ORIGINAL ARTICLES

Assessment of an artificial free-flow wetland system with water hyacinth (Eichhornia crassipes) for treating fish farming effluents¤

Evaluación de un humedal artificial con Jacinto de agua (Eichhornia crassipes) para el tratamiento de efluentes piscícolas

Avaliação de uma zona húmida artificial com Jacinto de água (Eichhornia crassipes) no tratamento de efluentes piscícolas

1 Centro de Investigación Piscícola, Facultad de Medicina Veterinaria y Zootecnia, Universidad de Córdoba.

2 Facultad de Ciencias Agrarias, Departamento de Producción Animal, BIOGEM, Universidad Nacional de Colombia Sede Medellín, Colombia.

* Corresponding author: Sandra C Pardo-Carrasco. Facultad de Ciencias Agrarias, Departamento de Producción Animal, BIOGEM, Universidad Nacional de Colombia, Sede Medellín. Calle 59A # 63-20, Bloque 50 oficina 314, Medellín, Colombia. E-mail: scpardoc@unal.edu.co.

(Received: May 30, 2013; accepted: November 16, 2013)

Summary

Background: fish farming effluents are mainly composed of organic matter and are considered a source of environmental pollution. Objective: to evaluate the efficiency of an artificial free-flow wetland system using water hyacinth (Eichhornia crassipes) to treat fish farming effluents under various hydraulic loadings. Methods: effluents generated from fingerling ponds of Oreochromis sp. and Piaractus brachypomus were passed through a constructed wetland system (40 m long and 7.7 m wide) to measure NO₂^-, NO₃^-, NH₄⁺, total phosphorus (TP), and Biochemical Oxygen Demand (BOD₅) removal efficiency. The hydraulic retention time was measured for six months in five assessment phases under real production conditions by using five hydraulic loadings (44.9, 45.3, 43.1, 41.6, 42.0 cm/day). Results: the hydraulic retention time of the constructed wetland system was 1.6 days, and its removal efficiency rates were: 67.9% for NH₄⁺, 32.1% for BOD₅, 27.1% for NO₂^-, 23.0% for TP, and 16.7% for NO₃^-. Removal rate was positively correlated with the loading rate of total inorganic nitrogen during the five phases of this study (r=0.956). Also, highest removal values and efficiency increase were reached in phase 5. Conclusions: the free-flow wetland with E. crassipes is efficient for removing nitrogen compounds, TP and BOD₅.

Resumen

Antecedentes: el efluente piscícola se compone principalmente de materia orgánica y es la principal fuente de impactos ambientales negativos. Objetivo: evaluar la eficiencia de un humedal artificial de flujo libre con Jacinto de agua (Eichhornia crassipes) en el tratamiento de efluentes piscícolas bajo diferentes cargas hidráulicas. Métodos: efluentes generados por estanques de alevinaje de Oreochromis sp. y Piaractus brachypomus se pasaron por el humedal artificial (40 m de largo y 7,7 m de ancho) y se determinó la eficiencia de eliminación de NO₂^-, NO₃^-, NH₄⁺, fósforo total (TP) y BOD₅ (Demanda Bioquímica de Oxigeno). Durante seis meses se determinó el tiempo de retención hidráulica en cinco fases de evaluación en condiciones reales de producción con cinco cargas hidráulicas (44,9, 45,3, 43,1, 41,6, 42,0 cm/día). Resultados: el tiempo de retención hidráulica del humedal artificial fue 1,6 días y registró eficiencias de eliminación de: NH₄⁺ (67,9%), BOD₅ (32,1%), NO₂^- (27,1%), TP (23,0%) y NO₃^- (16,7%). La velocidad de eliminación se correlacionó positivamente con la velocidad de carga del nitrógeno inorgánico total en las cinco fases de estudio (r=0,956); los mayores valores de eliminación y el incremento de su eficiencia se alcanzaron durante la fase 5. Conclusiones: el humedal artificial a flujo libre con E. crassipes es eficiente en la eliminación de compuestos nitrogenados, TP y BOD₅.

Palabras clave: agua de residuo piscícola, compuestos nitrogenados, macrófitas, remoción de nutriente.

Resumo

Antecedentes: o efluente piscícola se compõe principalmente de matéria orgânica e essa é a principal fonte de impactos ambientais negativos. Objetivo: avaliar a eficiência de uma zona húmida artificial de fluxo livre com Jacinto de água (Eichhornia crassipes) para o tratamento de efluentes piscícolas sob diferentes cargas hidráulicas. Métodos: efluentes gerados por viveiros de alevinagem de Oreochromis sp e Piaractus brachypomus se passaram pela zona húmida (40 m de comprimento e 7,7 m de largura) e foi determinada a eficiência da eliminação de NO₂^-, NO₃^-, NH₄⁺, fósforo total (FT) e BOD₅. Durante seis meses determinou-se o tempo de retenção hidráulica em cinco estágios de avaliação em condições reais de produção com cinco cargas hidráulicas (44,9; 45,3; 43,1; 41,6 e 42,0 cm/dia). Resultados: o tempo de retenção hidráulica da zona húmida artificial foi de 1,6 dias e registrou eficiências de eliminação: NH₄⁺ (67,9%), BOD₅ (32,1%), NO₂^- (27,1%), FT (23,0%) e NO₃^- (16,7%). A velocidade de eliminação correlacionou-se positivamente com a velocidade de carga do nitrogênio inorgânico total durante os cinco estágios de estudo (r=0,956); os valores de eliminação maiores e o incremento da sua eficiência se alcançaram durante o estágio 5. Conclusões: a zona húmida artificial a fluxo livre com E. crassipes é eficiente na eliminação de compostos nitrogenados, FT e BOD₅.

Palavras chave: agua de resíduo piscícola, compostos nitrogenados, macrófitas, remoção de nutrientes.

Introduction

Constructed wetland systems have been used for treating acid mine drainages, municipal surface water, industrial water, and livestock effluents. Constructed wetland systems have proven their ability to remove significant amounts of suspended solids, organic matter, nitrogen, phosphorus, trace elements, and microorganisms present in wastewater (Gentelini, 2007; Lee et al., 2009; Jing et al., 2001; Luna and Ramírez, 2004). The use of aquatic plants is becoming increasingly important for removing carbon compounds (measured as Biochemical Oxygen Demand (BOD₅)) form effluents and wastewater. Eichhornia crassipes is one of the most studied macrophytes used for the treatment of effluents. It is an outstanding species among freshwater hydrophyte communities from South America. This plant is widely used for treating wastewater, as it can assimilate and store pollutants, transport oxygen to the root area and foster a perfect medium for bacterial activity (Wedler, 1998). Moreover, it has high nutrient absorption capacity, especially for ammonium, nitrate and nitrite nitrogen (Wedler, 1998). Thus, constructed wetland systems are becoming increasingly used for the treatment of aquaculture effluents all over the world (Jing et al., 2001; Kadlec et al., 2000; Posadas, 2001; Lin et al., 2002; Lin et al., 2005). They are also a cleaner, more sustainable production alternative (Vinatea, 1999; New, 2003). Hence, this study evaluated the capabilities of water hyacinth in a constructed wetland system to treat fish farming effluents in northern Colombia. The efficiency of a free-flow constructed wetland system planted with water hyacinth was assessed for removing nitrogen compounds, total phosphorus, and BOD₅ while operating under commercial fingerling culture conditions.

Materials and methods

Ethical considerations

This study was approved by the Animal Experimentation Ethics Committee of the Fish Research Center (CINPIC), Universidad de Córdoba, Colombia (CINPIC 002 - May 18, 2008).

Layout

The wetland system was built in Aquacaribe farm (Córdoba, Colombia.), located at latitude 9°13'54'' N and longitude 75°49'11'' W, at 7 m.a.s.l. The annual mean temperature is 28 °C. The wetland was built on the ground (length: 40 m, width: 7.7 m; depth: 0.9 m). The bottom and slopes were covered with a 20 mm thick geomembrane (Figure 1) to isolate soil from water. The wetland was planted with water hyacinth (plants were the same size, age and color) collected from a nearby area. Plants were cleaned before planted in the wetland for further propagation. Water hyacinth was planted at a density of 5 ± 2 plants/m², and the average weight of plants was 2.2 ± 0.9 kg. Thus, a total of 1,500 macrophytes were planted during the study, reaching a biomass of 3.4 ton.

Wetland system operation

The system was monitored for 24 weeks (6 months) at two sampling locations (influent and effluent). During this time the fish farm effluent was added to the free flow wetland system (FWS) via gravity flow ditches. These entry points were not constant in the system, as they depended on the way in which water was replaced in the fish farm. The average flow rate of the wetland was estimated using weekly gauging with Bos equation (Bos, 1986), which uses the height and diameter of the stream to determine flow:

Where, D is stream diameter, H is stream height (m), and Q is flow (m³/s).

Efficiency assessment

The processes of pollution removal of the wetland were calculated using a first order kinetic model with piston flow, as established by the International Water Association (Kadlec et al., 2000).

Co/Ci = exp^(-K*t)

Where Ci is pollutant concentration in the influent (mg/L); Co is pollutant concentration in the effluent (mg/L); t is nominal hydraulic retention time (days) and K is the first order removal rate constant.

The loading rate of pollutants (LRP, g/m²/day) was estimated using the following equation:

LRP = HLR*Ci

Where HLR is the hydraulic loading rate (mg/L) and HRL was estimated using the following equation:

HRL = Qi*Aw

The pollutant removal rate (PRT, g/m²/day) was calculated using the following equation:

PRT = HLR*(Ci-Co)

The average rate of influent and effluent (Q) was also determined, and the hydraulic loading (q) was measured as the average flow rate (Q) divided by the wetland area. Finally, the nominal hydraulic retention time (t) was calculated by dividing the depth of the wetland system by the average flow rate.

Sample analysis

Physical and chemical parameters were measured three times per week via in situ readings. Parameters measured were pH, dissolved oxygen (DO), and temperature (T) at the influent and effluent points of the wetland. Measurements were taken once per day from 8:00 am to 10:00 a.m. with a digital oximeter (YSI, 550A, Yellow Springs, Ohio, USA) and a digital potentiometer (YSI, pH100), calibrated in accordance with the manufacturer's instructions. Measurements were taken at an average depth of 25 cm at the influent and effluent points of the system. Water was sampled at two points to measure nutrients using 2-L amber bottles, which were packed and sent to the Regional Water Institute of the University of Córdoba (IRAGUAS) where concentration of BOD₅, NO₂^-, NO₃^-, NH₄⁺, and total phosphorus (TP) were determined using procedures and techniques given by the 24 American Public Health Association (APHA, 1980). Results on total ammonia nitrogen (TAN) were determined as the sum of the values for NO₂^-, NO₃^-, and NH₄⁺.

A regression analysis between removal rate and loading rate of TAN was carried out to obtain the curve and best-fit equation.

Results

Wetland system setting

Vegetation grew rapidly. Its initial density was 5 ± 2 plants/m², and its density in phase 3 was 13 ± 3 plants/ m². In this phase, the differences between the values for nitrogen compound concentration, total phosphorus and BOD₅, as well as the difference between the influent and effluent were low. However, an increase in the removal rate of nitrogen compounds was observed in phase 5.

Similarly, from phase 3 onward, the studied parameters had a removal trend, with hydraulic loadings ranging from 41.6 to 45.3 cm/day (Table 2). This removal rate became observable after week 7 for inorganic nitrogen and BOD₅, and after week 10 for total phosphorus.

Nitrogen Removal

Pollutant concentration in the influent during the study period ranged from 0.005 to 0.029 mg/L for NO₂^-; 0.8 to 1.8 mg/L for NO₃^-; and 0.99 to 1.45 mg/L for NH₄⁺.

The average value for NH₄⁺ removal was 0.36 ± 0.38 g/m2/day, which represents 67.9% efficiency throughout the entire study. Phase 1 had the highest removal value for NH₄⁺: 0.88 g/m²/day, while the lowest value, -0.03 g/m²/day, was observed during phase 3 (data not shown).

The average rate of NO₃^- removal was 0.08 ± 0.05 g/m²/day, which corresponds to 16.7% efficiency. Additionally, a peak in this rate was observed in phase 4, with a value of 0.15 g/m²/day, which is equivalent to 31% efficiency (values calculated using data from Table 2). On the other hand, NO₂^- had an average of 0.001 ± 0.002 g/m²/day, that is, 27.1% efficiency. Negative values were recorded during phases 2 (-0.001 g/m²/day) and 3 (-0.003 g/m²/day). However, the highest removal value was reached in phase 5: 0.004 g/m²/day, corresponding to 37.9% efficiency.

The concentration of NO₂^- and TAN in the effluent was highly correlated with the loading rate of the pollutant, their r-values being 0.697 and 0.546, respectively, during the evaluation period. However, nitrogen compounds maintained low values in the effluent: less than 0.11 mg/L for NO₂^-; less than 1.8 mg/L for NO₃^-; and less than 2.9 mg/L for NH₄⁺. The TAN loading rate for each study phase had a high correlation with the TAN removal rate: r=0.956 (Figure 2).

TP concentration in the influent during the study ranged from 0.29 to 0.48 mg/L. The average P removal was 0.034 ± 0.038 g/m²/day, which represents 23% efficiency. The removal data were negative in phases 1 (-0.002 g/m²/day) and 2 (-0.010 g/m²/day), and the last phase had the highest TP removal rate: 0.077 g/m²/day, corresponding to 33.9% efficiency. The efficiency values of the system had a strong tendency to improve as the pollutant rate of hydraulic loading decreased.

There was a low positive correlation (r=0.271) between rate of hydraulic loading and TP concentration in the effluent. A low linear correlation (r=0.07) was observed during all phases of the study when relationships between variations in rate of loading and rate of TP removal in the wetland were established. However, a high correlation (r=0.96) was observed when the same correlation did not include values from phases 1 and 2 or from the stabilization period.

BOD₅ removal

Pollutant concentration in the influent during the study period ranged from 3.36 to 7.62 mg/L for BOD₅. The average BOD₅ removal during the period was 0.70 ± 0.44 g/m²/day, which corresponds to 32.1% efficiency. Although this value was negative in phase 1 (-0.04 g/m²/day), it ranged from 0.67 to 0.91 g/m²/day in the subsequent phases, and reached its maximum in phase three (1.11 g/m²/day), which corresponds to 47.6% removal efficiency.

A positive linear correlation (r=0.494) was observed when comparing the loading rate with BOD₅ concentration in the effluent. The average removal rate of BOD₅ was greater for higher hydraulic loadings, with values approaching 3 mg/L.

Discussion

Hydraulic conditions

Constructed wetland systems function better with a constant water flow allowing for appropriate removal of entering pollutants (Kadlec et al., 2000; Lin et al., 2002). Water inflow in the wetland depended upon the actual production conditions of the fish farm where the study took place. Thus, influent and effluent were affected by emptying, drying and replacement processes. In this study, the wetland flow ranged from 127.2 m³/day to 136.8 m³/day, and the estimated average value was 132.7 ± 5.1 m³/day. The average flow decreased 6.08 m³/day in the section between influent and effluent. The hydraulic retention time (HRT), also known as hydraulic residence time or t (tau), is a measure of the average time that a soluble compound remains in a constructed wetland.

Wetland system stabilization

A constructed wetland system generally requires an extended period of time to stabilize bacterial communities and removal processes (Kadlec et al., 2000). The soil added to the wetland bed provides ideal environments for the development of microbial processes such as nutrient nitrification, denitrification, and mineralization.

In this study, a maximum density of 13 ± 3 plants/m² was achieved at the end of phase 4, five months after water hyacinth were planted. This correlates with higher nutrient removal efficiency rates. It was suggested that plants did not cover the system completely at the beginning and the associated bacterial communities did not have favorable conditions for organic matter decomposition. Lin et al. (2002) planted Ipomoea aquatica on the front side of a FWS and Paspalum vaginatum on the other side under controlled conditions. They obtained good retention rates between the second and third months, but the densities used were greater than 30 plants/ m². Likewise, Lin et al. (2002) explained that their wetland system required approximately seven months to reach vegetation coverage approaching 80%, whereas in our study the vegetation covered more than 80% within five months, given the characteristics of the species used (E. crassipes) and the local weather conditions. This suggests that performance level established for pollutant removal may be achieved without full vegetation coverage, confirming the work by Lin et al. (2002). The results of this study suggest that replacing the plants five months after planting them would maintain the pollutant removal tendency and create macrophyte-free zones followed by plant-covered zones, thus allowing for high levels of dissolved oxygen, thereby increasing nutrient removal values.

Posadas (2001) stated that constructed wetland systems with 25% of the total size of the production ponds and 2-day retention time significantly improved channel catfish (Ictalurus punctatus) production. In this study, the size of the constructed wetland system was only 0.9% of the total size of the farm's reflecting pool, yet good removal rates were achieved due to climatic conditions and effluents had the largest loading of nitrogen compounds because they originated from fish and fingerling feeding processes. This is consistent with the results reported by Schwartz and Boyd (1995), who estimated that the area to treat aquaculture effluents should be 0.7 to 2.7 times the size of the pond.

Removal of nitrogen compounds

NH₄⁺ removal reached 67.9% efficiency throughout the study, suggesting that the system attained a certain balance due to macrophytes growth and their interrelationship with the microbial biofilm associated to the roots and the organic matter at the bottom of the pond. Phase 4 showed a reduction of 99.1% of NH₄⁺. These results are consistent with those reported by Spieles et al. (2000), who suggested that NH₄⁺ removal of a wetland system can range from 25 to 85%. In the present study, it was necessary to wait for two months after planting the water hyacinth in order to obtain sufficient NH₄⁺ removal rates. Moreover, it is possible that at the observed temperature (27 to 30 °C) macrophyte growth might have fostered the development of aerobic nitrifying organisms that could have performed the biological conversion from ammonium to nitrate.

Nitrite is an intermediate product of the nitrification process, which turns ammonium into nitrate (Fernandez et al., 2005). It is removed from wetlands mainly through oxidation by microorganisms of the Nitrobacter genus. The mean value of NO₂^- obtained in the influent for the entire process was 0.012 ± 0.01 mg/L, while the value in the effluent was 0.009 ± 0.01 mg/L, which corresponds to 27.1% efficiency. Phase 5 showed the best removal efficiency (37.9%). Therefore, it can be assumed that at this stage the wetland system became stable and bacteria developed, which enabled removal of pollutants and water selfpurification.

Nitrate remains in the water column, where it can be either assimilated by the macrophytic bacteria/microorganisms, or reduced to nitrogen gas by heterotrophic anaerobic bacteria. In this study, an average of 1.2 ± 0.2 mg/L NO₃^- was observed in the influent. The effluent, in turn, had a value of 1.0 ± 0.1 mg/L; hence, the removal rate was 14.2%. However, nitrate removal rate ranged from 26.7 to 30.8% in phases 4 and 5. Other authors conducting experiments under similar conditions reported similar values (Lin et al., 2005; Schulz et al., 2004). This proves that this parameter may have low removal efficiency under real production conditions. Comparing these data with studies conducted under similar conditions, we suggest that anaerobic bacteria communities in the bed of the system –responsible for the denitrification process– had limited growth. On the other hand, Nitrosomonas and Nitrobacter, responsible for the nitrification process that ultimately produces nitrate, experienced much better growth.

Total phosphorus removal

BOD₅ removal

BOD₅ determines the availability of dissolved oxygen in the influent and the type of microorganisms participating in the organic matter degradation process (Fernandez et al., 2005). If such microorganisms are aerobic, then the reactions will be quick and efficient. Conversely, if the organisms are anaerobic, the reactions will be slow and inefficient (Fernandez et al., 2005). Therefore, oxygen availability is crucial for the biodegradation of organic matter. The average BOD₅ removal efficiency observed was 19.1% throughout the study. These removal values are lower than those reported by Ramírez et al. (2005), but higher than those reported by Lin et al. (2005), who used Phragmites australis under controlled conditions to obtain averages of 4.6 mg/L in the influent and 4.1 mg/L in the effluent. This constitutes a removal of 10.9% of the BOD₅. Low BOD₅ removal rates could be due to the fact that the transfer by diffusion of oxygen from the submerged parts of the macrophytes (aerenchyma) to the effluents was low. Likewise, the vegetation reduced the rate of atmospheric air incorporation by physical processes.

The results of this study allow us to conclude that the constructed wetland system with water hyacinth (E. crassipes) can remove BOD₅, NH₄⁺, NO₃^-, NO₂^- and TP when it is operated with an average flow of 132.7 ± 5.1 m³/day, a mean hydraulic retention time of 1.6 ± 0.1 days, and a mean hydraulic loading of 43.4 ± 1.7 cm/day. This assumes assessed conditions for an area equivalent to 0.9% of the fish production ponds.

Acknowledgements

This study was fully sponsored by the Universidad de Córdoba. The authors would like to thank Aquacaribe Ltda. for participating in this study.

Notes

¤ To cite this article: Díaz A, Atencio V, Pardo S. Assessment of an artificial free-flow wetland system with water hyacinth (Eichhornia crassipes) for treating fish farming effluents. Rev Colomb Cienc Pecu 2014; 27:202-210.

References

American Public Health Association. APHA, AWWA, WPCF Standard Methods for the Examination of Water and Wastes. 1980.         [ Links ]

Bos M, Replogle J, Clemmens A. Aforadores de caudal para canales abiertos. International Institute for Land Reclamation and Improvement/ILRI. Wageningen: The Netherlands; 1986.         [ Links ]

Fernández J, de Miguel E, de Miguel J, Curt M. Manual de fitodepuración. Filtros de macrófitas en flotación; 2005. p.61-77.         [ Links ]

Gentelini A. Tratamiento de Efluente de Piscicultura Orgânica utilizando macrófitas acuáticas Eichhornia crassipes (Mart. Solms) e Eigeria densa (Planchon) [Maestría en Ingeniería Agrícola]. Cascavel: (Bra) Universidade Estadual do Oeste do Paraná; 2007.         [ Links ]

Henry-Silva G, Camargo A. Efficiency of aquatic macrophytes to treat Nile tilapia pond efluents. Scientia Agricola 2006; 63:433-438.         [ Links ]

Jing S, Lin Y, Lee D, Wang T. Nutrient removal from pullet river water by using construted wetland. Bioresour Technol 2001; 76:131-135.         [ Links ]

Kadlec R, Knight R, Vymazal J, Brix H, Cooper P, Haberl R. Constructed Wetlands for Pollution Control Processes, Performance, Design and Operation. IWA publishing; 2000.         [ Links ]

Lee C, Fletcher T, Sun G. Nitrogen removal in constructed wetland systems. Eng Life Sci 2009; 9:11-22.         [ Links ]

Lin Y, Jing R, Lee Y, Wang W. Nutrient Removal from Aquaculture Waste Water Using a Constructed Wetland System. Aquaculture 2002; 169:184.         [ Links ]

Lin Y, Jing S, Lee D, Chang Y, Chen V, Shih K. Performance of a constructed wetland treating intensive shrimp aquaculture wastewater under high hydraulic loading rate. Environ Pollut 2005; 134:411-425.         [ Links ]

Llagas A, Gómez E. Diseño de humedales artificiales para el tratamiento de aguas residuales en la UNMSM. Rev Inst investig Fac Minas Metal Cienc 2006; 9:17.         [ Links ]

Luna V, Ramírez H. Medios de soporte alternativo para la remoción de fósforo en humedales artificiales. Rev Int Contam Ambient 2004; 20:31-38.         [ Links ]

New M. Responsible aquaculture: is this a special challenge for developing countries? World Aquac 2003; 34:49-52.         [ Links ]

Pardo-Carrasco S, Suárez H, Jardim-Manso P, Soriano-Sierra E. Diseño de una herramienta de evaluación del desempeño ambiental en las granjas piscícolas. Rev MVZ-Córdoba 2005; 10:602-613.         [ Links ]

Posadas B. Comparative economic analysis of using different sizes of constructed wetlands in recirculating catfish pond production. J Appl Aquac 2001; 11:1-19.         [ Links ]

Ramirez E, Robles E, Bonilla P, Sainz G, Lopez M, De La Cerda J, Warren A. Occurrence of pathogenic free-living amoebae and the indicators in a constructed wetland treating domestic wastewater. Eng Life Sci 2005; 5:253-257.         [ Links ]

Schulz C, Gelbrecht J, Rennert B. Constructed wetlands with free water surface for treatment of aquaculture effluents. Journal Appl Ichthyiol 2004; 20:64-70.         [ Links ]

Schulz C, Gelbrecht J, Rennert B. Treatments of rainbow trout farm effluents in contructed wetland with emergent plants and subsurface horizontal water flow. Aquac 2003; 217:207-221.         [ Links ]

Schwartz M, Boyd C. Contructed wethlands for treatment of channel catfish pond effluents. Prog Fish-Cult 1995; 57:255-267.         [ Links ]

Spieles D, Mitsch P. The effects of season and hydrologic and chemical loading on nitrate retention in constructed wetlands: A comparison of low and high nutrient riverine systems. Ecol Eng 2000; 14:77-91.         [ Links ]

Tilley D, Badrinarayan H, Rosati R, Son J. Contructed wetlands as recircultation filters in large-scale shrimp aquaculture. Aquac Eng 2002; 26:81-109.         [ Links ]

Troell M, Neori A, Chopin T, Buschmann AH. Biological wastewater treatment in aquaculture more than just bacteria. World Aquaculture 2005; 36:27-29.         [ Links ]

Vinatea L. Aqüicultura e desenvolvimento sustentável Brasil: UFSC; 1999.         [ Links ]

Vinatea L. Sostenibilidad en acuicultura: Desafíos. En: memorias V seminario de Internacional de acuicultura, II congreso de investigaciones acuícolas; Bogotá 2005. Universidad Nacional de Colombia; 2005. p.1-4.         [ Links ]

Wedler E. Introducción en la acuicultura con énfasis en los neotrópicos. Barranquilla: Litoflash; 1998.         [ Links ]

Yang L, Chou L, Shieh W. Biolfiter treatment of aquaculture water for reuse applications. Wat Res 2001; 35:3097-3108.         [ Links ]