ERA-NET: Bluebio Microbial management in RAS for sustainable aquaculture production

Recirculating aquaculture systems (RAS) has many advantages compared to traditional flow-through systems and is now the preferred technology for land-based production of smolts. The reuse of water in RAS entails an extensive water treatment. Biofilters are a mandatory part of the treatment. Here, slow-growing microbes convert toxic ammonium to nitrate. Bacteria are thus essential for the chemical water quality. The concentration of nitrate is usually controlled by water exchange. A good microbial water quality must also be maintained to secure healthy environments for the fish. This involves the absence of pathogenic and opportunistic bacteria. However, as long as substrates are present, bacteria will grow in the RAS water, and disinfection of the water in the RAS loop will not be sufficient to prevent bacterial growth. Management of the microbial water quality is therefore not a trivial task. This project was part of the Bluebio project RASbiome, with DTU-Aqua (Denmark) and the University of Gent as academic partners. An important aim for the project was to explore the possibilities for improving the management of nitrogen (N) compounds in RAS by using alternatives to nitrification in the biofilter, specifically by using microbial processes leading to removal of N from the water: partial nitrification combined with anammox (PNA) or heterotrophic N assimilation. For the Norwegian research activities in the project, the overall aims were to: 1) Map the microbial communities and conversions in commercial RAS with smolt, 2) Adapt a biofilter based on PNA to conditions relevant to RAS, and thereafter to implement it in an experimental RAS, and 3) Assess the microbial water quality in RAS where we had implemented new approaches for microbial conversions of ammonia. We examined to identical RAS in a facility for smolt production. Water samples were analyzed for particles, carbon, N and phosphorus compounds and bacterial communities. For biofilm samples (trickling filter, moving-bed and fixed-bed biofilters) we characterized bacterial and archaeal communities. We found high relative abundances (20-30%) and high diversity of Nitrospira in the biofilm samples, including the trickling filter. The abundance of ammonia-oxidizing bacteria was low. However, characterization of the archaeal communities indicated that the archaeon Nitrosopumilus might a key ammonia oxidizer in the biofilters’ biofilm. To examine strategies to adapt the PNA process to low ammonia concentrations, we used moving bed bioreactors (MBBRs) with active PNA biofilm carriers. The carriers from a treatment process with very high concentrations of ammonia and organic matter (800 mg TAN/L og 3000 mg COD/L, respectively). We succeeded in obtaining a stable PNA process in a MBBR operated at only 10 mg TAN/L (TAN: total ammonia N). To the best of our knowledge, efficient N removal by PNA at such low TAN concentrations has not been reported previously. Characterization of biofilm communities revealed high relative abundances and very high diversities of Planctomycetes. This phylum encompasses all anammox (AMX) bacteria described so far. Surprisingly, the PNA was obtained when feeding the carriers with low TAN concentrations (10 mg/L) from the start. In another MBBR, where the TAN concentrations were gradually decreased, nitrification developed instead of PNA. Furthermore, we aimed at achieving pure AMX processes in two MBBRs operated under strictly anaerobic conditions. A gradual adaptation to low TAN concentrations again led to full nitrification and loss of AMX, even under strict anaerobic conditions, while in the reactor directly operated at low TAN concentrations, a stable AMX process was obtained. Next, we tried to implement the PNA MBBR into a simple recirculating system without fish, but this led to full nitrification and loss of the AMX activity, probably due to high O2 concentrations. A further development of the PNA process for applications in RAS needs to address the challenges related to management of the O2 concentrations. An important part of the project was to evaluate the microbial water quality in RAS where alternatives to nitrification had been implemented. At DTU Aqua, a new process for biological water treatment in RAS was developed. They applied biopellets to support microbial biofilm growth and demonstrated that ammonia could be removed, and potentially harvested, by heterotrophic N assimilation. We examined the effects on the microbial water quality by flow-cytometry. By staining the RNA content of bacterial cells, we could estimate the fraction of rapid-growing, potentially opportunistic bacterial populations. This proved to be an efficient approach, and we found that the fraction of opportunistic bacteria was low in a traditional RAS with a nitrifying biofilter, while it was significantly higher in RAS with a biofilter based on Het-N. This should be taken into consideration in any future development of the Het-N technology.

We obtained a detailed data set for two commercial RAS with salmon fry, sampled at three dates, including the concentrations of dissolved and particulate nitrogen, carbon, and phosphorus compounds, the abundance and size distribution of particles, and a detailed description of bacterial communities suspended in the RAS water, in addition to archaeal and bacterial biofilm communities in trickling filters, and moving-bed and fixed-bed biofilters. The results demonstrate key roles in nitrification for Nitrospira and ammonia-oxidizing archaea classified as Nitrosopumilus. Thus, our results emphasize the need for including Archaea when mapping nitrifying communities in RAS. We successfully adapted a microbial process based on partial nitrification and anammox (PNA) for conversion of ammonia to nitrogen gas at low ammonia concentrations (10 mg/L). This was obtained in a moving-bed bioreactor that was operated for over six months. We proved that an adaptation strategy based on gradual decrease of ammonia concentrations led to full nitrification and loss of the anammox activity, even under anaerobic conditions. However, we failed in implementing this treatment process in RAS. A main challenge is the high concentrations of dissolved oxygen, and this needs to be addressed in further developments. We consider the technology to have reached a TRL 2. Still, this development might be applicable to other treatment processes, yet to be identified. We also achieved a detailed characterization of the bacterial biofilm communities of the PNA and anammox reactors. This brought new insight about anammox communities, such as the extreme diversity of Plactomycetes (probably more than 2000 populations). We demonstrated that a simple and quick flow-cytometry-based approach could be used to assess the microbial water quality in aquaculture systems. This has a large applied potential, which will be further examined in new projects in collaboration with relevant industry partners.

In RASbiome, we aim at improving the sustainability of fish production in freshwater recirculating aquaculture systems (RAS). We will investigate microbiomes in RAS in relation to chemical and microbial water quality in commercial RAS, and implement two fundamentally distinct biological water treatment strategies to improve the management of nitrogen compounds in RAS. The first strategy involves anaerobic ammonia-oxidizing (anammox) bacteria in a process expected to remove nitrogen from the water without addition of organic carbon, at reduced energy and water consumption and CO2 production. The second strategy takes advantage of heterotrophic bacteria which assimilates nitrogen. This approach allows for harvesting nutrient rich sludge and is therefore compatible with recovery and recycling of nitrogen from in RAS water. Both biological treatment processes will capture dissolved N and lead to a reduced N-footprint. Furthermore, we will optimize microbial water quality by reducing the risk for blooms of opportunistic/pathogenic bacteria by implementing these new biological treatment approaches in well-considered treatment designs. The project is highly transdisciplinary, and involves experts from Belgium, Denmark, and Norway in fields like environmental engineering, biotechnology, microbiology, microbial ecology, and aquaculture. Industry partners, including two large commercial producers of salmon smolt and rainbow trout, and Krüger Kaldnes, supplier of RAS technology, play a crucial role in the project. We expect the proposed project to promote sustainable fish production by improving fish welfare and productivity due to stable and optimized chemical and microbiological water quality, by reducing environmental impact through nitrogen removal from discharged water, and by reducing operational costs. Thus, the project contributes in developing European aquaculture in a direction characterized by the 3R principle: reducing, reusing and recycling of waste material.