Accurate estimation of fish weight is essential for sustainable aquaculture; however, despite advances in sensing technologies, commercial aquaculture still relies primarily on manual sub-sampling, which may reduce estimation accuracy. Bioenergetics models provide a complementary approach for dynamic fish weight estimation, yet their parameters are often calibrated offline using non-site-specific data. This study presents the Bioenergetics–Extended Kalman Filter (BE-EKF). This real-time data assimilation framework simultaneously estimates fish weight and adaptively refines model parameters using periodic manual measurements. The framework integrates a discretized bioenergetics growth model with an augmented-state EKF, while numerical Jacobian approximations address the analytical challenges posed by nonlinear equations. The framework was validated on two groups of Nile tilapia (Oreochromis niloticus) in a recirculating aquaculture system using near-complete population measurements as reference data. Compared with literature-based bioenergetics models, BE-EKF reduced current estimation and future and full-cycle prediction errors by 20–60% under small sampling and by 50–80% under larger samples. A receding-horizon feeding optimization framework was further developed to evaluate the economic value of parameter fidelity. While data assimilation frameworks are typically evaluated only in terms of estimation accuracy, this study quantitatively evaluates how parameter fidelity affects objective-function performance and economic outcomes. Using EKF-calibrated and literature-based parameter sets as high- and low-fidelity models, respectively, low-fidelity optimization resulted in 10–40% lower profits, with further reductions under delayed parameter convergence—a critical limitation of the EKF. These findings establish parameter accuracy as a quantifiable economic variable and support data assimilation as a foundational component of precision aquaculture systems.
Recirculating aquaculture systems offer sustainable fish production but face challenges related to ammonia levels. Ammonia, particularly the un-ionized fraction, NH3, can negatively impact fish growth and health. Traditional ammonia estimation methods, including offline analyses and affordable, yet inaccurate, in-situ measurements, lack the ability to provide reliable real-time insights. Data assimilation combines simulation models and in-situ measurements to provide more accurate estimations. In this study, we demonstrate a novel approach using data assimilation to enhance real-time ammonia estimation in RAS. First, ammonia dynamics are described by forming equations that constitute a simulation model, based on the feeding amounts and the biofilter removal rate parameters. Then, an extended Kalman filter is presented and customized to integrate the simulation model and total ammonia nitrogen measurements for estimating NH3, NH4+, and biofilter parameters. We validated our method through synthetic and laboratory case studies and demonstrated its superior estimation capability as compared to in situ measurements or simulation models. Furthermore, improved ammonia estimation led to improved current and future fish weight estimations, which can be essential for reliable RAS management. The proposed approach facilitates wider adoption of DA in challenging estimations in aquaculture.
Hydrophobic carriers were examined for geosmin and 2-methylisoborneol removal from water derived from an aquaculture system. A combination of adsorption and biodegradation was found to underlie the removal of the off-flavor compounds. Adsorption of these compounds by the carriers was unaffected by the presence of organic matter in the water to be treated. A model based on adsorption/desorption and first-order degradation kinetics provided an accurate prediction for experimentally determined 2-methylisoborneol removal rates. Steady removal of geosmin and 2-methylisoborneol as well as nitrate reduction were observed during long-term operation of the plug-flow reactors with water derived from an aquaculture facility. Metagenomic analysis of the microbial community on the carriers during long-term operation of the reactors revealed a predominance of denitrifying bacteria. It was found that geosmin and 2-methylisoborneol led to statistically significant changes in the abundances of 21 contigs that contained genes involved in terpene degradation. This study shows that at low ambient concentrations of geosmin and 2-methylisoborneol in nitrate and organic-rich water, such as found in aquaculture systems, their biodegradation can be accomplished by terpene-degrading denitrifiers that develop on hydrophobic carriers used for filtration of the contaminated water.
The secondary metabolites geosmin and 2-methylisoborneol (MIB) are known to taint fish with an undesirable, earthy-muddy taste and odor. In an earlier study on a zero-discharge recirculating aquaculture system (RAS), it was found that geosmin and MIB were removed by microbial communities residing in sludge from the digestion basin of the system. In the present study, 16S amplicon sequencing was used to identify changes in relative abundances of bacterial taxa in geosmin and MIB-enriched crude sludge. The removal of geosmin and MIB by the sludge was accompanied by increased abundances of 12 operational taxonomic units (OTUs). The most prominent increase in abundances was recorded for OTUs affiliated with bacterial genera known to harbor denitrifiers. Among these were the Betaproteobacteria genera Thauera , which utilizes terpenes to fuel denitrification, and Comamonas , which was previously isolated from the digestion basin of the same system and is capable of growth on geosmin and MIB as sole carbon and energy sources. Thus far, denitrification has been associated with bacteria capable of utilizing terpenes other than geosmin and MIB. The significant increase in the abundance of denitrifying bacterial genera in sludge in which geosmin and MIB comprised only 0.06% of the total carbon content might indicate that such bacteria play a major role in the removal of these compounds in anoxic environments.
We studied the effect of rearing densities of Pacific white shrimp, Litopenaeus vannamei in three densities with three replicate treatments [T1: 0.4 million post-larvae (PL) ha−1, T2: 0.5 million PL ha−1, T3: 0.6 million PL ha−1] and water cutback approach on rearing environment, water use efficiency, water footprint and production performance. Conditional water exchange was carried out based on water quality parameters. Water quality suitability index was very good (7.5–9.0) up to 13th, 10th and 5th week of culture in T1, T2 and T3, respectively; which was attributed to rearing density, smaller-sized shrimp and low early feed input. Optimum rearing density of 50 PL m-2 (T2) led to total water use of 3.25 × 104 m3. It was seeming as a way to improve shrimp productivity (10.58 t ha−1 120 d−1), consumptive water use index (1.72 m3 kg-1 biomass), total water footprint (1229 m3 t−1 biomass) and net consumptive water productivity (USD 1.28 m-3). L. vannamei culture with low to moderate water exchange as in T2, helped uphold water quality suitable for the shrimp growth, improved water use efficiency (0.58 kg biomass m-3 water), minimized sediment load (45.3 m3 t-1 biomass), effluent outputs (0.63 × 104 m3), pumping cost (USD 30.1 t−1 biomass produced), and ratio of output value to the cost of cultivation (1.97). The findings and advancement in knowledge would offer the basis to augment shrimp rearing efforts and the water management approaches will help in preventing the production of waste and effluent while increasing water use efficiency and production performance.
The accumulation of geosmin and 2-methylisoborneol in culture water and fish is a common problem encountered in recirculating aquaculture systems (RAS). In these systems, like in other aquaculture systems, abatement of this problem is usually conducted by purging the fish with clean water prior to their marketing. In indoor RAS, mainly heterotrophic prokaryotes underlie the production of geosmin and MIB. While production of these off flavor compounds has been linked with the organic-rich parts of these systems, no further information on factors that promote the growth of the geosmin and MIB-producing microorganisms under these conditions is currently available. Thus far, geosmin and MIB removal from RAS has mainly been conducted by ozonation, albeit with limited success. Biodegradation of geosmin and MIB might serve as an additional method for removal of these compounds from RAS. Geosmin and MIB degradation has been detected in the latter systems, yet factors which promote the activity of geosmin and MIB degraders in these systems remain largely unknown. In the present review, the current knowledge on geosmin and MIB production and removal in RAS is presented. Emphasis is placed on the biodegradation of these off-flavor compounds, a process which has received little attention thus far.
In fluidized bed rnactors, applled for nitrate removal in an aquaculture unit, denitrificatlon of nitrate was found to result in an accumulation of nitrite. Isolation of denltrifylng bacteria from these reactors revealed that most of these isolates produced nitrite from nitrate under controlled laboratory. conditions. Two of the isolates, one with accumulated nitrite while respiring on nitrate, and the other not, were subject to further studies. It was found that nitrite accumulation in one of these isolates could be adequately explained by differences in the relative rates of nitrate and nitrite reduction. Inhibiting concentrations of nitrite were lower for the isolate which did not accumulate nitrite. Incubation of the nitrite-accumulating isolate with different carbon sources resulted ln differences in the relative nitrite accumulation (as compared with the runount of nitrate reduced). KEY\o.URDS: denitrlfylng bacteria, fluidized bed reactors 78
Microbial studies were done on a semi-closed recycling intensive fish culture system, equipped with an aerobic (blofllter) and an anaerobic {denitrifier) water treatment units. Total number of bacteria {TBN), the frequency of dividing bacteria cells (FDC) and the percentage of bacteria showing respiratory activity (as measured by INT reaction) were determined at nine sites in the culture system. The bacteria population of both the aerobic and anaerobic treatment units were also compared using scanning electron microscopy. TBN differed by orders of magnitude between sample sites in spite of intensive hydrolic mixing within the system (e.g. 106 bacteria ml-l in water flowing into the fish basin and 10 bacteria ml-l in the sand-water mixture within the anaerobic treatment unit). Dividing bacteria cells (FDC) were difficult to detect anywhere in the system, even where high loads of organic material should provide abundant bacterial nutrients. Water from the outflow of the aerobic treatment unit showed similar results of TBN (1 x 108 bacteria ml-l) and portions of !NT-active bacteria (12 13%) as did effluent from the denitrifying unit. HIT-active bacteria were most abundant inside the H. Sich and J. van Rijn biofilter (18% of TBN) and in fish basin effluent (19% of TBN). Microbial parameters varied only slightly during a 24 h cycle. In the two treatment units microbial colonization differed in both composition and quantity. Probable colonies of ammonia-oxydizing bacteria were detected within the biofilter. No such colonies were seen in the anaerobic unit. Nitrogen compounds (NH4-N, N02-N and 3 -N) were sampled simultaneously at three sites. Anmonia showed a periodicity corresponding to the expected metabolic activity of the fish, varying from 0.02 0.46 mg L-l in the morning to 0.54 2.45 mg L-l in the evening. During the 24 h cycle transformation of NH 4-N through N02-N to 3 -N was inefficient (1.0 3.5 mg L-l N02-N and 5.0 23.6 mg L -l No 3 -N). For their efficient removal, both the aerobic, but especially the anaerobic, treatment units should be optimized. Neither the additions of fish feed or fluctuations in ammonia levels had a noticeable effect on the total number of bacteria or the proportion of dividing or respiring bacteria cells. However, some specialized groups of microorganisms, may have been affected.
Recirculating aquaculture systems (RAS), offering many economic and fish husbandry benefits, are characterized by an accumulation of dissolved organic matter (DOM) and, specifically, humic substances (HS). As reported in a number of studies, HS may affect biological activity in both invertebrates and vertebrates. Given the accumulation of HS in RAS, it is therefore of great interest to characterize DOM and, specifically, its HS fraction in the RAS. The present study was aimed at characterizing long-term changes in fluorescent DOM composition in the culture water of RAS systems, which were operated in a novel, zero water exchange mode. Two such zero-discharge recirculating systems (ZDS) were examined: a freshwater system, stocked with hybrid tilapia (Oreochromis aureus x Oreochromis niloticus) and a marine system, stocked with gilthead seabream (Sparus aurata). Excitation-emission matrices (EEMs) of fluorescence, coupled with parallel factor analysis (PARAFAC), were used to characterize and quantify the different DOM components in the ZDS. In the culture water, one tryptophan-like and four HS-like components were identified. The fluorescence intensities of three of the HS-like components as well as the tryptophan-like component increased at comparable rates during ZDS operation while a much slower accumulation of these compounds was observed in a parallel operated, flow-through, freshwater aquarium. The ZDS examined in this study comprised a sludge digestion stage where a considerable accumulation of all fluorescent components was detected. A HS-like components and a tryptophan-like component in blood of tilapia from the freshwater ZDS were similar to components found in the culture water. Blood levels of both components were higher in fish cultured in the DOM-rich ZDS than in fish raised in the control, flow-through freshwater aquarium. Fluorescence of the HS-like component found in the fish blood increased also with time of ZDS operation. The finding that fish blood contains a HS-like fluorescent component may have important implications for the understanding of the physiological effects of HS in fish and the possible benefits of these substances in aquaculture.
Nitrate may reach surface and ground waters as a consequence of agricultural activity and discharge of domestic and industrial waste. Among the various methods used for nitrate removal, denitrification, a process in which nitrate is biologically reduced to elemental nitrogen, is relatively reliable and inexpensive as compared to other physical and chemical nitrate removal processes. Denitrification is generally conducted with biofilters in which bacteria are either immobilized to the surface of insoluble carriers or are entrapped within an immobilization matrix. We examined the use of non-toxic and biodegradable natural hydrocolloids for entrapment of denitrifying bacteria. Gel beads containing starch and alginate were used for this purpose. Three types of gel beads were examined: (1) wet gel beads; (2) porous wet gel beads; and (3) freeze-dried gel beads. With respect to nitrate removal, wet and dried gel beads showed similar removal capacities. Porous beads demonstrated an advantage over regular gel beads only during the first of the approximately 3-month incubation period. The viability of the immobilized bacteria was only slightly affected during prolonged refrigerated storage of the beads. Compared to freeze-dried beads, production costs of wet beads are significantly reduced, so it is anticipated that such carriers will eventually lead to a method that can be applied on an industrial scale.
When challenged with atypical Aeromonas salmonicida subsp. salmonicida, exposure of the common carp (Cyprinus carpio L.) to different humic-rich compounds resulted in a significant reduction in infection rates. Specifically, in fish exposed to (i) humic-rich water and sludge from a recirculating system, (ii) a synthetic humic acid, and (iii) a Leonardite-derived humic-rich extract, infection rates were reduced to 14.9%, 17.0% and 18.8%, respectively, as compared to a 46.8% infection rate in the control treatment. An additional set of experiments was performed to examine the effect of humic-rich components on the growth of the bacterial pathogen. Liquid culture medium supplemented with either humic-rich water from the recirculating system, the synthetic humic acid or the Leonardite humic-rich extract resulted in a growth reduction of 41.1%, 45.2% and 61.6%, respectively, as compared to the growth of the Aeromonas strain in medium devoid of humic substances. Finally, in a third set of experiments it was found that while the innate immune system of the carps was not affected by their exposure to humic-rich substances, their acquired immune system was affected. Fish, immunized against bovine serum albumin, displayed elevated antibody titres as compared to immunized carps which were not exposed to the various sources of humic substances.
The exposure of guppy fish (Poecilia reticulata), infected with the monogenea Gyrodactylus turnbulli and Dactylogyrus sp. to humic-rich culture water and feed, reduced both the infection prevalence (% of infected fish) and the infection intensity (parasites per fish) of the two parasites. Specifically, among fish exposed to: (a) humic-rich water and sludge from a recirculating system (RAS) and (b) synthetic humic acid (HA), infection prevalences of Gyrodactylus turnbulli were 17% and 25% respectively, as compared with an infection prevalence of 52% in the control group. The lower infection prevalence was accompanied by a significant reduction in the infection intensity: from 3.8 in the control group to 0.2 and 0.3 parasites per fish in the RAS and HA treated fish respectively. The infection prevalence and intensity of Dactylogyrus sp. were significantly lower (infection prevalence: 2.5%; infection intensity: 0.3 parasites/fish) in guppies exposed to RAS water and sludge than in the control group (infection prevalence: 50%; infection intensity: 0.8 parasites/fish).
Intensive pond cultivation of Gracilaria may achieve maximal yields by enrichment of seawater with ammonia and CO2. In the present study, we examined the use of organic waste material as a source of ammonia and inorganic carbon. A biofiltration system was constructed, consisting of several fermentation tanks and a fluidized bed reactor for the required bioconversion of organic wastes. Gracilaria conferta was cultured in a 30 m(2) pond from which seawater was circulated through the biofiltration system for a period of nine months. As compared to the filtered seawater used for water supply, the outlets of the fermentation tanks and the fluidized bed reactor showed significant increases in dissolved inorganic carbon and ammonia concentrations and significant decreases in oxygen concentrations and redox potentials. The most efficient organic waste compound tested had the highest C/N ratio. No significant differences were found between the Gracilaria yields of the control and biofiltration pond systems. These results might contribute a fundamental improvement in the economy of Gracilaria pond cultivation by water recycling through such a biofiltration system. Statement of relevance: The system might contribute a fundamental improvement in the economy of Gracilaria pond cultivation.
Off-flavor in fish poses a serious threat for the aquaculture industry. In the present study, removal of 2-methylisoborneol (MIB), an off-flavor causing compound, was found to be mediated by adsorption and bacterial degradation in sludge derived from an aquaculture system. A numerical model was developed which augmented Langmuir equations of kinetics of adsorption/desorption of MIB with first order degradation kinetics. When laboratory-scale reactors, containing sludge from the aquaculture system, were operated in a recirculating mode, MIB in solution was depleted to undetectable levels within 6 days in reactors with untreated sludge, while its depletion was incomplete in reactors with sterilized sludge. When operated in an open flow mode, removal of MIB was significantly faster in reactors with untreated sludge. Efficient MIB removal was evident under various conditions, including ambient MIB levels, flow velocities and sludge loads. When operated in an open flow mode, the model successfully predicted steady MIB removal rates with time. During steady state conditions, most of the MIB removal was found to be due to microbial degradation of the adsorbed MIB. Findings obtained in this study can be used in the design of reactors for removal of off-flavor compounds from recirculating aquaculture systems.
Background: Furan fatty acids (F-acids) are valuable minor fatty acids which are appraised for their protective role against lipid oxidation of polyunsaturated fatty acids (PUFAs). The most relevant dietary source for F-acids is fish with the predominant occurrence of up to five dimethyl- or monomethyl-substituted homologues. During the screening of fish from a zero discharge aquaculture (ZDA) system we noted the potential presence of unusual F-acids. Methods: We developed a method by gas chromatography with mass spectrometry operated in the selected ion monitoring mode for elucidation of the structures of the uncommon F-acids. Results: Carp from the ZDA system contained seven non-methylated F-acids with dominance of 8-(5-hexylfuran-2-yl)-octanoic acid (8F6). Non-methylated F-acids have never been detected before in fish. Subsequent analysis of other fish species and a batch of the fish feed confirmed the presence of non-methylated F-acids. Conclusions: F-acids in fish are derived from the feed. Our investigation indicates that more emphasis should be put on the F-acid concentrations in fish from aquaculture, which appears to depend on the quality of the fish feed.
Reuse of grey water (GW) enables to reduce fresh water consumption, but a treatment is required to prevent potential transmission and propagation of pathogenic organisms. This study presents results on the removal of pathogenic bacteria from GW as well as reduction of turbidity, TSS COD, and BOD by a novel treatment system. Compared to previous studied methods, three new elements are presented in the current treatment of GW: (1) A granulated complex of micelles of the organic cation octadecyltrimethylammonium (ODTMA) with montmorillonite was employed in filtration of GW. This complex was efficient in purifying GW due to its large surface area, positive charge and existence of hydrophobic domains. The granulated complex enabled flow when present exclusively in the filter; (2). A moving bed reactor for decomposition of part of the organic matter in the GW. This pretreatment stage, prior to the micelle-clay filter, was also efficient in removing pathogenic bacteria; (3) A regeneration stage of the micelle-clay filter conducted by passing either dilute solutions of Na-hypochlorite or HCl through the micelle-clay complex, or by heating the complex. Incubation of GW for either two weeks or one day in the pretreatment stage yielded a 10- and 7-fold enhancement in the volume filtered, which did not contain fecal coliforms, i.e., 300 and 210L for 40g of complex, respectively. The capacity of purified volume per gram of the complex increased further several-fold (>23L/g) for filters filled exclusively with granules. Regeneration of the complex in the filter further enhanced the capacity.