Hybrid wastewater treatment system using anaerobic microorganisms and reed (Phragmites communis)

A small hybrid wastewater treatment system consisting of a settling tank in series with an anaerobic filter-reed (Phragmites communis) treatment cell was evaluated and compared with a similar plant-free system. Data demonstrated that by combining anaerobic filters, also referred to as attached film filters, and vascular aquatic plants a synergistic effect is produced which increases the treatment efficiency of each individual system. The plant-free system reduced the BOD₅ from 114 to 31 mg/l in 6 h as compared to a reduction of 110 to 9 mg/l in the anaerobic filter-reed system in the same length of time. The BOD₅ and TSS after 24 h in each component of plant-free system were reduced from 114 mg/l to 14 mg/l and 51 to 15 mg/l, respectively. Under the same conditions, the hybrid system reduced the BOD₅ from 110 to 3 mg/l and the TSS from 68 to 6 mg/l. The hybrid system also reduced the total Kjeldahl nitrogen (TKN) from 16.1 to 2.9 mg/l, total phosphorus (TP)from 4.4 to 2.0 mg/l, and the ammonia (NH₃-N)from 12.4 to 0.6 mg/l after 24 h of exposure while the plant-free system demonstrated insignificant reduction of these components.     

Treatment of landfill leachate using UASB-MBR-SHARON–Anammox configuration

Leachate treatment is a challenging issue due to its high pollutant loads. There are several studies on feasible treatment methods of leachate. In the scope of this study, high organic content of young leachate was eliminated using an upflow anaerobic sludge blanket (UASB) and a membrane bioreactor (MBR) in sequence and effluent of the system was given to single reactor for high activity ammonia removal over nitrite (SHARON) and anaerobic ammonia oxidation (Anammox) reactors to remove nitrogen content. All reactors were set up at lab scale in order to evaluate the usage of these processes in sequencing order for leachate treatment. COD and TKN removal efficiencies were over 90 % in the combined processes which were operated during the study. The biodegradable portion of organic matter was removed with an efficiency of 99 %. BOD₅ concentration decreased to 50 mg/L by UASB and MBR in sequence even the influent BOD₅ concentration was over 8,000 mg/L. Although high nitrogen concentrations were observed in raw leachate, successful removal of nitrogen was accomplished by consecutive operations of SHARON and Anammox reactors. The results of this study demonstrated that with an efficient pretreatment of leachate, the combination of SHARON–Anammox processes is an effective method for the treatment of high nitrogen content in leachate.     

Generating “Tide” in Pilot‐Scale Constructed Wetlands to Enhance Agricultural Wastewater Treatment

The operation of tidal flow was studied using a pilot‐scale system treating high strength piggery wastewater. Located on a farm in Staffordshire, UK, the system consisted of five wetland treatment stages vegetated with common reeds of Phragmites australis. Wastewater samples were collected from the inlet and outlet of each stage and analyzed for BOD₅, COD, NH₄‐N, NO₃‐N, NO₂‐N, SS, PO₄‐P and pH. Average hydraulic and organic loadings on the system were 0.12 m³/m² d and 240 g BOD/m² d, respectively, which is considerably higher than the typical loadings on conventional subsurface flow systems. On average, BOD₅ and COD were reduced by 82 % and 80 % from initial concentrations of 2000 mg/L and 2750 mg/L, respectively, across the whole system. The first‐order kinetics constant for BOD₅ removal (K_BOD in m/d) in this tidal flow system is approximately 2.5 times the rate constant obtainable in a typical horizontal flow system, demonstrating a more efficient removal of organic matter in tidal flow wetlands. The overall efficiency of the system was found to increase with time before stabilizing towards the end of a start‐up period. Straight‐line correlations were established between the loading and removal of BOD₅ and COD. Contributions by individual stages to the overall treatment were analyzed. SEM images of wetland media demonstrated the formation of biofilms and microbial activities inside the matrices of the wetland system, which accounted for the degradations of organic pollutants.     

The Potential of Canna lily for Wastewater Treatment Under Indian Conditions

Low cost treatment of polluted wastewater has become a serious challenge in most of the urban areas of developing countries. The present study was undertaken to investigate the potential of Canna lily towards removal of carbon, nitrogen, and phosphorus from wastewater under sub-tropical conditions. A constructed wetland (CW) cell supporting vegetative layer of Canna lily was used to treat wastewater having high strength of CNP. Removal of biological oxygen demand (BOD₃) and chemical oxygen demand (COD) varied between 69.8–96.4% and 63.6–99.1%, respectively. C. lily could efficiently remove carbon from a difficult to degrade wastewater at COD:BOD ratio of 24.4. Simultaneous reduction in TKN and nitrate pointed to good nitrification rates, and efficient plant assimilation as the dominant nutrient removal mechanism in the present study. Suitable Indian agro-climatic conditions favored plant growth and no evident stress over the Canna plant was observed. High removal rate of 809.8 mg/m²-day for TKN, 15.0 mg/m²-day for nitrate, and 164.2 mg/m²-day for phosphate suggests for a possible use of Canna-based CW for wastewater treatment for small, rural, and remote Indian communities.     

Biomonitoring of wastewaters in treatment plants using ciliates

The aim of this project has been to study and compare the ciliate populations present in roptating biological reactors treating three different wastewaters. Wastewaters chosen were a maize mill (nejayote), a sugarcane/ethyl alcohol plant (vinasses) and a recycled paper mill (whitewaters). The initial dissolved organic contents, measured as soluble chemical oxygen demand (COD) and biochemical oxygen demand in five days (BOD₅), were 2040±150 mg COD L^–1 and 585±5 mg BOD₅L^–1 for nejayote; 2000±20 mg COD L^–1 and 640±5 mg BOD₅ L^–1 for vinasses and 960±200 mg COD L^–1 and 120±10 mg BOD₅ L^–1 for whitewaters. Results obtained indicate that ciliated protozoa proliferated in the different chambers of each rotating biological reactor (RBR). Saprobity indices, as a quantitative evaluation parameter, indicate that there are no universal species of ciliates associated with specific BOD₅ concentrations. Therefore, the number of species of ciliates present in the effluent indicate qualitatively the efficiency of removal of pollution from the wastewaters during treatment in the rotating biological reactors.     

Experience with establishment and operation of reed bed treatment for small communities in the UK

The design and development of constructed reed beds for secondary, tertiary and storm outflow treatment is described from the experience of a major water utility in the Midland region of the UK. There were only two trial sites in 1987, but 125 sites by June 1995. The problems and their resolution with the establishment ofPhragmites in gravel filled beds are described.Two secondary treatment systems with horizontal subsurface flow beds arranged in terraces show that dilute sewage can be treated to a good secondary standard, but that nitrification is likely to be only partially complete. The example of four tertiary treatment applications showed that, for beds sized at 0.7 to 1.2 m²/pe, effluents consistently averaged better than 5 mg/l BOD₅ and 10 mg/l TSS. Three sites showed improved removal of ammonium nitrogen (N) after the first year, with better than 50% removal. One site, treating highly nitrified secondary effluent, showed a small net increase in ammonium-N even after 4 years. Samples taken by the regulatory authority from 79 sites confirm the ability of the system to meet tight standards for BOD₅. A comparison made of the performance of storm reed beds during three storm events showed relatively consistent removal of BOD, TSS, ammonia nitrogen and TON.     

Appraising bacterial strains for rapid BOD sensing—an empirical test to identify bacterial strains capable of reliably predicting real effluent BODs

The measured response of rapid biochemical oxygen demand (BOD) biosensors is often not identical to those measured using the conventional 5-day BOD assay. This paper highlights the efficacy of using both glucose–glutamic acid (GGA) and Organisation for Economic Cooperation and Development (OECD) BOD standards as a rapid screen for microorganisms most likely to reliably predict real effluent BODs when used in rapid BOD devices. Using these two synthetic BOD standards, a microorganism was identified that produced comparable BOD response profiles for two assays, the MICREDOX® assay and the conventional 5-day BOD₅ test. A factorial experimental design systematically evaluated the impact of four factors (microbial strain, growth media composition, media strength, and microbial growth phase) on the BOD response profiles using GGA and OECD synthetic standard substrates. An outlier was identified that showed an improved correlation between the MICREDOX® BOD (BOD_sens) and BOD₅ assays for both the synthetic standards and for real wastewater samples. Microbial strain was the dominant factor influencing BOD_sens values, with Arthrobacter globiformis single cultures clearly demonstrating superior rapid BOD_sens response profiles for both synthetic and real waste samples. It was the only microorganism to approach the BOD₅ response for the OECD substrate (171 mg O₂?L^?1), and also reported BOD values for real waste samples that were comparable to those produced by the BOD₅ test, including discriminating between filtered and unfiltered samples.     

The performance of a multi-stage system of constructed wetlands for urban wastewater treatment in a semiarid region of SE Spain

This paper describes the results obtained in an experimental multi-stage system of created wetlands in Mojacar, in semiarid SE Spain, operating from June to October 1997. We compare the removal efficiency of four different series of treatments each consisting of three stages, using different flow rates of sewage, flow regimes, types of substrate and influents. Pretreated water from an anaerobic stabilization pond and treated water from the last pond of a lagoon system were used, the latter to test the system's suitability as a complementary system for removing nitrogen and phosphorus. In spite of the initial high wastewater concentrations, the effluent conforms to the strictest European norms (directive 91/271) for primary and secondary retention. A net treatment area of 2.3 m²/PE showed a high performance for SS (90–96%), COD (87%) and BOD₅ removal (90%) during the early stages of operation; however, nutrient removal was lower than was expected as compared with other studies. The addition of iron to the substrate improved phosphorus retention significantly (from 55 to 66%). The decrease of the net treatment area to 1.2 m²/PE did not significantly affect the wetland performance, with the exception of COD removal (78%). Series fed with treated water from the lagoon system (1.6 m²/PE) noticeably improved the quality of the effluent (average values of 7 mg/l total-N and 3 mg/l total-P).     

CONSTRUCTED WETLAND CONTROL OF BOD LEVELS IN AIRPORT RUNOFF

A surface water treatment system consisting of an aeration reservoir and pond (holding capacities 45,000 and 19,000 m³) and a network of 12 horizontal subsurface flow gravel-filled constructed wetland cells of different sizes (total surface area 2.08 ha) and planted with Phragmites australis, was commissioned at Heathrow Airport, London, United Kingdom, in the winter of 2002. Ongoing monitoring of the treatment system has shown significant reductions in the biochemical oxygen demand (BOD₅) throughout the system with levels decreasing by up to 76.7% across the constructed wetland cells following high anti- and de-icing fluid applications. However, continued exposure to BOD₅ concentrations exceeding the design target has resulted in anaerobic conditions in the wetland. The addition of nutrients to the treatment system has resulted in improved removal efficiency for elevated BOD₅ loadings in the aerated reservoir from 25.5% to 47.5%, The addition of different nutrient dosing regimes to complementary pilot-scale planted and unplanted vertical flow columns showed average but statistically insignificant BOD₅ removal percentage increases from 61.9 ± 21.1% to 70.8 ± 26.5%, respectively, in planted columns over a 7-day period. There is an overall improvement in the performance of the system, but operational reviews are continuing.     

Feasibility study of a cyclic anoxic/aerobic two-stage MBR for treating ABS resin manufacturing wastewater

This study investigated the feasibility and the treatment efficiency of a cyclic anoxic/aerobic two-stage MBR for treating polymeric industrial wastewater. The anoxic/aerobic hybrid MBR was operated without sludge withdrawal except sampling during the study. The results showed that the highest COD organic loading rate of 8.7 kg COD/m³ day from bioreactor was obtained at phase 3. The system achieved 97% BOD₅ and 89% COD removal. It also revealed that 93% of COD removal was contributed by bioreactor at phase 3 and the similar results happened to phases 1 and 2. The highest TN and TKN removals for each phase were 60, 74, 80% and 61, 74, 81%, respectively and limited by nitritation step. SEM images of nascent and fouled membranes were offered to evaluate the cleaning method. The system was operated for 174 days, resulting in high degradation rate, flexibility towards influent fluctuations and limited sludge production.     

Biochemical mediator demand – a novel rapid alternative for measuring biochemical oxygen demand

The biochemical oxygen demand (BOD) test (BOD₅) is a crucial environmental index for monitoring organic pollutants in waste water but is limited by the 5-day requirement for completing the test. We have optimised a rapid microbial technique for measuring the BOD of a standard BOD₅ substrate (150 mg glucose/l, 150 mg glutamic acid/l) by quantifying an equivalent biochemical mediator demand in the absence of oxygen. Elevated concentrations of Escherichia coli were incubated with an excess of redox mediator, potassium hexacyanoferrate(III), and a known substrate for 1 h at 37 °C without oxygen. The addition of substrate increased the respiratory activity of the microorganisms and the accumulation of reduced mediator; the mediator was subsequently re-oxidised at a working electrode generating a current quantifiable by a coulometric transducer. Catabolic conversion efficiencies exceeding 75% were observed for the oxidation of the standard substrate. The inclusion of a mediator allowed a higher co-substrate concentration compared to oxygen and substantially reduced the incubation time from 5 days to 1 h. The technique replicates the traditional BOD₅ method, except that a mediator is substituted for oxygen, and we aim to apply the principle to measure the BOD of real waste streams in future work. Received: 2 August 1999 / Received revision: 6 December 1999 / Accepted: 12 December 1999     

Polishing low-biodegradable and saline industrial effluent in a full-scale horizontal subsurface flow constructed wetland: evaluation of bio-treatability and predictive power of kinetic models

Abstract

This study evaluates the bio-treatability performance and kinetic models of full-scale horizontal subsurface flow constructed wetland used for the tertiary treatment of composite industrial effluent characterized by high-salt content ranging from 5830 to 10,400 µS/cm and biochemical oxygen demand (BOD₅): chemical oxygen demand (COD) ratio below 0.2. The wetland vegetated with Phragmites australis was operated in a semi-arid climate under an average hydraulic loading rate of 63?mm/d. The results of a 4-year operation calculated based on the concentration of pollutants showed that the average removal efficiency of COD, BOD_5, and total suspended solids (TSS) were 17.5, 5.1, and 11.2%, respectively. The system reduced up to 6.5?±?0.7% of electrical conductivity presenting poor phyto-desalination potential without considering the contribution of evapotranspiration in water balance in contrast to satisfying performance for heavy metals reduction. The comparison of the kinetics of organic matter removal obtained by the first-order and Monod models paired with continuous stirred-tank reactor and plug flow regime showed that Monod-plug flow model provided the best fit with the constants of 2.01?g COD/m²·d and 0.3014?g BOD₅/m²·d with the best correlation coefficient of 0.610 and 0.968 between the predicted and measured concentrations, respectively. The low kinetic rates indicate that the process is capable of effluent polishing instead of purification due to the presence of organic compounds recalcitrant to biodegradation and a high level of salinity.     

Pollution control in pulp and paper industrial effluents using integrated chemical–biological treatment sequences

The main objective of the present study was to improve the quality of pulp and paper industrial wastewater of two local mills RAKTA and El-Ahlia, Alexandria, Egypt, and to bring their pollutant contents to safe discharge levels. Quality improvement was carried out using integrated chemical and biological treatment approaches after their optimization. Chemical treatment (alum, lime, and ferric chloride) was followed by oxidation using hydrogen peroxide and finally biological treatment using activated sludge (90 min for RAKTA and 60 min for El-Ahlia effluents). Chemical coagulation produced low-quality effluents, while pH adjustment during coagulation treatment did not enhance the quality of the effluents. Maximum removal of the tested pollutants was achieved using the integrated treatment and the pollutants recorded residual concentrations (RCs) of 34.67, 17.33, 0.13, and 0.43 mg/l and 15.0, 11.0, 0.0, and 0.13 mg/l for chemical oxygen demand (COD), biochemical oxygen demand (BOD₅), tannin and lignin, and silica in RAKTA and El-Ahlia effluents, respectively, all of which were below their maximum permissible limits (MPLs) for the safe discharge into water courses. Specific oxygen uptake rate (SOUR) and sludge volume index (SVI) values reflect good conditions and healthy activated sludge. Based on the previous results, optimized conditions were applied as bench scale on the raw effluents of RAKTA and El-Ahlia via the batch chemical and the biological treatment sequences proposed. For RAKTA effluents, the sequence was as follows: (1) coagulation with 375 mg/l FeCl₃, (2) oxidation with 50 mg/l hydrogen peroxide, and (3) biological treatment using activated sludge with 2,000 mg/l initial concentration and 90 min hydraulic retention time (HRT), while for El-Ahlia raw effluents, the sequence was (1) coagulation with 250 mg/l FeCl₃, (2) oxidation with 45 mg/l hydrogen peroxide, and (3) biological treatment using activated sludge with 2,000 mg/l initial concentration and 60 min HRT. In conclusion, results confirmed that the application of the proposed sequential treatments removed almost all COD, BOD₅, high molecular weight compounds, and silica from RAKTA and El-Ahlia influents and produced high-quality effluents, thus achieving the main objective of this study.     

Performances of constructed wetlands for municipal wastewater treatment in rural mountainous area

A global performance evaluation of an experimental Horizontal SubSurface Flow Constructed Wetlands (HSSF) was made after 6 years of functioning. This wetland is situated in French prealpine mountain, at 720 m elevation. The HSSF process treatment consists in a three-stage system dimensioned for 350 People Equivalent. Different helophytes were planted such as Typha latifolia, Phragmites australis and Scirpus maritimus. The mean hydraulic residence time for sewage was closed to 4–5 days, but in summer the mean pollutant residence time increases to 6 days due to an important evapotranspiration. There is no clogging of the gravel matrix and the hydraulic conductivity was very good and stabilized. Removal pollutant efficiency was determinate at each stages. There was a high removal of total suspended solids (TSS) all year around with an average of 95.6% (±3.6). More than 80% of removal occurred in the first stage. Physical processes (decantation, filtration) associated with biological oxidation were the principal factors of this removal. For COD and BOD₅, removal efficiency in the first stage were close to 60% on average and more than 90% at the outlet of the wetland. These results are similar to those observed with SSFW in many cases. Influence of temperature seems very weak because there were no significant seasonal variations of the process efficiency. Minimum effluent quality standards (30 mg l^–1 TSS; 120 mg l^–1 COD; 40 mg l^–1 BOD₅) were always respected. In cold periods, nutrients uptake was reduced but remained up to 60% in average. Mean bacterial removal efficiency was about two order of magnitude (99%) but can reach up to five order of magnitude in summer. These cyclic variations follow a sinusoidal variation around an annual mean. Pollutants removals were correlated to their respective loadings and no limits has been observed except for nitrogen. These results confirm that SSFW, an ecotechnology, should be considered as an alternative to conventional treatment methods (activated sludge, fixed biofilm) for small communities even in mountainous area.     

Performance of an anaerobic baffled reactor and hybrid constructed wetland treating high-strength wastewater in Nepal—A model for DEWATS

Centralized wastewater treatment systems require sophisticated technologies and skilled manpower for their operation and maintenance (O&M). These systems have huge construction as well as O&M costs. Therefore, a Decentralized Wastewater Treatment System (DEWATS) rather than a centralized system might be especially beneficial in developing countries. A model for DEWATS is developed in Nepal with Anaerobic Baffled Reactor (ABR) and hybrid Constructed Wetland (CW). The DEWATS treats high-strength wastewater from 80 households (400 PE). This paper summarizes the performance of the DEWATS from July 2006 to August 2007 in the removal efficiencies of TSS, BOD₅, COD, NH₄–N, TP and FC. The ABR is very effective in the removal of organic pollutants and could achieve TSS removal up to 91%, BOD₅ up to 78% and COD up to 77%. The average removal efficiencies of the DEWATS is 96% TSS, 90% BOD₅, 90% COD, 70% NH₄–N, 26% TP and 98% FC.     

Enhanced denitrification and organics removal in hybrid wetland columns: comparative experiments

This study investigated three lab-scale hybrid wetland systems with traditional (gravel) and alternative substrates (wood mulch and zeolite) for removing organic, inorganic pollutants and coliforms from a synthetic wastewater, in order to investigate the efficiency of alternative substrates, and monitor the stability of system performance. The hybrid systems were operated under controlled variations of hydraulic load (q, 0.3-0.9 m³/m² d), influent ammoniacal nitrogen (NH₄-N, 22.0-80.0 mg/L), total nitrogen (TN, 24.0-84.0 mg/L) and biodegradable organics concentration (BOD₅, 14.5-102.0 mg/L). Overall, mulch and zeolite showed promising prospect as wetland substrates, as both media enhanced the removal of nitrogen and organics. Average NH₄-N, TN and BOD₅ removal percentages were over 99%, 72% and 97%, respectively, across all three systems, indicating stable removal performances regardless of variable operating conditions. Higher Escherichia coli removal efficiencies (99.9%) were observed across the three systems, probably due to dominancy of aerobic conditions in vertical wetland columns of the hybrid systems.     

Sewage treatment in a combined up-flow anaerobic sludge blanket (UASB)–down-flow hanging sponge (DHS) system

The performance of up-flow anaerobic sludge blanket (UASB) in combination with down-flow hanging sponge (DHS) system for sewage treatment at an average wastewater temperature of 15 °C has been investigated for 6 months. The results showed that a combined system operated at a total HRT of 10.7 h and total SRT of 88 days represents a cost effective sewage treatment process. The average COD_total and BOD_5 total concentrations measured in the final effluent of the total system (UASB + DHS) amounted to 43 and 3.0 mg/l, respectively, corresponding to the overall removal efficiency of 90% for COD_total and 98% for BOD_5 total. The total process provided a final effluent containing a low concentration of 12 mg/l for TSS.Eighty-six percent of ammonia was eliminated at space loading rate of 1.6 kg COD/m³ d and HRT of 2.7 h. The calculated nitrification rate of the DHS system according to the nitrate and nitrite production amounted to 0.18 kg/m³ d. The removal of F. coliform in the UASB reactor only amounted to 0.86 log₁₀. On the other hand, the F. coliform concentration dropped substantially, i.e. by 2.6 log₁₀ in the DHS system resulting only 2.7 × 10³/100 ml in the final effluent.The calculated average sludge production for UASB operated at an HRT of 8.0 h amounted to 30 g TSS/d, corresponding to sludge yield coefficient of 0.2 g TSS/g total COD removed, while it was indeed very low only 6.0 g TSS/d corresponding to sludge yield coefficient of 0.09 g TSS/g total COD removed, for DHS system.The DHS profile results revealed that in the first and second segment of DHS system, the COD_total, BOD_5 total and TSS was eliminated, followed by the oxidation of ammonia in the next segments.     

Partial nitrification and nutrient removal in intermittently aerated sequencing batch reactors treating separated digestate liquid after anaerobic digestion of pig manure

The performance of an intermittently aerated sequencing batch reactor (IASBR) technology was investigated in achieving partial nitrification, organic matter removal and nitrogen removal from separated digestate liquid after anaerobic digestion of pig manure. The wastewater had chemical oxygen demand (COD) concentrations of 11,540 ± 860 mg/L, 5-day biochemical oxygen demand (BOD₅) concentrations of 2,900 ± 200 mg/L and total nitrogen (TN) concentrations of 4,041 ± 59 mg/L, with low COD:N ratios (2.9) and BOD₅:COD ratios (0.25). Synthetic wastewater, simulating the separated digestate liquid with similar COD and nitrogen concentrations but BOD₅ of 11,500 ± 100 mg/L, was also treated using the IASBR technology. At a mean organic loading rate of 1.15 kg COD/(m³ d) and a nitrogen loading rate of 0.38 kg N/(m³ d), the COD removal efficiency was 89.8% in the IASBR (IASBR-1) treating digestate liquid and 99% in the IASBR (IASBR-2) treating synthetic wastewater. The IASBR-1 effluent COD was mainly due to inert organic matter and can be further reduced to less than 40 mg/L through coagulation. The partial nitrification efficiency of 71–79% was achieved in the two IASBRs and one cause for the stable long-term partial nitrification was the intermittent aeration strategy. Nitrogen removal efficiencies were 76.5 and 97% in IASBR-1 and IASBR-2, respectively. The high nitrogen removal efficiencies show that the IASBR technology is a promising technology for nitrogen removal from low COD:N ratio wastewaters. The nitrogen balance analysis shows that 59.4 and 74.3% of nitrogen removed was via heterotrophic denitrification in the non-aeration periods in IASBR-1 and IASBR-2, respectively.     

A three-stage experimental constructed wetland for treatment of domestic sewage: First 2 years of operation

Hybrid constructed wetland systems have recently been used to treat wastewaters where high demand for removal of ammonia is required. However, these systems have not been used too often for small on-site treatment systems. This is because in many countries ammonia is not limited in the discharge from small systems. Hybrid systems have a great potential to reduce both ammonia and nitrate concentrations at the same time. In our study we employed a three-stage constructed wetland system consisting of saturated vertical-flow (VF) bed (2.5 m², planted with Phragmites australis), free-drained VF bed (1.5 m², planted with P. australis) and horizontal-flow (HF) bed (6 m², planted with Phalaris arundinacea) in series. All wetlands were originally filled with crushed rock (4-8 mm). However, nitrification was achieved only after the crushed rock was replaced with sand (0-4 mm) in the free-drain wetland. Also, original size of crushed rock proved to be too vulnerable to clogging and therefore, in the first wetlands the upper 40 cm was replaced by coarser fraction of crushed rock (16-32 mm) before the second year of operation started. The system was fed with mechanically pretreated municipal wastewater and the total daily flow was divided into two batches 12 h apart. The evaluation of the results from the period 2007 to 2008 indicated that such a system has a great potential for oxidation of ammonia and reduction of nitrate. The ammonia was substantially reduced in the free-drained VF bed and nitrate was effectively reduced in the final HF bed. The inflow mean NH₄-N concentration of 29.9 mg/l was reduced to 6.5 mg/l with the average removal efficiency of 78.3%. At the same time the average nitrate-N concentration rose from 0.5 to only 2.7 mg/l at the outflow. Removal of BOD₅ and COD amounted to 94.5% and 84.4%, respectively, with respective average outflow concentrations of 10 and 50 mg/l. Phosphorus was removed efficiently despite the fact that the system was not aimed at P removal and therefore no special media were used. Phosphorus removal amounted in 2008 to 65.4%, but the average outflow concentration of 1.8 mg/l is still high. The results of the present study indicate very efficient performance of the hybrid constructed wetlands, but optimal loading parameters still need to be adjusted. The capital cost of the experimental system is comparable to the conventional on-site treatment plant but the operations and maintenance costs are about one third of the conventional plant.