Nitrate removal effectiveness of fluidized sulfur-based autotrophic denitrification biofilters for recirculating aquaculture systems

There is a need to develop practical methods to reduce nitrate–nitrogen loads from recirculating aquaculture systems to facilitate increased food protein production simultaneously with attainment of water quality goals. The most common wastewater denitrification treatment systems utilize methanol-fueled heterotrophs, but sulfur-based autotrophic denitrification may allow a shift away from potentially expensive carbon sources. The objective of this work was to assess the nitrate-reduction potential of fluidized sulfur-based biofilters for treatment of aquaculture wastewater. Three fluidized biofilters (height 3.9 m, diameter 0.31 m; operational volume 0.206 m³) were filled with sulfur particles (0.30 mm effective particle size; static bed depth approximately 0.9 m) and operated in triplicate mode (Phase I: 37–39% expansion; 3.2–3.3 min hydraulic retention time; 860–888 L/(m² min) hydraulic loading rate) and independently to achieve a range of hydraulic retention times (Phase II: 42–13% expansion; 3.2–4.8 min hydraulic retention time). During Phase I, despite only removing 1.57 ± 0.15 and 1.82 ± 0.32 mg NO₃–N/L each pass through the biofilter, removal rates were the highest reported for sulfur-based denitrification systems (0.71 ± 0.07 and 0.80 ± 0.15 g N removed/(L bioreactor-d)). Lower than expected sulfate production and alkalinity consumption indicated some of the nitrate removal was due to heterotrophic denitrification, and thus denitrification was mixotrophic. Microbial analysis indicated the presence of Thiobacillus denitrificans, a widely known autotrophic denitrifier, in addition to several heterotrophic denitrifiers. Phase II showed that longer retention times tended to result in more nitrate removal and sulfate production, but increasing the retention time through flow rate manipulation may create fluidization challenges for these sulfur particles.     

Temperature-dependent and surface specific formaldehyde degradation in submerged biofilters

This study investigated formaldehyde removal in submerged fixed media biofilters in commercial and pilot scale recirculation aquaculture systems. Steady removal of formaldehyde (F) was observed immediately after simulated therapeutic treatment in closed systems and complete removal occurred within 1–4 days depending on water temperature. Formaldehyde removal was dependent on available biofilter surface area, and comparable rates of surface specific removal (SSR) were observed in two different systems. SSR was positively correlated to temperature (Q₁₀ = 3.4) with estimates of 2.1 mg F/(m² h) at 5.7 °C to 6.5 ± 0.2 mg F/(m² h) at 14.5 °C. The estimates for SSR of formaldehyde can be used to predict actual treatment and effluent concentration with more accuracy. Furthermore, the results allow calculation on biofilter removal capacity of formaldehyde, applicable for developing biofilters ensuring sufficient formaldehyde removal in effluent water.     

Degradation of urea,ammonia and nitrite in moving bed biofilters operated at different feed loadings

This study investigated how removal rates of urea, ammonia, and nitrite in laboratory scale moving bed biofilters were affected by long-term feed loading. To generate different loadings, five identical freshwater flow-through systems (100 l/h) with rainbow trout (Oncorhynchus mykiss) were fed increasing fixed rations of a commercial diet. The filtered effluent from each system was lead through a moving bed biofilter installed end-of-pipe. After an acclimatization period of four months, the moving bed biofilters were spiked separately with urea, ammonia and nitrite in batch mode in three successive trials to investigate degradation kinetics. Results showed that urea, in addition to ammonia and nitrite, was degraded although the substrate limited/dependent removal rate of urea (first order kinetic) was lower than that of ammonia and nitrite. Degradation of urea could be described as first order kinetics below 2.5 mg N/l. Degradation of total ammonia nitrogen (TAN) and nitrite was substrate independent (zero order kinetic) above 2 mg N/l and subsequently substrate dependent as substrate concentrations in the bulk water declined. The transition zone from zero to first order degradation was elevated with increase in long-term biofilter loading. For ammonia and nitrite, a significant increase in the zero order removal rate constants related to long-term loading were observed up to a long-term feed loading of 207 g/d, corresponding to 69 g feed/m² filter media/d and an TAN + urea-N concentration of 2.70 mg N/l. Long-term feed loading had no obvious effect on first order removal rate constants of any of the three nitrogenous compounds. Degradation of urea resulted in generation of ammonia demonstrating that urea degradation contributes to the ongoing nitrification activity in aquaculture biofilters. For all three types of spiking (urea, ammonia and nitrite) accumulation of nitrate was observed in the moving bed biofilters, sustaining that nitrification had occurred.     

Structure optimization of CycloBio fluidized sand biofilters based on numerical simulation

To improve the removal efficiency for dissolved wastes within CycloBio (CB) fluidized sand biofilters (FSBs) in recirculating aquaculture systems, we investigated their structural design and optimization using computational fluid dynamics (CFD) modeling tools, an orthogonal test method, and experimental verification. Results showed that the effects of structural parameters on bed expansion from large to small were: cone height, cone diameter and slot width. The best combination was: cone height 60 mm, cone diameter 165 mm, and slot width 1.0 mm. The solid phase was well distributed not only in the radial direction, but also in the axial direction in the optimized CB FSB. The bed expansion (40%–120%) was increased about 13%. Energy savings were 21%–28% at the same bed expansion. When the optimized CB FSB was used to treat synthetic aquaculture wastewater, with three bed expansions and four levels of C/N, total ammonia nitrogen removal rate expressed per unit of expanded bed volume was high, from 629 to 881 g m⁻³ day⁻¹. All results indicated that the structure of the optimized CB FSB was more reasonable and that the combination of CFD simulation and the orthogonal test method could be successfully applied to structural optimization.     

End-of-pipe denitrification using RAS effluent waste streams: Effect of C/N-ratio and hydraulic retention time

Environmentally sustainable aquaculture development requires increased nitrogen removal from recirculating aquaculture systems (RAS). In this study, removed solids from a large commercial outdoor recirculated trout farm (1000 MT year⁻¹) were explored as an endogenous carbon source for denitrification. This was done by (1) a controlled laboratory experiment on anaerobic hydrolysis of the organic matter (from sludge cones, drumfilter, and biofilter back-wash) and (2) an on-site denitrification factorial experiment varying the soluble COD (COD_S)/NO₃-N ratio from 4 to 12 at hydraulic retention times (HRT) from 50 to 170 min in simple 5.5 m³ denitrification reactors installed at the trout farm.The lab-experiments showed that the major part of the readily biodegradable organic matter was hydrolyzed within 14 days, and the hydrolysis rate was fastest the first 24 h. Organic matter from the sludge cones generated 0.21 ± 0.01 g volatile fatty acids (VFA) g⁻¹ total volatile solids (TVS), and the VFAs constituted 75% of COD_S. Analogously, 1 g TVS from the drum filter generated 0.15 ± 0.01 g VFA, constituting 68% of the COD_S. Comparison of the laboratory hydrolysis experiments and results from the on-farm study revealed as a rough estimate that potentially 17–24% of the generated VFA was lost due to the current sludge management.Inlet water to the denitrification reactors ranged in NO₃-N concentration from 8.3 to 11.7 g m⁻³ and COD_S from 52.9 to 113.4 g m⁻³ (10.0 ± 1.2 °C). The highest NO₃-N removal rate obtained was at the intermediate treatments; 91.5–124.8 g N m⁻³_reactor d⁻¹. The effect of the C/N ratio depended on the HRT. At low HRT, the variation in C/N ratio had no significant effect on NO₃-N removal rate, contrary to the effect at the high HRT. The stoichiometric ratio of COD_S/NO₃-N was 6.0 ± 2.4, ranging from 4.4 (at the high HRT) to 9.3 (at the low HRT). A simple model of the denitrification reactor developed in AQUASIM showed congruence between modeled and measured data with minor exceptions. Furthermore, this study pointed to the versatility of the NO₃-N removal pathways expressed by the bacterial population in response to changes in the environmental conditions; from autotrophic anammox activity presumably present at low C/N to dissimilatory nitrate reduction to ammonia (DNRA) at high C/N, besides the predominate “normal” heterotrophic dissimilatory nitrate reduction (denitrification).     

Evaluating the chronic effects of nitrate on the health and performance of post-smolt Atlantic salmon Salmo salar in freshwater recirculation aquaculture systems

Commercial production of Atlantic salmon smolts, post-smolts, and market-size fish using land-based recirculation aquaculture systems (RAS) is expanding. RAS generally provide a nutrient-rich environment in which nitrate accumulates as an end-product of nitrification. An 8-month study was conducted to compare the long-term effects of “high” (99 ± 1 mg/L NO₃-N) versus “low” nitrate-nitrogen (10.0 ± 0.3 mg/L NO₃-N) on the health and performance of post-smolt Atlantic salmon cultured in replicate freshwater RAS. Equal numbers of salmon with an initial mean weight of 102 ± 1 g were stocked into six 9.5 m³ RAS. Three RAS were maintained with high NO₃-N via continuous dosing of sodium nitrate and three RAS were maintained with low NO₃-N resulting solely from nitrification. An average daily water exchange rate equivalent to 60% of the system volume limited the accumulation of water quality parameters other than nitrate. Atlantic salmon performance metrics (e.g. weight, length, condition factor, thermal growth coefficient, and feed conversion ratio) were not affected by 100 mg/L NO₃-N and cumulative survival was >99% for both treatments. No important differences were noted between treatments for whole blood gas, plasma chemistry, tissue histopathology, or fin quality parameters suggesting that fish health was unaffected by nitrate concentration. Abnormal swimming behaviors indicative of stress or reduced welfare were not observed. This research suggests that nitrate-nitrogen concentrations ≤ 100 mg/L do not affect post-smolt Atlantic salmon health or performance under the described conditions.     

Survey of large circular and octagonal tanks operated at Norwegian commercial smolt and post-smolt sites

A survey was conducted to determine the geometry, operating parameters, and other key features of large circular or octagonal culture tanks used to produce Atlantic salmon smolt and post-smolt at six major Norwegian Atlantic salmon production companies. A total of 55 large tanks were reported at seven land-based hatchery locations, i.e., averaging 7.9 (range of 4–12) large tanks per land-based site. In addition, one 21,000 m³ floating fiberglass tank in sea was reported. Culture volume ranged from 500 to 1300 m³ for each land-based tank. Most tanks were circular, but one site used octagonal tanks. Land-based tank diameters ranged from 14.5 to 20 m diameter, whereas the floating tank was 40 m diameter. Maximum tank depths ranged from 3.5 to 4.5 m at land-based facilities, which produced diameter-to-average-depth ratios of 3.6:1 to 5.5:1 m:m. The floating tank was much deeper at 20 m, with a diameter-to-average-depth ratio of only 2.4:1 m:m. All land-based tanks had floors sloping at 4.0–6.5% toward the tank center and various pipe configurations that penetrated the culture tank water volume at tank center. These pipes and sloping floors were used to reduce labor when removing dead fish and harvesting fish.Maximum flow ranged from 3 to 19 m³/min per land-based tank, with 400 m³/min at the floating tank, but tank flow was adjustable at most facilities. Land-based tanks were flushed at a mean hydraulic retention time (HRT) of 35–170 min. Maximum feed load on each land-based tank ranged from 525 to 850 kg/day, but the floating tank reached 3700 kg/day. Almost half of the large tanks reported in this survey were installed or renovated since 2013, including the three tank systems with the highest flow rate per tank (greater than 17.6 m³/min). These more recent tanks were operated at more rapid tank HRT’s, i.e., from 34.8 to 52.5 min, than the 67–170 min HRT typical of the large tanks built before 2013. In addition, flow per unit of feed load in land-based tanks that began operating before 2010 were lower (19–30 m³ flow/kg feed) than in tanks that began operating later (33–40 m³ flow/kg feed). In comparison, the floating tank operates at a maximum daily tank flow to feed load of 160 m³ flow/kg feed, which is the least intensive of all tanks surveyed. Survey results suggest that the recently built tanks have been designed to operate at a reduced metabolic loading per unit of flow, a tendency that would improve water quality throughout the culture tank, all else equal. This trend is possible due to the ever increasing application of water recirculating systems.     

Comparing clear-water RAS and biofloc systems: Shrimp (Litopenaeus vannamei) production,water quality,and biofloc nutritional contributions estimated using stable isotopes

Indoor shrimp aquaculture systems can be used to produce fresh, never-frozen, quality shrimp near metropolitan seafood markets regardless of season and climate. However, questions still remain regarding what type of production system is best suited to maximize indoor production. In this project, two types of systems were compared: clear-water (CW) RAS and biofloc (BF) systems. Three, 1.36 m³ tanks were assigned to each of the two treatments; CW tanks had external settling chambers, two foam fractionators, and external biofilters, all operated continuously. BF tanks had settling chambers and one foam fractionator which were operated as needed to control solids accumulation. Shrimp weighing 0.42 g were stocked in all tanks at 250 m⁻³ and grown for 55 days. Ammonia and pH levels were significantly (P < 0.05) higher in the CW treatment, while nitrite, nitrate, and turbidity were all significantly higher in the BF treatment, although all parameters remained within acceptable ranges for shrimp growth. Shrimp mean harvest weight was significantly higher, biomass (kg m⁻³) was significantly greater, and FCR was significantly lower in the CW treatment; there were no significant differences in survival between treatments. Isotope levels indicated that shrimp in the BF treatment obtained a portion of the C (18-60%) and N (1-18%) in their tissues from biofloc material; however, this effect did not positively influence production in that treatment. By nearly eliminating solids from the water and using an external biofilter, substantially better water quality was maintained in the CW systems, which may have been a major contributor to the improved shrimp production in that treatment.     

End-of-pipe single-sludge denitrification in pilot-scale recirculating aquaculture systems

A step toward environmental sustainability of recirculat aquaculture systems (RAS) is implementation of single-sludge denitrification, a process eliminating nitrate from the aqueous environment while reducing the organic matter discharge simultaneously. Two 1700 L pilot-scale RAS systems each with a 85 L denitrification (DN) reactor treating discharged water and hydrolyzed solid waste were setup to test the kinetics of nitrate and COD removal. Nitrate removal and COD reduction efficiency was measured at two different DN-reactor sludge ages (high θ_X: 33–42 days and low θ_X: 17–23 days). Nitrate and total N (NO₃⁻ + NO₂⁻ + NH₄⁺) removal of the treated effluent water ranged from 73–99% and 60–95% during the periods, respectively, corresponding to an overall maximum RAS nitrate removal of approximately 75%. The specific nitrate removal rate increased from 17 to 23 mg NO₃⁻-N (g TVS d)⁻¹ and the maximal potential DN rate (measured at laboratory ideal conditions) increased correspondingly from 64–68 mg NO₃⁻-N (g TVS d)⁻¹ to 247–294 mg NO₃⁻-N (g TVS d)⁻¹ at high and low θ_X, respectively. Quantification of denitrifiers in the DN-reactors by qPCR showed only minor differences upon the altered sludge removal practice. The hydrolysis unit improved the biodegradability of the solid waste by increasing volatile fatty acid COD content 74–76%. COD reductions in the DN-reactors were 64–70%. In conclusion, this study showed that single-sludge denitrification was a feasible way to reduce nitrate discharge from RAS, and higher DN rates were induced at lower sludge age/increased sludge removal regime. Improved control and optimization of reactor DN-activity may be achieved by further modifying reactor design and management scheme as indicated by the variation in and between the two DN-reactors.     

Nitrogen waste from rainbow trout (Oncorhynchus mykiss) with particular focus on urea

Particulate and dissolved nitrogen (N) waste components are removed in recirculating aquaculture systems (RAS) using different cleaning technologies, and to dimension and optimize their removal efficiency requires that the expected daily load of the different waste forms can be estimated. Using a laboratory, mass-balance approach, the current study examined the effects of commercially applied feeding levels on the loading of different N waste forms, including daily fluctuations in dissolved total nitrogen (TN), total ammonia nitrogen (TAN), urea-N, and non-characterized, dissolved N deriving from juvenile rainbow trout (Oncorhynchus mykiss). In addition, the study examined whether there was a removal of urea-N across a moving bed biofilter operated as end-of-pipe under commercial conditions. The laboratory, mass-balance study showed that there were no effects of feeding levels (1.3, 1.5 or 1.7% of the biomass per day ) on the excretion of dissolved N components, which constituted the majority of total N waste (>81.6% on average). The excretion of urea-N and non-characterized, dissolved N components constituted 12–13% and 9–11%, respectively of dissolved TN. The excretion of urea-N was largely constant and independent of the daily feeding practice, whereas that of non-characterized N appeared to reflect the daily feeding activity, following the trends in TN and TAN. The time limited feeding regime applied in the laboratory study resulted in a pulse in the excretion of TAN that a biofilter may be unable to fully level out, potentially resulting in unnoticed, critical water quality conditions in intensive RAS during certain times of the day. Particulate N waste constituted a minor fraction of total N waste (<18.4% on average), and the actual loading depended on the digestibility of dietary protein/nitrogen. Results from the commercially operated, nitrifying biofilter showed that urea-N was removed at a rate of 0.014 g N m² day⁻¹. Compared to the removal of TAN (0.208 g N m² day⁻¹), the moving bed biofilter was 1.07 times more active in removing dissolved N than immediately expected when only considering TAN.     

Performance of sequencing microbead biofilters in a recirculating aquaculture system

Biological filtration, or biofiltration, is the key technology in recirculating aquaculture systems. Sequencing microbead biofilters, in which the media maintains a continuous up-and-down movement, are based on traditional microbead filters but offer superior filtration properties. The performance characteristics of a sequencing microbead biofilter installed in a recirculating aquaculture system for rearing Barcoo perch at 29 ± 1 °C were examined. The total ammonia-nitrogen (TAN) concentrations and the nitrite-nitrogen concentrations during a 52-day culture period were maintained blow 1.6 mg/L and 0.9 mg/L. In order to ensure efficient biofiltration, the optimal actual application of hydraulic retention time was determined to be approximately 3–5 min. The water flow produced by the reciprocating motion of the media served to wash away suspended solids, ensuring the occurrence of optimal nitrification processes. Additionally, the reciprocating motion of the media enhanced ammonia treatment efficiency significantly by improving the transport of nutrients and nitrification activity. Compared to a static situation the ammonia removal rate increased by 27% based on the application of up-and-down reciprocating movement. The biofilm on the microbead forms as a compact, complex, and homogeneous structure, consisting of numerous microscopic thin sheets. Additionally, a multitude of pores, interstitial voids, and vertical channels were widely observed to convey obviously advantageous properties in support of fluid passage, thus enhancing mass transfer and ultimately contributing to biofiltration effectiveness. The optimum biofilm thickness for providing efficient biofiltration was determined to be approximately 70 μm for this filter.     

The “self cleaning inherent gas denitrification-reactor” for nitrate elimination in RAS for pike perch (Sander lucioperca) production

Denitrification reactors have proven their functionality in commercial recirculation aquaculture systems (RAS). Nevertheless, clogging occurs due to the low hydraulic loads necessary to accomplish anoxic conditions for a successful denitrification process in RAS, which hampers the adjustment of stable working conditions within fixed bed denitrification reactors. Reactors working on the basis of activated sludge demand careful hydraulic control and/or complex configurations for sludge retention.To develop a low-maintenance denitrification reactor, an enclosed moving bed filter, driven by recirculation of the inherent, oxygen poor gas was designed. A Self cleaning Inherent gas Denitrification reactor (SID-reactor) of 0.65 m³, which offered a moving bed volume of 0.39 m³ was connected with a RAS of semi-industrial scale for pike perch (Sander lucioperca) production. This species indicates suboptimal environmental conditions (as e.g. NO₃-N concentrations above approximately 68 mg l⁻¹) by prompt reduction of the feed intake. In different experimental series, the SID-reactor was operated with denatured ethanol, methanol, acetic acid or glycerin as carbon sources and changing operational modes.Clogging was prevented by a 40 second inherent gas recirculation twice an hour, which provided continuous, maintenance free operation with marginal energy demand. With inlet (RAS) and outlet NO₃-N concentrations in the range of 49 mg l⁻¹ and 12 mg l⁻¹, mean denitrification rates of 199 g to 235 g NO₃-N per m³ moving bed volume and day were determined for all tested carbon sources. Negative effects on the feed intake of the reared pike perch were detected with all carbon sources except methanol. Changing the mode of operation to continuous circulation of the filter bed at inlet NO₃-N concentrations of 26 mg l⁻¹, the denitrification performance reached 451 g NO₃-N per m³ moving bed volume and day. The SID-reactor allowed for the reduction of freshwater exchange in the pike perch RAS from 600 l to 70 l (−88%) and the sodium bicarbonate buffer from 182 g to 31 g (−83%) per kg of administered food. The easy and reliable operation of the SID-reactor could help to establish controlled denitrification as a routine purification step in RAS.     

Use of vegetated drainage ditches and low-grade weirs for aquaculture effluent mitigation: II. Suspended sediment

Total suspended solids are a priority pollutant under the Clean Water Act and a point of concern for aquaculture facilities. The use of ubiquitous vegetated ditches on the aquaculture landscape may serve as an environmentally and economically sustainable practice for reducing suspended sediment contributions to downstream environments. This study assessed effects of consecutive low-grade weirs on suspended solids retention and settling rates of aquaculture pond effluent in a single drainage ditch. Two control and nine treatment discharges were conducted in September and October 2012 at the Mississippi State University South Farm Aquaculture Facility. All discharges decreased total and volatile suspended solid loads. Total suspended solids were decreased 72–94%, with a significant removal rate of 0.02 ± 0.01 mg L⁻¹ min⁻¹ in both control (F = 6.12, P < 0.001) and treatment discharges (F = 16.02, P < 0.001). Volatile suspended solids comprised 2–80% of total suspended solids and had a significant removal rate of 0.02 ± 0.001 mg L⁻¹ min⁻¹ in both control (F = 10.46, P < 0.001) and treatment discharges (F = 6.28, P < 0.001). There was no significant difference in overall settling rates between control and treatment discharges; however, prior to weir 1, both total and volatile suspended solid concentrations increased in control discharges. Treatment discharges decreased both total and volatile suspended solids significantly (P < 0.001) prior to weir 1. Further analysis revealed flow rate to be a significant (P < 0.001) variable in total suspended solid removal while initial concentrations affected reduction rates of volatile suspended solids significantly (P < 0.001). These results suggest that the use of low-grade weirs could be a viable best management practice that easily integrates within the aquaculture landscape and creates hydraulic conditions conducive to sediment retention.     

A novel approach for ammonia removal from fresh-water recirculated aquaculture systems,comprising ion exchange and electrochemical regeneration

A new physico-chemical process for ammonia removal from fresh-water recirculated aquaculture systems (RASs) is introduced. The method is based on separating NH₄⁺ from RAS water through an ion-exchange resin, which is subsequently regenerated by simultaneous chemical desorption and indirect electrochemical ammonia oxidation. Approach advantages include (1) only slight temperature dependence and no dependence on bacterial predators and chemical toxins; (2) no startup period is required and the system can be switched on and off at will; and (3) the fish are grown in much lower bacterial concentration, making the potential for both disease and off-flavor, lower. A small pilot scale RAS was operated for 51 d for proving the concept. The system was stocked by 105 tilapia fish (initial weight 35.8 g). The fish, which were maintained at high TAN (total ammonia nitrogen) concentrations (10–23 mgN L⁻¹) and fish density of up to 20 kg m⁻³, grew at a rate identical to their established growth potential. NH_3(aq) concentrations in the fish tank were maintained lower than the assumed toxicity threshold (0.1 mgN L⁻¹) by operating the pond water at low pH (6.5–6.7). The low pH resulted in efficient CO₂ air stripping, and low resultant CO_2(aq) concentrations (<7 mg L⁻¹). Due to efficient solids removal, no nitrification was observed in the fish tank and measured nitrite and nitrate concentrations were very low. The system was operated successfully, first at 10% and then at 5% daily makeup water exchange rate. The normalized operational costs, calculated based on data derived from the pilot operation, amounted to 28.7 $ cent per kg fish feed. The volume of the proposed process was calculated to be ∼13 times smaller than that of a typical RAS biofilter. The results show the process to be highly feasible from both the operational and economical standpoints.     

Hydrodynamic characterization and performance evaluation of an aerobic three phase airlift fluidized bed reactor in a recirculation aquaculture system for Nile Tilapia production

The hydrodynamic characterization and the performance evaluation of an aerobic three phase fluidized bed reactor in wastewater fish culture treatment are presented in this report. The objective of this study was to evaluate the organic matter, nitrogen and phosphorous removal efficiency in a physical and biological wastewater treatment system of an intensive Nile Tilapia laboratory production with recirculation. The treatment system comprised of a conventional sedimentation basin operated at a hydraulic detention time HDT of 2.94 h and an aerobic three phase airlift fluidized bed reactor AAFBR operated at an 11.9 min HDT. Granular activated carbon was used as support media with density of 1.64 g/cm³ and effective size of 0.34 mm in an 80 g/L constant concentration. Mean removal efficiencies of BOD, COD, phosphorous, total ammonia nitrogen and total nitrogen were 47%, 77%, 38%, 27% and 24%, respectively. The evaluated system proved an effective alternative for water reuse in the recirculation system capable of maintaining water quality characteristics within the recommended values for fish farming and met the Brazilian standards for final effluent discharges with exception of phosphorous values.     

Effect of different C/N ratios and hydraulic retention times on denitrification in saline,recirculating aquaculture system effluents

Data on operation and performance of cost-effective solutions for end-of-pipe removal of nitrate from land-based saltwater recirculating aquaculture systems (RAS) are scarce but increasingly requested by the aquaculture industry. This study investigated the performance of a (semi)commercial-scale fixed-bed denitrification unit using single sludge for treating effluent from a commercial, saltwater RAS used for production of Atlantic salmon (Salmo salar). A fixed-bed denitrification reactor was fed continuously with 3-days hydrolyzed sludge from the commercial RAS, and was operated at different hydraulic retention times (HRTs; 1.82, 3.64, 5.46, or 7.28 h) or influent C/N ratios (3, 5, 7, or 10). Twenty-four h pooled samples were collected from the inflowing RAS water and the hydrolyzed sludge as well as from the denitrification reactor outlet, and samples were analyzed for nutrients and organic matter content.Nitrate removal rates increased consistently with decreasing HRT (from 64.3 ± 5.2–162.7 ± 22.0 g NO₃-N/m³/d within the HRTs tested) at non-limiting C/N ratios, while nitrate removal efficiencies decreased (from 99.6 ± 0.3–58.2 ± 8.9 %). With increasing influent C/N ratios at constant HRT (3.64 h), nitrate removal rates increased until the removal efficiency was close to 100 % and nitrate concentration in the denitrification reactor became rate-limiting. A maximum nitrate removal rate of 162.7 ± 2.0 g NO₃-N/m³/d was achieved at a HRT of 1.82 h and an influent C/N of 6.6 ± 0.5, while the most efficient use of hydrolyzed sludge (0.19 ± 0.02 g NO₃-N removed/g sCOD supplied) was obtained with a HRT of 3.64 h and a C/N ratio of 2.9. Removal rates of organic matter significantly and consistently increased with decreasing HRT and increasing C/N ratio. In addition, reducing HRT and increasing C/N ratios significantly improved removal of total phosphorus (TP) and PO₄-P.In conclusion, optimal management of the operating parameters (HRT and C/N ratio) in a single-sludge denitrification process can significantly reduce the discharge of nitrogen, organic matter, and phosphorous from land-based saltwater RAS and thus contribute to increased sustainability.     

Fixed-bed adsorption of Tricaine Methanesulfonate onto pyrolysed paper mill sludge

The removal of Tricaine Methanesulfonate from water in a fixed-bed column packed with pyrolysed paper mill sludge was evaluated for application in Recirculating Aquaculture Systems. Breakthrough adsorption curves were determined for this pharmaceutical, under different operating conditions. In a laboratory-scale approach, with synthetic effluent, a bed adsorption capacity of 125 mg g⁻¹ was obtained under a flux of 56 L min⁻¹ m⁻², with 400 mg L⁻¹ of MS-222 with 4.5 cm bed depth. The performance of the continuous adsorption in a column packed with biological paper mill sludge-based carbon, using real aquaculture wastewater as matrix, was not significantly affected. Then, scale-up was carried out and the bed depth service time model was applied and provided good predictions. Finally, the thermal regeneration of the exhausted adsorbent was assessed, which revealed satisfactory results for one cycle reusability after saturation; however, considering the nature of its precursor, the use of new adsorbent could be more profitable than its regeneration, functioning also as an additional option for the management of such wastes.     

Feed and treat: What to expect from commercial diets

Basic data describing the physical characteristics of fish fecal waste are important in the design of effective solid waste management in aquaculture, especially in land-based facilities such as recirculating aquacultural systems (RAS).This study describes the physical properties of feces from rainbow trout fed eight different commercially available and widely used diets in Germany. Additional data from an earlier but unpublished study pertaining to feces derived from two rather extreme all-vegetarian diets are also presented for consideration of the settling properties. The diets were tested on duplicate groups of 50 rainbow trout in a flow-through aquaculture system. The effects of the diets on the physical properties of fecal particles such as particle size distribution (PSD), modeled settling velocity and rheological character were examined and the effects of each diet on fish health, growth and feed utilization were determined. Specific growth rate (SGR) and feed conversion ratio (FCR) for the different diets ranged from 0.98% d⁻¹ ± 0.012% d⁻¹ to 1.39% d⁻¹ ± 0.012% d⁻¹ and 0.97 ± 0.017 to 1.61 ± 0.017 (mean ± S.E.), respectively. The density of presoaked feces was significantly lower than that of intestinal feces and ranged from 1.01013 ± 0.00692 g cm⁻³ to 1.04547 ± 0.00692 g cm⁻³ (mean ± S.E.). Stability data were in the range from 390.12 ± 29.4 Pa to 1214.79 ± 29.0 Pa for elastic modulus and from 62.12 ± 6.1 Pa s to 232.68 ± 6.0 Pa s for dynamic viscosity. Based on the stability and PSD data theoretical efficiencies for removal of fecal waste using a drum filter showed remarkable variation, ranging from 82.5 to 95.9% (60 μm gauze). Based on the same data, theoretical removal by a sedimentation basin with routinely using overflow rates of 0.057 cm s⁻¹ to 0.394 cm s⁻¹ ranged from 62.8 to 93.8%. Both fecal density and PSD have an exponential impact on settling performance. Increasing fecal density improves the removal efficiency of a sedimentation basin by about 20%, however sedimentation was seen to be a less robust and efficient removal technique than drum filtration. Sedimentation systems also experience additional problems with respect to leaching. Turbulence that was mimicked in this study reflects to an optimal fish farm, which means disintegrating effects are mainly caused by fish motion. If disintegrating units e.g. pumps are used, which are known to promote further particle breakdown the effects would be amplified.The results demonstrate the central importance of density of suspended solids in defining removal efficiencies and suggest that manipulation of fecal density might offer a new and effective means of managing and optimizing waste output from aquaculture operations. This study describes the basic properties of fecal wastes generated by commercial diets and can be used as a basis for further research.     

Start-up performance of a woodchip bioreactor operated end-of-pipe at a commercial fish farm—A case study

There is a need for simple, maintenance-free technologies for removing nitrogen (N) from aquaculture effluents. Denitrifying woodchip bioreactors have been used successfully to remove nitrate-N (NO₃-N) from ground and surface waters and may potentially be applied to dilute aquaculture effluents as well. Real-life applicability in commercial, outdoor fish farms including practical start-up issues such as e.g. time till stable performance and potential leaching are, however, unknown to the industry.This case study consequently investigated the temporal performance of a woodchip bioreactor (12.5 m³) during start-up. The bioreactor was operated end-of-pipe at a commercial, outdoor rainbow trout (Oncorhynchus mykiss) farm in Denmark operated at low recirculation intensity. Applying an empty bed contact time (EBCT) of 5 h, the specific objectives of the study were to resolve: i) how fast the bioreactor would start to remove NO₃-N; ii) how fast steady state was achieved; iii) which NO₃-N removal rates could be attained at the relatively low effluent temperature (∼8 °C) and iv) to which extent any concomitant leaching of phosphorous (P), ammonia or organic matter would occur.In- and outlet grab samples were obtained every 6 h until the bioreactor was in steady state (2 weeks) followed by weekly 24 h pooled samples for another 3 weeks (5 weeks in total). Additional grab samples were obtained from 9 sampling ports within the bioreactor on 3 consecutive days during steady state. Samples were analyzed for dissolved nutrients (total N, nitrate, nitrite, ammonium, total phosphorous, ortho-phosphorous, BOD₅ and COD). In addition, oxygen, temperature and pH were logged every 30 min while sampling and alkalinity were measured once a week.Removal of NO₃-N started immediately and remained stable at 7.06 ± 0.81 g NO₃-N/m³/d (n = 6) throughout the sampling period. Increased effluent NO₂-N concentrations (peaking at 1.14 mg NO₂-N/l after 4–5 days) were transiently observed during the initial 11 days. After that, the woodchip bioreactor was largely in steady state with respect to N-balances corroborated by a close match between filtered total-N (TN_diss) and NO₃-N removal rates. Measurements within the bed showed that the majority of the influent dissolved oxygen (DO) was consumed within the first part of the bioreactor and that NO₃-N removal thereafter proceeded gradually with distance within the bed. Leaching of non-structural, dissolved organic compounds were observed just after startup, causing a short-term (1 week) increase in effluent concentrations of COD, BOD₅, P and ammonium.Additional measurements carried out until 147 days after start-up showed that the woodchip bioreactor continued to remove TN_diss at an average removal rate of 7.81 ± 0.82 g N/m³/d, and that the initial leakage of P stopped altogether.In summary, the study demonstrated that woodchip bioreactors can effectively remove NO₃-N from dilute aquacultural effluents at low temperatures and commercial conditions and that stable performance is achieved within a few weeks.     

The effect of density on sea bass (Dicentrarchus labrax) performance in a tank-based recirculating system

Sea bass (Dicentrarchus labrax) (135 ± 4 g) were reared under tank-based recirculating aquaculture system for a 63-day period at four densities: 10, 40, 70, 100 kg m^?3. Fish performance, stress indicators (plasma cortisol, proteonemia plus other blood parameters—Na⁺, K⁺, glucose, pH, total CO₂^?) and water quality were monitored. At the end of the 63-day period, resistance to infection was also studied by a nodavirus challenge. A 25-day test was performed on fish from two extreme densities (10 and 100 kg m³) and one intermediate density (40 kg m³).With regards to the different density treatments, there was no significant difference between the daily feed intake (DFI) and the specific growth rate (SGR) up to a density of 70 kg m^?3. No significant difference was found between treatments concerning the feed conversion ratio (FCR) and the mortality rate. No density effect was observed on the fish stress level (plasma cortisol) or on sensitivity to the nodavirus challenge. Under these experimental rearing conditions, the density above 70 kg m^?3 has an impact on growth performance (DFI and SGR) indicators and also some blood parameters (CO₂) at the highest density tested (100 kg m^?3).This study suggests that a density up to 70 kg m^?3 has no influence on sea bass performance and welfare. At 100 kg m^?3, average specific growth rate was decreased by 14% without welfare deterioration according to the welfare indicators monitored.