Microbial challenges in recirculating aquaculture systems (RAS) for Atlantic salmon (Salmo salar) smolt and post-smolt

Aquaculture is one of the fastest growing food producing sectors today. Norway is the world’s largest producer of farmed Atlantic salmon (Salmo salar), and the aquaculture industry is an important contributor to value creation and employment nationally. The land based production phase in Norwegian salmon farming has the past years been extended to include post-smolt for an increasing number of farms. This is a consequence of problems with salmon lice and diseases in open net pens, escapes, pollution, industry public relations and economic aspects of regulations and concessions. Increased production time on land in recirculating aquaculture systems (RAS) with larger fish demand more comprehensive water treatment to maintain good water quality. Higher biomass and feeding generate more intensive organic loads and particles in RAS. Furthermore, prolonging the production on land to include post -smolt may involve introduction of salt water to the systems. Organic matter and salinity will affect the water treatment significantly, in particular nitrification and the microbial water quality. Bacteria are key players in the nutrient fluxes in RAS to maintain high water quality. The motivation for this thesis was to provide more knowledge on operation and rearing regimes in RAS for salmon smolt and post-smolt production, with a special focus on microbial challenges related to organic matter and salinity. Our first experiment evaluated the effects of enhanced particle removal with membrane filtration in RAS on concentrations of organic matter and its consequences for water quality and microbial conditions. This experiment was furthermore used to make a carbon and nitrogen mass balance. We evaluated the dynamics and fate of C and N input to RAS, and removal efficiencies of the water treatment, including a membrane, for C and N compounds. The results showed that the system with membrane filtration had higher microbial diversity, lower and shorter bacterial blooms and generally lower bacterial densities in the water than in the system without membrane filtration. The mass balance showed that membrane filtration reduced the fraction of input C and N ending up as particles in RAS. The membrane directly removed particles, reducing accumulation of C and N compounds which resulted in better water quality. The better physicochemical and microbial water quality in combination with higher temperatures led to better appetite of the fish and as a consequence, this system had less feed waste and better fish growth than the system without membrane filtration. High organic matter loadings did not impact the nitrification efficiency negatively due to total ammonia nitrogen (TAN) limitation. This implies that as long as TAN is limiting and there is sufficient oxygen concentrations in the biofilter, increased loadings of organic matter in post-smolt production with larger fish will not suppress nitrification. Membrane filtration has shown to be a suitable technology for removal of the smallest particles and bacteria in RAS to improve water quality. However, cost-benefit analyses with membrane filtration at different life stages during Atlantic salmon production remains to be done to determine economic feasibility for the fish farmers. Our second experiment studied how two different regimes for salinity increase in RAS affected the RAS microbiota, nitrification capacity and performance of fish. One regime was a gradual increase in salinity in a brackish water RAS with post-smolt, the other was a direct transfer of post-smolts from a low salinity brackish RAS to a high salinity/seawater RAS, both groups with subsequent transfer to sea. The results showed that salinity was a driver for bacterial succession in RAS water. This included a combination of physiological salinity adaptation processes and succession causing change in community structure and introduction of new species. We showed that it was possible to successfully increase the salinity in an operating RAS with fish without exceeding toxic concentrations of TAN and nitrite. We hypothesize this was due to the salinity history of the system and halotolerant nitrifying bacteria embedded in the biofilter biofilm. Whether one salinity adaptation strategy was better than the other in respect to the fish still remains unknown as there were no clear positive indications in either of the fish groups in the two salinity adaptation regimes both on land and at sea. The third experiment investigated the start-up of nitrifying biofilms in freshwater and brackish water MBBR biofilters. The development of the nitrifying community assembly in the biofilm and nitrification capacity were compared in the two reactors. We observed that after 60 days of start-up, the brackish water biofilm had half the nitrification capacity of the freshwater biofilm during stress-tests, with less diverse microbial communities and lower proportion of nitrifiers. However, low ammonia and nitrite concentrations with rapidly increasing nitrate concentrations indicated that complete nitrification was established in both reactors. The results suggest that nitrification developed in comparable time in brackish and freshwater, and brackish start-up can be a strategy for bioreactors with varying salinity, like in post -smolt production