World-wide coral reefs are in a state of decline as a result of many local and regional factors. Recent global mass mortalities of reef corals due to record warm sea temperatures have led researchers to consider global warming as one of the most significant threats to the persistence of coral reef ecosystems over the next 100 years. It is well established that elevated sea surface temperatures (SSTs) cause widespread coral bleaching, yet confusion lingers as to what facet of extreme temperatures is most important. Utilizing long-term in situ datasets, nine thermal stress indices were calculated and their effectiveness at segregating bleaching years a posteriori for multiple reefs on the Florida Reef Tract was tested. Simple bleaching thresholds based on deviations above the climatological maximum monthly SST were just as effective at identifying bleaching years as complex thermal stress indices. Near real-time bleaching alerts issued by NOAA's Integrated Coral Observing Network (ICON) are now based upon a running 30-day average SST, such that alerts are only issued when the running 30-day average SST exceeds the estimated bleaching threshold for a particular site. In addition to three widespread, mass-coral bleaching events, the Florida Reef Tract was impacted by three tropical storms and 12 hurricanes from 1997-2005. Sea surface cooling associated with the high frequency of hurricanes that impacted Florida in 2005 likely acted to ameliorate the severity and duration of bleaching. Nonetheless, hurricane-associated cooling is not expected to nullify the proposed effects of climate change on coral reefs. The role of thermal stress in coral bleaching has been extensively studied for eastern tropical Pacific (ETP) coral reefs. However, the ETP presents a unique opportunity as this region has sub-optimal conditions for coral reef development because of upwelling of carbon dioxide-enriched deep water along the shallow thermocline. This upwelling results in a depressed aragonite saturation state which is likely an additional factor in the poor reef development throughout the ETP. The highest aragonite saturation documented in this study occur in the Gulf of Chiriquí, which corresponds with the greatest reef development of the entire ETP. Seasonal upwelling had a significant effect on the carbonate chemistry of surface waters in Pacific Panamá. This regionally-depressed aragonite saturation state of the ETP appears to result in corals with a less dense skeleton. Density values of poritid corals from the Galápagos, where aragonite saturation was the lowest documented in this study, were significantly less dense relative to those from Panamá and the Great Barrier Reef. The density of non-living pocilloporid framework components were no different across the ETP aragonite saturation gradients. This could be a result of the activity of boring sponges removing the primary carbonate material within the dead coral skeleton, thus lowering density, albeit physical-chemical dissolution cannot be ruled out. These studies provide real-world examples of the ramifications of global climate change upon coral reef ecosystems.

The goal of this research was to develop statistically robust ecosystem-based approaches, while optimizing data acquisition on relatively unexploited fish species in South Florida reefs, i.e., parrotfishes, Family Scaridae, in Biscayne Bay (with seasonal roller frame beam trawl surveys, 1996-2000) and Florida Keys (with annual Reef Fish Visual Censuses, 1997-2001), by following these steps: (I) analysis of information gaps for the stocks, including systematics, biogeography, population dynamics, reproductive ecology, trophodynamics, habitat use, and fisheries dynamics of Western Atlantic parrotfishes; (II) determination of primary research objectives from prioritization in Step I; (III) determination of essential fish habitats, ontogenetic shifts, migrations, and reef-seagrass habitat, from integration of stratified sampling design for fisheries-independent surveys, habitat selection theory-based analyses, and length-based analyses; (IV) estimation of population dynamics and fisheries-specific parameters encompassing life history demographics from empirical data or comparisons to theoretical expectations adapted to local conditions; (V) simulation modeling of a realistic range of fishing scenarios and demographic characteristics to evaluate the efficacy of potential traditional fisheries and spatial management strategies; and (VI) application of sampling optimization procedures and fisheries ecology approaches. Four scarid species had an estimated combined abundance of ca. 36.8 x 106 individuals in the Florida Keys. Connectivity among seagrass beds, coral reefs and deep waters had three major patterns: seagrass dwellers, reef dwellers, with inshore-to-offshore ontogenetic , and a seagrass-reef connection, using Biscayne Bay as an important recruitment ground. Marine protected areas of the Florida Keys National Marine Sanctuary did not show effects on abundance, size composition or spatial distribution of any parrotfish studied. Simulations suggested relatively short longevities (5-10 years), moderate body growth curvature, high instantaneous natural mortality rates (0.3-0.6 y super minus one), and low annual survival rates (27-54%). Simulated estimates of fishing mortalities ranged from 0.3 to 0.6 y super minus one, indicating low levels of exploitation, but low Spawning Potential Ratios (SPR = 23.5-26%). Proposed potential exploitation based on a legal minimum size equal to their size at first maturity and fishing rates equal or below to their natural mortality should secure SPR values at 45-48%.

Marine invertebrate have been found to be a rich source of bioactive secondary metabolites. During the course of our investigation on the bioactive chemical constituents from marine invertebrates, twenty-eight metabolites have been isolated from two soft coral Lobophytum crassum and Dendronephthya griffini. Investigation on L. crassum has led to the isolation of fourteen compounds, including seven new cembranoids, crassumolides A¡VG (1¡V7), and three new glycolipids (2R)-1-hydroxy -3-hexadecyloxy-propyl-£]-D-arabinopyranoside (8), (2R)-1-hydroxy-3- octadecyloxy-propyl-£]-D-arabinopyranoside (9), and (2R)-1-acetoxy-3- hexadecyloxy-propyl-£]-D-arabino-pyranoside (10), coupled with four known compounds, lobohedleolide (11), 17-dimethylaminolobohedleolide (12), sinulariol A (13), and denticulatolide (14). Ten new steroids, griffinisterones A¡VJ (15¡V24), were isolated from the soft coral D. griffini, while a known, 15-chlorogriffinsulfate (25), and three new polychlorinated acyclic compounds¡Agriffinsulfate (26), 15-chlorogriffinol (27), and griffinol (28), were also purified from the same organism. Structures of these metabolites were identified as new natural products by extensive spectroscopic methods. Except for the use of 2D NMR, the 24-epimers of 15 and 16 were identified by a single-crystal X-ray crystallography on 15. Lobohedleolide (11), obtained in large quantity in L. crassum, has also been modified to 29¡V34 by chemical conversion. Oxidation with meta-chloro-peroxybenzoic acid (MCPBA) afforded compounds 29 and 30, and with selenium dioxide led to the formation of 31. Compound 34 is the product of O-coupling reaction of 11 with 1-hydroxybenzotriazole (HOBt). Over 99¢Mdiastereoselectivity was observed in the process of Henbest hydroxyl-directed epoxidation on 11 to yield 32. EDC-coupling with an aid of HCl salt of DMAP afforded methylester 33 in high yield, and proved the absolute stereochemistry of 3. The cytotoxicity and anti-inflammatory activities of the natural and modified compounds were also discussed herein.

Coral reefs, seagrass beds and mangroves constitute a complex mosaic of habitats referred to as the tropical seascape. Great gaps exist in the knowledge of how these systems are interconnected. This thesis sets out to examine ecological connectivity, i.e. the connectedness of ecological processes across multiple scales, in Zanzibar and Mafia Island, Tanzania. Paper I examined the current knowledge of interlinkages and their effect on seascape functioning, revealing that there are surprisingly few studies on the influences of cross-habitat interactions and food-web ecology. Furthermore, 50% of all fish species use more than one habitat and 18% of all coral reef fish species use mangrove or seagrass beds as juvenile habitat in Zanzibar. Paper II examined the seascape of Menai Bay, Zanzibar using a landscape ecology approach and studied the relationship between fish and landscape variables. The amount of seagrass within 750m of a coral reef site was correlated with increased invertebrate feeder/piscivore fish abundance, especially Lethrinidae and Lutjanidae, which are known to perform ontogenetic and feeding migrations. Within patch seagrass cover was correlated with nursery species abundance. Paper III focused on a seagrass-dominated seascape in Chwaka Bay, Zanzibar and showed that small-scale habitat complexity (shoot height and density) as well as large-scale variables such as distance to coral reefs affected abundance and distribution of a common seagrass parrotfish Leptoscarus vaigiensis. Paper IV studied the connectivity and functional role of two snappers (Lutjanus fulviflamma and L. ehrenbergii) using stable isotopes (δ15N and δ13C) and found that connectivity between habitats was maintained by ontogenetic and foraging migrations by these species. The thesis concludes that ecological connectivity and multi-habitat usage by fish is a general and important characteristic in the Western Indian Ocean and should be considered in management planning. / <p>At the time of the doctoral defense, the following papers were unpublished and had a status as follows: Paper 2: Submitted. Paper 4: Submitted.</p>

Over the last decade, coral reef remote sensing research has focused on habitat map development. Advances in field methods, spatial and spectral resolution of remote sensing data, and algorithm development have led to more detailed map categories and to heightened map accuracy. Studies have provided guidance for practitioners in areas such as imagery selection, algorithm application, and class selection methods, but the product has remained relatively unchanged – a habitat map showing the spatial distribution of a range of substrate classes, classified primarily on the basis of their spectral signature. However, the application of such a product in a management context has not been elaborated by the remote sensing community. The research described in this thesis addresses the challenge that the application of remotely-sensed coral reef information in a coral reef management environment elicits. In such an environment, the coral reef manager asks: "What can the map do to help me?", while the remote sensing scientist asks: "What type of information do you need?". The research described here aims to reconcile these two points of view, by answering the research question of this thesis: How can coral reef remotely-sensed information address stakeholder-specific coral reef management objectives? This question was answered through the development of a four-stage strategy. The strategy includes: 1) developing a traditional habitat map, 2) investigating stakeholder receptivity to the habitat map, 3) linking stakeholder interests with habitat data, and 4) illustrating the linked habitat data in what we term a management map. The strategy was applied on Bunaken Island, Indonesia, and involved the collection of both qualitative and quantitative data sets. The research was relevant to the communities on Bunaken Island, as they are directly responsible for the management of the coral reef resources surrounding Bunaken Island, and they are regularly planning and implementing coral reef management projects. The effectiveness of the four-stage strategy was evaluated in a framework that compares potential and actual uses of habitat maps and management maps in coral reef management projects. It was shown that management maps are superior to habitat maps for a wide range of management purposes. This research has provided two main contributions to the field of coral reef remote sensing and management. The first is the four-stage strategy that results in the development of management maps, and the second is the framework for evaluating the effectiveness of the management maps. This research seeks to traverse the gap between producers and users of coral reef remotely-sensed information. The recommendations made from this research addresses coral reef management procedures, action research, and cross-cultural communication. Each recommendation is founded on collaboration between scientist and manager. Such collaboration is crucial for successful application of remotely-sensed information to management.

Over the last decade, coral reef remote sensing research has focused on habitat map development. Advances in field methods, spatial and spectral resolution of remote sensing data, and algorithm development have led to more detailed map categories and to heightened map accuracy. Studies have provided guidance for practitioners in areas such as imagery selection, algorithm application, and class selection methods, but the product has remained relatively unchanged – a habitat map showing the spatial distribution of a range of substrate classes, classified primarily on the basis of their spectral signature. However, the application of such a product in a management context has not been elaborated by the remote sensing community. The research described in this thesis addresses the challenge that the application of remotely-sensed coral reef information in a coral reef management environment elicits. In such an environment, the coral reef manager asks: "What can the map do to help me?", while the remote sensing scientist asks: "What type of information do you need?". The research described here aims to reconcile these two points of view, by answering the research question of this thesis: How can coral reef remotely-sensed information address stakeholder-specific coral reef management objectives? This question was answered through the development of a four-stage strategy. The strategy includes: 1) developing a traditional habitat map, 2) investigating stakeholder receptivity to the habitat map, 3) linking stakeholder interests with habitat data, and 4) illustrating the linked habitat data in what we term a management map. The strategy was applied on Bunaken Island, Indonesia, and involved the collection of both qualitative and quantitative data sets. The research was relevant to the communities on Bunaken Island, as they are directly responsible for the management of the coral reef resources surrounding Bunaken Island, and they are regularly planning and implementing coral reef management projects. The effectiveness of the four-stage strategy was evaluated in a framework that compares potential and actual uses of habitat maps and management maps in coral reef management projects. It was shown that management maps are superior to habitat maps for a wide range of management purposes. This research has provided two main contributions to the field of coral reef remote sensing and management. The first is the four-stage strategy that results in the development of management maps, and the second is the framework for evaluating the effectiveness of the management maps. This research seeks to traverse the gap between producers and users of coral reef remotely-sensed information. The recommendations made from this research addresses coral reef management procedures, action research, and cross-cultural communication. Each recommendation is founded on collaboration between scientist and manager. Such collaboration is crucial for successful application of remotely-sensed information to management.

Coral diseases have become more prevalent and destructive over the past 20 years, possibly due to an increase in stressful environmental factors that may weaken corals defenses against disease. Aspergillosis is a disease caused by the fungus Aspergillus sydowii, which apparently infects only two species of gorgonian corals in the Caribbean Ocean (Gorgonia ventalina and G. flabellum). We hypothesized that the differential resistance to infection is caused by differences in chemical defenses among gorgonians. Freeze-dried gorgonian powders and extracts deterred fungal growth, but potencies varied among gorgonian species and among fungi. Extracts and powders generated from G. ventalina all strongly inhibited fungal growth. Since G. ventalina was predicted to have weak antifungal chemical defenses compared to gorgonians not known to suffer from aspergillosis, we concluded that gorgonian susceptibility to fungal infection is determined by factors other than, or in addition to, chemical defenses. In order to investigate specific gorgonian antifungal strategies, we attempted to use bioassay-guided fractionation to isolate antifungal compounds from four gorgonians: Gorgonia ventalina, Briareum asbestinum, Eunicea succinea, and Pseudopterogorgia americana. We succeeded in isolating two antifungal compounds, diastereomers of 9,11-seco-24-hydroxydinosterol, from the gorgonian Pseudopterogorgia americana. This compound was previously identified by other groups, but this study is the first to establish its antifungal activity. At natural concentration, one diastereomer of 9,11-seco-24-hydroxydinosterol inhibited the growth of three different fungi, suggesting that at least this diastereomer may possess broad-spectrum antifungal activity. The results from our survey of gorgonian chemical defenses indicate that susceptibility to aspergillosis cannot be explained by chemical growth inhibition alone. Further areas of investigation include induction of gorgonian chemical defenses, examination of growth-inhibiting mechanisms of antifungal metabolites, and identification of non-chemical factors affecting gorgonians vulnerability to fungal infection.

A transect of cores ranging from 798 m to 1585 m water depth in the South Atlantic Ocean document the relative intermediate water mass nutrient geometry and stable isotopic variability of AAIW during the Younger Dryas cooling event. The data reveal concurrent delta^13 C and delta^18 O excursions of 0.59 ppt and 0.37 ppt within the core of Antarctic Intermediate Water (AAIW) centered at 11,381 calendar years before present based on radiometric age control. A portion of the delta^1 3C variability (0.22 ppt) can be explained by a shift in thermodynamic equilibrium concurrent with a drop in temperature of 1.8°C at the locus of AAIW formation. The remaining 0.37 ppt increase in delta^13 C most likely resulted from increased wind velocities, and a greater coupling between the ocean and the atmosphere at the locus of AAIW formation (increased efficiency of the thermodynamic process). Deepwater coral mounds are aggregates of corals, other organisms, their skeletal remains, and sediments that occur on the seafloor of the world’s oceans. In the Straits of Florida, these features have been referred to as lithoherms. We use digital, side-scan sonar data collected from the submarine NR-1 from an 10.9 km^2 area at ~650 m water depth to characterize quantitatively aspects of the morphology of 216. Their lengths, widths, heights, areas, orientations and concentration on the seafloor have been determined. Analysis indicates that the outlines of relatively small to medium sized lithoherms can be effectively described with a piriform function. This shape is less applicable to the largest lithoherms because they are aggregates of smaller lithoherms. Nearly all of the lithoherms studied have axes parallel to the northward flowing Florida Current, and the heads of 80 percent of these features face into the current. The shape and orientation of the lithoherms, and evidence of megaripples and scouring in the sonar data suggest that these features are formed by a unidirectional current. Following an extensive investigation of over 200 lithoherms via side-scan sonar imagery and direct observation, we have developed a qualitative model for the formation of the lithoherm type of deep-water coral mounds in the Straits of Florida. Lithoherm formation can be characterized by four main stages of development: nucleating, juvenile, mature singular, and fused. Fused lithoherms can form via transverse and/or longitudinal accretion, however, transverse accretion at the head of the mound is likely the most efficient mechanism. A comparison of lithoherm spatial relationship to local bathymetry agrees with previous observations of deep-water coral mound formations along the levied margins of density flow scour channels.

This research focuses on diterpenoids from Taiwanese soft corals Xenia umbellata Lamarck, Junceella juncea Pallas and Junceella fragilis Ridley. Twelve diterpenoids in addition to one sesequiterpenoid were isolated. Our investigation of the soft coral X. umbellata Lamarck afforded five natural products, including two new xenicane diterpenes, xenibelatols A-B (1-2), together with two known xenicane diterpenes, 7,8-oxido- isoxeniolide (3), 9-hydroxyxeniolide-F (4), and a cadinene sesequiterpene, xenitorin A (5). Chemical investigation of the gorgonian J. juncea Pallas, has resulted in isolation of a new briarane diterpene, juncenolide H (6). Continuing our investigation of the gorgonian J. fragilis Ridley, we isolated seven briarane diterpenes, including four new briaranes, flajunolides A-D (7-10), along with three known briaranes, junceellolide E (11), umbraculolide A (12), 11This research focuses on diterpenoids from Taiwanese soft corals Xenia umbellata Lamarck, Junceella juncea Pallas and Junceella fragilis Ridley. Twelve diterpenoids in addition to one sesequiterpenoid were isolated. Our investigation of the soft coral X. umbellata Lamarck afforded five natural products, including two new xenicane diterpenes, xenibelatols A-B (1-2), together with two known xenicane diterpenes, 7,8-oxido- isoxeniolide (3), 9-hydroxyxeniolide-F (4), and a cadinene sesequiterpene, xenitorin A (5). Chemical investigation of the gorgonian J. juncea Pallas, has resulted in isolation of a new briarane diterpene, juncenolide H (6). Continuing our investigation of the gorgonian J. fragilis Ridley, we isolated seven briarane diterpenes, including four new briaranes, flajunolides A-D (7-10), along with three known briaranes, junceellolide E (11), umbraculolide A (12),11£\, 20£\-epoxy-4-deacetoxy junceellolide D (13). The new compounds 1,2 and 6-10 possess xenicane-type and briarane-type skeletons respectively. The structures of new compounds were determined by 1D-, 2D-NMR spectroscopic analysis and physical methods such as optical rotation, UV, IR, mass spectrum, as well as comparison with the spectroscopic data reported for related compounds. Compounds 1 and 2 are geometric isomers of compounds 3 and 4. The only difference between them resides in the side chain. The geometry of the side chain influenced the relative spatial proximity of H-12, H-13, H-14 to the carbonyl at C-3, and consequently the extent to which these protons are subjected to the anisotropic effects of the carbonyl. Compounds 6-10 have acetyl groups at C-2, C-9, C-12, C-14 positions. Because of structural difference appears in briarane skeleton, they showed different chemical shifts in specific positions. Biological activity test¡Arevealed that compound 5 exhibited moderate cytotoxic activity against KB and WiDr cancer cell lines with ED50 values at 5.9 and 9.9 £gg/ml respectively.

An ecosystem always involves some implicit relations between habitat environment and inhabitants, whose reciprocal links can not be identified easily. Three sets of ecological monitoring data were analyzed in this study, including coral reef, algae (Thalassia hemprichii Aschers) in Kenting National Park, and Formosan landlocked salmon (Oncorhynchus masou formosanus) in the basin of Chichiawan Stream. Two data-oriented analysis techniques, which are Habitat Evaluation Procedure (HEP) and Group Method of Data Handling (GMDH), were applied to retrieve the embedded patterns from these data sets. Eventually, for each data set, a forecasting model based on the technique of combined forecasting were developed, which is to integrate the results from HEP and GMDH, for improving the overall modeling precision. The results of this study show that the data-oriented analyses, such as HEP and GMDH, are useful for finding valid information from the ecological data. Furthermore, the combined forecasting technique can really improve the performance of model prediction even for the ecological research. In order to acquire the most important habitat environmental factors affecting the inhabitants, this study also performed sensitivity analysis of the models. The contributions of this study are to identify effective knowledge for future ecological research and to provide reasonable suggestions for formulating conservation strategy.