New measures to improve water quality with protein skimming, biofiltration, and ozone (ProBiOzon)

This report describes the main results from the EMFAF project “New measures to improve water quality with protein skimming, biofiltration and ozone (ProBiOzon)”. The project was led by DTU Aqua and included public research organizations (DTU Aqua Hirtshals & Lyngby), private technology suppliers (OxyGuard Int. A/S), commercial fish farm (Danish Salmon and Nr. Vium model trout farm), and the Danish Aquaculture Organization.

The project focused on applied research with test and development in collaboration with the aquaculture sector, with experimental activities from spring 2022 to fall 2023. The research activities included three types of water treatment approaches: protein skimming, ozone treatment, and biofiltration. The activities combined treatment and removal efficiency on several water quality parameters with veterinarian investigations of fish used in the experiment. The experiments were done under controlled conditions and at two commercial RAS and included several prolonged experiments from 2 weeks up to 40 weeks. Main findings of the project work packages are listed below.

Protein skimming
Prolonged experiments with protein skimmers were conducted at two different commercial aquaculture facilities to assess removal efficiency and improve operation. At the saltwater RAS Danish Salmon, two full-scale protein skimmer models were tested and compared under commercial operation, with controlled ozone dosages and hydraulic retention time. Both models showed similar removal efficiencies based on single-pass water analysis. The hydraulic retention time did not affect removal efficiency, while increasing the ozone dosage significantly improved the removal efficiency of microbes and microparticles, with up to 60% removal of microbial activity in a single pass. An ozone dose of between 7 and 10 g of O₃ per kg of feed was determined to be ideal for this specific salmon RAS.

Another study focused on testing and developing a new concept: a sustainable passive protein skimmer (PPS) prototype designed for model trout farms. This prototype uses existing airlifts to generate foam. Three PPS prototypes were constructed and tested in different raceways at freshwater model trout farm, Nr. Vium Dambrug. Testing and sampling campaigns were conducted during winter and spring and demonstrated a large removal capacity of organic matter, microparticles, and microorganisms.

Extrapolation based on prototypes covering the area of one airlift showed that up to 70% of the available particulate organic matter (measured as BOD_5part) could be removed within a day. The removal efficiencies of microbial activity and microparticles were also substantial, reducing their abundance by more than half in the raceway. The current hurdle to implementing advanced filtration technologies is the cost. However, the novel PPS prototype uses the existing airlifts, eliminating additional operational expenses and allowing for low cost construction. This together with the very high removal efficiency makes PPS a potential key technology to address some of the current challenges in aquaculture.

Ozone treatment
To evaluate the impact of ozone treatment, a continuous 4-week field trial was conducted at a type 3 model trout farm. Ozonated water was compared with untreated water from the same fish farm. The trial aimed to document the effects of ozone treatment on aquaculture water and simultaneously test and verify its influence on several water quality parameters. Despite fluctuations in the water matrix over the 30-day experimental period, ozonation led to the overall improvements in water quality, without negatively affecting fish welfare. Ozone production was measurable, and a distinct ozone gradient was observed within the system; however, no ozone was detected in the fish tanks. Ozone had a significant effect on several key water quality parameters. It improved UVT and reduced turbidity, resulting in more transparent water. A substantial reduction of microbial activity and biomass was observed with a single pass in ozone-treated systems., Although TAN and nitrite concentrations were high and fluctuating during the trial, ozone effectively oxidized nitrite to the non-toxic nitrate. Due to the short duration of the treatment, the effects of ozone on COD and BOD₅ remained inconclusive.

Fish health and pathogens
The effect of foam fractionation (FF) treatment and ozone treatment on fish health was evaluated through a series of trials. First, a brief pilot study was set up by exposing 10 groups of 40 naïve rainbow trout fry to increasing concentration of organic waste material (foam fraction) for four days, followed by a four-day recovery period. Differences in feeding behaviour were observed in the two highest concentrations (1:6 and 1:3 dilution ratios). Importantly, the pathogen Flavobacterium psychrophilum was detected in the organic waste introduced to the fish. Following this pilot trial, a more comprehensive experimental trial was set up to couple organic waste exposure with FF devices, to analyse the changes in microbial water quality and fish health upon treatment. In parallel, a known bacterial pathogen (Yersinia ruckeri) was introduced to assess the ability of the FF treatment to modify transmission dynamics and concentrate the pathogen within the foam fraction. The trial was conducted in eight semi-closed RAS systems (30 L tanks; four conditions set in duplicates) over a 11-week period. The survival probability assessed using (Kaplan-Meier estimates was slightly improved in the treated tanks, both in the absence (p=0.055) and presence (p=0.066) of pathogen infection. Y. ruckeri was detected in higher numbers among co-habitants in untreated tanks compared to the those treated with FF. A similar trend was observed regarding F. psychrophilum, a pathogen that was introduced with organic waste equally in all tanks. Molecular analysis of water indicated that Y. ruckeri concentrations were approximately five times higher in untreated tanks compared to FF-treated ones, suggesting the ability of FF to lower the circulating pathogen loads. Furthermore, when comparing water and foam fraction from the same system, Y. ruckeri concentration were consistently higher in the foam fraction at the most time points, reinforcing FF’s potential to lower the amount of circulating pathogen in the systems.

The effect of water ozonation was assessed in the previously described experimental set-up (Oxyguard). Groups of rainbow trout (average weight 20 gr) were placed in four experimental tanks at a stocking density of 14 kg/m 3 of water. Conditions were set in duplicates to compare ozonation treatment to the untreated control systems. Over a one-month period, the observed survival was close to 100% in both conditions, with an average mortality of 1.2% in ozonated tanks and 2% in untreated tanks, and no statistically significant difference being observed (p=0.16, Kaplan-Meier survival probability estimation). Fish were sampled (N=10-14 per condition) at the start, midpoint and endpoint of the experiment. Growth was assessed by measuring weight and length, with no significant differences observed between the ozonated and untreated groups at any time point. Internal organs were examined for bacterial infections and compared with rainbow trout from the farm raceways connected to the experimental tanks. In the raceways, three known bacterial pathogens (F. psychrophilum, A. salmonicida, Y. ruckeri) were detected. In the experimental tanks, F. psychrophilum was found both in ozonated and untreated tanks; however, by the final sampling point, no pathogens were detected in fish from the ozonated tanks. Further analyses are ongoing to study the changes in the total microbial communities both in the water and on the host.

Biofiltration
A new biofiltration concept utilized compressible bioelements made of polyurethane, referred to as “French press biofilters,” was developed and tested. These biofilters were installed in a 20 m³ RAS with 250 kg trout and operated over an 18-week period. The biofilters were tested at two different surface loading rates (high and low hydraulic retention time) and removal efficiencies of dissolved N (TAN and nitrite), organic matter, bacterial load, UVT, and turbidity were assessed. The biofilters demonstrated high treatment efficiency, ease of use, and immediate applicability. Single-pass TAN removal efficiency reached up to 80%, averaging 70% under high flow conditions and 40% at low flow conditions. Substantial reductions of organic matter were consistently observed. When the biofilter elements were squeezed/compressed, significant amounts of organic matter could be easily and safely retrieved.

Following compression and drainage, the biofilter removal performance was either unaffected or slightly reduced, emphasizing that frequent backwashing can be implemented without compromising efficiency. The concept was validated under controlled conditions and is expected to have several practical applications use, e.g. during the start-up of RAS or as a new tool/treatment method to mitigate TAN and nitrite accumulation in RAS.

Nitrification performance during start-up in saltwater was tested in a comprehensive experimental setup. The study tested four different types of biofilter elements -PR Plast PP, PR Plast PE, Mutag PE bioelements, and Levapor polyurethane foam- installed in six biofilters connected to a commercial seawater salmon RAS. Regular spiking experiments were conducted to quantify TAN and nitrite removal under controlled conditions. Volumetric TAN and nitrite removal rates were investigated at both low and high substrate concentrations (1° and 0° order conditions) over a 40-week period. PP and PE bioelements were colonized very slowly during the first two months after start-up. However, their volumetric TAN removal rates (VTR) increased linearly thereafter, stabilizing at 300-350 g TAN/m³/d by the end of the 40-week trial. In contrast, foam bioelements showed substantial removal rates as early as one week into the trial, peaking after 12 weeks with TAN removal rates reaching up to 575 g/m³/day at 13°C. However, performance declined over time, and by the end of the trial, VTR for foam bioelements fell below those of the other materials tested. The study provides new information to the aquaculture industry and identify new research question to achieve and maintain high and stable biofilter performance.