image. The segmentation and object oriented classification provide good classification results in 10-40 m water depth with accuracy values of >80 % when classifying two classes and >60% when classifying three classes. This section of research focuses on the analysis of seabed habitats through the use of algorithmic calculations at multiple scales to quantitatively delineate distinct seabed regions based on texture parameters. The textural characteristics of three habitat classes are explored and tested onsidescan sonar data in two case studies. Segmentation results are validated using underwater video transects and statistical techniques. The classified sidescan acoustic images are spatially characterised using fragmentation statistics. These results are a leap forward providing a methodology and a terminology to describe the distribution of shallow rocky reef, combined with a classified map and an uncertainty map generated by the object oriented technique. Fuzzy classification techniques are used to characterise the two dimensional structure of shallow rocky reefs from multi-beam bathymetric data. The results from two morphometric classification techniques are contrasted and compared. Many physical and biological processes acting on the seabed are highly correlated with bathymetric features, such as ridges and channels. Examples of these include the selection of habitat by many commercially fished species, such as rock lobster, abalone and reef associated fish species. These physical attributes can therefore often be key predictors of habitat uitability, community composition and species distribution and abundance. These methods greatly improve insight into classification and related uncertainties of morphometric classification. There are many potential applications of seabed habitat mapping for which estimates of uncertainty will provide additional crucial information. As habitat mapping becomes more common in Australian coastal waters and datasets build up over time, compatibility between different sets of information will be essential. Monitoring and change detection analysis requires detailed baseline data with uncertainty estimates to ensure that features that display change are reliably detected. The accuracy of marine habitat maps and their associated levels of uncertainty are extremely hard to convey visually or to quantify with existing methodologies. The new techniques developed in this research provide a rigorous alternative to current practices. The methods used in this research integrate existing techniques in a novel way to improve insight into classification and related uncertainty for seabed habitat maps which will progress and improve resource management for Australian coastal waters.

In marine habitat mapping, single beam echo sounders are widely used to derive information about the geophysical properties of the seabed, while underwater video can provide supplementary information about the physical structure of the seabed and associated marine biological communities. In this thesis, data from both systems are integrated and used to classify seabed habitats. The habitat classification is based on categories within a hierarchical system that is conducive to information from different instruments or collected at different spatial scales. The classification of single beam echo sounder data at different levels of the hierarchical classification is the focus of the first half of the thesis. The first data chapter examines the effect of depth, bottom slope, prevailing weather conditions, and vessel speed on measured acoustic return from the seabed, and the subsequent capacity to classify this data at the substrate level. The following three chapters are case studies that progressively develop techniques for classification of single beam acoustic data at lower levels of the hierarchical classification including identification of soft sediment habitats in commercial scallop fishing grounds; mapping the distribution of urchin barrens on rocky reefs; and detection and mapping of sub-surface giant string kelp, Macrocystis pyrifera. Each of the case studies develops analysis and classification techniques that are applicable for mapping at levels below substrate in the hierarchical classification. At lower levels of the hierarchical classification, biological communities and species distributions are commonly used as habitat descriptors. The second half of this thesis focuses on extracting information from video for the classification of biological communities. In the first of these chapters, methods are compared for extracting estimates of algal cover on temperate rocky reef substrates from towed underwater video. The algal cover data is then used to examine the capacity of a towed video to detect changes in algal community structure at two spatial scales. The following chapter describes the design, construction and evaluation of a stereo video system developed to measure sponge morphological metrics. These metrics are then used to establish a quantitative classification of sponge functional morphology. Differences in sponge functional morphology are examined between sponge communities in two different regions with differences detected in both the composition of functional groups and the size of those functional groups. The thesis presents a framework and methodologies for extracting both physical and biological information from single beam echo sounder and underwater video systems. These methods can easily be incorporated into existing seabed mapping programs, and provide information that will improve our understanding of the spatial distribution of subtidal habitats. This information is directly beneficial to marine resource management, including marine protected area planning and fisheries management, and will allow baseline documentation of habitats for future climate change research.

This thesis explores learning in and through the emergence of a network of communities who participated in the B.C. Coastal Eelgrass Stewardship Project (the Project). I draw on a two-year ethnographic investigation of 20 community groups who were trained to map and monitor eelgrass habitat and carry out education and stewardship-related activities. People from a multitude of backgrounds, including scientists and non-scientists, and a diversity of places, from small coastal communities to urban centers, worked towards the collective goal of mapping and conserving the extent of eelgrass habitat along the coast. Using cultural-historical activity theory, I develop an alternative framework for understanding learning and change in a network of communities. The collection of three main chapters, shows that learning, emergence, and stabilization of the network arose through the following dialectical relations: individual/collective, social/material, and agency/structure. This thesis shows that viewing and supporting the Project as a dynamic learning network makes it more stable.

Eelgrass (Zostera marina) is a critical nearshore marine habitat for juvenile Pacific salmon (Oncorhynchus spp.) as they depart from their natal streams. Given the poor marine survival of Coho (O. kisutch) and Chinook (O. tshawytscha) salmon juveniles in recent decades, it is hypothesized that deteriorating eelgrass habitats could contribute to their low survival. The primary goal of this research was to investigate the possible long-term spatial-temporal trends in eelgrass habitat in the Salish Sea and was addressed by two main objectives: (1) Define a methodology for mapping eelgrass habitats using UAV imagery to create a baseline for long-term mapping; and (2) Assess changes in eelgrass area coverage and fragmentation over the period of 1932-2016 using historic aerial photographs and Unmanned Aerial Vehicle (UAV) imagery, and assess the relationship between eelgrass and residential housing density and shoreline activities. Three study sites in the Southern Gulf Islands of the Salish Sea were chosen for analysis. The overall accuracies of eelgrass delineation from UAV imagery were 95.3%, 88.9%, and 90.1% for Village Bay, Horton Bay, and Lyall Harbour, respectively. The UAV method was found to be highly effective for this size of study site, however results were impacted by the environmental conditions at the time of acquisition, namely: sun angle, tidal height, cloud cover, water clarity, and wind speed. The results from the first objective were incorporated into a long-term dataset of historic aerial photography and used to evaluate changes in eelgrass area and fragmentation. All three eelgrass meadows showed a deteriorating trend in eelgrass condition. On average, eelgrass area coverage decreases by 41% while meadow complexity as indicated by the shape index increases by 76%. Shoreline activities (boats, docks, log booms, and shoreline armouring) and residential housing density increased markedly at all sites over the study period. By using a linear correlation model, it was revealed that eelgrass areal coverage and fragmentation (Shape Index) were, in general, very strongly correlated to these landscape-level coastal environmental indicators. While this correlation model is not meant to show a direct causative impact on eelgrass at these sites, these results suggest an overall deterioration of coastal environmental health in the Salish Sea due to a dramatic increase in the use of the coastal zone, as well as likely declines in water quality due to urbanization.
Graduate

For the west coast of North America, from northern California to southern Washington, a habitat suitability prediction framework was developed to support wave energy device siting. Concern that wave energy devices may impact the seafloor and benthos has renewed research interest in the distribution of marine benthic invertebrates and factors influencing their distribution. A Bayesian belief network approach was employed for learning species-habitat associations for Rhabdus rectius, a tusk-shaped marine infaunal Mollusk. Environmental variables describing surficial geology and water depth were found to be most influential to the distribution of R. rectius. Water property variables, such as temperature and salinity, were less influential as distribution predictors. Species-habitat associations were used to predict habitat suitability probabilities for R. rectius, which were then mapped over an area of interest along the south-central Oregon coast. Habitat suitability prediction models tested well against data withheld for crossvalidation supporting our conclusion that Bayesian learning extracts useful information available in very small, incomplete data sets and identifies which variables drive habitat suitability for R. rectius. Additionally, Bayesian belief networks are easily updated with new information, quantitative or qualitative, which provides a flexible mechanism for multiple scenario analyses. The prediction framework presented here is a practical tool informing marine spatial planning assessment through visualization of habitat suitability.
Graduation date: 2012

Η παρούσα διδακτορική διατριβή αποτελεί ένα συνδυαστικό ερευνητικό προϊόν που στοιχειοθετείται από την ανάπτυξη υπολογιστικών εργαλείων επεξεργασίας και ανάλυσης θαλάσσιων γεωφυσικών δεδομένων και την εφαρμογή τους σε πρωτογενή δεδομένα, συλλεγμένα από το Εργαστήριο Θαλάσσιας Γεωλογίας και Φυσικής Ωκεανογραφίας (Ε.ΘΑ.ΓΕ.Φ.Ω) του πανεπιστημίου Πατρών, κατά το διάστημα 2005 – 2011. Τα πεδία στα οποία συγκεντρώνεται το κέντρο βάρους της διατριβής είναι: 1) τα συστήματα ακουστικής ταξινόμησης πυθμένα και 2) η χαρτογράφηση και παραμετροποίηση εμφανίσεων των πολύ σημαντικών θαλάσσιων ενδιαιτημάτων της Ποσειδώνιας και των κοραλλιογενών σχηματισμών στο Ιόνιο και στο Αιγαίο πέλαγος. Έτσι αναπτύχθηκαν και παρουσιάστηκαν εκτενώς τα λογισμικά εργαλεία SonarClass και TargAn, που αναφέρονται αντίστοιχα στην αυτόματη ακουστική ταξινόμηση πυθμένα και την παραμετροποίηση περιοχών ενδιαφέροντος σε εικόνες ηχοβολιστών ευρείας σάρωσης και εφαρμόστηκαν για την χαρτογράφηση λειμώνων ποσειδώνιας στη Ζάκυνθο και κοραλλιογενών σχηματισμών (τραγάνας) στις Κυκλάδες νήσους. Παράλληλα και επεκτείνοντας το εύρος των ερευνητικών προϊόντων αυτής της διατριβής, αναπτύχθηκαν επίσης: 1) το λογισμικό SBP-Im-An για τη γεωαναφορά και ψηφιοποίηση παλαιών αναλογικών καταγραφών τομογράφων υποδομής πυθμένα, 2) το λογισμικό χωροστάθμησης θαλάσσιων γεωμαγνητικών δεδομένων MagLevel και 3) το λογισμικό ποσοτικοποίησης αλιευτικών ιχνών σε δεδομένα ηχοβολιστών ευρείας σάρωσης PgStat, με αντίστοιχες σημαντικές εφαρμογές σε πρωτογενή δεδομένα. Η παρούσα διατριβή επιδεικνύει πρωτοτυπία τόσο σε επίπεδο ανάπτυξης νέων μεθόδων ανάλυσης και επεξεργασίας γεωφυσικών δεδομένων όσο και σε επίπεδο παρουσίασης εφαρμογών τους σε περιοχές μελέτης με ιδιαίτερο περιβαλλοντικό ενδιαφέρον αλλά και σε πεδία έρευνας για τα οποία το ενδιαφέρον της σύγχρονης θαλάσσιας επιστημονικής κοινότητας βρίσκεται στο απόγειό του.
The present PhD thesis is a combinational research product concerning the development of software tools for the processing and analysis of marine geophysical data and their application to original data, collected by the Laboratory of Marine Geology and Physical Oceanography (L.M.G.P.O), university of Patras, Greece, during the period 2005-2011. The fields that this thesis focuses on are: 1) the Acoustic Seabed Classification Systems and 2) the mapping and quantification of very important marine habitats that specifically are the Posidonia Oceanica Prairies and the Coralline formations in the Aegean and Ionian seas. The software tools SonarClass and TargAn, that respectively refer to the Acoustic Seabed Classification and the quantification of Regions Of Interest in swath sonar imagery are presented and applied to the cases of Posidonia Oceanica in Zakinthos Isl. (Ionian Sea) and Coralline formations in Cyclades Isl. (Aegean Sea). Additionally and extending the range of the research products of this thesis, other software tools that are presented are: 1) the SBP-Im-An for the recreation (georeferencing and digitization) of old analog Sub Bottom Profiler recordings, 2) the MagLevel for the tie line leveling of marine geomagnetic data and 3) the PGStat for the quantification of trawl marks in swath sonar imagery, all of them with significant applications to original data. This thesis demonstrates originality due to both the development of new methods for the analysis and processing of marine geophysical data and the applications to study areas with particular environmental interest and research fields for which the attention of the marine scientists is at its peak.