Prior to the beginning of the experiment, 75% of the total volume of the aquaculture subsystem was filled with artesian well water. The remaining 25% was filled with water from 60-day mature inoculums from RAS and BFT systems previously performed. These inoculums were used to ensure that the microbial communities of the MBBR and bioflocs were already mature. For the maintenance of the biofloc microorganisms and the C:N ratio of the water at 15:1, liquid molasses was added to the BFT and DFP fish tanks three times a week as a complementary carbon source. The amount of molasses was calculated based on the input of N by fish feed. Calcium hydroxide was added to the fish tanks when the alkalinity was under 80 mg L− 1.Nile tilapia juveniles were purchased from a commercial hatchery and acclimatized for 7 days after arriving at Caunesp. In all treatments, 114 masculinized juveniles were stocked, totaling an initial biomass of 0.43 kg m− 3 and density of 300 fish m− 3 in each fish tank. After 4 weeks of culture, the number of fish per tank was managed in order to readjust the densities. Thus, 70 fish per tank were kept, resulting in a density of 184 fish m− 3 and a calculated biomass of 3.6 kg m− 3. During the 56-day trial, the fish were hand-fed with the test diets four times a day at 08:30, 11:00, 14:30 and 18:00 h.
The amount of feed was calculated based on the percentage of body weight recommended by a commercial feed industry,ranging from 12 to 5% according to the average fish weight. A sample of at least 20% of the total number of fish in each tank was weighed weekly to adjust the amount of feed in all treatments. At the end of the experiment,vertical grow all tilapia juveniles were counted and weighed and 15 fish from each tank were individually measured. Final individual body weight,total fish length,weight gain,yield,specific growth rate,feed conversion ratio and survival were assessed.For this, three different samples were taken: 30 individuals from the initial fish population; 10 from each repetition per treatment at the end of the experiment; and 15 g of each diet. The fish were anesthetized and euthanized. Subsequently, the whole bodies were weighed, packed and frozen at − 20 ◦C for later analysis. The frozen fish were ground, homogenized in a meat grinder and lyophilized. The lyophilized matter was used to determine the percentage of dry matter and crude protein,according to the methodology of A.O.A.C., Association of Official Analytical Chemists. The same analyses were applied to determine the composition of the diets. All analyses of the proximal composition of the tilapia tissue were made in duplicate.Two trials of butter lettuce production in different phases were carried out. In Cycle 1, the seedling phase was evaluated in a 14-day trial. For this, hydroponic seedlings at 7 days after sowing and 0.59 ± 0.08 g were grown until 21 d.a.s. In Cycle 2, the final production phase was performed, in which new hydroponic seedlings at 21 d.a.s. and 2.04 ± 0.57 g were planted and cultivated for 21 days until harvest. In both cycles, 8 plants were distributed in each hydroponics subsystem with a density of 19 plants m− 2.These four lettuces per tank were selected randomly at the beginning of the trials and the same were weighed weekly throughout each trial.
At the end of Cycle 1, all plants were weighed and the following growth parameters were evaluated: leaf and root height,total wet weight,total dry weight,number of leaves per plant,productivity and specific growth rate. At the end of Cycle 2, the following growth parameters were evaluated for seven lettuces from each plant tank: leaf and root heights,wet leaf and root weights,dry leaf weight,number of leaves per plant,and productivity. Also, in both cycles, a visual analysis was applied to identify the non-marketable plants. Plants that contained up to 33% of abnormalities on the leaf surface, i.e., with a yellowish color, burns or wrinkles, were considered non-marketable. For the control of plant pests, twelve traps were distributed through the greenhouse. A visual scan of the presence of plant pests or diseases was performed daily on all plants and no sign of them was seen during the trials.Prior to the beginning of both trials, 50% of the total volume of each plant tank was filled with water from the aquaculture subsystem and, for the remaining 50%, artesian well water was used. The tanks of HP treatment were only filled with artesian well water. In the DAPS plant tanks, the water was collected from the upper-middle portion of the MBBR of the DAPS aquaculture subsystem. In the DFPs treatment, the water of each DFP fish tank underwent a decantation and filtration process before being directed to the plant tanks. This means that the water was pumped into the RFS and remained there for 20 min until the biofloc particles were decanted. After that, the RFS supernatant was directed to a bag filter and then to the PTs. The initial volume of water taken from each aquaculture subsystem, which was used to supply two PTs,was replaced with artesian well water. Between the two cycles of plant production, all the PTs were emptied, cleaned and all the aforementioned procedures for filling the PTs were repeated.
When the EC in all PTs was stable, the seedlings were planted. For plant nutrition during the experiment, water from each aquaculture subsystem or well water was added manually to the PTs, at the proportion of 2% of the initial volume of the PT since it was the estimated volume of water evaporation in the PTs. In the DFP systems, the biofloc decantation and filtration procedures were always carried out and the decanted bioflocs returned to the fish tanks, except for samples collected in the beginning, middle and end of the experiment. In both plant cycles, the commercial fertilizer solution was added according to the equation above and only if the registered EC values were below the expected ranges. In Cycle 1, the EC was maintained between 1.1 and 1.3 mS cm− 1 and a unidirectional water flow between fish and plant tanks occurred once a day on alternate days. In Cycle 2, the EC was between 1.6 and 1.8 mS cm− 1 and a unidirectional water flow occurred once a day, six days per week. Aiming to maintain the pH in the PTs at between 5.5 and 6.5, diluted phosphoric acid was added when the pH exceeded 6.5.FLOCponics systems were run in a decoupled layout with the aim of enabling proper management of each subsystem; thus, taking advantage of the nutritional benefits of the biofloc-based culture to produce tilapia juveniles and lettuce. The findings of this study suggest that some critical points usually associated with FLOCponics systems were addressed by individualizing the aquaculture and hydroponic subsystems. For instance, the difficulty of maintaining a low concentration of solids in the hydroponics subsystems and, at the same time, providing a sufficient amount of bioflocs in the fish tanks,has been reported as an issue of coupled FP systems. Another challenge of coupled FP is regulating the water pH within the appropriate range for fish, bioflocs and plant growth. The trade-offs related to solids concentration and water pH were tackled in the present study through the use of decoupled FLOCponics. The physical-chemical parameters of the water were monitored in each subsystem of all treatments to interpret the production results.
Except for the maximum values of settleable solids,the other results for water quality in the aquaculture subsystems were within the acceptable range for tilapia and also for BFT and RAS microorganisms. The mean results of settleable solids were within the recommended range of 5 to 50 mL L− 1 for tilapia ; however, in some measurements the values exceeded 100 mL L− 1. No issues regarding a high amount of solids in the aquaculture subsystem have been reported so far in the research on FLOCponics. As already mentioned, low concentrations of bioflocs have usually been indicated as a drawback of coupled FLOCponics systems. Hargreaves stated that the values of settleable solids above the recommended value of 50 mL L− 1 do not favor fish growth or nutrition, but might result in oxygen depletion and a higher electricity demand in biofloc-based cultures. The high amount of settleable solids in the fish tanks, associated with the unexpected increase in environmental temperature, was probably the main factor that caused the sudden drop in DO and the unviability of the DFP-36 treatment. For the other treatments, the DO values were always higher than 3 mg L− 1 despite the recorded values of settleable solids. The accumulation of solids was probably a result of the methodology adopted for regulating the C:N ratio, in which the addition of the carbon source was performed periodically based on the amount of N inputted by the feed. Given the need for minimizing the risks associated with high solids concentrations, it is recommended to test the methodology based on the concentration of ammonia in the water to regulate the C:N ratio or to remove and reuse the solids in FP systems. The differences found for the mean values of DO, pH, EC and TDS in the aquaculture subsystem were a result of the different input of N and carbon source in each treatment. The input of the carbon source seems to be the main factor in these results, indoor growers since in the DAPS the mean values of pH, EC and TDS were distinct from those recorded in the BFT and DPF-32, even though all of them received the diet with 32% crude protein.
For the nitrogenous compounds and orthophosphate results,it is hard to conclude whether or how the dietary protein or integration with plant production affected the variation of these nutrients during the experiment. Further studies with a focus on the nutrient flows between the BFT and hydroponic subsystems and the carrying capacity of DFP systems are still required to understand the efficiency of recovering nutrients from the BFT effluents by plants. In the hydroponic subsystems, except for the pH values in Cycle 1, the other parameters of water quality remained within the expected ranges in both cycles. The pH values should remain between 5.5 and 6.5 to enable higher bioavailability of nutrients, whether they come from the aquaculture subsystems or from the extra commercial fertilizer. In Cycle 1, the pH was above the recommended range in all treatments and the highest values were recorded in the DFP treatments. Phosphoric acid was added in the plant tanks to regulate the pH when it exceeded 6.5. Another factor that interfered with the pH was the buffering in biofloc-based cultures. Despite these issues with pH, they exerted no negative effects on the growth of lettuce seedlings. As expected, the results for tilapia growth demonstrated that the well-known benefits of BFT for juvenile nutrition are also found in the DFP systems. Tilapia juveniles fed with 32% CP and grown on both biofloc-based treatments grew 22.7% more than those in DAPS also fed with 32% CP. Luo et al.,Long et al.,and Hisano et al. showed the same tendency of improved zootechnical performance for tilapia grown in BFT compared to RAS, although both were fed with the same amount of CP. They indicated the uptake of the microbial bioflocs as a complementary feed by tilapia as the main reason for these results. Not finding differences in FCR among the treatments is somewhat surprising, since better feed conversion is usually related to biofloc-based culture compared to RAS. Nevertheless, the results of PER,PPV,and CPwg show the highest efficiency in using the dietary protein in the fish produced in the biofloc-based system compared to DAPS with 2.93, 40.11% and 13.70%, respectively. These results suggest that even in an integrated system the in situ food present in biofloc-based systems is used by tilapia juveniles to complement their dietary protein needs. The similar results for tilapia growth in DAPS and in DFP fed with 8% lower CP reinforce this statement. The zootechnical results of tilapia fed with lower CP suggest positive economic and environmental implications of DFP. Since protein is usually the most expensive component in the diets,the use of lower CP levels will result in lower feed costs.