Several papers have reported on the development of antibiotic resistance and implications for human medicine but fewer deal with environmental impacts of antibiotic use. Marine sea cage aquaculture in SE Asia is often established close to coral reef ecosystems. Large amounts of antibiotics are used in the cultivation of fish and lobster and hence released directly into the environment. This study investigates the antibiotic practices in sea cage farms producing fish and spiny lobster in Vietnam, mainly for the domestic market. There are approximately 3500 sea cage farms in Vietnam and we performed semi-structured interviews with 109 sea cage farmers asking them if they use antibiotics and if so; what sort/when/how often/how much. We found that the Vietnamese cage farmers are using antibiotics in an unstructured way, which seems to have little or no effect on the survival of the stock, or profit of the farm. The fact that the farmers live at their farm and use the sea next to the cages both for fishing and collecting filter-feeding bivalves for direct consumption, as well as a toilet, poses an additional risk for the spreading of human antibiotic resistant pathogens. Thirteen different antibiotics were found in the study. Eighty-two percentage of the lobster farmers and 28% of the fishfarmers used antibiotics. The average amounts used were over 5 kg per produced ton of lobster and about 0.6 kg per ton of fish, which is much higher than in other studies. Several antibiotic substances listed as critical and highly important for human medicine by WHO were used prophylactically and routinely with little control and enforcement of regulations. We tested and detected antibiotic resistance to Tetracycline, Vancomycin and Rifampicin in the coral associated bacteria Bacillus niabensis as far as 660m from fish farms with resistance decreasing with distance from the cage farms. The antibiotics are likely to have negative effects on the coral-symbiont relationship adding further risks to an already stressed environment.

Global marine seaweed aquaculture is growing rapidly. In Zanzibar, Tanzania, seaweed farming, primarily conducted by women, is the main coastal aquaculture activity. Many types of aquaculture are linked to a specific ecosystem (e.g. shrimp-mangrove), and understanding if such a coupling exists for seaweed farming important for further development. A prerequisite to understand if farming affects coastal habitats is the need to know where, and on which habitat, the farms are located. In this study, we investigated the habitat preferences of seaweed farmers by interviews, field observations and satellite imagery analysis. We found that the majority of the farms were distributed in a narrow corridor (380-600 m from shore) along the coast where water depth (x) over bar = 2 m) and tidal regime (+/- 2 m) allow for a suitable environment for both the algae and the farmers. Within this corridor, thus defined by depth, the water is deep enough for the algae not to be overexposed to sunlight but also sufficiently shallow for the women to access and work on the farms at low tide. The farmers accordingly expressed depth as the major limiting factor when choosing the site for their farms, and the preferred habitat type was seagrass beds. Most farms (92%) were partly located on seagrass meadows, but also other habitats, such as sand. The total area of the studied seaweed farms was 65.6 ha, with 39% of this being seagrass meadow, which is significantly more than the seagrass cover in the farming corridor. The farms also covered 43% sand; however, the interviews indicate that a substantial part of the sandy areas was, in fact, also recently covered by seagrasses. Our findings are relevant for improved management, conservation, and marine spatial planning, as we show where and on which habitats seaweed farms are preferably located. This information can be used to further investigate the ecological impact on the habitats and their associated fauna and in order to provide more effective management actions. Furthermore, this is much-needed baseline information for investigating the increased production of seaweed, i.e. if the habitat has any effect on the algae growth.

A majority of the sea cage farms in South East Asia are located close to coral reefs. This causes a conflict between conservation and food production since sea cage aquaculture has a number of negative impacts on coral reefs. The aim of this investigation was to assess the drivers causing the sea cage farmers to place their farms close to reefs and to examine some potential farming effects in detail i.e. usage of coral reef fish for grow out farming and feed. For some 3500 Vietnamese fish and lobster farms, we measured; the distance to the closest coastal city (proxy for infrastructure access), satellite derived Chl a (proxy for water quality), wind fetch, and the adjacent coastal slope and elevation. We also performed 159 semi-structured interviews with fish and lobster cage farmers from three regions in Vietnam. The interviews revealed that the choice of farming site is mainly determined by access to infrastructure, wind and wave shelter, and water quality. Although the farmers used coral reef services, e.g. coral reef derived seedlings, they were in general not aware of coral reef presence or did not find it important for selection of site. Both coral reefs and sea cage farms were found close to steep rocky coasts, which are favorable for corals, and provide sufficient depth for sea cages. Sea cages were always found on the leeward side of the coast where the wind fetch is low enough for the floating farms and their inhabitants. Most of the farms were located within 20 km from a coastal city confirming the importance of access to infrastructure. With few exceptions, sea cage farms were located in areas with good water quality, where also coral reefs are present. The study showed that several of the coral associated species groups farmed were dependent on wild caught seedlings and that 22% of the feed used at farms was trashfish of coral reef associated species. We consider the spatial correlation between sea cage farms and coral reefs as circumstantial and suggest that shared environmental preferences explain the farm distribution pattern, rather than access to ecosystem services provided by the nearby reef itself. We found no evidence that it is necessary for sea cage farms to be located near coral reefs and strongly recommend that sea cages are moved further away from coral reefs, but to areas still providing clear water, shelter and access to infrastructure.

Aquaculture is of growing importance in the global seafood production. The environmental impact of aquaculture will largely depend on the type of environment in which the aquaculture system is placed. Sometimes, due to the abiotic or biotic conditions of the seascape, certain aquaculture systems tend to be placed within or near specific ecosystems, a phenomenon that in this thesis is referred to as aquaculture system - ecosystem links. The exposed ecosystems can be more or less sensitive to the system specific impacts. Some links are known to be widespread and especially hazardous for the subjected ecosystem such as the one between the shrimp aquaculture and the mangrove forest ecosystem. The aim of this thesis was to identify and investigate causes and consequences of other spatial links between aquaculture and ecosystems in the tropical seascape.

Two different aquaculture system - ecosystem links were identified by using high resolution satellite maps and coastal habitat maps; the link between sea cage aquaculture and coral reefs, and the one between seaweed farms and seagrass beds. This was followed by interviews with the sea cage- and seaweed farmers to find the drivers behind the farm site selection. Many seaweed farmers actively choose to establish their farms on sea grass beds but sea cage farmers did not consider coral reefs when choosing location for their farms. The investigated environmental consequences of the spatial link between sea cage aquaculture and coral reefs were considerable both on the local coral reef structure, and coral associated bacterial community. Furthermore, coral reef associated fish are used as seedlings and feed on the farms, which likely alter the coral food web and lower the ecosystem resilience. Unregulated use of last resort antibiotics in both fish- and lobster farms were also found to be a wide spread practice within the sea cage aquaculture system, suggesting a high risk for development of antibiotic resistant bacteria. The effects of seaweed farms on seagrass beds were not studied in this thesis but have earlier been shown to be rather substantial within the borders of the farm but less so outside the farm.

Further, a nomenclature is presented to facilitate the discussion about production system - ecosystem links, which may also be used to be able to incorporate the landscape level within eco-certifying schemes or environmental risk assessments. Finally - increased awareness of the mechanisms that link specific aquaculture to specific habitats, would improve management practices and increase sustainability of an important and still growing food producing sector - the marine aquaculture.

In the South China and Java Seas, cage farming is a recent regional activity, which since the year 2000 has experienced an annual growth of 29%. The region holds the highest diversity of marine life, which is partly or completely dependent on coral reefs. The increasingly growing coastal human population in the area relies on ecosystem goods and services provided by the reefs that are threatened by anthropogenic activities. Sea cage farming is one of the stressors negatively impacting coral reefs by being point sources of nutrients and other effluents. To date no systematic information is available on the physical location of marine farms in relation to the coral reefs. Little is known about the distance where impact from the farms can be detected on nearby coral reefs. The present survey aimed to fill this gap by assessing to what extent marine cage farms in South East Asia are placed in the vicinity of the reefs and at which distance stress indicators from the farms are observed. We used Google Earth satellite images to investigate the extension and spatial distribution of sea cage aquaculture in relation to the presence of coral reefs. The stress indicators were locally assessed in Central Vietnam by recording turf algal overgrowth, coral mortality, live coral and branching coral cover at increasing distances from the farms. We found that 90% of sea cage farms throughout the region clustered closer than 5 km from coral reefs and 50% of them closer than 1 km from reefs. In Taiwan, 71% of the cages were located within 100m from a reef. This pattern is nonrandom and could not be explained by the natural distribution of coral reefs; only 5% of the Vietnamese coast harbors coral reefs, and sea cage farms are present in these areas only. This indicates that the farms require similar conditions as the reefs including clear and shallow waters and protection against storms and wave action. We found that turf algal overgrowth decreased at 287 m +/- 54 m, dead coral at 1446m +/- 154 m, live coral cover increased at 566 +/- 221 m and branching corals increased at 867 m +/- 140 m from the cage farms. We conclude that proximity to coral reefs should be considered when planning future developments of sea cage aquaculture, and recommend that distances of at least 1.5 km should be kept. Statement of relevance: Consider coral reefs when planning sea cage aquaculture site.