Báo cáo Nghiên cứu khoa học Towards Zero Discharge of Wastewater from Floating Raceway Production Ponds (Milestone No. 5)

Ministry of Agriculture & Rural Development PROGRESS REPORT Intensive in-pond floating raceway production of marine finfish (CARD VIE 062/04) MILESTONE REPORT NO.5 Development of a zero-discharged system Report Author: Michael Burke, Tung Hoang & Daniel Willet December 2007 1 AUSTRALIA COMPONENT 2 Towards Zero Discharge of Wastewater from Floating Raceway Production Ponds (Milestone No. 5) D.J. Willett1, C. Morrison1, M.J. Burke1, L. Dutney1, and T. Hoang2 1Department of Primary Industries and Fisheries, Bribie Island Aquaculture Research Centre, Bribie Island, Queensland, Australia. 2Nha Trang University, International Centre for Research and Training, NHATRANG City, Vietnam Correspondence: Daniel Willett, Bribie Island Aquaculture Research Centre, PO Box 2066 Bribie Island, Queensland, 4507 Australia. daniel.willett@dpi.qld.gov.au EXECUTIVE SUMMARY A major problem with intensified pond-based aquaculture production systems has been managing water quality and discharge quotas due to the accumulation of waste nutrients. This is exacerbated in the current CARD project which demonstrated the very high production capability of in-pond raceways in excess of 35 ton/ha of combined mulloway and whiting. While the current operation managed water quality through exchanging water (approximately 10% per day on average – see MS No.4), it is recognised that with water conservation issues and environmental nutrient discharge impacts, flushing pond water to waste is a less desirable solution. One of the original goals of this project was to investigate strategies that limited water discharge to show that raceway production of fish could be sustainable. This report summarises details of water remediation strategies investigated to progress towards zero water discharge. Waste sumps were installed into the raceways as a proposed means for collecting and concentrating uneaten feed and faeces, thereby reducing nutrients entering the ponds. A trial tested the effectiveness of these solids traps by comparing Total Solids, TN and TP collected in the sump with those flowing out of the raceway through the end screens. Results showed that the waste sumps are generally not effective at concentrating solids for periodic removal. This was primarily due to flow dynamics within the raceways causing eddies to form that keep solids from going down into the collector. In addition, fish within the raceways continually stir up and resuspend particulate waste, allowing it to be expelled into the pond. However, the sumps may be useful as a discharge point in a remediation system which recirculates pond water via an external treatment pond. 3 An original objective of the project was to investigate the culture of the red marine macrophyte Harpoon Weed (Asparagopsis armata) as a nutrient sink. While much previous research at BIARC has looked to develop seaweed biofilters for pond-based aquaculture, the culture of A. armata was novel and offered advantages over commonly used green seaweed species, according to new literature. Several attempts to collect seed stock and culture the specific tetrasporophyte phase of this species however proved problematic and the seaweed failed to thrive and eventually died. Specific factors responsible are discussed. Concurrent research at BIARC is developing technologies that overcome many of the common impediments to seaweed culture and these are discussed in light of future work evaluating A. armata as a biofilter. Recent international research has demonstrated the successful use of bacterial-based processes (termed Bio-floc treatment) for water quality management in pond-based aquaculture. The concept involves manipulating substrate Carbon:Nitrogen ratios to promote heterotrophic nutrient assimilation. A series of experiments were conducted to determine whether bio-floc treatment may be incorporated effectively as part of the raceway production system, specifically as an external component of a recirculating system. The trial defined a Carbon dose rate that achieved almost complete elimination of toxic N species (TAN and NOx) from raceway effluent within 12 hours and prolonged the period prior to remineralisation. A successful shift from a phytoplankton-dominated waste stream to a bio-floc community was also achieved by applying this carbon dose in a replicated continuous-flow treatment system. The bio-floc community was characterised by lower, stable pH (8.0-8.2) and DO (6.9-8.8) levels, increased biomass and a decreased proportion of phytoplankton present. This demonstrated that effluent treated in an external bio-floc pond would be suitable for recirculation, and a schematic of a proposed integrated production system is presented. Of the wastewater remediation strategies investigated in this project, it is evident that bio- floc treatment was the most promising technology to progress towards zero water discharge. INTRODUCTION A major goal of this CARD project was to develop a pond-based fish production system that is both sustainable and profitable, designed to increase production and improve stock 4 management efficiencies and ultimately make better use of existing unprofitable aquaculture pond infrastructure in Australia and Vietnam. The development of low-cost in- pond Floating Raceways (FRs) in this project has demonstrated an innovative approach to larval rearing, juvenile nursery and fish growout. As reported in Milestone No.4, the FR system within a pond a Bribie Island Aquaculture Research Centre demonstrated production capability in excess of 35 ton/ha of combined mulloway and whiting. An inherent problem of any pond-based production system is the accumulation of residual organic matter (uneaten feed, faeces) and toxic inorganic nitrogen (specifically ammonia). Even the best practices cannot avoid this since it has been shown that fish and shrimp only assimilate on average about 25% of ingested food – the rest being excreted into the water column predominately as ammonia (Boyd & Tucker 1998; (Funge-Smith and Briggs 1998; Hargreaves 1998). This feeds phytoplankton blooms which are at best only a partial nutrient sink in ponds stocked at densities above 5 ton/ha (Avnimelech 2003; Brune et al. 2003). Dense phytoplankton blooms can cause lethal DO and pH fluctuations and their overgrowth can lead to bloom crashes and subsequent release of ammonia (Krom et al. 1989; Boyd 1995; Boyd 2002; Ebeling et al. 2006). Water exchange is usually required to alleviate this problem and maintain suitable pond water quality; however with water conservation issues and environmental nutrient discharge impacts, flushing pond water to waste is becoming a less desirable solution. Clearly, production of fish in the order of 35 ton/ha as demonstrated in this project cannot be maintained without a means to remediate or exchange water. The current project managed water quality using secchi depth as gauge of appropriate conditions and by exchanging water (approximately 10% per day on average – see MS No.4). One of the original goals of this project was to investigate strategies that limited water discharge. A number of strategies were proposed, including the culture of Harpoon Weed (Asparagopsis armata) as a nutrient sink; partitioning ponds to into ‘fish culture’ and ‘remediation’ zones; and manipulating Carbon:Nitrogen ratios to promote bacterial nutrient processing. This report will summarise details of water remediation strategies investigated, with particular emphasis on partitioned bacterial nutrient processing as it became evident that this was the most promising technology to progress towards zero water discharge. Strategy 1: Raceway sump to trap solids 5 Background: Reducing direct nutrient input into production ponds reduces pressure on biological remediation processes. The regular removal of uneaten feed and faeces directly from raceways before it is allowed to enter the pond will prevent further nutrient release and mineralisation from this waste source over the production period. The amounts of these settleable solids within floating raceways will vary depending on feeding rates and efficiencies. In turn, the ability to harvest these solids depends on flow dynamics within the raceways and the design of the solids trap. A preliminary experiment was designed to gauge the effectiveness of a solids trap built into the raceways as a means for reducing nutrients entering the ponds. Methods: Plastic stormwater drain sumps were inserted into the tail end floor of each raceway as a solids trap (Fig 1). These sumps were connected via a flexible hose to a pump on a timer which periodically (twice daily) pumped collected waste to a holding tank for evaluation of nutrient content. On monthly occasions between February and October 2006, water leaving the raceways through the end screen was also sampled and nutrient data was compared with that from the sump waste to determine differences. Water quality analyses evaluated Total Solids (TS), Total Nitrogen (TN) and Total Phosphorous (TP), and were determined using validated laboratory protocols based on standard methods (American Public Health Association 1989) and nutrient analysis equipment at BIARC. 6 Figure 1. Design and configuration of the solids trap inserted within the nursery raceways. A plastic grate cover (not shown) prevented fish from entering the sump. Results & Discussion: Nutrient analyses showed some small differences in concentration between water pumped from the sump and water leaving the raceways through the end screen (Table 1.) The greatest difference was with TS, where the sump captured on average 16% more solids than water discharged from the pond. Differences in TN and TP between sump and raceway screen were smaller but still showed a marginally greater average nutrient removal via the sump. This data cannot be statistically validated however because monthly data from the raceway was from a single water sample (due to budgetary constraints) whereby no measure of error rate can be determined. Regardless, the sump was designed to trap and concentrate solids into a thick sludge that could be periodically removed from the pond. It was clear that only a slightly more concentrated effluent was captured by the sumps and their role in preventing nutrients entering the pond from the raceways was limited. This suggests that the waste sumps are not effective at collecting solids for periodic removal. However, they may be useful as a discharge point in a remediation system which recirculates pond water via an external treatment pond. It is an advantage, in this instance, to discharge the most concentrated effluent as possible into the 7 treatment pond, and this was employed in subsequent bio-floc remediation trials (see below). Similar waste removal systems were employed by Koo et al. (1995) in in-pond raceways developed for channel catfish, and likewise their waste removal system showed poor performance. The primary problem was due to inefficient settling of waste in the solids collectors. A known difficulty with raceways is that when solids reach the end of the tank, the hydraulic forces do not efficiently concentrate the solids around the drain. Water reflected off the end wall generates turbulence, causing eddies to form that may keep solids from going down into the collector (Van Wyk, 1999). In addition, fish within the raceways continually stir up and resuspend particulate waste, allowing it to be expelled into the pond. Table 1. Differences in water collected from the solids trap and water leaving the raceway through the end screen, over seven months (n=7). Constituent Mean concentration in water expelled from raceway (mg/L) Mean concentration in water from sump (mg/L) Total Solids 15.4 18.35 Total Nitrogen 2.07 2.33 Total Phosphorous 0.78 0.83 Strategy 2: Evaluation of Harpoon Weed Summary: The concept of using seaweeds as biofilters for removing waste nutrients from fish and shrimp aquaculture operation is well known, with a seminal review by Neori et al (2004) describing the state of the art of this technology. Presently, the most commonly proposed and researched biofilters are green seaweeds from the genus Ulva and the red seaweed Gracilaria. Yet, in practice most seaweed-based remediation systems have proven not to be economically viable, mainly due to the low value of the produced seaweed and the high labour and area requirements for its cultivation. Other physical impediments to the culture of seaweeds in effluent from aquaculture ponds include their susceptibility to epiphytism (Friedlander et al., 1987), infestation by grazers such as amphipods, and 8 competition for available nutrients with phytoplankton (Palmer 2005). These difficulties are compounded by the accumulation of effluent particulate matter on the seaweed’s surfaces. The result therefore in practice, is that growth rate of the seaweeds (and their corresponding value as a nutrient sink) is very often limited and nutrient removal efficiencies are below optimum rates achieved in scaled trials under more favourable conditions (Palmer 2005; previous BIARC research). The present CARD project proposed to investigate the performance of the red seaweed Asparagopsis armata (also known as Harpoon Weed) as a sink for waste nutrients generated in raceway production system. This species was selected on the basis of new work by Schuenhoff & Mata (2004) which suggested that it had considerably greater market value than other seaweeds due to its high concentration of halogenated organic metabolites. Once extracted, these halogenated compounds are used for antifouling and in the cosmetic industry as fungicides. Schuenhoff & Mata (2004) suggest that these compounds are also responsible for limiting epibiota and epiphytes in culture – an advantage over other cultured seaweeds. In addition, its reported removal rate of ammonia is superior to that of Ulva species and it is also a native species to Australia (Fig 2). Figure 2. Harpoon weed (Asparagopsis armata) growing on rocks in Moreton Bay, S.E. Qld. Photo by Marine Botany Group, University of Qld (2003) 9 A proposal was drafted to collect harpoon weed from Moreton Bay as a seed stock to trial its growth rate and nutrient uptake under effluent conditions generated in the raceway pond at BIARC. In particular, it is the tetrasporophyte phase of the plant that is reported useful for biofiltration. Several collecting expeditions were mounted in conjunction with marine botanists from the University of Qld. Only a small amount of harpoon weed in its tetrasporophyte phase was located. It was transferred to a production unit at BIARC and supplied with pond effluent in order to cultivate larger quantities for use in a replicated bioremediation trial. Unfortunately, the harpoon weed failed to thrive and eventually died preventing the trial being conducted. It is uncertain whether seasonal or effluent-specific factors were responsible. Given the previous considerable work conducted at BIARC evaluating seaweed biofilters and the difficulty in locating, collecting and culturing this specific macrophyte, plans for further trials were terminated for the current project. Future work in evaluating this species as a biofilter, however, is planned as part of ongoing BIARC wastewater remediation studies. Based on current research at BIARC on seaweed biofilters, to effectively incorporate seaweeds into a bioremediation system for pond-based aquaculture it appears that pre- treatment of the effluent would be necessary so that competing plankton levels, fouling organisms and suspended materials are reduced, and so that nutrients are converted into forms available for direct plant uptake. Current work at BIARC, outside of the CARD project, is assessing the role of polychaete-aided sand filtration as one such pre-treatment option (Palmer 2007). Strategy 3: Bacterial nutrient processing Background: There is now recognition that promoting a swing from autotrophic (phytoplankton-based) to heterotrophic (bacterial-based) processing of residual pond nutrients has many advantages for water remediation. Sewage effluent treatment has long employed bacterial digestion of organic matter in activated sludge systems (Arundel 1995) and more recent studies have shown that suspended growth systems, where heterotrophic- dominated processes regulate water quality, have great application for limited-water- exchange shrimp and tilapia production (Avnimelech 1999; Burford, et al. 2003; Erler et al. 2005). In aquaculture, these heterotrophic-dominated growth systems are generally termed Bio-floc systems. 10 The challenge is to determine the best configuration for incorporating biofloc treatment as part of the raceway production system. Two approaches are possible: in-pond biofloc treatment or external biofloc treatment as part of a recirculating system. Most studies on using bio-floc water remediation for aquaculture have advocated floc formation within the culture pond as a supplementary source of dietary protein (Avnimelech 1999; McIntosh et al. 2001; Erler et al. 2005) in addition to controlling water quality. While increased feed utilisation is ideal, the excessive turbidity and high oxygen demand created by bio-flocs may have a negative effect on fish cultured within floating raceways. The high DO demands of the floc colony in addition to those of the cultured species means that cultured stock are even more vulnerable in the event of any aeration failure, especially in intensive production systems such as floating raceways. High suspended solids levels can foul the gills of cultured animals and lead to bacterial, protozoan and fungal infections (Boyd 1994). In addition, not all cultured species will access or target the additional protein source provided by the bacterial flocs – especially higher order species (non filter feeders). Alternatively, establishing a bio-floc zone as a component of a treatment system external to the culture pond (i.e. post-production) is a new approach for this technology and may be more suited to FR production for the reasons detailed above. Waste nutrients potentially could be captured within bio-flocs, which in turn are periodically harvested from the water in isolation from the cultured stock. Significantly cleaner supernatant could then be returned to the culture pond. While sedimentation ponds are routinely used in Australia to treat post-production wastewater, local studies have shown they are generally ineffective at reducing Total Nitrogen, mostly due to remineralisation and inadvertent discharge of the dominating phytoplankton (Preston et al. 2000; Palmer 2005). Directly harvesting phytoplankton is difficult and generally cost prohibitive to farmers, so a need exists for a new approach to enhance the performance of post-production treatment ponds. For a Bio-floc Pond (BFP) to effectively operate as a post-production wastewater remediation system there must be mechanisms for converting phytoplankton-dominated wastewater into a bio-floc community which packages nutrients into the more harvestable ‘floc’ form. A key mechanism for promoting heterotrophic assimilation of waste nutrients is through the manipulation of substrate carbon:nitrogen (C:N) balance. Heterotrophic 11 bacteria utilise organic carbon as an energy source, which is required in conjunction with nitrogen to synthesize protein for new cell material (Avnimelech 1999). For the bacteria to metabolise available nitrogen efficiently into the floc, carbon must not be limiting. Therefore, maintaining an appropriate C:N ratio by adding carbonaceous material is necessary. Theoretical carbon requirements can be calculated based on the C:N ratio of bacterial biomass, bacterial carbon assimilation efficiency and the bio-available N levels in the pond water (Hargreaves 2006). While a quantitative rationale for estimating C additions was described by (Avnimelech 1999), his equation was based on total ammonia nitrogen (TAN) residue. A complication is that TAN is not the only form of nitrogen available to heterotrophic bacteria. Dissolved organic nitrogen (DON) in particular, but also nitrite and nitrate can constitute a varying but substantial portion of bio-available N in aquaculture wastewater (Preston et al. 2000) and bacteria may scavenge these in addition or in preference to ammonia (Jorgensen et al. 1994). Therefore, C additions based solely on TAN level may be under-dosing. Calculating real-time (i.e. on-the-day) bio-available N levels is difficult (particularly for DON which requires laboratory digestion and analysis) whereas daily in-the-field testing of TAN is standard practice, so we acknowledge the validity of Avnimelech’s (1999) suggestion to use TAN as a convenient reference to gauge C requirements. The objective of this study was to refine C dosing requirements based on real-time TAN readings for more complete nutrient assimilation in discharged wastewater. A further objective was to assess the ability to convert plankton-dominated wastewater into a bio-floc community using these established C dose rates, within pilot-scale external treatment ponds. Methods: A series of experiments were carried out at BIARC during 2006. The wastewater source was the discharge from the sumps of the FRs containing the mulloway and whiting. Molasses (37.5% C) was the carbohydrate source used to adjust substrate C:N ratios in both experiments because it contains simple sugars, negligible nitrogen, is readily available and relatively inexpensive. Experiment 1 This trial investigated the effect of molasses addition at two application rates on wastewater nutrient levels over a 48 hour period. Nine 3L tanks were filled with common 12 wastewater and supplied with continuous aeration to ensure thorough mixing. The experiment was conducted in the dark to prevent photosynthesis. Three treatments in triplicate were tested: Control, Molasses 1 and Molasses 2. Molasses doses were based on the following equation (adapted from Avnimelech 1999): Cadd = Nww x ([C/N]mic/E) Where: Cadd is the amount of C required Nww is the bio-available N in wastewater [C/N]mic is the C:N ratio of bacterial biomass [typically about 5 (Moriarty 1997; Hargreaves 2005)] E is the bacterial C assimilation efficiency [assumed to be 0.4 (Avnimelech 1999)] Therefore: Cadd = Nww x 12.5 According to this equation, 12.5 g C is needed to convert 1 g bio-available N into bacterial biomass. Given that molasses is 37.5% C, 33.3 g of molasses is needed to convert 1 g bio- available N. A stock solution of molasses was prepared (100 g molasses L-1 = 37.5 g C L-1) to aid addition to the experimental tanks. Molasses 1 treatment was a single molasses dose based on Nww = the real-time TAN level measured in the wastewater immediately prior to filling experimental tanks. 'Molasses 2' treatment was based on double the amount of Molasses 1 to account for the extra ‘unmeasured’ bio-available N present. No molasses was added to the Control treatment. After molasses addition, two 50mL water samples (one filtered [0.45um] & one unfiltered) were taken from each tank at regular intervals (0, 3, 6, 12, 24, 48 hrs). Nutrient concentrations in the water samples were measured including Total Nitrogen [TN], Total Phosphorus [TP], Total Ammonium Nitrogen [TAN], Nitrate/Nitrite [NOx], and Dissolved Inorganic ortho-Phosphate [DIP]), Dissolved Organic Nitrogen [DON] and Dissolved Organic Phosphorus [DOP]. Measurements were conducted using validated laboratory 13 protocols based on standard methods (American Public Health Association 1989) on a Flow Injection analyser at BIARC. Data was statistically analysed using Arepmeasures with treatment and time as parameters on Genstat 8th Ed Software. Experiment 2 This trial tested the efficacy of shifting a plankton-dominated wastewater stream to a Bio- floc community, using previously established C dose rates in a pilot-scale treatment system. Wastewater was distributed into four concrete raceways (each 8.6m x 2.7m x 0.8m; Volume: 19,000L). Two raceways were established as replicate Bio-floc Ponds (BFPs) and the remaining two as replicate Passive Settlement Ponds (PSP) (see Figure 3). A two-day effluent retention time was tested. This is equivalent to a water exchange rate of 20% of production pond water per day into a treatment system that occupies 30% of farm pond area (as this is typical of many Australian aquaculture farms using ponds), and represents the most challenging, realistic demand a treatment system is likely to experience. Flow of effluent through the treatment raceways was continuous to enable more accurate monitoring. Figure 3. Simulated post-production treatment ponds in the remediation trial showing Bio-floc Pond (BFP) on left and Passive Settlement Pond (PSP) on right. 14 To simulate real conditions in the Passive Settlement Pond (PSP), there was no additional aeration or stirring provided and wastewater discharged from the surface through a standpipe. The Bio-floc Pond (BFP) used vigorous aeration with diffusers to ensure thorough mixing and to restrict anaerobic zones within the raceway (Fig 3). Organic carbon was added proportional to influent ammonia level as required to maintain prescribed C:N ratios (as determined in Experiment 1), and averaged 200 ml of Molasses every 2 days. Weekly monitoring involved assessing untreated (influent) and treated discharged water quality. A YSI multiprobe meter measured the Standard parameters (pH, temperature, salinity, dissolved oxygen [DO]) during the experiment. Methods for determining nutrient concentrations, total suspended solids [TSS], and Chlorophyll A [Chl-a] were as described for Experiment 1. Measurements assessed differences between bio-floc treatment and standard phytoplankton-dominated PSP treatment. In addition, differences between the (untreated) influent and post-treatment water were measured to assess the efficiency within each treatment system. Changes in water quality parameters were statistically analysed using Arepmeasures with treatment type and time as parameters on Genstat 8th Ed Software. Results: Experiment 1 Results for each constituent tested are described in detail in the paragraphs below and displayed graphically in Figure 4. Nitrogen TAN levels in the un-dosed Control treatment increased significantly (p>0.01) during the trial period. In contrast, at just three hours after a single addition of C, TAN levels in the two molasses treatments had fallen by over 35% and were significantly (P>0.01) lower than the control. By six hours TAN removal remained consistent between the two molasses treatments with over 65% of TAN removed from the water. However, beyond six hours TAN in the lower dose (Molasses 1) treatment began to rise again, suggesting the exhaustion of available C supplies before complete ammonia assimilation occurred. The 15 higher dose (Molasses 2) continued to decrease significantly (p>0.01) so that after 12 hours, ammonia was virtually eliminated (96% removal). TAN levels began to increase significantly (p>0.01) again after 24 hours in Molasses 1 and after 48 hours in Molasses 2, presumably due to degradation of senescing phytoplankton not accounted for Initially (3-6 hrs) the un-dosed Control treatment experienced a significant (p>0.01) release of DON before maintaining the elevated level for the duration of the experiment. In contrast, the addition of C provided a subdued and delayed (6-12hr) release of DON. However 24 hours after C addition DON was significantly (p<0.01) reduced by 30% with the lower C dose treatment (Molasses 1) and 85% with the higher dose (Molasses 2). The DON levels returned to similar levels at the conclusion of the experiment 48 hrs after C addition, suggesting an exhaustion of the available C The TN levels were not significantly influenced (p>0.05) by C addition for the experimental period. This Suggests the C addition can significantly influence the nutrient processes without impacting the nutrient budget. NOx levels were tested however the levels were negligible or below detectable levels throughout the experimental period. High C:N ratios typically inhibit nitrification and nitrifying bacteria are often out-competed by heterotrophic bacteria. Phosphorus The DIP levels followed the same trends as the TAN levels. The un-dosed Control treatment increased significantly (p>0.01) during the trial period. Again, 6 hours after the addition of C, DIP levels remained consistent between the two molasses treatments (with 50% of DIP removed), but after 12 hours the lower dose (Molasses 1) commenced rising while the higher dose (Molasses 2) continued to decrease significantly (p>0.01) to almost completely eliminating DIP (93% removal). DIP levels also began to rise significantly after 24 hours in Molasses 1 and 48 hours in Molasses 2 as seen in the TAN levels. DOP levels were significantly (p>0.05) lower in the Control samples but the level of C dose did not significantly (p<0.05) effect the response. 16 Similarly to TN levels, C addition did not significantly effect (p<0.05) TP levels during the experimental period. Again suggesting the C addition can significantly influence the nutrient processes without affecting the nutrient budget. 0.0 0.5 1.0 1.5 2.0 2.5 3.0 0 6 12 18 24 30 36 42 48 HOURS TA N m g/ L Control Molasses 1 Molasses 2 0.0 0.1 0.2 0.3 0.4 0.5 0 6 12 18 24 30 36 42 48 Hours D IP m g/ L Control Molasses 1 Molasses 2 0 2 4 6 8 10 0 6 12 18 24 30 36 42 48 HOURS D O N m g/ L Control Molasses 1 Molasses 2 0.0 0.5 1.0 1.5 0 6 12 18 24 30 36 42 48 Hours D O P m g/ L Molasses 2 Molasses 1 Control 0 4 8 12 16 20 0 6 12 18 24 Hours TN m g/ L Control Molasses 1 Molasses 2 0.0 1.0 2.0 3.0 4.0 5.0 0 6 12 18 24 Hours TP m g/ L Control Molasses 1 Molasses 2 Figure 4: Nutrient levels over the experimental period in controls and at two molasses doses. Experiment 2 Standard Parameters Phytoplankton-dominated PSP treatment systems are characterised by the high pH (<8.5) and DO (<8 mg/L) levels measured during the trial (See Figure 5). In the BFP treatment both the DO & pH levels were significantly (p>0.05) lower compared to the PSP which suggests a successful shift away from a phytoplankton dominated community (Funge- Smith and Briggs 1998). Significant (p>0.05) fluctuations within the PSP (pH 8.14-9.08; 17 DO 9.74-19.16) treatment demonstrated the dangerous bloom/crash cycling typical in this type of community (Hargreaves 2006). While the BFP (pH 8.00-8.17; DO 6.86-8.80) system maintained consistent levels during the experimental period. 7.0 7.5 8.0 8.5 9.0 9.5 1 2 3 4 5 6 7 8 9 10 11 12 Week pH PSP BFP 6 8 10 12 14 16 18 20 22 1 2 3 4 5 6 7 8 9 10 11 12 Week D O m g/ L PSP BFP Figure 5: Water Quality measurements for pH and Dissolved Oxygen (DO) Temperature and Salinity remained within biological limits for both systems. As expected, the temperature was similar in both systems (15.3 – 21.0 OC) on most occasions. Salinity showed significant (p>0.01) fluctuations over time for both treatments due to rain events. The salinity of the BFP was significantly(p<0.01) lower than PSP on a number of occasions probably due to the more effective mixing of rain water which can float on top of still seawater in the PSP. Nutrient Analyses In general, both treatments significantly (p<0.05) lowered the dissolved nutrients levels present in the untreated water. The inorganic nitrogen (TAN and NOx) was effectively eliminated from the untreated water by the BFP treatment. The BFP treatment preformed significantly better than the PSP treatment for NOx (p<0.01) and DIP levels (p<0.01). Importantly, this suggests a more efficient removal of the toxic components of wastewater occurs in the BFP treatment (See Figure 6). TN &TP levels in the BFP treatment were significantly (p<0.01) higher than levels present in PSP. The BFP treatment also significantly (p<0.01) increased the TN levels from the untreated water (influent). In contrast, the PSP significantly reduced the TN levels of the Untreated water suggesting PSPs are more efficient at overall nutrient removal at this stage. The high levels of TN & TP suggest efficient processing and assimilation of nutrients to biomass. 18 0.0 0.5 1.0 1.5 1 2 3 4 5 6 7 8 9 10 11 12 WEEK g/ L N O x m UNTREATED PSP BFP 0.0 0.2 0.4 0.6 0.8 1 2 3 4 5 6 7 8 9 10 11 12 WEEK TA N m g/ L UNTREATED PSP BFP 0.0 0.1 0.2 0.3 0.4 0.5 1 2 3 4 5 6 7 8 9 10 11 12 WEEK DI P m g/ L UNTREATED PSP BFP 0.0 1.0 2.0 3.0 4.0 5.0 1 2 3 4 5 6 7 8 9 10 11 12 WEEK TN m g/ L BFP UNTREATED PSP 0.0 0.5 1.0 1.5 1 2 3 4 5 6 7 8 9 10 11 12 WEEK TP m g/ L BFP UNTREATED PSP 0 20 40 60 80 1 2 3 4 5 6 7 8 9 10 11 12 WEEK TS S m g/ L BFP UNTREATED PSP 0.0 0.5 1.0 1.5 2.0 1 2 3 4 5 6 7 8 9 10 11 12 WEEK D O N m g/ L BFP PSP UNTREATED 0 20 40 60 80 100 120 140 1 2 3 4 5 6 7 8 9 10 11 12 WEEK C hl A u g/ L BFP UNTREATED PSP Figure 6: Nutrient levels during the experimental period in untreated influent and from bio-floc ponds and passive settlement ponds. Two characteristics of the BFP system explain the elevated nutrients levels. Firstly the BFP suspends and digests the organic matter (nutrients) within the water column. Secondly, the formation of bio-flocs (with the efficient digestion of nutrients) means that nutrients can become concentrated within water column of the BFP thus providing the elevated TN & TP levels. As there were significantly (p<0.05) higher DON levels detected 19 in the BFP treatment than in the Untreated water, N may be accumulating in a refractory DON form as suggested by other researchers such as (Erler et al. 2005). In contrast, within the PSP system organic material (nutrients) settles out of the water column, but later reminerialises causing the elevated levels normally seen in PSPs later in the season (Preston et al. 2000). Improved containment of the bio-floc (separation from water column) will dramatically increase the efficiency of the BFP treatment and is discussed later. Further research into whether DON accumulates will also assist to address this issue. TSS, another indicator of water column biomass, confirmed the trend that the BFP treatment significantly (p<0.01) increased biomass (TSS levels) present compared to both Untreated and PSP samples. Figure 6 displays results for all nutrients. Interestingly, Chlorophyll A (ChlA) levels in the BFP treatment were significantly (p<0.05) higher than ChlA levels present in Untreated samples on most occasions and was significantly higher than the PSP treatment during the final three weeks (See Figure 6). A heterotrophic community in a BFP treatment might be expected to have less photosynthetic material (ChlA) than the phytoplankton dominated communities present in the untreated water or PSP system. However, others have observed that C addition did not affect ChlA levels in production system (Avnimelech 2001; Erler et al. 2005; Hari et al. 2006). The higher ChlA levels in the BFP treatment can be explained by the retention of phytoplankton within the floc material and thus within the system (i.e. concentrating the phytoplankton). Hargreaves (2006) described suspended organic material in BFPs as primarily made up of senescing algal cells colonised by bacteria. It is therefore, more appropriate to look at the proportion of phytoplankton within the whole community structure. Although the ChlA levels are higher in the BFP system, the community structure has a lower proportion of phytoplankton than the PSP (See Figure 7). Phytoplankton biomass can be estimated from the ChlA levels using the relationship: 1 mg ChlA = 200mg dry weight (Pagand et al. 2000). Estimates of the contribution by phytoplankton to the TSS levels recorded for each system were calculated. The graphs below demonstrate the difference in community structure achieved by the applied treatment. The PSP community was dominated by phytoplankton (57%) with a low percentage (43%) of other particulates (including bacteria, and zooplankton etc.). In 20 contrast, the BFP community had a relatively low percentage of phytoplankton (41%) and was dominated by other particulates (59%) presumably bacterial biomass. PSP 0% 20% 40% 60% 80% 100% 1 2 3 4 5 6 7 8 9 Week g/ L TS S m Other Phytoplankton BFP 0% 20% 40% 60% 80% 100% 1 2 3 4 5 6 7 8 9 Week TS S m g/ L Other Phytplankton Figure 7: Proportion of phytoplankton present during the experimental period Discussion: Increasing the C dose in BFPs to 30g C L-1 achieves almost complete elimination of dissolved nutrients within 12 hours and extends the period before a significant remineralisation or release of these dissolved nutrients occurs. This suggests that with higher C dosing, treatment systems require only 12 hours retention time to process available dissolved nutrients and exceeding 24 hours will complicate the system with remineralisation and reduce efficiency. The data also suggests that carbon plays a part in the processing of DON, however the data was inconclusive and further work in this area is required. The subsequent experiment included the application of C at this higher dose rate to demonstrate the effect on a phytoplankton-dominated waste-stream in a continuous flow pilot-scale treatment system. By applying the higher C dose and BFP principles to phytoplankton-dominated influent we demonstrated a clear shift to a bio-floc community. A Bio-floc community can be characterised by the following criteria: o Low levels of photosynthesis occurring indicated by lower and more stable pH levels due to the release of carbon dioxide into the water column and lower DO levels due to uptake of available oxygen (Hargreaves 2006). o High nutrient levels (Burford, Thompson et al. 2003) o High levels of organic matter (which can be measured by TN & TP) and low levels of dissolved nutrients due to assimilation (Avnimelech 2003; Ebeling, et al. 2006). 21 o A high level of water column suspended material and a low proportion of phytoplankton present in the community biomass (Burford et al. 2003). The shift to a bio-floc community was indicated by the differences in the standard parameters of DO and pH, which were lower in the predominantly ‘heterotrophic’ BFP system compared to the primarily ‘photosynthetic’ PSP. Both systems maintained all standard parameters within biological and EPA limits throughout the trial period. The effect of adding a carbon source to lower pH has been previously discussed in many papers (Pote et al. 1990; Avnimelech 2003). Our work confirms these findings and also achieved consistency in DO and pH levels by adding molasses to the BFP system. It is well accepted that the key to water quality management for production systems is stability (DPI&F 2006) and this study shows the BFP system is successful in providing both acceptable water quality and stability. Both photosynthetic and Bio-floc communities assimilate dissolved inorganic nutrients and the significant reduction in each of the dissolved inorganic nutrients is evidence that assimilation occurred in both treatment systems trialled in this experiment. However the BFP system did perform better in reducing the potentially toxic nitrogen species TAN and NOx. Toxicity of un-ionised Ammonia is dependant on high pH, and temperature (Hargreaves 1998). Therefore, low TAN levels in conjunction with the lower pH levels, greatly reduces the risk of toxic un-ionised ammonia in BFP systems. Nitrite is also a potentially toxic form of nitrogen and may accumulate due to incomplete nitrification processes (Hargreaves 1998). The effective reduction of NOx (Nitrate+Nitrite) to low levels compared to the Untreated water, suggests that assimilation rather than nitrification is occurring in the BFP treatment. Assimilation reduces the presence of both Nitrate and Nitrite and prevents nitrification, which can result in the accumulation of the toxic nitrite intermediate. This study demonstrated the potential of bio-floc treatment as an external component in a recirculating production system. There is no need to discharge wastewater to the environment so long as the toxic components of the water can be removed. As such, higher TN and TP levels in a production system are not a concern to fish health while there is limited TAN and NO2, and while DO levels can be maintained. This trial demonstrates that those conditions can be achieved with bio-floc treatment. High TSS can be detrimental to 22 fish health as discussed earlier so the preferred production model would be an external biofloc treatment as part of a recirculating system. For most effective performance, a means to separate or exclude bio-flocs from the supernatant would permit the return of treated water back to the production pond without a high BOD or TSS load. Schneider et al. (2007) also reported a similar conclusion when trying to apply a bacteria reactor to clear Recirculating Aquaculture System wastewater. Such a bio-floc exclusion device needs further research but may be in the form of a mechanical particle filter such as a screen or drum filter. Figure 8 shows a schematic representation of the proposed recirculating system, which offers scope to grow and additional crop of prawns (or similar detritivore) within the bio-floc pond, which graze on the nutrient-rich bio-flocs and have the added benefit of helping to keep flocs in suspension. Supernatant returned to production pond Floc excluder Figure 8. Schematic representation of the proposed recirculating system, with external bio- floc pond for water treatment. Conclusion Of the wastewater remediation strategies investigated in this project, it is evident that bio- floc treatment, particularly as a component of an integrated recirculating production system, is the most promising technology to progress towards zero water discharge. Acknowledgements This milestone report forms part of the Project ‘Intensive In-Pond Raceway Production of Marine Finfish’ CARD VIE 062/04 funded by CARD (Collaboration for Agriculture Research and Development) program through the Ministry of Agriculture and Rural Aeration – F7 or similar for O2 delivery and particle suspension Production Pond Bio-floc Pond New water input for evaporation losses Drain for periodic sludge removal Floating raceways Banana prawns stocked at low densities and unfed – graze on flocs/ keep flocs suspended Paddlewheel Organic-rich wastewater removed from raceways to Bio-floc Pond 23 Development of Vietnam. The research team would like to thank the Queensland Department of Primary Industries and Fisheries, in particular Adrian Collins, Ben Russell and Blair Chilton for their efforts in establishing the project. We also thank our Vietnamese research colleagues ably led by Dr Tung Hoang (Director, International Centre for Research & Training, Nha Trang University) for their valuable help and support throughout this project. References American Public Health Association (1989). Standard Methods for the examination of water and wastewater. L. S. Clesceri, A. E. Greenberg and R. R. Trussell. Washington, Port City Press: 10-31 - 10-35. Avnimelech, Y. (1999). Carbon/nitrogen ratio as a control element in aquaculture systems. Aquaculture 176(3-4): 227-235. Avnimelech, Y. (2003). "Control of microbial activity in aquaculture systems: active suspension ponds." World Aquaculture Dec: 19-21. Boyd, C. E. (1995). Chemistry and efficacy of amendments used to treat water and soil quality imbalances in shrimp ponds. In: Swimming through troubled water - Proceedings of the special session on shrimp farming, San Diego, The World Aquaculture Society. Boyd, C. E. (2002). Understanding pond pH. Global Aquaculture Advocate June: 74-75. Brune, D. E., G. Schwartz, et al. (2003). Intensification of pond aquaculture and high rate photosynthetic systems. Aquacultural Engineering 28(1-2): 65-86. Burford, M. A., P. J. Thompson, et al. (2003). Nutrient and microbial dynamics in high- intensity, zero-exchange shrimp ponds in Belize. Aquaculture 219(1-4): 393-411. DPI&F (2006). Australian Prawn Farming Manual - Health Management for Profit. Nambour, Queansland Complete Printing Services. Ebeling, J. M., M. B. Timmons, et al. (2006). Engineering analysis of the stoichiometry of photoautotrophic, autotrophic, and heterotrophic removal of ammonia-nitrogen in aquaculture systems. Aquaculture 257(1-4): 346-358. Erler, D., P. Songsangjinda, et al. (2005). Preliminary investigation into the effect of carbon addition on Growth, water quality and nutrient dynamics in Zero-exchange shrimp (Penaeus monodon) culture systems. Asian Fisheries Science 18. 24 Schuenhoff, A. and L. Mata (2004). Seaweed provides both biofiltration, marketable product. Global Aquaculture Advocate February: 62-63. Funge-Smith, S. J. and M. R. P. Briggs (1998). Nutrient budgets in intensive shrimp ponds: implications for sustainability. Aquaculture 164(1-4): 117-133. Koo, K.H., Masser, M.P. & B.A. Hawcroft (1995) An in-pond raceway system incorporating removal of fish wastes. Aquacultural Engineering 14:175-187 Hargreaves, J. A. (1998). "Nitrogen biogeochemistry of aquaculture ponds." Aquaculture 166(3-4): 181-212. Hargreaves, J. A. (2006). Photosynthetic suspended-growth systems in aquaculture. In: Aquacultural Engineering: Design and Selection of Biological Filters for Freshwater and Marine Applications 34(3): 344-363. Hari, B., B. Madhusoodana Kurup, et al. (2006). The effect of carbohydrate addition on water quality and the nitrogen budget in extensive shrimp culture systems. Aquaculture 252(2-4): 248-263. Jorgensen, N. O. G., N. Kroer, et al. (1994). Utilization of Dissolved Nitrogen by heterptrophic bacterioplankton: Effect of Substrate C/N ratio. Applied and Environmental Microbiology 60(11): 4124-4133. Krom, M. D., J. Erez, et al. (1989). Phytoplankton nutrient uptake dynamics in earthen marine fishponds under winter and summer conditions. Aquaculture 76(3-4): 237- 253. McIntosh, D., T. M. Samocha, et al. (2001). Effects of two commercially available low- protein diets (21% and 31%) on water and sediment quality, and on the production of Litopenaeus vannamei in an outdoor tank system with limited water discharge. Aquacultural Engineering 25(2): 69-82. Neori, A., T. Chopin, et al. (2004). Integrated aquaculture: rationale, evolution and state of the art emphasizing seaweed biofiltration in modern mariculture. Aquaculture 231(1- 4): 361-391. Obaldo, L. G. and D. H. Ernst (2002). Zero-exchange shrimp production. Global Aquaculture Advocate June: 56-57. Pagand, P., J.-P. Blansheton, et al. (2000). The use of high rate algal ponds for the treatment of marine effluent from a recirculating fish rearing system. Aquaculture Research 31: 729-736. Palmer, P. J., Ed. (2005). Wastewater remediation options for prawn farms. Brisbane, DPI&F Publications. 93pp. 25 Palmer, P. (2007) Sand worms trialled at prawn farm. Qld Aquaculture News 30:5 Pote, J. W., T. P. Cathcart, et al. (1990). Control of high pH in aquacultural ponds. Aquacultural Engineering 9: 173-186. Preston, N. P., C. J. Jackson, et al. (2000). Prawn farm effluent: composition, origin and treatment. CSIRO. CRC & Fisheries Research & Development Corporation: 1-71. Schneider, O., V. Sereti, et al. (2007). Kinetics, design and biomass production of a bacteria reactor treating RAS effluent streams. Aquacultural Engineering 36(1): 24- 35. Van Wyk, P. (1999). Chapter 4 - Principles of Recirculating System Design, Harbour Branch Oceanographic Institution: 59-99. 26 VIETNAM COMPONENT 27 Integrated production of the tiger prawn and different marine fish fingerlings in a zero-discharged coastal pond Tung Hoang1*, Michael Burke2, Quyen Banh1 & Daniel Willett2 1 Nha Trang University, Vietnam. Email: htunguof@gmail.com 2 Department of Primary Industries and Fisheries, Queensland, Australia. Abstract An integrated model with intensive nursing of marine fish in floating raceways and low-density prawn farming in the reservoir pond was developed and tested. Results showed that pond water quality was good and stable with no exchange for four months during which several batches of barramundi, grouper and cobia were nursed in raceways. The cultured prawns reached premium size after four months of culture with high feeding efficiency. Other emerging challenges such as predation of escaped fish from the raceways, difficulties in promoting Artemia biomass culture in the reservoir pond and possible technical damage of the air supply system were identified and addressed. This current study establishes important steps to further development of the proposing integrated model, which allows water reuse and thus imposes no environmental impacts on the surrounding environment. Key words: integrated farming, marine finfish, prawn and bioremediation. 1. INTRODUCTION Advanced nursing of fingerlings of barramundi (Lates calcarifer) in SMART floating raceways has been conducted successfully in coastal pond, formerly used for shimp farming in Khanh Hoa Province, Vietnam by the CARD VIE062/04 Project “In-pond intensive floating raceway production for marine finfish” (Hoang et al. 2007). Although the primary objectives of this project have been achieved (i.e. increased the production of large-size fish seed for local farmers in creased and utilized abandoned shrimp ponds in Khanh Hoa Province), it is important to further develop a farming protocol that require no water exchange with the surrounding environment, hereby called zero-discharged system, to minimize the risk of diseases for cultured species and at the same time any negative environmental impacts caused by this innovative farming model. 28 When small fingerlings of barramundi (total length 20 ÷ 30 mm) are nursed in floating raceways, fish wastes and unused feed are driven out of the raceways by the effluents. The removal of these wastes should rely on natural nutrient recycling the absence of detritus feeders. Their decomposition’s products will be partly utilized by phytoplankton and partly accumulated in the pond sediment. The most apparent limitation of this nursing system is the nutrient load (from the raceways) is not continuous and keeps changing all the time, making pond water quality less stable due to “bloomed and crashed” growth of pond phytoplanton. In this curent research, Artemia is used to feed on organic matter released from the raceways. Theoretically, the establishment of an Artemia population in the pond, apart from utilize fish wastes and uneaten feed, can bring in more advantages. If managed properly, Artemia biomass both young and adults will be pumped from the pond water into the raceways, providing live preys at different sizes to barramundi fingerlings. However, in order to maintain a nutrient input for algal growth (that feeds Artemia), we considered to stock the pond with tiger prawn (Penaeus monodon) at low density. It has been reported consistantly that tiger prawns when farmed at low densities between 5 ÷ 15 individiuals/m2 often grow fast, have large sizes and much higher value. This by-product is expected to bring in an additional income for fish-nursing farmers. Low feeding rate and shrimp wastes will help maintain nutrient inputs to sustain algal growth in the pond, improving water quality (Hoang et al. 2007b). This paper reports on the design and results of our preliminary trial on the integrated production of Penaeus monodon and advanced nursing of three marine species including barramundi (Lates calcarifer), Malaba grouper (Epinephelus malabaricus) and cobia (Rachycentron canadum), using floating raceways as the key element. 2. MATERIALS & METHODS 2.1 Trial design The trial was conducted in a 2000-m2 pond, located 1 km from Nha Phu Bay in Khanh Hoa Province (Figure 1). This reservoir pond was partioned in the middle by plastic sheet in order to create an internal flow driven by a 2-HP paddle wheel (Made in Taiwan) which operated four hours a day from 05:00 ÷ 07:00 and from 15:00 ÷ 17:00. Six SMART-1 (3 m3 each) and one SMART-2 (6 m3) floating raceways were placed at one end of the pond (Figure 2). These were used for advanced nursing of barramundi fingerlings. Postlarvae of Penaeus monodon and Artemia were cultured in the reservoir pond while barramundi fingerlings were nursed inside raceways. Covering nets were used in all the raceways to 29 prevent fish escape as barramundi has been well-known as one the major predators to prawns. Figure 1: The experimental site. The pond that used for the trial is on the left. Figure 2: Raceway set-up (left) and pond preparation (right) Prior to the trial the pond was emptied and sun dried for 7 days. Agriculture lime was then applied to the pond bottom as a rate of 7 kg/100 m2. Next, the pond was filled up by pumping water from the nearby canal and left undisturbed for three days before chlorine (25 ppm) was applied for disinfection. A week later the pond sediment was disturbed by dragging heavy iron chains across the pond. This stimulated the distribution of nutrients in the pond sediment into its water collumn, allowing algae to lightly bloom in one week. Artemia nauplii were then released into the pond as a density of 5 individuals/L in four consecutive times (every seven days). At the same time, prawn postlarvae were stocked at 15 individuals/m2. Feeding was conducted only from the second week since stocking. 2.2 Sources of trial animals Hatchery-produced barramundi (20 mm total length) were collected locally and transported by road in plastic bags at a density of 500 fish/bag to the experimental site. The 30 fish were all healthy and did not show any abnormal signs regarding their behaviour and external appearance. Prior to collection grading was done in nursing tank to ensure relative homogeneity of fish size. Transportation took about 1 hour and the fish were acclimatized with water temperature in the reservoir pond for 30 mins before being released into the raceways. Hatchery-produced fingerlings of the Malaba grouper (Epinephelus malabaricus, total length 5 cm) and cobia (Rachycentron canadum, total length from 2 ÷ 10 cm) were road transported for circa 24 hours from the North to Nha Trang. Stocking density was 2,000 fish per tank (500 L). Tank water was maintained at 22 ÷ 23oC and aerated all the time. No water exchange and feeding was conducted during the transportation. Prior to transportation fish were not fed for 24 hours. Several batches of fish were transported from Cat Ba and Cua Lo to Nha Trang. PL15 of the giant tiger prawn (Penaeus monodon) were sourced from a local hatchery. Checking for virus infection was done using PCR and revealed no positive sign. The postlarvae were transported in plastic bag at a density of 1,000 PLs/bag. Initially, Artemia cysts from Great Lake (USA) were hatched for the first two innoculations in the pond. However, Vinh Chau strain (Vietnam) was then used for two more innoculations as it was suspected that the Great Lake strain could not tolerate high temperature of the pond water. The selected cysts were disinfected by 200 ppm chlorine prior to incubation. After 24 hours of incubation, Artemia nauplii were harvested and released into the reservoir pond by spreading equally over the pond water surface. Figure 3: Sampling water for analysis (left) and feeding experimental fish (right) 2.3 Water quality monitoring 31 Key environmental factors such as water temperature, pH and dissolved oxygen (DO) were measured twice a day, at 06:00 ÷ 07:00 and 14:00 ÷ 15:00. Since no water exchange with the surrounding environment was made, salinity was checked every second day as it did not change much. Weekly, pond water samples were collected (Figure 3) and checked for NH3-N, NO2-, NO3- and PO43- using commercial test kits for ornamental tanks (Made in Germany). 2.4 Husbandry and growth monitoring The experimental fish were fed several times a day: every two hours from 06:30 to 17:30 using commercial pellets for marine fish (INVE, Thailand) and for prawns (GROBEST, Vietnam). INVE pellets (800 ÷ 1200 µm) were used only for the barramundi during the first week of advanced nursing due to their small size and the fact that these fish were used to feed on the same pellets in hatchery. GROBEST pellets for prawns (size 2 ÷ 3) are used mainly for feeding (Figure 3). Half an hour prior to feeding, the pellets were mixed with vitamin C and then coated with squid oil. Feeding rate was approximately 10 ÷ 18% body weight of the nursed fish and adjusted based on actual feeding and weather. Basically, fish size and quantity were measured and recorded as the begining and the end of each nursing trial. To monitor their growth, random sampling of 30 fish was done every week. Prawns were fed twice a day at 09:00 and 18:00. Feeding rate was 10% of the estimated standing biomass of prawns in the first month and 5% in the following three months. Feeds were spread evenly across the pond and monitored through the two feeding trays. Prawn health was examined by checking feeding trays and later on using cash net. 3. RESULTS 3.1 Transportation of fish As for barramundi fingerlings there was no problem with transportation from the hatchery to the experimental site as the distance is short. Due to the fact that fingerlings of the Malaba grouper and cobia were not produced locally yet, transportation of fingerlings of these two species have been conducted from Cat Ba (1,400 km from Nha Trang) and from Cua Lo (1,000 km from Nha Trang). It took at least 24 hours to complete the journey. Generally, the Malaba grouper tolerated this long transportation much better than cobia. Mortality was less than 5% for fish larger than 4 cm total length, mainly because cannibalism when transported at high densities. Fish started feeding immediately after stocking in the raceways. 32 The adopted density (2,000 4-cm groupers or 5-cm cobia or 1,000 10-cm cobia in 400L of seawater) appeared to be suitable. Although the fish were not fed 24 hours before transportation, fish wastes and “vomitted” materials were substantial and compromised water quality in the transportation tank. This prompts the need to design transportation tanks that have a simple continuous filter to remove these wastes from the water. The use of probioticsto partially remove NH3 is also recommended. Water temperature between 22 ÷ 24oC was good for both groupers and cobia. Cobia appeared to be much more sensitive to temperature and DO than the Malaba grouper. Cobia became less active at 20oC and mortality was much higher. We have trialed transporting 20-day-old larvae of cobia (about 2 cm total length) at two temperatures: 22 ÷ 24oC and 26 ÷ 27oC. The results was that all fish died at the lower temperatures just after 3 ÷ 4 hours in the transportation tank. About 42% fish survived at the higher temperatures, but they were all weak after 24 hours of transportation probably because of the physical trauma and the lack of live food during transportation. 3.2 Productivity of main products – fish fingerlings Seven nursing trials (two for barramundi, two for grouper and three for cobia) have been conducted during 4 months. The trialed fish grew well in raceways, but survival was different among species (Table 1). Like in a number of previous trials, 2-cm barramundi grew to 4 ÷ 5 cm in two weeks and 6 ÷ 8 cm in three weeks at water temperatures of 29 ÷ 31oC. Survival varied among trials, ranging from 51 to 78 % and appeared associated with the quality and homogeneity of the nursed fish. In the second trial for barramundi (T2), the fish were bought from Research Institute for Aquaculture No.3 and were not uniform in size. Cannibalism was, therefore, much higher and reduced survival rate to 51%, which is lower than expected. The Malaba grouper performed very well in the raceways and the second trial showed that nursing density can be increased to 5,000 ÷ 10,000 fingerlings per a 3-m3 raceway if investment is not a limiting factor. The fish adapted very well to the provided commercial pellets, particularly the Grobest prawn pellets after a short time of feeding with the expensive INVE pellets (for marine fish fingerlings). High survival and good health of the harvested fish was a promissing sign for the use of low cost feed as the INVE pellets for marine fish are five fold more expensive than the Grobest pellets for prawn. The former is, however, essential for small fish as they are nutritious, highly uniform in size and floats well in water. 33 Table 1: Performance of the nursed barramundi and grouper in raceways Barramundi Grouper Parameters Trial 1 Trial 2 Trial 1 Trial 2 Stocking size (cm) 2,0 ± 0,1 2,3 ± 0,8 5,2 ± 0,4 6,3 ± 0,5 Harvest size (cm) 6,2 ± 1,1 8,4 ± 1,0 11,0 ± 0,7 6,9 ± 0,4 Nursing period (day) 23 41 41 10 Average growth (cm/day) 0,182 0,149 0,141 0,086 Total fish stocked 30,000 20,000 2,000 5,000 Total fish harvested 23,400 10,202 1,921 4,960* Survival (%) 78,0 51,0 96,0 99,2* *recorded on day 7 of the trial, two days before the system crashed due to a cyclone on 05/08/2007 Table 2: Performance of cobia nursed in raceways Parameters Trial 1 Trial 2 Trial 3 Stocking size total length (cm) body weight (g) 5,1 ± 0,3 10,9 ± 1,1 7,7 ± 2,1 9,0 ± 1,0 Harvest size total length (cm) body weight (g) 10,2 ± 1,6 4,5 ± 2,1 51,0 ± 16,1 12,3 ± 1,5* Nursing period (day) 16 44 10 Total fish stocked 4,000 1,000 1,500 Total fish harvested 2,080 582 1,456* Survival (%) 52,0 58,2 97,1* *recorded on day 7 of the trial, two days before the system crashed due to a cyclone on 05/08/2007 Success with cobia was lower than with barramundi and the Malaba grouper. Fish grew fast in the raceways and fed well on the provided feed. However, chronic mortality was observed in all three batches of fish, particularly with the first two. It was suspected that fish died because of parasite infection (called “summer parasite” by RIA1’ staff at Cua Lo station) as they were infected when in hatchery. Both dead and weak fish were sent to the Department of Fish Pathology, Nha Trang University for dianogsis, but resulted in no confirmation of the causes or possible pathogen(s). Freshwater bath, as recommended by Mr. Nguyen Quang Huy – RIA1, was very effective in reducing mortality. Despite it a 34 small proportion of nursed fish kept dying daily. Infected fish were sluggish, stopped feeding and their gills are rotten. No other abnormal signs were observed. Fish died after three to four days in a highly skinny form, i.e. big head and slim body. Figure 4: clockwise from left top (a) feeding cobia in raceways, (b) active feeding behaviour of the nursed fish, (c) harvested fingerlings and (d) harvested Malaba grouper Figure 5: Dead cobia fingerlings with rotten gills 35 The trials also showed that cobia fingerlings are highly delicate and vulnerable to low DO. Fish health can be severly compromised or mortality will occur if fish is hold up in a poor DO environment. Five minutes in a 20-L bucket that contained 150 cobia fingerlings could be lethal to many of them. Transportation to and from in-pond floating raceways is, therefore, difficult particularly when fish reach larger sizes, e.g. 15 ÷ 25 cm total length. Advanced nursing of cobia in floating raceways should be conducted in open seawater, preferably in sheltered area. This option helps remove the associated costs for pond preparation and water treatment and maintenance. When SMART-2 floating raceways (self-floating, 6 m3 in volume; Hoang & Burke 2007) are used, the whole raceway can be towed easily by a small boat to grow-out cages making it convenient for fish transportation and stocking. More importantly, the fish are always in water and expose to no stress, but a gradual acclimatization to new living environment. Nonetheless, advanced nursing in coastal ponds is still a good option, especially for 4-cm fish to larger sizes. 3.3 Productivity of by-product: tiger prawn The by-product, tiger prawns, was harvested after about 114 days from stocking. The total amount of harvested prawns was 286 kg. Prawn size was large and highly uniform, averaging about 29.2 g each. This was exceptional in comparison with other farms in the same area where farmers often found it hard to grow Penaeus monodon to this large size in four months over the last several years. As the prawns were large, they fetched a high price of A$ 7.86/kg on the local whole-sale market. However, survival rate was only 32.6%. This was much than expected, i.e. 75 – 80% or 700 kg and was mainly because of predation by the barramundi escaped from the raceways. Figure 6: Cultured Penaeus monodon in the reservoir pond one month before harvest (left) and at harvest time after 114 days of farming (right) 36 Figure 7: Barramundi (Lates calcarifer) fingerlings escaped from the floating raceways became predators of cultured tiger prawn (Penaeus monodon) in the reservoir pond Circa 120 rather-larged barramundi (size between 110 and 340 g, the smallers were not counted) were recorded at harvest time. During the trial a similar number of barramundi were also removed from the reservoir pond by cast net. Out of these ten fish were disected to verify their predation on prawns showed 60% had prawn’s remainings in stomach (Figure) while the rest’s stomaches were full of prawn’s pellets. Overall, 344 kg of Grobest pellets had been used for feeding prawns (and other aquatic animals) in the pond. This accounted for a Feeding Efficiency (FE) of 0.83. The actual Feed Conversion Ratio (FRC) would be low as a certain amount of feed was consumed by the barramundi. 3.4 Financial analysis Apart from environmental stability the profitability of the proposing integrated farming system is considered of prime importance, particularly in comparison to either prawn farming or advanced nursing of marine fish alone. This financial analysis was conducted using actual costs and market values of the harvested fish. In the midnight of the 4th (to the 5th) of August, a cyclone hit Nha Trang and blew up the airpipe system causing surfocation and death to all fish in the raceways. This unexpected natural disaster severly compromised the profitability of the tested farming model. Therefore, we can discuss on three different scenarios. First, if the fish in the raceways were sold right before this disaster, the profit of the 4-month crop would be A$ 2,737 for a 2,000-m2 pond or equivalent to an extrapolated annual profit of A$ 27,374 for one ha of pond if two crops were conducted or A$ 41,061 if three crops were conducted. Second, if value of the dead fish was not accounted, the lost would be A$ 1,699 for a 4-month crop. However, should fish escape from the raceways be completely controlled, one would expect between 500 ÷ 700 kg of prawns, which valued at A$ 3,930 ÷ 5,502. This amount means an extra income 37 of A$ 1,682 ÷ 3,254 from the by-product that would well compensate the lost (of fish) and provide a minor profit. Third, the trial was just a first step to explore this exciting integrated farming system. Its improvements, once made, will help ensure stability and high profit as one of the two riskes observed in this trial – fish escape from raceways is manageable while the other – damage because of an unexpected cyclone will be minimal with a more solid piping system in commercial context. There is no doubt that the proposing integrated farming system is a good option for both fish and prawn farmers in Vietnam. Further improvements are, however, needed to fine-tune it. Table 2: Financial analysis of the proposing model for a 4-month crop Items Quantity Unit cost (A$) Total (A$) Actual operation costs Prawn postlarvae 30,000 pcs 0.0021 64.3 Feed for prawns 344 kg 1.57 540.6 Fish fingerlings Barramundi Grouper Cobia 50,000 pcs 7,000 pcs 6,500 pcs 2,857.0 3,000.0 3,214.3 Feed for fish INVE pellets Grobest pellets* 25 67 15.71 1.57 392.8 105.2 Artemia cysts 3 kg 80.00 240.0 Pond preparation 400.0 Electricity, fuel 4 months 200.00 800.0 Man power 4 man-months 200.00 800.0 Others (probiotics, consumables) 300.0 Pond rent 4 months 100.0 400.0 Depreciation of the raceway system 4 months 200.0 800.0 TOTAL COST 13,714.3 Income Barramundi 5,529.0 Grouper 5,109.6 Cobia 3,565.1 Prawn 2,248.0 TOTAL INCOME 16,451.7 Profit per a 4-month crop 2,737.4 Extrapolated profit (for 1 ha of pond per year; 2 crops) 27,374.1 Extrapolated profit (for 1 ha of pond per year; 3 crops) 41,061.2 38 3.5 Water quality Needless to mention that the proposing integrated system and its components were designed to achieve a stable culture environment where materials could be well recycled within the system mainly through biological means and natural processes. The fact that both fish and prawns were growing well in the raceways and the reservoir pond during 4 months of experimentation is a persuading evidence that this objective was achieved. Furthermore, water quality appeared to be suitable for both fish and prawns as shown by periodical monitoring for the reservoir pond (Table 3). The only disapointment was that biomass culture of Artemia appeared not possible despite several attempts of innoculation. It was suspected that Artemia nauplii were outweighed by natural copepods that had well established in pond water. As the main products of the system are large-sized fingerlings of marine finfish, the establishment of Artemia biomass culture in the reservoir pond is of prime importance. It can provide good live preys at different sizes for the fish and utilize both organic matters and algae, thus helps clean up the water and assist nutrient recycling. Further attempts to culture Artemia biomass in the reservoir pond may have to be done with water disinfection after the pond is filled up. Innoculation of algae and Artemia nauplii should then follow immediately. Table 3: Summary of water quality in the reservoir pond during the experimental period. ND: not detected Parameters Range Mean ± S.D. Temperature (oC) 26.4 ÷ 36.6 32.4 ± 2.1 pH 7.6 ÷ 8.6 Daily pH fluctuation 0.1 ÷ 0.7 DO (mg/L) at 7 am 3.3 ÷ 6.5 4.9 ± 0.8 DO (mg/L) at 2 pm 4.2 ÷ 14.8 9.7 ± 1.7 Alkalinity (mg CaCO3/L) 90 ÷ 110 96 ± 7.0 Secchi disk reading (cm) 27 ÷ 55 36.6 ± 7.2 Salinity (ppt) 28 ÷ 31 29.2 ± 1.2 NH3-N (mg/L) ND ÷ 0.2 NO2- (mg/L) ND 39 Even in the absence of Artemia, the quality of pond water was suitable and stable (Table 3). Daily fluctuation of pH varied between 0.2 and 0.7, which is ideal for prawn farming (MPEDA/NACA 2003). This is probably because the nutrient load to pond water from the raceways was not so high. It has been known that feeding fish nursed in SMART floating raceways is highly efficient thanks to the internal dynamics of water (within the raceway) and high densities of the nursed fish (Hoang et al. 2007a,b). Similarly, low biomass of prawns and feeding rate also helped maintain a stable environment. Sechii disk reading showed a light bloom of algae, which was stable and able to assimilate both NH3- N and NO2-, keeping these parameters very low. It is important to note that no water exchange was conducted during the trial, demonstating that this system is a zero- discharged one. 3.6 Management issues Despite a technical failure (due to natural disaster) that significantly reduced the trial’s profitability, this study has established an important step for the development of a zero- discharged system where intensive nursing of marine finfish is integrated with low-density prawn farming. Several lessons have been learnt from the trial and thus should be well addressed when and wherever this kind of system is applied or adopted: • Quality of fish fingerlings: it is important to ensure that fingerlings are free of pathogents and relatively uniform in size. Low survival of cobia in the trials was because of an unknown disease. Similarly, great size variation in one batch of barramundi (purchased from RIA3) resulted in a much higher mortality because of cannibalism. • Periodical treatment for disease prevention: one of the advantage of nursing fish in floating raceways is greater control over diseases and this should be made use of by periodical prevention. Freshwater bath for cobia and grouper should be done by pumping freshwater into the raceways rather than transfering fish into a freshwater holding tank in order to reduce stresses and physical trauma. It was found that when caught by scoop nets, the dorsal fin of one grouper can easily damage the eyes of some other ones. Similarly, cobia fingerlings showed signs of exhaustion when being handled too much. • Avoidance of unexpected problem: measures should be taken to avoid unexpected problems such as fish escape from the raceways and technical failure of the air supply system. It was found that barramundi can swim against strong currents and get out of the covered raceways through the airlifts. A net 40 that covers the opening of all airlits should be able to stop them from entering the pond, thus keeping the cultured prawns from predation. Technical failure of the air supply system is lethal to the nursed fish, which are normally at very high densities. The cyclone in early August had torn the supply air pipe right at the connection between the on-land PVC pipe and the soft pipe that ran to the supporting pontoon. Strong winds around 2 or 3 am of the 5th of August have pushed the pontoon a few meters away from its original position, thus damaged the pipe and cut off oxygen supply to the raceways. The responsible technicians hidding in a nearby building did not notice this until it was too late for any recover attempt. 4. RECOMMENDED FARMING PROTOCOL For those who are interested in this proposing integrated model the information provided hereby can be considered as a guide. System’s components and actual management can be changed in accordance with the local climate and production context. • Pond size: 2,000 m2 with an even water depth of at least 1.2 m. Higher depths can be an advantage. The pond should be partioned by plastic sheet in the middle to create an internal flow around it. • Floating raceways: Either SMART-1 (3 m3, supported by a pontoon) or SMART-2 (6 m3, self-floating) (Hoang & Burke 2007) can be used. Total working volume of the whole system should be around 30 ÷ 40 m3. This floating raceway system is driven by a central 3-hp air blower. Necessary measures should be taken so that the air piping system is solid enough again any possible damage and an alarming system is triggered when damage may happen. • Pond preparation: follows the standard preparation used for prawn farming. Once filled up, pond water should be treated with chlorine 15 ppm. Algal bloom is then promoted by the use of inorganic fertilizers. Artemia innoculation can follow immediately. A large net bag that can enclosure the whole floating raceway system should be used to prevent fish from entering the reservoir pond and become predators to the cultured prawns. • Stocking and farming of prawn: Postlarvae are stocked at a density of 15 individuals/m2. Stocking should be done, if possible, well before nursing marine finfish. This is to minimize possible losses because of predation of escaping fish as larger prawns escape from predators better than smaller ones. 41 No feeding is required for the first two weeks. From the third week onwards feeding can be conducted at a daily rate of approximately 5% of standing biomass. The cultured prawns should reach 22 g on average after three months and 30 g on average after four months of culture. For a standard 2,000-m2 pond a production of between 500 and 700 kg of prawns should be expected given good management of pond water quality and predation. • Nursing marine finfish in floating raceways: Species like barramundi and the Malaba grouper can be nursed from small sizes (as small as 1.5 cm total length for barramundi and 3.0 cm total length for grouper) to larger sizes (up to 8.0 ÷ 10.0 cm for barramundi and 15.0 cm total length for grouper). When nursing cobia it is important to plan ahead transportation methods for large fingerlings to grow-out cages to ensure fish are not stressed due to physical trauma or insufficient oxygen supply. Nursing density can be 3,000 ÷ 5,000/m3 for barramundi; 1,800 ÷ 3,000/m3 for the Malaba grouper. Initital stocking density of 4- or 5-cm cobia can be 1,000/m3, but should be reduced with time (particularly in terms of biomass) as the fish grow very fast, i.e. approximately 0.6 ÷ 1.0 cm per day. Feeding for the first seven to ten days should be done with INVE pellets (800 ÷ 1,200 µm). Over the last three to five days Grobest prawn pellets (2 mm) can be used to feed them. Before feeding the pellets should be well mixed with commerical Premix and squid oil (20 mL for 1 kg of feed) and air-dried for 30 mins. • Water quality management: Important parameters such as DO, pH, temperature, Secchi dish reading and salinity should be monitored daily whereas NH3-N, alkalinity and NO2- can be checked weekly. Every seven days a 500-g bag of Pondplus probiotics should be applied to pond water regardless how good the water quality could be. • System management: Advanced nursing of marine fish should be done as continuously as possible. This ensures that the nutrient load from the raceways into pond water is consistent, thus stablizing algal bloom. Continous attention should be paid on prawn farming as experience showed that this is an effective way for workers to stay focused on the system because nursing fish fingerlings often follows an on-and-off mode, particularly when supply and/or demand is not reliable. 42 • Technical assistance: can be sourced from local aquaculture expertise in accordance with your species of interest. Regarding the floating raceway and concepts of this integrated system, please contact Queensland Department of Primary Industries & Fisheries in Australia or the Ministry of Agriculture and Rural Development or Nha Trang University in Vietnam. ACKNOWLEDGMENT We would like to thank Khanh Hoa Fisheries Extension Center (KFEC) for allowing the project to use their facilities for the trials. Thanks are given to Mr. Huynh Kim Khanh (KFEC), Mr. Ngo Van Manh (Department of Mariculture), Mr. Nguyen Hong Hieu (N45, Nha Trang University), Dr. Do Thi Hoa and other researchers at the Department of Fish Pathology (Nha Trang University) for helping us implementing the trials. This project was not possible without strong and continuous supports of Queensland Department of Primary Industries & Fisheries (Australia), the Ministry of Agriculture & Rural Development of Vietnam and its CARD Office, and Nha Trang University. REFERENCES Hoang T., Luu T. P. & Huynh K.K. (2007) Trials of advanced nursing of barramundi Lates calcarifer in in-pond floating raceways. Journal of Fisheries Science and Technology 01/07: 12-18 (in Vietnamese). Hoang T., Huynh K.K., Banh T.Q.Q, Nguyen D.M & Burke M. (2007) Use of floating raceways for marine finfish fingerling production and potential for the development of an integrated farming system. In: Proceeding of IMOLA Symposium, Hue 19 – 20 April, 2007, pp 1 – 14. Hue University of Agriculture and Forestry. Hoang T. & Burke M. (2007) Floating raceways provide options for marine fish fingerling production. Global Aquaculture Advocate Jul-Aug: 54-55. MPEDA/NACA (2003) Shrimp Health Management Extension Manual. Prepared by the Network of Aquaculture Centres in Asia-Pacific (NACA) and Marine Products Export Development Authority (MPEDA), India, in cooperation with the Aquatic Animal Health Research Institute, Bangkok, Thailand; Siam Natural Resources Ltd., Bangkok, Thailand; and AusVet Animal Health Services, Australia. Published by the MPEDA, Cochin, India. 43