Effects of rainfall on water quality in six sequentially disposed fishponds with continuous water flow

An investigation was carried out during the rainy period in six semi-intensive production fish ponds in which water flowed from one pond to another without undergoing any treatment. Eight sampling sites were assigned at pond outlets during the rainy period (December-February). Lowest and highest physical and chemical parameters of water occurred in pond P 1 (a site near the springs) and in pond P 4 (a critical site that received allochthonous material from the other ponds and also from frog culture ponds), respectively. Pond sequential layout caused concentration of nutrients, chlorophyll-a and conductivity. Seasonal rains increased the water flow in the ponds and, consequently, silted more particles and other dissolved material from one fish pond to another. Silting increased limnological variables from P 3 to P 6 . Although results suggest that during the period under analysis, rainfall affected positively the ponds’ water quality and since the analyzed systems have been aligned in a sequential layout with constant water flow from fish ponds and parallel tanks without any previous treatment, care has to be taken so that an increase in rain-induced water flow does not have a contrary effect in the fish ponds investigated.


Introduction
Fishponds and fish breeding tanks are dynamic ecosystems, characterized by shallowness and continuous water flow, which directly affect the limnological variables throughout the day.Water flow is extremely important due to its transport of nutrients, microorganisms, addition of oxygen to the medium and organic material.In fact, ration is added daily to the water as fish feed and thus the fertilization of the water medium increases.Further, accumulated detritus on the bottom of the fishpond may be removed and relocated to the next pond.Introduced nutrients and those produced in the fishponds are used for organic production whilst others remain in solution or suspended in the water or even absorbed by the soil (BOYD, 2003).
Since the ponds receive a daily load of ration plus the ration originating from the previous pond, this fact may promote an increase in the water nutrient load.However, depending on flow rate and pond depth, continuous water flow may present a concentrating or diluting effect on the materials Acta Scientiarum.Biological Sciences Maringá, v. 32, n. 1, p. 9-15, 2010 contained in the water (SIPAÚBA-TAVARES et al., 2003).
The grow-out period in the southeast region of Brazil extends during the summer season, from November to March, when temperatures fluctuate with subsequent higher growth rates and yield.During this period, the amount of feed supplied as well as stock density per pond may exceed the capacity of such systems, which causes deterioration of water quality (MACEDO; SIPAÚBA-TAVARES, 2005).
Appropriate management, knowledge and control of the ecological, biological and socioeconomical aspects of aquaculture are important to obtain and maintain proper water quality that guarantees high biomass production.The above is mandatory to clarify issues related to water quality and minimize the negative impact caused by this activity.In fact, environmental concerns and limitation in water availability are some of the most important factors in aquaculture (METAXA et al., 2006).
The present work was conducted to monitor the physical and chemical parameters in the inlet water of six semi-intensive fishponds with continuous water flow, distributed sequentially during three distinct grow-out periods, before (October), during (January) and after (May) the grow-out fish period.

Study area and fish management
The current investigation was carried out at Aquaculture Center (21°15'S; 48°18'W), 'Universidade Estadual Paulista', Jaboticabal, São Paulo State, Brazil, in the inlet water of six ponds, with continuous water flow, provided 5% daily rate of rearing volume, during the months of May, October, and January.The six fishponds (P 1 -P 6 ) were laid out sequentially, or rather water from one fishpond flowed directly into another, while it received water from other tanks and smaller parallel fishponds, (with exception of P 1 -water supply), used by the nutrition, frog culture, fish breeding and shrimp culture sectors.The surface areas of the six ponds are 3,800 m 2 (P 1 ); 2,306 m 2 (P 2 ); 9,231 m 2 (P 3 ); 4,351 m 2 (P 4 ); 5,671 m 2 (P 5 ); and 2,507 m 2 (P 6 ).The ponds presented average depth varying from 1.41 m (P 1 ) to 1.87 m (P 3 ) (SIPAÚBA-TAVARES et al., 1991).
The first pond (P 1 ) receives water directly from a well (W), and actually functions as a water supply for the other ponds.Food management of P 2 , P 4 , P 5 , and P 6 , comprises 10 to 12 kg of commercial ration containing corn, soybean and meal fish with a daily addition of 28% of crude protein.However, the largest pond, P 3 , receives 20 kg day -1 of the same ration.In addition to the water outlet for the subsequent ponds, P 2 , P 3 and P 4 have another outlet.Pond P 2 has pipes to directly dispose of water into the frog culture ponds, while water is pumped in pond P 3 to irrigate the agricultural fields belonging to the University.In the case of pond P 4 , the water is distributed by pipes to irrigate hydroponic systems and into three more fishponds outside the area.

Physical and chemical parameters
Water was always measured at 9:00 am, during 15 consecutive days on each month (May, October and January) at the inlet water, with a Van Dorn bottle (5 L).A total of seven sampling sites were assigned: W = well, and P 1 to P 6 = ponds 1 to 6. Temperature, pH, electrical conductivity and dissolved oxygen were measured 'in situ' by a Horiba U 10 water quality checker.The nutrients were determined by spectrometer, according to Golterman et al. (1978) and Koroleff (1976) for nitrite, nitrate, total-phosphorus, and ammonia.The wind speed was measured by an anemometer Lutron AM-4201, whereas chlorophyll-a was measured according to Nush (1980).The alkalinity and inorganic carbon were determined according to Mackereth et al. (1978).

Statistical analysis
Two-way nested ANOVA was applied to physical and chemical parameters in order to compare sites (P 1 -P 6 and W) and periods (May, October and January) (SOKAL; ROHLF, 1981).Significance level was p < 0.05.

Results and discussion
The limnological variables tended to fluctuate during the period under analysis.Actually they depended on the water flow and the management employed.Since the water goes directly from one pond to another without prior treatment, a tendency exists to accumulate certain elements in the medium.However, the system dynamics and the management practiced in pond 4 (P 4 ), or rather, where the water is partially conducted to the plantation nearby, somewhat decreased the organic and inorganic material carried to the next sequential ponds (P 5 and P 6 ).
A two-level nested Anova demonstrated variation in the limnological variables analyzed comparing the sites (spatial variation) and periods (temporal variation).Limnological variables pH and total phosphorus varied (p < 0.01) among sites (W, P 1 -P 6 ) and periods (months); the opposite was reported for dissolved oxygen and nitrite (p > 0.01).
Although temperature did not vary among sites owing to their being geographically very close to one another, it varied according to the period (p < 0.01).Actually, the season affected variations; the same occurred with the wind (p < 0.05) variable, since it is highly associated with the climate (Table 1).Daily variation in water temperature that did not exceed 4 o C provided good conditions for freshwater culture.The wind varying from 1.31 to 4.32, 2.15 to 4.8, and 1.45 to 2.56 km h -1 , respectively in May, October and January (Figures 1 and 2).
The inlet water in the ponds was well oxygenated and revealed average concentrations above 4.6 mg L -1 .The highest dissolved oxygen concentrations were observed at P 3, during May, with 10.3 mg L -1 .The inlet water in site P 3 presented adequate oxygenation, varying from 5.6 to 10.3 mg L -1 during the period under analysis.Water flow from P 2 to P 3 was a sharp inclination, thus causing high water oxygenation.Actually it is a good way of introducing more oxygenated water into the fishponds when water flows from one pond to another without previous treatment (Figure 1).), bicarbonate (mg L -1 ), free CO 2 (mg L -1 ) and wind speed (km h -1 ) in the fishponds during the period under analysis, where P 1 = well, and P 2 -P 6 = ponds 1 to 6.
Concentration of dissolved oxygen should be kept above 4 to 5 mg L -1 (SIPAÚBA-TAVARES, 1995) for excellent fish growth.In fact, such pattern has been reported in the fishpond under analysis (Figure 1).
In general, during May the pH was acidic, ranging from 4.92 to 6.80.During the study period the well (W) was acidic, ranging from 4.92 to 5.90.In October and in January, pH was slightly acidic in ponds 1 to 3, ranging from 5.25 to 7.04.However, pH varied from 6.77 to 7.46 in P 4 (Figure 1).
Rise in pH is generally caused by the photosynthesis process which, through CO 2 consumption, causes an increase in pH rates.As a rule, shallow artificial systems with a continuous water flow may occasionally cause small variations in pH and alkalinity rates (SIPAÚBA-TAVARES et al., 1998).In fishponds, CO 2 and H + from nitrifying bacteria and other acids produced by the organism decrease pH rate (SIPAÚBA-TAVARES et al., 2003).
Although conductivity tended to increase from P 4 to P 5 , in general the highest conductivity was reported in P 4 in May (126.6 S cm -1 ) and January (121.7 S cm -1 ).In October, P 5 showed higher rates (109.3S cm -1 ) (Figure 1).
Fishponds with continuous water flow receive large amounts of allochthonous material, such as rations and manure, and, consequently, have rates above 100 S cm -1 .The fishponds are environments with great management requirements.In fact, such conductivity is associated with accumulative effect of the current caused by the water's continuous flow and consequently to the short residence time (SIPAÚBA-TAVARES et al., 2007).Currently higher rates of conductivity were reported in fishponds (P 4 -P 5 ) which are directly affected by organic and inorganic, as occurred in the case of P 4 which receives water originating from ponds P 1 , P 2 , and P 3 and from the frog culture, whose daily wash causes a great amount of organic matter being introduced into the pond.Pond P 5 receives water from the shrimp culture from which it receives a large amount of nutrients (Figure 1).
Water from the well (W) was acid and free CO 2 concentrations were higher than rates of the other sites (P 1 and P 6 ), which varied from 137.6 to 426.4 mg L -1 .In the sites P 1 to P 6 , bicarbonate rates were generally highest than those of free CO 2 , with the exception of those in May at sites P 1 , P 2 , and P 3 (Figure 2).Alkalinity was higher from site P 4 and the highest mean concentration was observed during May at site P 4 , with 181.67 mg L -1 .Alkalinity and bicarbonate rates were significantly different (p < 0.01) only between sites (Table 1, Figure 2).
Although alkalinity rates are above those recommended for fish culture (20 to 40 mg L -1 of Ca CO 3 ), the continuous water flow may have affected oscillations since higher concentration have been reported in the last ponds (P 4 -P 6 ).
Nutrients concentrations and conductivity rates may have been affected by fish stock and species density in the fishponds since these factors directly affected the limnology of shallow artificial systems (SIPAÚBA-TAVARES et al., 1998).
Only nitrite was not significant (p > 0.05) either during the periods and at the sites.On the other hand, nitrate and ammonia behaved differently, or rather, the former was different (p < 0.05) only for the period; the latter was different (p < 0.01) only for sites (Table 1).
The nitrification is not inhibited at high dissolved oxygen concentrations leading to complete nitrification with decrease nitrite concentration in the water (MA et al., 2008).In our study, high concentration of dissolved oxygen (> 4.6 mg L -1 ), favored nitrate predominance and nitrite decrease in the ponds under analysis.
In general, the lowest concentrations of nitrogen compounds were observed in January, or rather, during the rainy period and at the highest effect of water current.Nitrate and ammonia were predominant and varied between 87.2 and 1,243.7gL -1 , and between 9.8 and 416.7 g L -1 , respectively, during the experiment period (Figure 3).
Continuous water flow in ponds contributed to the loss of part of the nitrogenous compounds, whereas the movement caused by the entrance and exit of water caused a higher nitrification of the system with less nitrite concentrations when compared with nitrate rates.As a rule, highest ammonia and nitrate accumulation was reported in the last fishponds (P 4 -P 6 ) (Figure 3).
Ponds with water renewal tend to have nitrite concentrations above 37 g L -1 ; in non-renewal water concentrations are very low and show the absorption of this element by the phytoplankton and the activities of nitrifying bacteria in the environment (SIPAÚBA-TAVARES et al., 2000).Nitrite has generally remained below 20 g L -1 in fishponds, with the exception of peaks during May at P 5 and P 6 (93.7 and 73.4 g L -1 , respectively).Total phosphorus was significantly different (p < 0.01) between spatial (sites) and temporal (periods) variations (Table 1).Highest total phosphorus concentrations were observed in May with highest peak in P 3 (245.5 g L -1 ), keeping concentration above 138 g L -1 (Figure 3).Lowest concentration of this nutrient during January and October, generally, bellow 100 g L -1 , may be due to the fact that nutrient is easily chelated in the sediment.Discharge controls phosphorus dynamics in the environment with regard to its absorption in the sediment (AVNIMELECH, 1998).
There are physical, chemical and biological forces influencing the different total phosphorus components (SINDILARIU et al., 2008).Total phosphorus is removed primarily by cation exchange reactions with the substrate and algae community, even though climate factors such as heavy rainfall (January) may affect the availability of this nutrient in the water (SIPAÚBA-TAVARES et al., 2007).Although no significant differences were observed (p > 0.05) between periods with regard to chlorophyll-a, it tended to increase as from pond P 3 (p < 0.05).The highest and lowest concentrations of chlorophyll-a were 252.96 g L -1 (P 3 ) and 0.28 g L -1 (W) (Figure 3).
Chlorophyll-a concentration is directly related to continuous water flow and management (feed) which determine higher or less availability of nutrients in the water which, in turn, may cause or not an increase in algae biomass (SIPAÚBA-TAVARES, 1995).
May (final grow-out period) presented in general, the highest values for limnological variables.Nutrients, inorganic carbon compounds and alkalinity showed higher concentrations in the water column due to the accumulation of rations added to the fishponds during grow-out fish period (November -March) (Figures 2 and 3).
Whereas water quality is the key to a successful aquaculture activity, water supply is of paramount importance to maintain good farming conditions.In the current study, the well (W) that supplied water to pond P 1 contained a reasonable nutrient concentration, which indicated a possible contribution to the subsequent fishponds.This was probably due to the fact that around the area there were different agricultural crops which integrated the University's experimental research.
When compared to other sites (P 2 -P 6 ), the inlet water in P 1 generally had low concentrations of analyzed variables.It received water directly from well (W), which supplies the other fishponds (P 2 to P 6 ).Pond P 1 has a water surface rich in floating, emergent and sub-emergent aquatic macrophytes, acting as a natural biofilter and, consequently, improving the quality of the water flowing to pond P 2 .
Sipaúba-Tavares et al. (2003) verified that the inlet water, after passing through the floating macrophyte located in the first 30 meters, presented reduced nitrite, total-phosphorus, orthophosphate, conductivity, alkalinity, and inorganic carbon.
The sequential distribution of the ponds and high allochthonous materials from pond P 4 influenced some limnological variables in pond P 5 .The tendency to increase limnological variable concentrations from pond P 4 onwards showed the need to treat the water before pond P 4 and P 5 .Another management practice employed in these ponds was the emptying of the ponds for harvesting.This procedure is not appropriate since it releases a great part of the nutrients on the sediment into the water column.Emptying and filling pond P 5 for harvesting prior to the start of current investigation may have contributed towards the re-suspension of the sediment, since nutrients, chlorophyll-a and conductivity rates were higher in this pond.
When water is drained directly from one pond to another, it may affect the characteristics of subsequent fishponds.It is thus recommended that partial emptying of the fishpond, with approximate 10% of water left in the system (BOYD; QUEIROZ, 2001).
The grow-out fish period (January) did not present serious problems.In fact, the heavy rains that occurred during January diluted the material contained in the ponds.Since fishponds are shallow, the rain and the wind homogenized the water by favoring aeration in the ponds (SIPAÚBA-TAVARES et al., 2007).
The maintenance of good water quality boils down to a good product on the market.Since the water is passed from one pond to another without previous treatment, the need for adequate management practices to the specific conditions of each pond is evident.Thus the inlet water in each fishpond should either have a inclination similar to that in pond P 2 to P 3 or the wetland should be set in an appropriate place.Emptying the ponds for harvesting should be reevaluated, since it disturbs the sediments and damages the water quality.Therefore, the technique of maintaining about 10% of the water volume and emptying slowly the pond may be adopted.Water quality should be monitored more frequently, not merely during the fish growout period when food addition is more intense, but also just after this period, owing to the high nutrient concentration reported during May.

Table 1 .
Results of two-way nested Anova.