Dissolved oxygen , organic matter and nutrients in fish systems combined with bio-addition of friendly microorganisms Dissolved oxygen , organic matter and nutrients in fish systems combined with bio-addition of friendly microorganisms

Effluents from conventional intensive aquaculture, based on high water exchange, may strongly contribute to the acceleration of eutrophication processes, due to the significant nutrient load present in unconsumed foods and excreted waste. On the other hand, there are intensive and super-intensive systems for the cultivation of aquatic organisms that use water exchange associated with strong aeration and biological transformation of macronutrients. Such sustainable systems use the bio-addition of friendly microorganisms to increase the rate of degradation of macronutrients and maintain water quality. A mathematical adjustment was used in this study to verify the dynamics of organic matter, nitrogen and phosphate during the experiment period. A pulse dynamic at regular times occurred indistinctly for all indicators evaluated in this study, although the intensity of the pulses showed particularities among the different indicators of water quality. Although the behavior of regular pulses of the indicators occurred, this condition did not affect the growth of fish subjected to a simulated system of intensive production associated with the bio-addition of friendly microorganisms.


Introduction
The water in intensive production of fish, become more concentrated in pollutants due to the drainage procedure and also to the synergism existing between water quality-storage density -feed rate, because depending on the conditions, the tank environment can become toxic. According to Liu et al. (2016), Leal et al. (2018) and Coldebella et al. (2018), aquaculture effluents cause variations in water quality such as: decreased dissolved oxygen concentration, increased biological oxygen demand, nitrogen and phosphorus forms, increase in total suspended solids (TSS), and increased algal biomass (bloom), this scenario, the environmental degradation caused by aquaculture effluents should be one of the points observed carefully. Boyd (2003) suggests the application of a set of standards and procedures that constitute the so-called good management practices (GMP) as a way to improve the quality and reduce the volume of effluents.
Regarding the environmental impact of effluents, the importance of friendly or beneficial microorganisms in aquaculture tanks is emphasized by Liong (2015) and Adel and Dawood (2021).
The use of bio-addition is justified when the need for rapid degradation of polluting compounds is necessary. According to Divya et al. (2015), Manzoor (2016) and Deng et al. (2021) the use of an immobilized bacterial consortium represents a viable alternative for the in situ removal of nutrient residues in aquaculture systems due to the low cost of its application.
The aim of this study was evaluating the behavior of the organic matter and nutrients in effluents of aquaculture systems when submitted to bio-addition of beneficial and friendly microorganisms.

Materials and Methods
The experiment was conducted following the methodology of Verdegem and Beristain (2005) and Chen et al. (2020), which were modified for this study. And the study was conducted in Piracicaba São Paulo, in the fish farming sector, linked to department of non-ruminant animal husbandry of the Escola Superior de Agricultura "Luiz de Queiroz" -ESALQ/USP, in a greenhouse in order to control environmental variables. The lot of fish was acquired from a fish farm. The species used in this study was Oreochromis niloticus (L.) (Nile tilapia). The tilapia was chosen due to this species' availability in the needed size for the beginning of the study.

Conduct of the study
In the reception of the tilapia juveniles' batch, we adopted a sanitary protocol. The fish were stored in six 1 m -3 boxes with a useful volume of 0.8 m -3 and submitted to malachite green solution at the concentration recommended by Brown (2000) (10 mg 100 L -1 for three days), combined with the administration of the antibiotic oxytetracycline incorporated into the food portion (30 mg kg -1 of portion) for one week. The mean weight of tilapia individuals was approximately 120 g. Then, the fish were then transferred and permanently stored in the experimental tanks. Eighteen 1.0 m -3 polyethylene plastic boxes were used, with a useful volume of 0.8 m 3 tank -1 . The experimental tanks were individually equipped with constant aeration through air diffusers kept by a mechanical blower, and the oxygen rate was kept close to saturation. The system was provided with an individualized system of water inlet and outlet, supplied by a spring.
The drainage system was composed of PVC pipes placed internally to the tanks that allowed the flow of wastewater. The water temperature was kept at ± 25 ºC.
For each experimental tank, 40 juveniles.m -3 were stored with initial mean weight of 120 g. Therefore, with a storage density for the useful volume of tanks corresponding to 32 fish m -3 . tank -1 , which is equivalent to an initial density of 0.025 fish.m -3 and a biomass of 4.8 kg.m -3 tank -1 . Commercial portion was fed to the fish, which was skinned and extruded with a content of 32 % crude protein. The food was offered during seven days of the week, twice a day throughout the experiment period. Food was offered to fish ad libtum up to satiety.
Food intake data were collected weekly during the experiment (weighing of the consumed food). From the weighing data, the following information on the zootechnical performance of fish was obtained: initial biomass; final biomass; initial mean weight; final mean weight; initial length; final length; weight gain ( = − ); feed conversion (amount of food/weight gain); condition factor ( = ); feed efficiency index ( = ) and survival rate (%).
The hydraulic retention time (HRT) of the tanks was 44.4 hours, the flow rate adopted during the test was 18 liters.h -1 .
During the whole test we did not use the siphon for waste and unconsumed food.

The microbial inoculum
The composition of the inoculum, which we call the code AQ+ was: cereal bran 75 g -1 ; salt 15 g; Protease 9.6 g -1 ; Amylase 0.15 g; Lipase 0.07 g -1 ; hemicellulose 0.13 g -1 and Beta-glucanase 0.05 g -1 . According to the supplier, the bacteria concentration present in the inoculum was 1. The pipetted volumes of the stock solution for composition of the concentrations of the treatments were: T1 (160 µL -1 ), T2 (640 µL -1 ), T3 (1280 µL -1 ), T4 (2560 µL -1 ) and T5 (3840 µL -1 ), for the control, 160 μL -1 of the cereal filtrate was pipetted without the microbial inoculum, corresponding to 0.04 g L -1 . Then these volumes were diluted in 5 liters of water from the tanks and administered. The administration of the inoculums occurred from the 14 th day forward, being performed every two days.
For the control (T0), we pipetted 160 μL -1 of the cereal filtrate without the immobilized bacterial consortium. Water samples were collected every fifteen days during the experiment period (75 days).

Water quality
The water samples were collected directly from the tank outlet using a siphon. The flow system of the experimental tanks was composed of a 3/4'' PVC pipe covered by another 1" PVC pipe located inside each tank. The water samples for pH reading were performed by Orion Star potentiometer, the dissolved oxygen was determined by oxymeter. The electrical conductivity was determined by conductivimeter. The BOD was determined by ( = − ×120 ). The total dissolved solids (TDS) were determined by conductivimeter. Total ammonia was determined by colorimetry using the salicylic acid addition method, and spectrophotometer read (640 nm wavelength).
Nitrate was determined using the diazotization method, and the reading was made in a spectrophotometer (543 nm wavelength). Nitrate was determined by the cadmium reduction method and the reading was performed by spectrophotometer (220 nm wavelength).
The phosphate was determined by the persulfate acid digestion method), and the reading was performed in spectrophotometer (680 nm wavelength). The phosphate was determined by the ascorbic acid method, and the reading was performed in spectrophotometer. The chlorophyll "a" (phytoplankton biomass index) was determined by chlorophyll extraction by filtration, and the reading was performed in spectrophotometer (663 nm wavelength) (Golterman et al., 1978).

Data analysis
An analysis of variance (ANOVA) was performed between treatments for both water quality parameters (physicochemical and nutrients) as well as for zootechnical performance parameters. When significant difference was detected between treatments, mean tests (Student's t test) were performed. The adopted significance level was 5% (P < 0.05).

Adjustment models
The calibration of a mathematical adjustment consists of adapting the parameters of the equation to the real world, considering the physical, chemical and biological variables of this environment, adjusting the estimated data in the adjustment with the data observed in the field (Simonovic, 2002). For the calibration of the mathematical model presented in this work, 10 empirical models from the specialized literature were evaluated and the mathematical model that best correlated the experimental data was the model presented in Equation 1. It is important to emphasize that the mathematical model of Equation 1 was selected because it is the only one to satisfactorily correlate dissolved oxygen, organic matter (BOD) and nutrients (NH3, NO2, NO3 e PO4).

Method for adjusting the coefficients of the mathematical model
The methodology for obtaining the three adjustable parameters of Equation 1 consisted of using the experiment data of dissolved oxygen concentration, organic matter (BOD) and nutrients: ammonia, nitrite, nitrate and phosphate in the experiment system, as a function of the time for the decay of its rates. The coefficients a, b and c of Equation 1 were adjusted from the statistic software, using the mathematical method of adjustment of nonlinear functions of Levenberg-Marquardt.
The adjustment procedure consisted of an initial estimate for each of the three coefficients until adjusting them, using the highest value of the correlation coefficient R2 and the absolute mean deviation as a decision-making criterion for the better optimization of these coefficients.
Pulses of dissolved oxygen gradients were verified at regular intervals (12 days) from the beginning to the end of the experiment period among all treatments. including control (Figure 1). After a fall in the first 12-day interval. Oxygen dynamics were stable only in T5. which maintained the dissolved oxygen gradients in the intervals between 2.5 and 3.0 mg L -1 (Figure 1, Table 2).
Research, Society and Development, v. 11, n. 4, e26111427382, 2022 (CC BY 4.0) | ISSN 2525-3409 | DOI: http://dx.doi.org/10.33448/rsd-v11i4.27382 Apparently T3 presented the best behavior in the evolution of dissolved oxygen gradients over experiment time. It is important to mention that the biomass of fish stored in the different treatments started from an average weight of 4.8 kg m -3 and reached a final weight of 8.8 kg m -3 , a ratio twice higher than super-intensive aquaculture systems in Brazil. The hydraulic retention used in the system was 44.4 hours, which corresponds to a flow rate of 18 L h -1 . Regular pulses of dissolved oxygen at alternating times may be related to the nature of organic matter. The raw sewage load composed of unconsumed food and waste present in the experiment system presented a high fraction of carbon and assimilable nitrogen (high protein value) and when combined with temperature, accelerated the oxidation rate. The experimental condition allowed a higher coefficient of re-aeration due to both the ease of mixing and depth (Sperling, 2017), in addition to the greater surface turbulence induced by the mechanical aeration. Tables 3 and 4 show mathematical adjustment for the coefficients of dissolved oxygen, organic matter and nutrients.
Considering such conditions, the load of ammoniacal nitrogen present in the water column between treatments. For having presented a behavior in regular pulses of time intervals of 12 days, is similar to that verified by Nixon and Pilson (1983).
Altered water quality conditions promote increased resident microbial respiration. And also the benthic demand and nitrification can be significant enough to cause an imbalance of the dissolved oxygen rate in the system as a function of time. On the other hand, the re-aeration, verified over the experiment time, was probably more dependent on the mechanical aeration (that occurred in the system) than on the photosynthetic activity produced by algae and photosynthesizing bacteria.
To Yun et al. (2019) and Chen et al. (2020), the carbon load is particularly important due to its relationship with the BOD in the system. In this study, a low chlorophyll "a" gradient (0.94 mg m -3 ) was found (Figure 2), according to the modified Carlson classification, thus indicating a non-dominance of photosynthetic algae. On the other hand, Samer (2015) suggest that under aerobic decomposition, 50% of metabolized organic matter is converted into bacterial cells and, therefore, as Luo et al.
(2020) quotes, with the high load of intensive aquaculture systems, bacteria biomass is expected to be high. Source: Authors (2021).
The experiment system presented a dynamic for BOD in regular pulses from the 14 th day until the end of the experiment period. The kinetics of BOD followed a first-order reaction pattern, an expected behavior for the decay of organic matter in aquatic environments. This evolutionary behavior verified in the experiment system was compatible, as a function of time, to self-purification zones, even under conditions of a BOD with low gradients due to the necessary mechanical aeration for the maintenance purposes of stored fish. Through mathematical modeling, Tables 3 and 4 indicate a high correlation for the BOD coefficient (R 2 = 0.98) regarding its decay in T3.
Although in super-intensive aquaculture systems (high storage density) there is often a reduced self-purification capacity of the system (Adel and Dawood, 2021;Li et al., 2019), considering the experiment conditions, it was observed the temporal decay of organic matter in all treatments due to microbial activity, including the control the consortium of beneficial microorganisms inoculated at different concentrations in the treatments.
There was a synergistic evolution in time for nutrients, ammonia, nitrite and nitrate in the experiment system. A narrow and concentrated pulse was observed for nitrite ( Figure 4) around day 26 of the experiment while for nitrate the pulse occurred with greater dispersion. Both pulses were observed in T5. For nitrite and nitrate, all treatments decayed in time, and for nitrite T4 and T5 presented an even greater decay than the control, while for nitrate only T5 showed greater decay in relation to the control. Tables 3 and 4 show high correlations in T4 for the decay of ammonia (R 2 = 0.82) and nitrite (R 2 = 0.98) due to the bioaddition of the microbial consortium.
Ammonia is the final product of aerobic or anaerobic mineralization of organic nitrogen produced by the respiration of different heterotrophic microorganisms (Esteves, 2011). In this study, non-ionized ammonia behaved temporally in cycles for all treatments. Treatments 1 and 4 showed greater dispersion over time in relation to other treatments and control. Research, Society and Development, v. 11, n. 4, e26111427382, 2022 (CC BY 4.0) | ISSN 2525-3409 | DOI: http://dx.doi.org/10.33448/rsd-v11i4.27382 9 Table 3 -Coefficients obtained through equation 1 for the mathematical modeling of the decay rate of organic matter and nutrients during the experiment period.
Source: Authors (2021). Table 4 -Correlation coefficients and absolute mean deviation obtained through the mathematical modeling for nutrient decay rate as a function of bio-addition of a microbial consortium. T0  T1  T2  T3  T4  T5   R  The peak of ammoniacal nitrogen was observed at 50 days and the lowest value recorded at 38 days, around 0.40 mg L -1 . The trend to stabilize the ammoniacal nitrogen gradients was verified at 74 days with concentrations that fluctuated between 0.40 and 0.70 mg L -1 . The low concentration of chlorophyll "a" reveals that apparently the behavior of ammoniacal nitrogen ( Figure 3) was governed by the activity of heterotrophic microorganisms. This behavior may have occurred through the amonification-nitrification process, although phytoplankton is more efficient in ammonia transformation, when at low concentrations, as suggested by Hargreaves (1998), Klawonn et al. (2015) and Lukwambe et al. (2015). Unlike that verified by Hoang et al. (2018), there was no mortality and inhibition of the growth of stored fish in this study, during the experiment period, caused by the association of density with ammoniacal nitrogen gradients.

Treatment/statistic data
Ammonia toxicity to fish in both natural and confined systems is a function of high pH and temperature and, which did not occur in this study, where the gradients remained neutral for pH, and the temperature had an average gradient of 24.9 o C, which reveals a non-toxicity of ammonia to stored fish In fish, ammonia toxicity leads to hyperactivity, convulsions, loss of balance, lethargy and coma, in aquaculture systems, ammonia toxicity is commonly expressed by reduced growth and suppression of immunocompetence (Boyd, 2017;Hargreaves, 1998;Pinto et al., 2016). The stored fish presented a final biomass yield compatible with the control, allowing us to infer that there was no toxicity from the ammonia gradients present in the system.
On the other hand, the temporal dynamics of nitrite and nitrate (Figures 4 and 5) was guided by an expected pattern because its higher concentrations during the study occurred when the dissolved oxygen consumption in the medium was at its highest. This dynamic evidenced the nitrification process, ruled by heterotrophic bacteria responsible for oxidizing ammonia, genus Nitrossomonas and bacteria for oxidation of nitrite, genus Nitrobacter. On the other hand, a lower plankton activity may also result in a lower rate of ammonia assimilation by algae, which would increase the load breathed by nitrifying bacteria. Esteves (2011) notes that nitrification is an eminently aerobic process and only occurs in areas where oxygen is available. Throughout the experiment, however, it is observed the decay of nitrite in the medium, probably due to the breathing process, mediated by facultative anaerobic bacteria, in order to enable the oxidation of the remaining organic matter in the sediment, the denitrification process that may have occurred over the experiment period. Nitrite expresses its toxicity when hemoglobin is connected forming methemoglobin, a substance that does not have the capacity to carry oxygen. Kroupová et al. (2018) infer that nitrite toxicity occurs mainly in intensive and super-intensive production systems and that among the clinical signs described for nitrite toxicity are lethargy, weakening of the escape reflex, muscle contractions, agony and death. However, during the experiment period, there were no clinical signs such as the mentioned above as a response to the use of bio-addition of the beneficial microorganism's consortium.

12
The most prevalent inorganic form of phosphorus in wastewater and also in aquaculture systems is soluble phosphate or soluble reactive phosphorus. Phosphorus is an essential element for microorganisms responsible for the stabilization of organic matter (Sperling, 2017) and is primarily responsible for the cultural eutrophication of aquatic ecosystems (Esteves, 2011). The soluble phosphate gradients in this study showed a behavior close to linearity until day 38, that oscillated between 0.5 and 1.8 mg L -1 for treatments and control. In Tables 3 and 4, reactive phosphorus shows a high correlation (R 2 = 0.97) in T5 due to its decay with the bio-addition of the microbial consortium.
On the other hand, pulses also occurred in all treatments and the control from day 38 of the experiment, except in T5.
In T4 and in the control, gradients between 4.0 and 5.0 mg L -1 were observed, respectively. Boyd and Tucker (1992) and Rubinos et al. (2011) observed strong interaction between high phosphate gradients at pH of up to 8.0.
While Xu et al. (2016) found that the bio-addition of B. subtilis reduced pH and soluble phosphate in aquaculture systems. In this study, the pH among all treatments, including control, remained neutral and presented an average gradient of 7.56. Excess phosphorus in diets offered to fish in intensive aquaculture systems is common and consequently generating high levels of phosphorus excretion. The behavior in the time of reactive phosphorus (Figure 5) in the system may be directly related to the performance of metabolic growth of tilapia juveniles (initial growth) as Verdegem and Beristain (2005) mentions. The thin layer of sediment deposited in the substrate of the tanks acted as a sink of phosphorus, resulting in the deposition of solids. Phosphorus gradients may have been increased at 50 days by the re-suspension process, Li et al. (2020) mentions that a significant fraction of mineralized organic matter in the substrate can occur in the water column from the foraging activity of fish. David et al. (2017) and Flickinger et al. (2020) observed in multitrophic cultivation that the evolution in phosphorus concentrations in water and sediment are highly correlated with the daily and corrected supply of food for growth of the stored biomass. Therefore, phosphorus exchanges between the water column and sediment contribute less to explaining phosphorus pulses in water than feeding.

Conclusion
A pulse dynamic at regular times occurred indistinctly for all indicators evaluated in this study, although the intensity of the pulses showed particularities among the different indicators of water quality. Although the behavior of regular pulses of the indicators occurred, this condition did not affect the growth of fish submitted to a simulated system of intensive production associated with the bio-addition of friendly microorganisms.
The mathematical adjustment proposed to represent pulses at regular time intervals. The temporal behavior of dissolved oxygen, organic matter and nutrients in a simulated intensive aquaculture system submitted to bio-addition of beneficial microorganisms responded satisfactorily to their adjustment with correlations greater than 0.8, that shows that the model can be considered a model that satisfactorily correlates to the experimental data. And as future suggestions it is expected that this work will be a guide for a greater use of friendly microorganisms in aquaculture to reduce cleaning costs and longer production cycles due to better water quality.