Transformation of phosphorus in an experimental integrated multitrophic aquaculture system using the media filled beds method in plant cultivation

The objective of this study was to trace the transformations of phosphorus in an integrated multitrophic aquaculture (IMTA) system and to determine whether the method of plant breeding influenced the dynamics of these changes. In the experiment, the media filled beds (MFB) method of plant cultivation was applied. Fish tanks were stocked with 200 common carp Cyprinus carpio, and hydroponic terraces were planted with 49 zantedeschia Zantedeschia sp. bulbs. Water samples were taken directly from the fish tanks immediately after each type of filtration (mechanical, biological and hydroponic). The basic forms of orthophosphates in the IMTA system included some H2PO4 ions but mostly HPO4. The higher supply of reactive phosphorus that occurred over time in the experiment may have contributed to a decrease in calcium ion concentration due to the formation of Ca3(PO4)2 and CaHPO4 salts, thus inactivating some of the phosphorus available to plants. Phosphorus may have also been inactivated in the sediment due to the formation of Mg3(PO4)2 and MgHPO4 salts after decreasing the concentration of calcium ions as a result of their precipitation in sediments. Mineralization of organic matter took place under aerobic conditions. Organic matter was a source of biogenic substances in the IMTA system. Experimental results showed that IMTA systems have significant potential to reduce phosphorus in aquaculture wastewater and thus provide a good environment for fish farming by improving water quality.


INTRODUCTION
In view of the growing global human population and the consequent requirement to increase food production, there is a need to seek economically viable and environmentally safe solutions for food production. One of the problems that must be addressed is the availability of phosphorus for fertilizer production, as phosphorus is a key nutrient in the developing agricultural industry. The extraction and processing of phosphate is a highly energy-intensive process that results in a wide range of adverse envi-ronmental effects (Cerozi & Fitzsimmons 2017). From this perspective, the use of phosphorus-rich aquaculture waste for irrigation and soil fertilization makes both economic and environmental sense.
In aquaculture, widely used recirculating closedcircuit systems, such as the recirculating aquaculture system (RAS), are becoming increasingly popular worldwide. These types of systems can be improved through integration with hydroponic farming, thus providing additional economic and ecological benefits (Palm et al. 2018). The integration of aquaculture and hydroponic plant breeding has been analyzed ABSTRACT: The objective of this study was to trace the transformations of phosphorus in an integrated multitrophic aquaculture (IMTA) system and to determine whether the method of plant breeding influenced the dynamics of these changes. In the experiment, the media filled beds (MFB) method of plant cultivation was applied. Fish tanks were stocked with 200 common carp Cyprinus carpio, and hydroponic terraces were planted with 49 zantedeschia Zantedeschia sp. bulbs. Water samples were taken directly from the fish tanks immediately after each type of filtration (mechanical, biological and hydroponic). The basic forms of orthophosphates in the IMTA system included some H 2 PO 4 − ions but mostly HPO 4 2− . The higher supply of reactive phosphorus that occurred over time in the experiment may have contributed to a decrease in calcium ion concentration due to the formation of Ca 3 (PO 4 ) 2 and CaHPO 4 salts, thus inactivating some of the phosphorus available to plants. Phosphorus may have also been inactivated in the sediment due to the formation of Mg 3 (PO 4 ) 2 and MgHPO 4 salts after decreasing the concentration of calcium ions as a result of their precipitation in sediments. Mineralization of organic matter took place under aerobic conditions. Organic matter was a source of biogenic substances in the IMTA system. Experimental results showed that IMTA systems have significant potential to reduce phosphorus in aquaculture wastewater and thus provide a good environment for fish farming by improving water quality.
KEY WORDS: Integrated multitrophic aquaculture · IMTA · Media filled beds · MFB · Transformations of phosphorus compounds OPEN PEN ACCESS CCESS many times using different system layouts, plant species and species of aquatic organisms. The most suitable and effective aquacultural practice involves closed recirculation systems integrated with hydroponic systems, i.e. integrated multitrophic aquaculture (IMTA), because nutrient concentrations can be sufficiently maintained for proper plant growth and development (Seawright et al. 1998).
According to Cerozi & Fitzsimmons (2017), the transformation of phosphorus and its budget in IMTA systems is not fully understood. The dynamics of phosphorus metabolism in IMTA systems have been described by Seawright et al. (1998), but according to that study, the amount of phosphorus in the fish, plants and sediments was higher than the amount of phosphorus supplied with the feed. Phosphorus dynamics were also discussed by Cerozi & Fitzsimmons (2017), who created a model which can be used as a tool to maximize the use of phosphorus in the system and minimize the amount of phosphorus in sediments. In their experiment, they showed that phosphorus use by fish comprised 42.3% of the phosphorus available in the feed and that phosphorus use by plants amounted to 29.4%. They also found that 13.1% of the phosphorus was not available for the fish or plants.
The concentrations of different nutrients in IMTA sys tems vary depending on the production techno logy adopted (Seawright et al. 1998). These include 3 hydro ponic systems classified according to growth medium types. In the nutrient film technique (NFT), plants are grown in long, narrow plastic channels with a 1% gradient through which a thin layer of water constantly flows. The deep water culture (DWC) method is also widely used in aquaponic systems. In this system, plants float on plastic sheets in 30 cm deep tanks filled with water. The roots of the plants grow directly in the continuously flowing aerated water, and the rafts of plants floating on the water move as the plants grow (Love et al. 2015, Forchino et al. 2017. Finally, the media filled beds (MFB) system is probably the simplest and most common technique for small-and medium-scale systems. In this system, a terrace is filled with a special substrate that provides extra space for biofiltration and thus for efficient plant growth. This type of bed is flooded with water from the RAS system, in this way the nutrients necessary for the proper growth of plants are provided. Because of the mechanical support for the roots of the plants, this substrate can be used to grow various species of plants (Somer ville et al. 2014, Engle 2015, Li et al. 2019. Most IMTA systems are designed on a nitrogen budget, which does not guarantee that other nutrients, such as phosphorus, reach levels for sufficient plant growth and development. Phosphorus deficiency will inhibit plant growth, while excess phosphorus can lead to antagonistic interactions with microelements, especially zinc (Barben et al. 2010).
Taking into account a number of aspects related to the functioning of the IMTA system, such as production technology, the availability of nutrients for both fish and plants and ecological aspects, this study was undertaken to determine the intensity of phosphorus compound transformation in the experimental multitrophic aquaculture system using the MFB method of plant breeding.

Description of the experimental IMTA system
The experimental IMTA system was set up in the laboratory of the Department of Aquatic Bioengineering and Aquaculture of the West Pomeranian University of Technology, Szczecin, Poland, and had a volume of 5 m 3 and was run for 21 weeks from February to the end of May. The system consisted of 3 poly ethylene rearing fish tanks, each with a capacity of 1 m 3 , a mech an ical filter (0.5 m 3 ) filled with 0.1 m 3 of CALDNESS K1 mobile resin and a biological filter comprising a nitrification chamber (0.5 m 3 ) filled with 0.1 m 3 of Mutag BioChip mobile medium with a protective active area of 3000 m 2 m −3 . The entire IMTA system was initiated 3 wk before the start of the ex periment to allow the biofilm on the biological bedding to develop naturally, using water from an other system in operation. The entire system, including the water in the breeding tanks and the biofilter bed, was aerated using a Hi-Blow blower, which forced compressed air into membranes installed in the tanks. The remains of uneaten food and faeces were removed through an outflow located centrally at the bottom of the tanks, transferring contaminated water to the mechanical filter. The mechanical filter was rinsed every 2−3 d, depending on its condition.
The hydroponic terraces (0.5 m 3 ) were made of PVC foam sheets filled with perlite (420 l) used as a medium for planting in the MFB method. Tricolor lights (GROW LED UFO 90 W) were placed above each terrace. The lighting regime was 12 h light:12 h dark, with the plants illuminated by full light and the fish tanks remaining in the twilight. Water loss by evaporation, transpiration and sludge removal was compensated for by adding mains water at 1.5−2.0% of the system volume over the experiment.

Fish tanks
Each experimental tank was stocked with 200 common carp Cyprinus carpio L. The mean initial weight of each fish was 40 g. The fish were fed twice a day throughout the experiment; the daily feed ration was adjusted every 2 wk depending on the body weight of the fish. The feed used was initially the commercial feed Aller Primo, comprising 37% protein content, 12% fat, 32.5% nitrogen-free extracts (NFE) and 7% ash, with a total energy of 19.6 MJ and digestible energy of 16 MJ per kg. For practical reasons, this feed was replaced after 1 mo by the identically composed Aller Primo Float, as the use of recirculating floating feeds in the system reduced the risk of water quality deterioration caused by unused feed and gave a better level of feeding, allowing for better ongoing control of the condition of the fish. During the experiment, water temperature was maintained at an average of 22°C, forming optimal conditions for carp breeding.

Hydroponic terraces (filters)
At the beginning of March, the hydroponic terraces were planted with zantedeschia (Zantedeschia Spreng.) bulbs at a density of 49 plants terrace −1 . The bulbs were planted in the perlite medium, maintaining a distance of 10 cm from the edges of the tank and 20 cm between the bulbs. Water samples were taken once a week beginning in February, when the fish were introduced into the rearing tanks, until the end of May, when the flowers from the hydroponic terraces were harvested. Water samples from the fish tanks were taken from a half-tank depth, directly at the water outlet of the filters and the fish were transferred to the breeding in the Fisheries Experimental Station in Nowy Czarnów. A diagram of the IMTA system experiment is shown in Fig. 1.

Physico-chemical analyses of water
The concentrations of biogenic phosphorus compounds were measured in 3 replicates using a PHARO 300 MERCK UV-VIS spectrophotometer. Reactive phosphorus (mg dm −3 ) and total phosphorus (mg dm −3 ) were analysed at λ = 882 nm. Organic phosphorus (mg dm −3 ) was calculated from the difference between total phosphorus and reactive phosphorus, and dissolved oxygen concentration (mgO 2 dm −3 ) was determined using the Winkler method. Other parameters measured included pH, water temperature, alkalinity, bicarbonate, calcium and magnesium. Organic matter loading of the experimental IMTA system was calculated by determining the chemical oxygen demand (COD). Moreover, the ratio of bicarbonate ions to carbonate ions was calculated. All analyses used standard methods (APHA 1999).
The percentage of reduction after each filter was calculated, and results are presented as mean, minimum and maximum values. Results were analysed statistically using Statistica v.13.3 software. The significance of the correlation coefficients was tested using an R-Spearman non-parametric test; results were considered significant at p < 0.05.

Calculation of phosphorus balance
The equation expressing the phosphorus balance during the experiment is: P feed = ΔP fish + P water + P faeces + P plants + sediments (1) where P feed is the amount of phosphorus in the feed, ΔP fish is the increase in phosphorus accumulated in the fish during the experiment, P water is the amount of phosphorus accumulated in the water and P faeces is that in the fish faeces. All variables are given in the unit 'kg per kg'.
From Eq.
(1), we obtain the amount of phosphorus in the plants and sediments: P plants + sediments = P feed -(ΔP fish + P water + P faeces ) (2) ΔP fish can be evaluated considering the mass of the fish at the beginning (M fish (begin) ) and end (M fish (end) ) of the experiment and the concentrations of the phosphorus in the fish bodies (p) measured experimentally (p fish (begin) ) and (p fish (end) ): ( 3) Phosphorus supplied to the RAS (P feed ) comes from the feed (M feed ) and amounts to: All concentrations of the phosphorus: p fish , p feed and p water are given in unit 'kg per kg', with the exception of p feaces , which is taken from the data given by the feed producer in 'kg per 100 kg'. The concentration of the phosphorus is related to the average fish mass M fish and is equal to: while P water is determined based on the concentration of phosphorus measured experimentally after the experiment (p water ): P water = p water · V water (7) 3. RESULTS

Breeding
Fish were weighed every 3 wk and the feed conversion ratio (FCR) determined. During the experiment, the weight gain of fish was almost 7-fold. The greatest increase in fish weight was observed after the system had stabilized the FCR coefficient showed a downward trend, while in the second phase it stabilized at a constant level (Table 1).

Variability in the physical and chemical parameters of the water
Reactive phosphorus concentrations showed an upward trend during the experiment, both in the water of the fish tanks and in the water from the individual filters, ranging from 0.07−3.29 mg dm −3 (Fig. 2). In the final 2 weeks, a slight decrease in concentration to about 2.90 mg dm −3 was noted at all measurement points (Fig. 2). No significant differences in reactive phosphorus concentrations were ob served among the filters (Fig. 2). For the concentrations of this form of phosphorus, an average reduction level was determined in the water after each filter (Table 2, Fig. 3). The observed change in the reduction of reactive phosphorus concentrations in the water after each filter compared to the water after the fish tanks was statistically significant, which indicates that after each filtration step, the concentration of reactive phosphorus was lower than in the water of the fish tanks (Table 3).
The highest concentration of organic phosphorus (5.53 mg dm −3 ) was recorded on the last day of the experiment in the water from the fish tanks; this amount was twice as high as the previous determinations (Fig. 2). In the hydro ponic terrace water samples, the concentration of organic phosphorus was characterized by a similar variability to the biological and mechanical filter samples (Fig. 2). During the experiment, the concentration of organic phosphorus was reduced after all filters, and the levels of reduction fluctuated significantly (Table 2, Fig. 3). The correlation coefficients for the reductions in organic phosphorus for all filters were statistically significant (Table 3).
Week Total phosphorus concentrations ranged from 0.15− 5.96 mg dm −3 . A decrease in total phosphorus concentration to an average of 5.45 mg dm −3 was re corded at the end of the experiment (Fig. 2). The average reduction after biological filtration was −5.6%, while the water filtered by the hydroponic filter had an average reduction of −5.4% (Table 2, Fig. 3).
Dissolved oxygen concentrations showed a downward trend during the experiment, both in the water from the fish tanks and after filtration (Fig. 4A). The highest dissolved oxygen concentrations were observed in the initial phase of the experiment, when levels above 16.0 mg dm −3 were recorded (Fig. 4A). After the 3 rd week of the experiment, the amount of dissolved oxygen gradually decreased, reaching a minimum of 2.0 mg dm −3 in the last week of the experiment (Fig. 4A).
pH remained stable throughout the experiment, with an average value of 7.47 in the fish tank water, 7.43 after filtering with the mechanical filter, 7.48 after the biological filter and 7.46 after the hydroponic filter (Fig. 4B). The highest pH values were recorded in the first 2 weeks of the experiment, followed by a decrease to a stable level between 7.0 and 7.5 until the end of the experiment (Fig. 4B).
The highest value of general alkalinity was observed in the first 3 weeks of the experiment (Fig. 4C). Then a gradual decrease of alkalinity was ob served up to the 9 th week of the experiment, when it reached a minimum value (Fig. 4C). An increase in alkalinity to a level above 2.50 mval dm −3 was observed in the 14 th week of the experiment; alkalinity remained at this level until the end of the experiment (Fig. 4C).   (Fig. 5A). The highest concentrations of bicarbonates were observed in the initial phase of the experiment (values peaked up to 4 times in the first week) when they were above 140.0 mg dm −3 , followed by a decrease. The lowest bicarbonate concentrations were observed twice: in the 6 th week of the experiment and between the 8 th and 10 th weeks of the experiment. After 10 weeks, the bicarbonate concentrations ranged between 100.0 and 140.0 mg dm −3 (Fig. 5A).
The highest concentrations of calcium ions were recorded during the first 9 weeks of the experiment (Fig. 5B). In the 10 th week, concentrations decreased rapidly and continued to decline in the following weeks until the last week of the experiment (Fig. 5B).
Variability in the concentrations of magnesium ions followed a different pattern. The mean concentration was above 95.0 mg dm −3 , both in the water from the fish tanks and after all the filters. Magnesium ion concentrations ranged from 70.3−144.0 mg dm −3 (Fig. 5C) and remained at a constant level until the 8 th week of the experiment. In the 9 th week, there was a single increase in magnesium ion concentrations to a maximum value (Fig. 5C). In the following weeks, concentrations returned to the values observed in the first 8 wk of the experiment (Fig. 5C).
The ratio of bicarbonate ions to calcium ions ranged from 0.4−3.2 (Fig. 6). The ratio declined to a minimum in the 9 th week and then increased to a maximum value at the end of the experiment (Fig. 6).
Correlation coefficients were determined to illustrate the role of calcium ions in the ratio of bicarbonate to calcium ions. All relationships were statistically significant, which indicates the high affinity of calcium ions to bicarbonate ions (Table 4). Calcium ions combine with bicarbonate ions to form calcium bicarbonates in the system, which results in the loss of calcium ions from the system in the water from the fish tanks and after each filtration.   (Fig. 7). Lower index values were ob served in the first 5 weeks of the experiment (Fig. 7). Between Weeks 5 and 12, an increase in the index value to a maximum was observed (Fig. 7). In the 13 th and 15 th week, there was a decrease in the amount of organic matter in the system (Fig. 7). After the 15 th week, an upward trend was observed.
The re duction of organic matter after all filters fluctuated significantly, and the overall tendency indi- cates an increase in organic matter during the experiment (Table 2). On the other hand, there was less organic matter in the water after each filter relative to the amount of organic matter in the water after the fish tanks. This was confirmed by the significant correlation coefficients (Table 3). The relationship between the concentration of calcium ions and reactive phosphorus is shown in Fig. 8. Until the 9 th week of the experiment, when the concentration of reactive phosphorus remained be low 1.5 mg dm −3 , calcium ion concentrations re mained above 140.0 mg dm −3 . After the 9 th week, there was a sharp decrease in calcium ion concentrations and an increase in reactive phosphorus. This trend continued until the end of the experiment (Fig. 8).

Total phosphorus balance in the experimental multitrophic aquaculture system
We used the equations in Section 2.3 as well as data obtained from the experiment and provided by the feed producer to obtain the following values: M fish (begin) = 24.27 kg, M fish (end) = 159.95 kg, M feed = 165.841 kg, V water = 5000 dm 3 , p feed = 1%, p fish (begin) = 0.25%, p fish (end) = 0.27%, p water = 5.46 · 10 −6 kg dm −3 and p faeces = 0.39 kg. From these values, we calculated the amount of phosphorus accumulated in the plants and sediments as P plants + sediments = 0.947 kg, which is 57% of the phosphorus supplied to the RAS. The amount of phosphorus in the sediments can be estimated based on the phosphorus reductions obtained after individual filtration. The maximum reduction of phosphorus after the hydroponic filter was 21.2%, so it can be assumed that the amount of phosphorus accumulated from circulation in plants was just 21.2% (Table 2). However, we can assume that not all of the available phosphorus was used by the plants. Therefore, it was assumed that only half of the 21.2% of available phosphorus was accumulated in the plants. On this basis, the following result was obtained: P plants = 10.6% and P sediments = 46.4%. Fig. 9 shows the balance of total phosphorus delivered with the feed in the experimental IMTA system.

Transformation of phosphorus in an integrated aquaponic system
Phosphorus is a basic macroelement necessary for the growth and development of living organisms. It occurs in forms that are strongly dependent on envi-  Prabhu et al. 2007, Resh 2013. In our experiment, the pH was maintained at a level that allowed the formation of phosphorus in forms available to the plants; hence, based on the previously discussed results of pH determinations, we me that HPO 4 −2 dominated in the system, with a small amount of H 2 PO 4 − . The appropriate pH of the water is important not only to ensure the presence of phosphate forms available to plants but also to create appropriate conditions for the proper course of biochemical processes in the system (Espinal & Matulić 2019).
The primary source of phosphorus in an aquaponic system is fish feed (Eck et al. 2019). A stable and relatively low value of the FCR coefficient indicates optimal use of feed by the fish; the unused part of the feed is certainly a source of phosphorus in the IMTA system, even though phosphorus can be a limiting element and can impede plant growth (Graber & Junge 2009). We assume that a ratio of 1:15 nitrogen: phosphorus is optimal for building organic matter in the primary production process. The results from calculations obtained in our experiment showed phosphorus was a limiting element because the nitrogen: phosphorus ratio was 1:4.9.
According to Rafiee & Saad (2005), fish can use up to 15% of the phosphorus available in their feed. Cerozi & Fitzsimmons (2017), on the other hand, pointed out that in hydroponic lettuce production, the amount of phosphorus supplied from fish tanks can   (2017) reported a PUE of 29.4%. In our experiment, a 21% reduction of phosphorus was achieved. Thus, the reduction level obtained in our experiment was lower than previously described reduction levels. It can be assumed that the lower PUE was related to the MFB method of plant cultivation, where hydroponic terraces were filled with perlite. The filling of the hydroponic terraces and the oxy gen conditions in the experiment may have contributed to the formation of insoluble phosphorus compounds such as hydroxyapatites and their deposition in the sediments that formed at the bottom of hydroponic tanks; hence the inferior absorption of orthophosphates by plants using the MFB method. In addition, at relatively high calcium concentrations in the system, phosphorus may be excluded from circulation by forming Ca 3 (PO 4 ) 2 or CaHPO 4 salts at neutral pH (Yep & Zheng 2019). Phosphorus precipitation may also occur through the formation of the salts Mg 3 (PO 4 ) 2 and MgHPO 4 (Palm et al. 2018).
A higher supply of reactive phosphorus in the experiment may have contributed to a decrease in calcium ion concentration as a result of the formation of Ca 3 (PO 4 ) 2 or CaHPO 4 (Fig. 8). In the second phase of the experiment, reactive phosphorus precipitated in the form of Mg 3 (PO 4 ) 2 and MgHPO 4 . An important factor preventing the coupling of orthophosphates with calcium ions is the general alkalinity of the water and the associated concentration of carbonates and bicarbonates (Cerozi & Fitzsimmons 2017). The speciation of carbonates in the water is strongly pHdependent and is similar to that of orthophosphates. At high pH, carbonate ions dominate and, along with calcium ions, will form calcium carbonate. At a lower pH, calcium bicarbonate will be formed. It should be as sumed that the presence of bicarbonates at the level presented in this experiment also allowed binding of calcium ions with bicarbonates. A deficit of bi car bonates relative to calcium ions occurred between Weeks 5 and 11 of the experiment (Fig. 6). This was also confirmed by an increase in the concentration of reactive phosphorus after 6 wk of the experiment Therefore, after the 6 th week, there was more reactive phosphorus available for the plants because calcium was bound to bicarbonates. After the system stabilized, the supply of reactive phosphorus in creased. The balance of phosphorus supplied to the circulation with the feed clearly shows that both fish and plants use orthophosphates, al though a significant amount is excluded from circulation in the form of sediment.
The water pH also affects the possibility of phospho rus assimilation by plants. According to Yildiz et al. (2017), higher pH contributes to the creation of phos pho rus forms that are not accessible by plants.
With a stable level of pH, however, the main reasons for lower absorption of reactive phosphorus by plants were presumably the oxygen conditions and the formation of Ca 3 (PO 4 ) 2 and CaHPO 4 salts. Thus, loss of phosphorus from the cycle may have been mainly due to its entrapment in the resulting sediment. According to Schneider et al. (2005), 30−65% of the phosphorus from fish feed is lost due to chemical transformations, leading to the formation of phosphorus that is not accessible by plants and which is then removed from circulation by mechanical filtration. Yogev et al. (2016) estimated these losses to reach as much as 85%. A similar situation occurred in our experi ment, where the precipitation of phosphorus in sediments was estimated at 45%. Achieving stable conditions in the circulation and equilibrium between the concentrations of individual parameters in creases the availability of reactive phosphorus, which can be better ab sorbed by plants. Eck et al. (2019) proposed using ad ditional containers to the aquaponic system that would form a trap in which the phosphorus could be retained and then converted by aerobic and anaerobic processes into a form available to the plants and subsequently returned to the system. Despite the relatively high phosphorus losses, PUE was generally higher than nitrogen use efficiency (NUE) (Cerozi & Fitzsimmons 2017, Yep & Zheng 2019. Similar results were also ob tained in this experiment, where NUE averaged 16.6% (Tórz et al. 2021) and PUE averaged 21%.
The basic form of organic phosphorus in ecosystems is insoluble phytates. Enzymatic transformations with the use of phytases result in the release of these organic forms of orthophosphates, which are available to plants. Bacteria play a significant role in enzymatic transformations of phytates, showing significant activity in the rhizosphere. Increasing available phosphorus can be achieved through biochemical processes. Microorganisms also contribute to the control of pathogens (Jorquera et al. 2008, Lennard & Goddek 2019. Considering the method of plant breeding in our experiment (i.e. MFB), it should be as sumed that in this environment in the rhizosphere there could also be both microorganisms producing phytases (increasing the pool of available phosphorus in the system) and microorganisms controlling pathogens; therefore, to better understand the transformations of phosphorus in the IMTA systems, further experiments should be conducted that also take the microbiological aspect of the system into account.
Aerobic conditions play a decisive role in processes related to the transformation of biogenic compounds. Improvements in aquaponic systems can be achieved by introducing other microorganisms into the system. Fang et al. (2017) proposed introducing microalgae into the biological filter to increase NUE by 13.8%, increase dissolved oxygen and reduce nitrogen losses associated with N 2 O emissions. Moreover, micro algae can use the CO 2 produced by the bacteria, and they also contribute to a reduction in ammonium nitrogen in the water (Gilles et al. 2014). If there is an excessive accumulation of algae in the system, the microalgae can be a source of food for the fish, which is an additional benefit of their introduction into the system. Improving oxygen conditions through the use of microalgae can also contribute to improving phosphorus metabolism in the system to prevent loss of the forms of phosphorus available to the plants. The proposed solutions are interesting and promising and require further research.

Circulation of organic matter in an IMTA system
Accumulation of organic matter (feed residue, fish droppings, microorganisms developing in the sys tem) can adversely affect aquaponic systems. Some organic matter is necessary because it ensures the constant presence of nutrients in circulation (resulting from mineralization of the organic matter), but an excess can be harmful. Too much organic matter in the form of sludge creates conditions for an anaero bic environment and thus for the development of an aerobic bacteria that decompose the organic matter, contributing to the formation of carbon dioxide, ammonia or methane, which are harmful to fish (Yep & Zheng 2019). Moreover, if sediment accumulates around the roots of the cultivated plants, the re sulting anaerobic environment restricts penetration of oxygen into the root system, which is necessary for the proper uptake of nutrients by the plants (Yep & Zheng 2019). Anaerobic conditions can also contri bute to the formation of undesirable compounds such as geosmin and 2methylisoborneol produced by Streptomyces, for which organic matter containing phosphorus provides a desirable environment (Rurangwa & Verdegem 2015). Although these micro organisms are rarely found in aquaponics systems, primarily due to the aerobic conditions, the formation of compounds such as geosmin and 2-methylisoborneol can occur when there is a significant accumulation of organic sediment (Yep & Zheng 2019).
The increase in organic matter was related to the course of the experiment -along with the growth and development of the fish, the system was enriched with organic substances, and dissolved oxygen was consumed (Figs. 4A & 7). As mentioned earlier, there was no depletion of oxygen in the system at any time, so it should be assumed, based on Yep & Zheng (2019), that no undesirable compounds such as geosmin or 2-methylisoborneol were formed, which confirms that these substances occur sporadically in aquaponic systems. The highest reduction of organic matter occurred after the biological and hydro ponic filters, so it is most likely that in these filters, mineral forms of biogenic substances were released from the organic matter via biochemical processes. Thus, it can be concluded that the amount of organic matter found allowed for the constant presence of nutrients in circulation.
Depending on the manner of aquaponic breeding, variable amounts of both organic nitrogen and organic phosphorus can be obtained (Rakocy 2012, Rakocy et al. 2017, Yep & Zheng 2019). This does not change the fact that the amount of organic matter provided access to biogenic substances in the experiment. Without the presence of the organic matter and its mineralization, the aquaponic system would need to be enriched with the nutrients necessary for plant growth and development. Another important value of the presence of organic matter is that the microorganisms involved in its decomposition are antagonistic to the pathogens of the plant root systems and thus allow for healthy development of these systems (Yep & Zheng 2019). Mineralization of orga nic matter occurs throughout the entire system, whether in the mechanical, biological or hydroponic filters. As mentioned earlier, organic phosphorus is also a source of nutrients in aquaponic systems, and its reduction from the system is the result of mineralization and transformation into forms available to plants.

The calcium and magnesium cycle in an IMTA system
In aquaponic systems, biogenic macro-and microelements that may not be present in sufficient quantities for proper plant growth and development are calcium, potassium, magnesium and iron (Seawright et al. 1998). Reared fish usually have minimal re -quirements for these elements, hence their amounts are low in production feed. Calcium is important for the proper growth and development of plants. It is involved in the construction of cell walls, the permeability of cell membranes and is an important element in cell division. The availability of calcium also makes plants more resistant to bacterial and fungal infections (Liu et al. 2014). A deficit of calcium will manifest in reduced plant growth or deformation of the edges of young leaves (Maucieri et al. 2019). In turn, magnesium is part of chlorophyll, and a symptom of magnesium deficit is browning of the older leaves (Maucieri et al. 2019). As mentioned earlier, the primary source of all nutrients (including macroand microelements) in the aquaponic system is the remains of feed and fish excrement. An ideal situation would be if all feed added to the system was consumed by the fish. However, typically, part of the feed supplied (less than 5%) is not eaten by the fish and becomes a source of nutrients in the system (Yogev et al. 2016). The amount of feed left uneaten in an aquaponic system can be estimated from the FCR coefficient. Even a stable and relatively low value of the FCR does not guarantee full utilization of the feed, which makes it a source of all nutrients. The chemical composition of fish faeces depends on the fishes' diet, and it also affects the quality of the water in circulation (Goddek et al. 2015). The retention of nutrients such as macroelements in the system is closely related to the species and size of the fish, the method of feeding, the chemical composition of the feed and the water temperature in the system (Schneider et al. 2005). At higher temperatures, for example, fish metabolism is accelerated, resulting in a higher amount of nutrients in the form of dissolved salts in fish faeces (Turcios & Papenbrock 2014). According to Rafiee & Saad (2005), fish consume an average of 26.8% of the calcium and 20.3% of the magnesium available in the feed. The ionic forms of the macroelements available to the plants are formed by the mineralization of organic matter. However, individual macroelements may interact with each other and form insoluble forms that are not available to the plants. The balance between the available and unavailable macroelements is determined by the pH of the water. A well-known process in aquaponic systems is the precipitation of calcium in the form of Ca 3 (PO4) 2 , CaHPO 4 and CaCO 3 , as mentioned earlier (Peng et al. 2018). We assumed that the decrease in the concentration of calcium available for plant growth during our experiment was caused by the formation of insoluble salts. In the first phase, insoluble salts were formed from a combination of calcium and orthophosphates (Fig. 8). In the next phase, calcium bicarbonates could also be formed. Similar dependencies were described also by Cerozi & Fitzsimmons (2017). Calcium is more readily absorbed by plants at pH > 7 (Lennard & Goddek 2019). Considering that pH was kept above 7 during our experiment, it can be assumed that the pH of the water caused proper absorption of calcium by the plants. Villarroel et al. (2011) reported that in aquaponic cultures, magnesium concentration is about 4.0 mg dm −3 , while the optimal magnesium concentration for proper plant growth ranges from 20−50 mg dm −3 (Sonne veld & Voogt 2009). Over the entire experimental cycle, magnesium concentrations significantly exceeded the level observed by Villarroel et al. (2011), so it was available in sufficient quantity for plant development. As in the case of calcium, better assimilation of magnesium occurs at pH > 7, so in this case, the water pH caused proper absorption of magnesium by the plants. As presented in Villarroel et al. (2011), magnesium, similarly to calcium, can also be removed from circulation in the form of magnesium bicarbonates. In the present study, after the concentration of calcium ions had de creased, magnesium bicarbonates could be precipitated; however, this process did not significantly affect the availability of magnesium for plants, as during the whole experiment the concentration of magnesium ions was above 50 mg dm −3 .

CONCLUSIONS
The basic forms of orthophosphates in the IMTA system were H 2 PO 4 − and HPO 4 2− . A higher level of reactive phosphorus, which occurred over the course of the experiment, may have contributed to a decrease in calcium ion concentration due to the formation of Ca 3 (PO 4 ) 2 or CaHPO 4 , and thus inactivated part of the phosphorus available to the plants. Phosphorus may have been inactivated in the sediment due to the formation of Mg 3 (PO 4 ) 2 and MgHPO 4 in the second phase of the experiment. The water pH was maintained at 6.90−8.18, ensuring the efficiency of biochemical processes in the system and the proper use of nutrients by the plants. These con ditions were also suitable for carp breeding. The amount of organic matter gradually increased throughout the experiment, reaching a maximum of 204 mgO 2 dm −3 at the and of the experiment. Mineralization of organic matter took place under aerobic conditions. Organic matter was a source of biogenic substances in the IMTA. The range of general alkalinity and the observed bicarbonate concentrations suggest that the decrease in calcium ion concentrations was also caused by the formation of CaCO 3 and Ca(HCO 3 ) 2 . Magnesium ions did not show such a significant affinity for bicarbonate ions. Concentrations of calcium and magnesium ions were maintained at an optimal level for proper plant growth.
This experiment showed that achieving stable conditions in the aquaponic cycle and in the balance between the concentrations of individual parameters increases the availability of reactive phosphorus, which can be better absorbed by plants. Our experiment also showed that the selected method of plant breeding has an impact on the transformation of phosphorus in the aquaponic systems.