Dniel Mwendw Wmu, Ptrick G. Home, Jmes M. Rude, Stephen Ondimu

aDepartment of Soil, Water and Environmental Engineering, Jomo Kenyatta University of Agriculture and Technology, P.O. BOX 62000-00200, Nairobi, Kenya

bDepartment of Agricultural and Biosystems Engineering, Jomo Kenyatta University of Agriculture and Technology, P.O. BOX 62000-00200, Nairobi, Kenya

ABSTRACT

Recirculating Aquaculture Systems (RAS) offers a better option to increase aquaculture production with limited land and water resources while minimizing water pollution. The biggest challenges in RAS is to maintain favorable water quality for the fish to thrive. The practice of RAS in Kenya is minimal with the improper matching of RAS components and production densities for the existing few and in most cases, leading to a system failure. This study aimed at evaluating environmental and energy requirements for different production biomass of Nile tilapia in a RAS established in a controlled environment. Both production densities and water flow rates were varied while at the same time, the water quality parameters (dissolved oxygen, ammonia, pH, electrical conductivity and temperature) were monitored. The energy consumed for pumping and aeration was also monitored. Tilapia stocking biomass varied from 2.3 kg/m3 to 10 kg/m3 while flow rate was varied from 2.0 L/min and increased at intervals of 1.0 L/min to the maximum attainable flow rate of 10.0 L/min. Crushed pumice rock packed in a 1000 L tank and equipped with a bell siphon was used as the biofilter. Ammonia removal reduced with increasing flow rate with removal rates of 75% at 2.0–3.0 L/min to 9% at 8.0–10.0 L/min. The water pH increased with increasing flow rate with R2 ranging from 0.4 to 0.9. Electrical conductivity increased with flow rate from 112 mg/L to as high as 209 mg/L. Dissolved oxygen increased with flow rate and ranged from 1.0 mg/L to 7.0 mg/L. Energy consumption increased with flow rate and raged from 0.4 kW h at 2.3 kg/m3 stocking biomass to 2.4 kW h at 10 kg/m3 stocking biomass. High stocking densities and flow rates resulted to lower purification efficiencies and higher power consumptions than low stocking densities and flow rates. In order to maintain good RAS water quality, increased production and profits among farmers using RAS in Kenya,there is need to give proper advice on the right combination of stocking biomass, power, and water flow rate. In addition, similar studies should be carried out for other common fish species such as the African catfish.

ARTICLEINFO

Keywords:

Biofilter

Flow rate

Purification efficiency

Stocking biomass

Water quality

1.Introduction

Recirculating Aquaculture System (RAS) is a system that employs the principles of efficient water utilization and conservation with the aim of maximizing production of the target organism while minimizing pollution and water costs (Lekang, 2013). RASs used for farming aquatic organisms employs the principle of reusing the outlet water from the production systems instead of discarding it and getting new water for the system (Rahman, Verdegem, & Wahab, 2008). As a result, the quantity of new water required is reduced, thus reducing pressure on water supply systems. It is possible to recycle all the water from the production tanks such that the replacement of water will only be done to cater for evaporative loss or consumptive needs of the fish (Lekang, 2013).However, due to the high cost of treating and purifying the effluent from the production system, the possibility of 100 percent recycling is not usually achieved (Badiola, Mendiola, & Bostock, 2012).

RAS have been in existence as early as 1950’s although their potential to grow fish on a commercial-scale has only been realized in the last two decades (Badiola et al., 2012). Water treatment and testing technologies widely used in wastewater treatment have been extended to RAS upon the realization of the potential of RAS. The use of RAS around the world has been triggered by the increasing demand for white meat and decreasing trend of the fish harvested from the natural water bodies (FAO, 2016).

Issues of dwindling water and land resources can be addressed through the use of systems such as RAS that utilize less space and water to produce aquatic organisms (Avnimelech, 2006). According to Lekang(2013), due to the water conserving nature of RAS, aquaculture can be practiced in areas where water is a limiting factor. Moreover, production in established farms can be increased with the existing amounts of water. The quality and quantity of water leaving the production tanks differs from one system to another depending on the type of aquatic organisms being raised (Avnimelech, 2006; Kazimierz & Caffey, 1996).

This study aimed at determining the variation of environmental parameters with Nile tilapia (Oreochromis niloticus) production biomass and water flow rates in Recirculation Aquaculture Systems (RAS) and the energy requirements for environmental control of different production biomass.

Stocking densities of up to 15 kg/mhave been used successfully in cage systems and RAS studies (Gibtan, Getahun, & Mengistou, 2008;Ridha, 2006; Sri-uam et al., 2016). On the other hand, stocking densities below 3 kg/mhave been used in pond systems successfully.

A Recirculating Aquaculture System (RAS) includes the production unit which houses the aquatic organisms, a pump to transport the water around the system and water treatment system to remove contaminants from the effluent water, a pipe network joining, the production tank, the pump and the treatment system and sometimes an aeration component to add oxygen to the water (Lekang, 2013). The pump for recirculating water and the water filtration system for removing contaminants from the water are the items that make the RAS system distinct from traditional flow-through systems. Physical, chemical and biological processes are involved in the RAS water treatment system to improve the water quality to levels which the farmed species can tolerate and remain productive (Van Rijn, 2013). A good understanding of the RAS processes is important for one to design a good RAS. A good choice of biofilter may also prove very expensive to acquire.

RAS despite its many pros have a number of limitations (Pillay &Kutty, 2005). These includes the initial costs of installation, the operation costs and costs of maintenance.

With RAS,fish can be produced in the home backyards and the family white meat needs met at low costs and additionally, provide a source of income through the sale of the surplus. To achieve the efficiency and effectiveness of the RAS systems, water quality, which is one of the most limiting issue in all Aquacultural productions, needs to be addressed (Avnimelech, 2006; Pillay & Kutty, 2005).

Unfortunately, in Kenya, at present, there are no developed specifications and standard designs for various RAS stocking intensities making it difficult for potential aquaculture farmers to have reference guiding standards for an optimum RAS production system.

RAS is one type of the many modes of aquaculture production systems practiced in Kenya (KMFRI, 2017; Munguti et al., 2014). Most of the farmers practicing inland fish farming are using the through-flow systems which are usually extensive to semi-intensive. Therefore,increased production in aquaculture requires the use of systems with less water and land requirements such as the recirculating aquaculture systems.

2.Materials and methods

2.1.Methodology

The RAS components constructed and assembled included; a fish tank, a biofilter, connection of pipes and their fittings, pumps and aerators as well as installation of a greenhouse in which the complete set up was housed.

The laboratory setup of the recirculating aquaculture system was as shown in Plates 1 and 2. The RAS was housed in an 8 m by 15 m greenhouse.

Plate 1.The RAS systems in the greenhouse.

Plate 2.The RAS showing the production tanks, biofilter tank and sup tank among other components.

2.1.1.The production tanks

Three 1,000 L production tanks ma de of a good grade polyethylene were used. The RAS was operated with seven (7) stocking densities broken down into low (2.3, 3.5 and 4.0 kg/m), medium (5.0 and 7.0 kg/m) and high (9.0 and 10.0 kg/m) stocking densities. The system was run for one month before the stocking biomass was increased from one biomass to the next. At the end of the experiment at a given stocking biomass, the fish were weighed to determine their corresponding weight. The deficit weight to attain the next stocking biomass was then calculated and the fish of that weight introduced into the tanks. The production tanks were fitted with flush out pipes at the side near the bottom to allow for the removal of any solids from fish waste and uneaten feeds. Once in a week, 50–100 L of the water in the production tanks was drained to flush out settleable solids in the production tanks and then replaced with clean water (Badiola et al., 2012).

Upon completion of setting up the RAS components, the production tanks were stocked with Nile Tilapia (195 ±15 g). Fig. 1 presents the schematic layout of the RAS components.

Fig. 1.A schematic layout of the RAS components.

2.1.2.Feeding

The fish were fed with a 3 mm pelleted (25% protein, 7% fat and 7% fiber) feed. The feeding was done twice in a day at 9:00 a.m. and 4:00 p.m. by broadcasting the feed by hand on the water surface. The fish were fed approximately 4% of their body weight. Any excess feed floating on the water surface was scooped with a net and removed to reduce on the organic waste in the water.

2.1.3.The biofilter

Based on the expected ammonia production from the constructed RAS and the Biofilter nitrification rate estimated at 0.000857 g/m/day(Santamarina, Klein, Wang, & Prencke, 2002), the volume of the biofilter required was determined. This was achieved by considering the highest stocking biomass to be used for this study. The biofilter material used was packed in a 1000 L tank and operated as a trickling filter.Webb, Hart, Hollingsworth, and Danylchuk (2015) observed that an oversized biofilter be used to provide an additional space for aeration and degassing as the water goes through the biofilter media. Porous lava rock (pumice) was the substrate of choice that was used as a biofilter material for this study. Pumice is light in weight, highly porous and provides sufficient porosity for movement of water as compared to sand,gravel and other alternative substrates. The high density of pores in pumice provides a high specific surface area for growth of bacteria films.The biofilter was also inoculated with raw water from a nearby fish pond to accelerate the growth of the nitrifying bacteria (Gutierrez-Wing &Malone, 2006; Helfrich & Libey, 1991). The design of the biofilter was aimed at keeping the ammonia levels below 0.05 mg/L and nitrite levels below 0.5 mg/L (Badiola et al., 2012). To prevent entry of organic solids such as the fish droppings and uneaten feed that could have sunk in the production tank, a receiving sump installed at the point of entry of water into the biofilter was used to trap such solids. This sump was periodically removed, cleaned and returned into place.

2.1.4.Pump and aerator and connection pipes

A 0.5hp (0.37 kW) submersible pump was used to pump water from the sump after purification back into the production tanks. A gate valve fitted just after the pump was used to regulate the flow. Polypropylene(PPR) pipes of 25.4 mm internal diameter were used to connect the various tanks to allow for water recirculation. The choice of pipe diameter selected follows pump’s designer’s recommendation that, the used pipe diameter correspond to the pumps’ intake and discharge diameters to avoid reducing the pumps working efficiency (Larralde &Ocampo, 2010). Aerators with aeration rates of 15 L/min were used to pump air into the production tanks in order to boost oxygen availability.The pump’s air outlets were fitted with an atomizer (air stones) to produce microbubbles. The recirculated water was released from a height of about 0.5 m above the level of water in the production tanks so as to splash the water in the production tanks thus increasing infusion of oxygen into the water.

2.1.5.Measurement of flow rate, energy and water environmental parameters

Once the system was fully functional with all components in place and stocked with fish, the water quality parameters such as temperature,ammonia, dissolved oxygen (DO), pH and electrical conductivity (EC)were measured. Three (3) 250 ml samples were collected into a sampling cups from the center of the water surface in the production tanks and at the point of water exit from the biofilter tank on a daily basis (between 4:00–5:00PM). The parameters were measured within 15 min of sample collection for the different stocking densities and water flow rates. A newly acquired HQ40d HACH™ multimeter, a PHC101 pH probe,ISENH3181 ammonia probe, a LDO101 dissolved oxygen probe and a CDC401 electrical conductivity probe were used to take the readings.The procedures of measurement are as described in the HACH manual for the multimeter used (Hach Company, 1992). Flow rate was varied from 2.0 L/min and increased at intervals of 1 L/min to a flow rate of 10 L/min. The water flow rate was varied by use of a gate valve positioned just before the biofilter. The desired flow rate was then determined using the stopwatch and bucket method. This was done repeatedly until the desired flow rate was achieved. Each flow rates was maintained for 24 h(one day) and repeated three times. On the other hand, electricity consumption was measured by an installed electric meter. The pump would only start pumping when the float switch attached to the pump had been raised beyond a given level. The pumping then continued until the float switch drops to a minimum levels after which the sump tank starts to fill again. The difference in power consumption at different flow rates emanated from the fact that high flow rates would lead to rapid fill of the sump tank and hence more frequent pumping as compared to low flow rates. All the collected data was recorded in excel sheets. Purification efficiencies (PE) of biofilter at different flow rates were then computed for each set of flow rate for the various stocking densities on the basis of the amount of ammonia removed as presented in Equation(1) (Fletcher, Jones, Warren, & Stentiford, 2014)

Line graphs, bar graphs and measures of central tendency were generated from the collected data. Analysis of variance (ANOVA) of the environmental parameters were also conducted for the stocking densities and the corresponding flow rates. The ANOVA were conducted based on the randomized complete block design (RCBD) with the two primary factors being measured flow rate at corresponding stocking biomass. For every stocking biomass, all the flow rate levels were run and each parameter tested. The fish weight was measured during stocking and at the end of month, for each stocking biomass. The average weight gain and feed conversion ratio (FCR) was also computed from the weight data.

3.Results

See Tables 1–9 and Figs. 2–7.

4.Discussion

4.1.Water quality

The water quality was ascertained every time before more water was injected into the production tanks. The water parameter levels werewithin the ranges shown in Table 1. Compared to the ideal/recommended values, the water quality was found to be within the acceptable levels for culture of Nile tilapia.

Table 1Raw water quality parameter levels.

4.2.Environmental parameters for different production densities and water flow rates

4.2.1.Ammonia

Tables 2 and 3 show the concentrations of ammonia at different flowrates and stocking densities before and after the biofilter respectively.On the other hand Fig. 2 shows the variation of ammonia with flow rate at 9 kg/mstocking biomass. In most of the cases, the ammonia concentration increased with increase inflow rate. This is a clear indication of reducing ammonia removal with increasing flow rate. At low flow rates, the hydraulic retention time is longer as compared to that at higher flow rates leading better removal of ammonia at lower flow rates as compared to when the higher flow rates.

Table 2Average ammonia concentration (mg/L) in the production tanks at different flow rates and stocking densities.

Table 3Ammonia concentration (mg/L) after biofilter tank at different flow rates and stocking densities.

Fig. 2.A plot of ammonia versus flow rate at 9.0 kg/m3 in the production tank and after the biofilter.

Fig. 3.Variation of dissolved Oxygen with flow rate at 2.3 kg/m3 stocking biomass in the production tank and after the biofilter.

Fig. 4.Variation of electrical conductivity with flow rate at 4.0 kg/m3 stocking biomass.

Fig. 5.Variation of pH with flow rate at 2.3 kg/m3 stocking biomass.

Fig. 6.Variation of purification efficiency with flow rate at 4 kg/m3 stocking biomass.

Fig. 7.Variation of Energy for pumping and aeration with flow rate at 10.0 kg/m3 stocking biomass.

Table 4Dissolved oxygen concentration (mg/L) after biofilter tanks at different flow rates and stocking densities.

Table 5Electrical conductivity (mg/L) in the production tanks at different flow rates and stocking densities.

Table 6pH in the production tanks at different water flow rates and stocking densities.

Table 7Purification efficiency values (%) of the pumice biofilter.

Table 8Energy (kWh) consumed by the pump and the aerator.

Table 9Weight data offish at different stocking densities.

Tables 2 and 3 also shows a tendency of increase in ammonia concentration with increase in stocking biomass. This is because high stocking densities lead to higher ammonia generation as compared to low stocking densities.

The increase in ammonia concentration with flow rate tend to fit a linear function with the mean square value being as high as 0.88 at 7 kg/mand 9 kg/mstocking densities. The ammonia concentration levels recorded were substantially below the lethal levels (0.05 mg/L) for most flow rates in both the production tanks and the biofilter (Chen, Ling, &Blancheton, 2006). Similar observations were made by Ngugi, Bowman,and Omolo (2007) and Sri-uam, Donnuea, Powtongsook, & Pavasant,2016. From the analysis of variance at 95% confidence level, both variation of stocking biomass and flow rate had a significant effect on the ammonia concentration before and after biofilter. Sun, Wang & Liu(2016) also found out that, ammonia nitrogen was significantly affected by flow rate. In another study, Rahman et al. (2008) found out that additional stocking of Nile Tilapia led to increased concentration of Nitrogen and phosphorous nutrients in RAS water. In a study to test the efficiency of two biofilter media (polypropylene plastic chips and polyethylene blocks) media, Ridha and Cruz (2001) observed that the two filters were efficient in removing toxic ammonia and in maintaining the quality parameters within the acceptable and safe limits for the growth and survival of Nile Tilapia.

4.2.2.Dissolved oxygen

Fig. 3 shows the variation of DO with flow rate at 2.3 kg/m. Dissolved oxygen increased gradually with increasing flow rate at a given stocking biomass. The tendency of the dissolved oxygen to increase with flow rate can be attributed to the mixing of the waters in the production tank as the water plunges back into the production tanks with increasing force and velocities. . According to Downing and Truesdale (1955), there was a much more rapid increase in rate of oxygen dissolution with increasing rate of stirring or wind velocity.

Moreover, the reduced detention time with increasing flow rate leads to reduced oxygen uptake by the nitrifying bacteria; a phenomenon which would also lead to increase in oxygen concentration with increasing flow rate. The aerator which operated at a pumping rate of 15 L/min maintained the dissolved oxygen levels above 2.3 mg/L for most stoking densities. However, there was an observed decrease in oxygen concentration with increasing stocking biomass especially at higher stocking densities as shown in Table 4. According to Ngugi et al.(2008) and Sri-uam et al. (2016), the Nile Tilapia are a bit hardy and can survive low oxygen concentrations to levels below 3 mg/L. This makes the oxygen levels achieved in this study to fairly appropriate for the survival of Nile Tilapia. However, the growth rate at a given oxygen level was better at low stocking biomass as compared to higher stocking biomass.

Consequently, the RAS maintained favorable oxygen levels in the production tanks (5.0 ± 2.11 mg/L) and after the biofilter (4.38 ± 1.73 mg/L) respectively for the fish to grow and thrive except at 9.0 kg/mand 10 kg/mwhere the oxygen levels went below 2.3 mg/L and the fish seemed to gasp for oxygen especially at midday when the temperatures were highest (Mallya, 2007). Moreover, the oxygen concentration in the production tanks was relatively higher than that in the sump tank after the water flows through the biofilter (Zhang et al., 2011). This is illustrated by the case of 2.3 kg/mproduction biomass as shown in Fig. 3.This is because the nitrification process is an aerobic process in which oxygen is consumed to facilitate for the conversion of ammonium (NH)into nitrite and then into nitrates. The rate of dissolution with flow rate showed an approximately linear increase (Downing & Truesdale, 1955).Both variations inflow rate and stocking biomass had a significant effect on the oxygen concentrations of the RAS water as observed from analysis of variance at 95% confidence level. A significant difference in oxygen consumption by Nile tilapia at different stocking densities in a recirculating aquaculture was also observed by García-Trejo et al. (2016).

4.2.3.Electrical conductivity

The levels of electrical conductivity increased with increase inflow rate and stocking biomass with Rranging between 0.8 and 0.95 as shown in Fig. 4. This was attributed to the decreasing conversion of ammonia into nitrites and nitrates with increasing flow rates (Van Rijn,1996). Table 5 shows increasing EC with increasing stocking biomass.

The levels of EC were not different before and after the biofiltration.However a slight increase in EC of the water after passing through the biofilter was observed at higher flow rates as compared to lower flow rates (Zhang et al., 2011). From the analysis of variance at 95% con fidence level, both stocking biomass and flow rates showed a significant influence of the RAS water electrical conductivity. Kabir Chowdhury, Yi,Lin, & El-Haroun (2006) observed a significant effect on biomass growth which was an indication of reducing carrying capacity in Nile tilapia with rising salinity. However, in this study, the increase in electrical conductivity did not show any influence on the fish behavior and feeding habits.

4.2.4.pH

pH increased gradually with increase inflow rate both before and after biofilter. At a stocking biomass of 2.3 kg/m, it can be seen that the pH was higher in the production tank and lower in the sump tank after the biofilter as shown in Fig. 5. This is an indication of ammonium removal by the biofilter (Zhang et al., 2011). This is because during conversion of ammonium to nitrite and nitrates hydrogen ions are produced which then combine with the hydroxyl radicals leading to the lowering of pH. Ammonia varies proportionately with pH (Wurts, 2003).At lower ammonia levels, the pH is lower and at higher ammonia levels the pH is also higher. As more Ammonia is produced, more hydrogen(H) ions are taken up leaving hydroxyl radical (OH) to dominate and hence a rise in pH. On the other hand as ammonia gets removed in the water, ammonium breaks down to create an equilibrium (Miron et al.,2008). This breakdown process releases hydrogen ions thereby leading to a decrease in pH as presented in Equation 2.7 (Masser, Rakocy, &Losordo, 1999, p. 452).

Table 6 shows pH levels in the production tank at different flow rates.It was observed that, the pH levels in the production tank and after the biofilter were within the acceptable range 7.7 ± 0.49 and 7.49 ± 0.28 respectively for the fish and the nitrifying bacteria to thrive. From the analysis of variance at 95% confidence level, stocking biomass had a significant influence on pH as compared to flow rate.

4.2.5.Purification efficiency

The purification efficiency decreased sharply with increasing flow rate as shown in Fig. 6 at a stocking biomass of 4 kg/m. Ammonia removal reduced with increasing flow rate with a Rranging from 0.3 to 0.7 for most stocking densities. Table 7 shows how purification efficiencies vary with flow rates at the different stocking densities.

It was also observed that purification efficiencies decreased with increasing stocking biomass at a given flow rate. This was attributed to the increased ammonia production with increasing stocking biomass and reducing nitrification due to reducing detention time as the flow rates increased. Ebeling, Rishel, and Sibrell (2005) reported TSS removal efficiencies as high as 99% using polymers as flocculation aids in RASs.Similarly, Ebeling, Sibrell, Ogden, and Summerfelt (2003) reported TSS and phosphorous removal efficiencies of 89% and 93% while using alum and ferric chloride coagulants in a RAS system respectively. In a study using immobilized Ba-alginate and Ca-alginate beads to remove ammonia, 94% and 87% of loaded ammonia were removed within 3.4 h of hydraulic retention time (Kim et al., 2000). ANOVA for purification efficiency at 95% confidence level showed that, both variations inflow rate and stocking biomass had a significant influence on the purification efficiency.

4.2.6.Energy requirements for environmental control

The amount of energy consumed by the pump and the aerators increased progressively with flow rate as shown in Fig. 7 and Table 8.

From the observations, the right combinations of stocking densities and flow rate leads to lower cost of production and favorable environment for the Nile tilapia to thrive. Lower stocking densities require low flow rates and minimum aeration while higher stocking densities require higher flow rates and aeration to maintain favorable environment for the fish to thrive and minimize production costs. Park, Kim, Kim, and Jo(2008) in a study to determine the energy consumption and changes in water quality in a pilot-scale RAS monitored for 155 days reported total power consumption of 3925 kW. This power consumption included power for heating aeration and pumping. Approximately 11.22 kW and 22.44 kW of energy was consumed to produce 1 kg offish using RAS tanks and Biofloc technology (BFT) tanks respectively (Luo et al., 2014).

4.2.7.Fish growth

The weight data is as presented in Table 9. The percentage weight gain was slightly higher at low stocking densities than at high stocking densities (García-Trejo et al., 2016; Gibtan et al., 2008). Conversely, FCR increased with increasing stocking biomass. The lower weight gain at high stocking densities was attributed to stress and poorer water quality at higher stocking densities (Verster, 2017). The RAS maintained optimal water temperatures (26.5 ± 2.55C) for the Nile tilapia to survive (Santos, Mareco, & Dal Pai Silva, 2013).

5.Conclusion

•For each stocking biomass, high flow rates led to poor environmental conditions than at lower flow rates. Ammonia removal was highest at lower flow rates with percentage removals above 70% and reduced gradually to as low as 9.5% at high flow rates. Dissolved oxygen reduced progressively with increasing stocking densities from as high as 7.0 mg/L at 2.3 kg/mto as low as 1.2 mg/L at 10 kg/mpH levels ranged between 6.9 and 8.5 for all the flow rates. Electrical conductivity increased with increasing stocking densities and flow rates to values above 190.0 mg/L at 9.0 L/min and 10.0 L/min.

•Energy consumption increased with increase inflow rates for each stocking biomass from 0.5 kwh at 2 L/min to 2.0 kwh at 10 L/min.

CRediT authorship contribution statement

Daniel Mwendwa Wambua: Conceptualization, Methodology,Investigation, Formal analysis, Formal analysis, Investigation, Writing -original draft. Patrick G. Home: Supervision, Funding acquisition,Project administration. James M. Raude: Supervision. Stephen Ondimu: Supervision.

Declaration of competing interest

The authors declare that there is no conflicts of interest.

Acknowledgements

We acknowledge the financial support from Japan International Cooperation Agency (Africa-ai-Japan project) grant numbers, IPIC/03/17 and Interdisciplinary 2 which enabled us to accomplish this research.