Water Quality Assessment of a Trout Farming Effluent , Bocaina , Brazil

The concern in mitigating the negative impact generated by the discharge of nutrients in the receiving water body is a challenge for the sustainable development of Brazilian fish farms. Thus, the purpose of this study was to evaluate the water quality and environmental impact caused by trout farming system effluent with focus on discharge of phosphorus. Sampling was performed on a weekly basis in triplicate from September to November 2010. Sample sites were distributed according to the water flow: upstream from trout farming system, water supply, effluent, artificial wetland, mixing zone and downstream (60 m from effluent). In the field, pH, conductivity, dissolved oxygen, water temperature and turbidity were measured. In laboratory, nitrogen and phosphorus series, chlorophyll a, total solids suspended and their organic and inorganic fractions were analyzed. For the good growth of trout in production system, the abiotic factors described in the water quality monitoring demonstrated acceptable values. Environmentally, after passing through the production system and artificial wetland, there was an increase in concentrations of total phosphorus, total nitrogen, orthophosphate, ammonium, chlorophyll a, total solids suspended and their organic and inorganic fractions (P < 0.05). These results are related with the quality of feed, feeding management and the inefficiency of the artificial wetland. Therefore, it is necessary to use best quality feed to meet the nutritional requirements of trout, maintaining an optimal feed conversion and reducing pollution generated by effluent.


Introduction
Although Brazil is situated in a region of tropical and subtropical climates, trout has been adapted well in the mountainous regions where temperatures are lower, close to 22˚C in the hottest month and for at least four months is greater than 10˚C [1].In these areas, the rivers are small, with an insufficient flow for the deployment of large trout farming systems [2].In Brazil from 2008 to 2010, there was an increase of 40% in the trout production.It was produced about 3660 tons in 2008, 4381 tons in 2009 and 5122 tons in 2010 [3].Fish farming, as other animal production, causes impact on the environment [4] and has a potential trend of freshwater eutrophication, due to excessive discharge of nutrients [5] [6].The pollution is generated from the feed introduced into the farming system, directly by the dispersion of the feed not eaten and/or by the metabolic products excreted by fish [4] [7].The feed produced in Brazil uses waste of fish, with low levels of protein and high ash contents.Depending on the characteristics of the raw material, the feed contains a large amount of mineral material from the bones and scales [8].
The effect of discharge of the effluent of the fish farming in the ecosystem is related to the concentration and characteristics of the pollutant and the assimilative capacity of the receiving water body [9].This effect can vary according to the production system, management practices, and the type of supplied food [10] [11].
Studies that assess the level of pollution in the effluent and in the receiving water bodies are extremely important in order to propose appropriate measures to mitigate the impacts generated by trout farming.Several authors have focused on this theme [12] [13], that remains inedited in Brazil.
Currently, the biggest challenge of trout farming and other productions in freshwater ecosystems is to mitigate the impacts generated by the productive system in the environment.The purpose of this study was to evaluate the water quality and environmental impact caused by trout farming system effluent with focus on discharge of phosphorus.

Study Area and Fish Ponds Management
This study was conducted in a trout farming system located in the National Park of Serra Bocaina, 35 km from the Bananal city, São Paulo State, in the Paraiba do Sul River Basin situated 1155 m above sea level at coordinates 22˚50'03.92"S and 44˚25'46.33"W (Figure 1).
The trout farming used raceways system (adapted) getting constant water flow with estimated average flow rate at 40 L/s, keeping the residence time around 1 h 46 min of the entire system.The masonry tanks were divided into three sectors (Figure 2) with 4 tanks each, total volume of 315 m 3 , with an average density 2.45 Kg/m 3 and production capacity estimated at 3.5 t/year.
The initial stock was 50,000 individuals of rainbow trout (Oncorhynchus mykiss) at different stages of growth, consisting of fingerlings (2 g), juveniles (20 -140 g) and adults (220 g), totaling 776 Kg in the system.Production in intensive feeding regime used commercial extruded feed 36% -42% crude protein (CP), ranging from 1.5% to 6% of the fish biomass.The estimated fish biomass was calculated by sampling 100 individuals per tank, totaling 400 individuals per sector, and extrapolated to the whole system of production.

Limnological Parameters
Sampling was performed on a weekly basis in triplicate from September to November 2010 in six sites (n = 234).The Figure 3 shows the distribution of the sampling sites in the production system.
In the field, with Multiprobe HORIBA U-22 model, were determined the variables pH, conductivity (Cond), dissolved oxygen (DO), water temperature (T) and turbidity (Turb).The values of DO were transformed in percentage of oxygen saturation.In the laboratory, the concentrations of total phosphorus (TP) and total nitrogen (TN) were determined according to the method proposed by Valderrama [14].The orthophosphate (P-PO 4 ) followed the methodology described in Strickland and Parsons [15].Ammonium (N-NH 4 ), nitrite (N-NO 2 ) and nitrate (N-NO 3 ) were measured according to the methods described in Apha et al. [16].The concentrations of total suspended solids (TSS) and their organic and inorganic fractions were determined using the gravimetric method described in Teixeira et al. [17], with modifications based on Wetzel and Likens [18], who consider the maximum temperature of 60˚C during 24 h to obtain the organic fraction.Glass fiber filters and Millipore AP-40 paper with porosity from 0.7 µm to 1.4 µm were used to retain particulate fraction [16].The chlorophyll a was de-   termined according to the method described by Sartory and Grobellar [19].
The flow rate was calculated from the product of the channel section and average water velocity, measured by the float method [20].The discharges of effluent released were determined by the product between water flow and the nutrient concentration.

Statistical Analysis
In order to characterize the zootechnical cycle environmental parameters were described in terms of average value and confidence interval of 95%.Site variations on the variables were analyzed through a Kruskal Wallis test after ( ) transformation, followed by a SNK test.The level of significance adopted was 0.05 [21].
Considering the current legislation CONAMA resolution [22] (Class II, lotic environment) and compared to the usual monitoring water quality parameters for production, only total phosphorus demonstrated not acceptable value at S4 point.The turbidity values were below 10 NTU at all sites and the nitrate values were below 0.10 µg/L.The trout faming system exported 9.80 kg TP/month e 120.00 kg PT/year with discharge of 13.21 g TP/kg of feed in a volume of feeding of 25.00 kg/day (Table 2).

Discussion
The results presented here constitute the first contribution to trout farming systems in tropical and subtropical regions.Despite the system analyzed is considered small, due to the water flow, it represents 70% of Brazilian trout farming systems, where production is up to 20 t/year due to water availability [26].
The abiotic factors described in the monitoring water quality such pH, conductivity, dissolved oxygen, water temperature and turbidity demonstrated acceptable values for trout production.The pH remained slightly acid, favorable condition to the welfare of trout, because when in alkaline pH, ammonia excreted by the fish becomes more toxic.The conductivity values lower than 39 μS/cm indicate low rate of decomposition of organic matter related to the low residence time of water in the production system.During the production cycle, the water remained well oxygenated with more than 100% of oxygen saturation values.The limit for percentage of oxygen saturation recommended for trout is 80% [23].The water temperature remained at acceptable values for the best growth of the animals, not exceeding 18.63˚C, value within the classification of highland tropical climate (10˚C -22˚C) [1].Regarding the turbidity, low values were obtained (<10 NTU).
The impact of intensive trout farming on a river depends on the size of the farm, the farming practices, the nature and volume of the wastes produced [10].In Brazil, in 2010, from the 264,000 tons of feed were produced for aquaculture, 80.4% went to feed fish [27] [28].Many industries have produced low quality rations to be marketed at lower prices.This fact leads to the use of diets of low digestibility [29].This situation results in the deterioration of water quality due to excessive excretion of nutrients coming from fish and remains of feed.
Environmentally, changes have occurred in the parameters related to the eutrophication process.After passing through the production system (S3 and S4) it is possible to observe that there was an increase in concentrations of nutrients, with significant differences for TP, TN, P-PO 4 , N-NH 4 , CHL, TSS and their organic and inorganic fractions.High feeding rates increase the output of organic matter from farms either as uneaten feed or faeces [30].At the S5 and S6 sites there was a reduction in concentrations when compared to the values of S3 and S4 due to the dilution process of the receiving water body.
The increases in the concentrations of TP and TN attributed to the feeding and metabolism of organisms were due to use of low quality feed and corroborated by the conversion rate considered inadequate (2.12).Other studies in intensive system of production of trout, with higher water flow, production and use of feed showed values of feed conversion ratio closer to 1:1, indicating better management, fact which caused minor discharge of phosphorus compared to the present study [7] [24] [25].The higher phosphorus loads on trout farming investigated in this study seemed to be associated with the high phosphorus content of feeds used and insufficient feed management.
In countries where strict environmental regulations govern the operation of trout farms, phosphorus content of trout feed and FCR are not permitted to exceed 0.9% -1.0% and 1.0, respectively [31] [32].In Brazil, there is no legislation that deals with the content of phosphorus in the feed composition for the purpose of reducing environmental impact.Brazilian legislation regulates only the limits of allowable concentrations of phosphorus in the effluent to be discharged in accordance with framing of the water body (Table 1).
Despite to the orthophosphate, when the percentage of P-PO 4 in relation to TP is <10% indicates an accumulation of P in phytoplankton and debris [33].In this study, the percentage of P-PO 4 relative to TP was >90% indicating that this element was exported directly in effluent.Average concentrations after passage by trout farming (Site 3) showed no decrease in the mixing zone (Site 5) and downstream (Site 6) (P > 0.05), demonstrating that the distance of 60 m was insufficient to observe the activity of resilience on the receiving water body.
The increase of N-NH 4 is clearly related to feeding and excretion of the fish.The mean concentration of N-NH 4 in effluent of the trout farming system (217.32µg/L) increased by 104% from water supply (106.34 µg/L).However, the N-NH 4 concentration do not exceeded the recommended limit for effluents [22].The mean concentration of N-NO 2 in effluent of the trout farming system (0.58 µg/L) increased by 152% from water supply (0.23 µg/L) and the N-NO 2 concentration in effluent do not exceeded the recommended limit for effluent [22], dynamic similar to the N-NH 4 .The mean concentration of TSS and their organic and inorganic fractions in effluent of the trout farming system (1.58; 0.46; 1.12 mg/L) increased by 222%, 100% and 348% from water supply (0.49; 0.23, 0.25 mg/L), respectively.
Most studies on evaluation of total suspended solids do not describe their inorganic and organic fractions [34].Due to their contribution in nutrient input, the estimated concentration of total solids in suspension (TSS) coming from of production system may be an appropriate variable for evaluating the effluent from a fish farming.In farming systems with good management practices, the main source of suspended material originates from the fish feces [35].Therefore, in this study, the results can be associated with feeding practices, such as frequency, stocking density, quality and amount of feed that interfere in excretion rates.
The constant renewal of water in the trout farming hindered the establishment of phytoplankton, indicated by the low values of chlorophyll a.
After passing through the artificial wetland there was an increase in the concentration of TP, TN, P-PO 4 , N-NO 2 and organic TSS.Although not statistically significant, indicates that the wetland is not efficient.Among the different aspects for improving this treatment system must consider the proportional dimension of wetland to the production.

Conclusion
Finally, long-term sustainability of many fish farms may be dependent on their ability to reduce their waste outputs.Therefore, it is necessary to use best quality feed to meet the nutritional requirements of trout, maintaining an optimal feed conversion and reducing pollution generated by effluent.

Figure 1 .
Figure 1.Map of South America (Brazil) highlighting the location of the trout farming system.

Figure 2 .
Figure 2. Representation of the trout farming system on Serra da Bocaina (SP), organized into three sectors: fingerlings, juveniles and fattening, where each sector is divided into four tanks.Arrows = flow of water supply to the tanks.Black bar = channel supply.

Figure 3 .
Figure 3. Representation of sampling sites of the study place on Serra da Bocaina-São Paulo, where: S1-upstream from trout farming system; S2-water supply; S3-effluent; S4-artificial wetland; S5-mixing zone and S6dowstream (60 meters from the effluent); arrows = flow of water through the trout farming system and returning to the river that supplies the production.

Table 1 .
Average and confidence interval (95%) of parameters analyzed during the study cycle of trout farming system.Means followed by the same letters in the line do not differ by SNK test (0.05).