Fish weight was recorded monthly by extraction of all animals from each tank. In this process, changes in total biomass of fish were determined to calculate different growth parameters as average daily growth rate , specific growth rate and feed conversion ratio , according to Lobillo et al. . Average fish weight was calculated by dividing the total fish weight by the number of animals. ADGR was calculated as the ratio between average fish weight increment and elapsed time. SGR was calculated as 100 x. FCR was computed as the ratio between consumed feed and total fish weight increment . Tench were fed with a trouts’ starter feed from the company “Biomar” with 54% protein and 18% fat. Animals were fed twice a day applying an amount equivalent to 1.5% of total biomass, therefore modifying the quantity of feed provided after each fish weighting. The following parameters were analysed daily: ambient and water temperature by means of a maximum-minimum thermometer and the volume of water replacement due to evaporation and transpiration water losses. Daily water replenishment rate was calculated as the ratio between the average daily consumption and the total water running in each system. Water samples were collected weekly from the fish tank in order to measure pH, electrical conductivity and nitrates. The pH and EC were determined with a pH-meter GLP 22 and an EC-Meter BASIC 30 ; respectively. When the pH of the water dropped rapidly to values close to 7, sodium and potassium hydroxides were added to the sump until it reached a value above 7 again. Nitrates concentration was obtained by means of an RQflex 10 plus . Dissolved oxygen levels were determined using colorimetric test kits . Aquaponic systems usually show low concentration levels of elements such as K, Fe or Ca. Therefore, the plants were periodically examined to observe the occurrence of nutrients insufficiency. To alleviate these potential deficits, K2SO4 at 1.5% was foliarly applied in all aquaponics systems. The applications were performed twice a week , first thing in the morning with a manual sprayer.
Chelated iron solution was directly added to the water , 0.1 L for each system. This was done on Fridays every fortnight. The effect of the substrate type used in the aquaponic production of strawberry,rolling benches bare roots and rockwool, on studied variables was assessed by means of an analysis of variance . Since multiple, repeated measurements were made in an experimental unit in our experimental design, a “repeated measures structure” was considered. In this design, the observations can no longer be considered to be independent, and as a result, we assumed that there were correlations in the residual errors among time periods. A mixed general linear model , considering substrate type as fixed factor and sampling time and block as random factors, was performed with the STATGRAPHICS Centurion XVIII statistical package . Previously, the normality and homoscedasticity were checked with the SmirnovKolmogorov test and the Levene test, respectively. The best transformation meeting normality criteria was obtained by means of Box-Cox transformation using the same software. In all the cases, with this power transformation, normality and homocedasticity were met. Post hoc analysis was performed using the HSD Tukey test and differences were considered significant when P < 0.05. When interaction between factors was found to be significant, the effect of main factors could not be discussed since this means that a substrate has an effect that changes along the sampling period. In these cases, the evolution along sampling period was described and compared between the two substrates studied. Despite the fact that the air temperature presented important variations within the greenhouse , water temperature in fish tanks underwent lower deviations, ranging between 17 and 28.1 ◦C throughout the period in which this test was performed . In general terms, the three systems studied were very similar with regards to the studied parameters of water quality. Nitrate content of the water increased from initial values around 20–30 mg L− 1 to values higher than 70 mg L− 1 after 42 days from the start of the test .
In this same period, pH decreased from initial values close to 8, similar to those corresponding to the available water source, to values between 7.2 and 7.6 as a consequence of the activity of nitrifying bacteria. In order to stabilize both parameters, the accumulated amount of water added to the system was increased from 49 days since planting . At the end of this trial, the daily water replenishment rates for systems 1, 2 and 3 were 2.16%, 2.21% and 2.22%; respectively. In relation to the electrical conductivity of water, the variations observed in Fig. 5 were analogous to those followed by the nitrate content as a consequence of the above referred water replacement operations. Weekly measurements of the dissolved oxygen levels showed stable values , which are considered adequate values for both fish and plants. Initial fish biomass in each of the three tanks was around 700 g, with 122 animals per tank and average weights per fish were between 5.71 and 5.76 g. After an acclimatization period of 30 days, average fish weights increased to 6.46–6.47 g, reaching 10.93–10.97 g at the end of the trial. Average daily gain ratios ranged from 0.023 and 0.025 g day− 1 at the initial period to 0.074–0.081 at the end, with a global value for the full period of 0.057 g day− 1 . Specific growth ratios recorded at the same dates varied from 0.383 to 0.408–0.749–0.833% day− 1 , with global values of 0.707–713% day− 1 . Feed amounts supplied to the animals were increased as they grew, so that during the first month they were given 195 g, in the second 359 g and in the third 420 g, which meant a total consumption of 974 g per tank. According to the feed consumption and the weight gained by the fish, the feed conversion ratios were calculated, with values in the range 2.22–2.32 during the first 30 days and 1.48–1.61 at the end of the trial, with global values of 1.53–1.56. The mortality rate of the fish was very low, since only one animal died in one of the three tanks studied in this trial. As happened with the water quality parameters between the three tanks shown above, there was a great uniformity in the results obtained for the different parameters related to the growth of fish in the three tanks. No discordant data were found for average weights at different ages, average daily growth rate, specific growth rate; mortality, or feed conversion ratios. Almost all the plants in the trial were able to produce fruits during the study period, as it can be observed in Table 2. Total production per plant varied between 0 and 105 g, where zero values indicate that the plants did not produce fruits at an early stage.
Taking into account that the total area occupied by the crop was 7.2 m2 a density of 10 plants m− 2 was estimated, so that the productions obtained for the different blocks of 6 plants were between 140 and 565 g m− 2 . The quality variables and SPAD values were only affected by the elapsed time since transplanting . Overall, SPAD values tended to be higher at the end of the period, with a more marked trend and lower average values in new leaves when compared with old leaves . On the other hand, the effect of the differences between the two substrate types studied on productive variables varied along the growing period studied as revealed by the significant interaction between this factor and time. The number of flower buds and the number of flowers were initially higher for the plants cultivated with bare roots than with rockwool, but at the end of the trial both values were increased for the plants in rockwool . On the contrary, the number of fruits with bare roots was slightly lower at the beginning, but at the end of the trial it showed the highest values, whereas the accumulated weight of fruits that was initially lower for bare roots eventuallyequalled those corresponding to rockwool at the end of the period . Our systems allowed to maintain stable water conditions throughout the time period of the test, resorting to simple pH control operations by adding alkalis, supplementing the deficit nutrients for the plants of these systems in the form of EDDHA, potassium sulphate, and sodium and calcium hydroxides; and reduction of nitrate levels based on water changes. The total time devoted to these tasks did not significantly increase the time normally required for operations related to the hydroponic cultivation of strawberries, where the pH and levels of nutrients in the nutritive solution must also be controlled and regulated. In this sense, pH regulation is a fundamental operation to optimize the extraction of nutrients by the roots of plants, the development and growth of fish and the performance of the bacterial community in the biofilters. This can be attributed to the use of a low fish density in our aquaponic facilities, which allowed the use of a simple system of aquaculture circuits . This avoids the need for a clarifier and settler, ebb and flow bench while the simple biological filters used reduced the investment since there is no requirement of extra aeration devices or powerful pumps for recirculation of water. However, if the density of fish was to be increased or the duration of the test prolonged, it is likely that these common elements in RAS facilities would have to be introduced.
The initial adjustment that was made of the ratio between the area dedicated to strawberry cultivation and the daily amount of feed provided to the fish was not fully appropriate, probably due to the fact that most of the bibliographic references used for the initial adjustment were based on the use of tilapia, a fish species that requires lower protein contents in the compound feed supplied. Tench requires higher levels of crude protein, so the total contribution of nitrates generated was somewhat higher in our systems. This explains the substantial increase in the daily rate of water replacement from day 49 to keep nitrate levels stable around 70 mg L− 1 . Despite this increase in water consumption, the final daily water replacement rate was around 2.2%, well below the 5–10% required in recirculating aquaculture systems. Regarding this water saving potential, in an aquaponic tench-lettuce production system similar to this one, the authors obtained daily water replacement rates of 1.18% in a 65 days period . This should be highlighted, since in the conditions in which strawberry productions are carried out in Spain, water is a scarce resource. In this sense, the results obtained by Van Ginkel et al. comparing traditional agricultural productions in California with hydroponic and aquaponic productions, pointed out that these systems are recommendable strategies to save water, since consumption was 66 and 8 times lower than in the traditional systems, respectively. As previously mentioned, the growth of the animals in the three tanks was quite homogeneous, although the weight increases were not of great magnitude, something that is characteristic of this species. The survival rate of animals was very high , within the range described by other authors for animals with similar weights .This confirms the high adaptability of this species to the aquaponics system. In order to keep the water quality parameters stable over time, regardless of the changing weight of the fish with time elapsed, it is recommended to use several connected fish tanks . In each of these tanks there would be fish of different ages, which would be harvested in stages, thus maintaining a similar proportion of animals in different growth stages. With this system, fingerlings should be restocked each time a tank is harvested . This arrangement of tanks would also allow increasing the density of fish in a simple way, improving the productivity associated with fish farming. In view of the current findings, the slow growth of the tench is verified, which supposes an important obstacle for the extension of its cultivation in recirculating aquaculture systems, except to support sport fishing favouring the repopulation of the water courses in which naturally it is found.