Evaluation of a hybrid constructed wetland for treating domestic sewage from individual housing units surrounding agricultural villages in South Korea

Abstract

The treatment efficiency of 2- and 3-stage constructed wetlands (CWs) was evaluated for treating domestic sewage from houses surrounding agricultural villages. The optimum depth of filter media was 90 cm. The optimum volume ratio of vertical flow (VF) and horizontal flow (HF) beds was 1:2, and the optimum filter medium was broken stone in the VF-HF 2-stage hybrid CWs. Based on the above optimum conditions, removal efficiency of biochemical oxygen demand (BOD), chemical oxygen demand (COD), suspended solids (SS), total nitrogen (T-N), and total phosphorus (T-P) were 99, 98, 99, 68, and 72%, respectively. However to utilize constructed wetlands (CWs) for treating domestic sewage for an individual house, would require downsizing of the 2-stage hybrid CWs. In addition, the low removal efficiency of T-N and T-P in 2-stage hybrid CWs would require improvements necessary to meet acceptable water quality discharge standards. Thus, to reduce the CWs’ area and improve the T-N and T-P removal efficiencies, VF-HF 2-stage hybrid CW was modified into VF/HF(I)-HF(II) and VF/HF(I)-HF(III) 3-stage hybrid CW. The optimum reduced size of 3-stage hybrid CW was the VF/HF(I)-HF(II) configuration which also increased T-N removal. Using this system, removal efficiency of BOD, COD, SS, T-N, and T-P were 99, 98, 99, 83, and 75%, respectively. In VF/HF(I)-HF(II) CW, the removal velocity of BOD, COD and SS was rapid on the order of VF (1^st stage) ≫ HF(II) (3^rd stage) ≥ HF(I) (2^nd stage), VF (1^st stage) ≫ HF(II) (3^rd stage) > HF(I) (2^nd stage) and VF (1^st stage) ≫ HF(I) (2^nd stage) > HF(II) (3^rd stage), respectively. The removal velocity of T-N and T-P in VF/HF(I)-HF(II) CWs was rapid on the order of HF(I) (2^nd stage) > HF(II) (3^rd stage) ≥ VF (1^st stage) and VF (1^st stage) > HF(I) (2^nd stage) > HF(II) (3^rd stage), respectively.

---

Introduction

On a global scale, sewage is the main point-source pollutant.¹ Between 90 and 95% of the sewage produced in the world is released into the environment without any treatment.^2–4 Eutrophication has been recognized as a pollution problem in European and North American lakes and reservoirs since the mid-20th century.⁵ Impact of sewage discharge on water quality has become more widespread in recent year, causing deterioration of aquatic environments and disrupting ecosystems. Since 1990, eutrophication of lakes and reservoirs in Korea has become more of a pollution problem.⁶ ILEC/Lake Biwa Research Institute⁷ reported that 54% of lakes in Asia are eutrophic; in Europe, 53%; in North America, 48%; in South America, 41%; and in Africa, 28%.

Eutrophication of water bodies is often a result of increased nutrient from wastewater, sewage effluent, run-off from lawn fertilizers, and domestic sewage form individual housing units.⁸ Eutrophication is the result of nutrient enrichment of aquatic systems. Eutrophication can be caused by phosphorus or nitrogen containing compounds entering lakes, rivers and reservoirs. Nitrogen (N) and phosphorus (P) are most limiting nutrient governing productivity in estuaries, coastal ecosystems and most freshwater systems.⁹ Therefore, we should limit the amount of nitrogen, phosphorus and organic matter in wastewater from being discharged into aquatic ecosystems.

Constructed wetlands are considered as low-cost alternatives for treating municipal, industrial, domestic and agricultural wastewater. Over the last two decades, several studies have reported the potential use of wetlands for removal of nutrients from wastewater.^10,11 Constructed wetlands have been used for decades for the treatment of domestic or municipal sewage. The most common constructed wetland systems are designed as horizontal flow (HF) CWs system but vertical flow (VF) CWs are also becoming popular. Because such systems do not provide both aerobic and anaerobic conditions, single-stage constructed wetlands such as horizontal flow (HF) CWs or vertical flow (VF) CWs cannot achieve high total nitrogen and total phosphorus removal.¹² In particular, the VF CWs has a very high nitrification capacity, whereas the HF CWs shows high denitrification efficiency.¹³ Therefore, combinations of constructed wetlands systems for treating domestic sewage are necessary to achieve higher treatment efficiency, especially for nitrogen removal.^14,15 In addition, unlike nitrogen, phosphorus is retained by filter media mainly in the mechanism of adsorption or precipitation in constructed wetlands.^16,17 Once the substrate media is saturated, the material must be either washed or replaced. For this reason, selection of filter medium for use in constructed wetlands is an important parameter to effectively remove phosphorus from wastewater.¹⁸

In Korea, most individual housing units surrounding agricultural villages lack structured sewer systems as compared to urban areas that utilize large-scale municipal sewage treatment plants. Most individual housing units have a single or a combined septic tank system for treating domestic sewage.⁶ However, the removal efficiency of pollutants in a single or a combined septic tanks are low.¹⁹ Especially, domestic sewage from individual housing units contains high concentrations of nitrogen and phosphorus.⁶

This study evaluated the use of constructed wetland system with a higher treatment efficiency as compared to the existing individual and/or combined septic tank system. Such constructed wetland system when applied to treating sewage from individual housing units surrounding agricultural village, must be small scale and easily constructed. The systems should also to harmonize with the surrounding scenery. It is also necessary to develop constructed wetland for individual housing units that has high treatment efficiency for T-N and T-P in order to satisfy established water quality standards.²⁰ Therefore, to develop a hybrid constructed wetland for individual housing unit for treating domestic sewage, information on optimum operating conditions such as depth, volume ratio, filter medium and configuration is needed.

The main goal of this study was to evaluate a hybrid constructed wetland system for possible use in treating domestic sewage from detached housing units surrounding agricultural villages in South Korea. The specific objectives were: (1) to identify the optimum size and depth of VF and HF beds in 2-stage hybrid CWs; (2) to identify the optimum volume ratio of VF and HF beds in 2-stage hybrid CWs; (3) to identify the optimum configuration in 2 or 3-stage hybrid CWs; (4) to quantity the removal efficiency of pollutants at VF (1^st stage), HF(I) (2^nd stage) and HF(II) (3^rd stage) beds in VF/HF(I)-HF(II) 3-stage hybrid CWs.

Materials and methods

Charaterization of materials

Domestic sewage used in the study was collected from a dormitory located at Gyeongsang National University, South Korea. Domestic sewage from this dormitory had a pH of 7.3 ± 0.2, EC of 0.49 ± 0.03 dS/m, BOD, COD, SS, T-N, and T-P of 122 ± 28, 76.2 ± 23.3, 91.4 ± 19.3, 45.5 ± 9.8, and 4.19 ± 1.21 mg/L, respectively. Analyses of pH, EC, DO, BOD, COD, SS, T-N, and T-P were performed in accordance with Standard Methods.²¹

The filter media (coarse sand and broken stone) evaluated in the study was collected from an aggregate quarry in Korea. Steel slag obtained from a steel company (POSCO, Korea) was also used in this study. Steel slag is remnants from steel furnaces (steel converter). The filter media used in the study were coarse sand, broken stone, steel slag and mixed filter media (coarse sand: broken stone: steel slag = 1:1:1). The physico-chemical characteristics of the filter media used in the small-scale hybrid CWs are listed in Table 1. Chemical composition and physical properties of the materials were analyzed as follows; pH (1:5 water extraction), EC (1:5 water extraction), organic matter (walkley and black method), T-N (kjeldahl method), T-P (vanadomolybdate method). The concentrations of K, Ca, Mg, Fe and Al in the materials were analyzed by atomic absorption spectrophotometer (Shimadzu AA-680, Japan) following acid digestion with H₂SO₄–HClO₄ (n = 3) according to the method of soil analysis.²²

Table 1 Physico-chemical characteristics of filter media used in the study

Filter media	Porosity	Bulk density	d10	d60	d60/d10^a	pH	EC	O.M	T-N	T-P	K	Ca	Mg	Fe	Al
	(%)	(g/cm^3)	(mm)	(mm)		(1:5)	(dS/m)	(%)	(mg/kg)
a Uniformity coefficient.
Coarse sand	29	1.58	1.2	3.5	2.92	7.9	0.05	0.54	74.0	118	1638	2212	489	2428	45
Broken stone	38	1.47	1.3	2.7	2.07	7.8	0.04	4.13	3.9	184	1394	1031	1942	4955	23
Steel slag	45	2.64	1.5	3.0	2.0	12.1	4.78	0.18	23.2	2,324	89	286 594	46 310	170 228	5769

---

The particle-size distribution of the filter media on a weight basis was analyzed in triplicate by conventional dry-sieving techniques. The grain-size distribution plots were used to estimate d₁₀ (10% of the sand by weight is smaller than d₁₀) and d₆₀ (60% of the sand by weight is smaller than d₆₀). The uniformity of the particle-size distribution (the uniformity coefficient) was calculated as the ratio of d₆₀ to d₁₀. Porosities were determined from the amount of water, needed to saturate a known volume of filter media (n = 3), and the bulk density of the filter media (g/cm³) was based on the ratio of the dry weight to the bulk volume of the filter media (n = 3).

Small-scale hybrid constructed wetlands experiment

The small-scale hybrid constructed wetlands (located in Gyeongsang National University, South Korea at 35°16′N latitude and 127°56′E longitude) evaluated in the study consisted of either 2-stage or 3-stage beds (CWs) containing filter media (coarse sand, broken stone, steel slag and mixed filter media). The beds representing various configuration (2- or 3-stage), and consisting of vertical flow (VF; aerobic condition) and horizontal flow (HF; anaerobic condition) beds are shown in Fig. 1–3.

Fig. 1 Diagram of 2-stage VF-HF hybrid constructed wetlands used for treating domestic sewage at different depth of VF and HF beds.

Fig. 2 Diagram of 2-stage VF-HF hybrid constructed wetlands used for treating domestic sewage at different volume ratio of VF and HF beds.

Fig. 3 Diagram and configuration of 3-stage hybrid constructed wetlands used for treating domestic sewage. System A: VF-HF CWs (control), System B: VF/HF(I)-HF(II) CWs, and System C: VF/HF(I)-HF(III) CWs.

i. 2-stage hybrid constructed wetlands experiment. The pilot scale VF-HF 2-stage CWs for obtaining the optimum depth was 0.5 m (width) x 0.5 m (length) x 1.0 m (height) for the VF bed (0.25 m³ total volume) and 0.7 m (width) x 0.35 m (length) x 1.0 m (height) for the HF bed (0.25 m³ total volume) (Fig. 1). The beds were constructed from 8 mm thick plastic sheets. In the VF bed, a ventilation pipe was installed at the bottom at 0.5 m in order to maintain natural ventilation. The HF bed was also divided into twelve sections to maximize the hydraulic retention time in the bed. Domestic sewage was added to the VF bed using a vertical flow method (flow direction from top to bottom), and the water leaving the bed flowed into the HF bed by horizontal flow. The horizontal bed was always saturated with water.

The filter media used in VF-HF 2-stage CWs was coarse sand (CS), broken stone (BS), steel slag (SS) and mixed filter media (MF; coarse sand: broken stone: steel slag = 1:1:1). The VF-HF 2-stage CWs represented a combination of filter media in the following configuration: CS-CS, BS-BS, SS-SS and MF-MF. For each VF and HF bed, 0.5, 0.7 and 0.9-m layer of filter media were placed in the bottom of the beds in order to obtain the optimum depth. Phragmites australis plant was transplanted in the VF bed, and Iris pseudoacorus plants was transplanted in the HF bed. Our previous studies showed that Phragmites plant was best adapted for VF beds and Iris for the HF beds.²³ The aerobic system was pre-operated with O₂ supplied by a ventilation pipe in the VF bed. Anaerobic condition was maintained for one month by not supplying oxygen to the HF bed.

In order to obtain the optimum volume ratio of VF and HF beds (based on the results from the optimum depth in beds), the volume ratios of VF and HF beds were maintained by 1:1, 1:2 and 1:3 using four filter media (coarse sand, broken stone, steel slag and mixed filter media) in the VF-HF CWs, respectively (Fig. 2). The size of VF bed constructed from 8 mm thick plastic sheet was 0.5 m (width) x 0.5 m (length) x 1.0 m (height) (0.25 m³ total volume). The size of HF beds were 0.7 m (width) x 0.35 m (length) x 1.0 m (height) (0.25 m³ total volume), 0.7 m (width) x 0.7 m (length) 1.0 m (height) (0.5 m³ total volume) and 0.7 m (width) x 1.07 m (length) x 1.0 m (height) (0.75 m³ total volume), respectively. The 2-stage VF-HF CWs treatment were established using the same condition as the 2-stage VF-HF CWs including size, filter media and wastewater application method.

The efficiency of nutrient removal from the wastewater by the 2-stage VF-HF CWs was measured. The removal efficiencies for BOD, COD, SS, T-N, and T-P were calculated using the following equation:

The applied hydraulic load was added at the rate of 0.6 m³/m²/day. The hydraulic load was calculated for first stage bed (VF bed) in the 2-stage VF-HF CWs. Samples of the influent and effluent were taken every fifth day and analyzed for a period of 12 months. The analyses of BOD (5-day BOD test), COD (closed reflux, titrimetric method), SS (suspended solid dried at 103–105 °C), T-N (organic nitrogen + NH₄–N + NO_x–N), and T-P (ascorbic acid method) were performed in accordance with Standard Methods.²¹

ii. 3-stage hybrid constructed wetlands experiment. In order to improve T-N and T-P removal and reduce the CWs area based on the results from the optimum conditions in VF-HF 2-stage hybrid CWs (system A), 3-stage units was constructed using a combination of vertical flow (VF) and horizontal flow [HF(I), HF(II), and HF(III)] beds (Fig. 3).

The combinations were system B [VF/HF(I)-HF(II)] and C [VF/HF(I)-HF(III)]. The size of the VF bed was 0.7 m (width) x 0.35 m (length) x 1.0 m (height) (0.25 m³ total volume), and the HF beds such as HF(I), HF(II) and HF(III) were also 0.7 m (width) x 0.35 m (length) x 1.0 m (height) (0.25 m³ total volume). The filter media used in these beds were coarse sand and broken stone. The system B and C were established using the same condition as the 2-stage VF-HF CWs including flow wastewater application method.

The applied hydraulic load was at the rate of 0.6 m³/m²/day. The hydraulic load was calculated for first stage bed (VF bed) in system A, B and C. Samples of the influent and effluent were taken every fifth day and analyzed over a period of 6 months. The analyses of BOD (5-day BOD test), COD (closed reflux, titrimetric method), SS (suspended solid dried at 103–105 °C), T-N (organic nitrogen + NH₄ߝN + NO_xߝN), and T-P (ascorbic acid method) were performed in accordance with Standard Methods.²¹

Removal velocity of pollutants in VF/HF(I)-HF(II) 3-stage constructed wetlands

In order to quantify removal of BOD, COD, SS, T-N and T-P, effluent water samples were collected at the following hydraulic retention times 0 (inlet), 0.14, 0.28, 0.56, 0.84, 1.12, and 1.40 days and analyzed for a period of 1 month after pre-operation. The removal efficiency of COD, T-N and T-P were determined using the linear equation ln(C/C_o) = −kt based on the COD, T-N and T-P concentration in the influent and effluent water, respectively. The retention time or efficiency of removal was controlled by amount of wastewater pumped through the system.

Removal efficiency of pollutants was estimated using the linear velocity equation. Linear velocity is expressed by V =kC. The reaction was related to pollutant concentrations, which is the momentary reaction velocity or time which C changes. Thus the dc can be expressed by −dc/dt. Therefore, the linear velocity equation (2) can be expressed as a −dt/dc at V when V =kC.^17,18 Linear velocity equation (3) is an integrated linear velocity equation (2).

where C is the concentration of pollutants in water at effluent (mg/L), C_o is the concentration of pollutants in raw water (mg/L), k is the pollutant's removal velocity constant, and t is hydraulic retention time (days).

Statistical analysis

Statistical analysis of data was conducted using SAS software (version 8.02, 1999–2001, SAS Institute, Cary, NC). Simple linear regression analysis using PROC REG (SAS 9.1, SAS Institute Inc. Cary, NC, USA) was conducted to determine if the slope was significantly different from a theoretical model (P = 0.05).

Results and discussion

1. 2-stage hybrid constructed wetlands experiment

Removal efficiency of pollutants at different depth of VF and HF beds in VF-HF 2-stage hybrid constructed wetlands. Many VF-HF systems are based on the original hybrid systems developed by Seidel at the Max Planck Institute in Krefeld, Germany.¹⁵ In the early and mid-1980s, several hybrid systems of the Seidel's type were constructed in France and UK. In the 1990s and early 2000s, VF-HF systems were constructed in additional European countries (Norway, Austria and Ireland). The system is now receiving more attention around the world.^24–26

The removal efficiency for BOD, COD, SS, T-N and T-P from wastewater was evaluated in order to identify the optimum depth of each bed. Fig. 4 presents the results of the removal efficiency of the BOD, COD, SS, T-N, and T-P in VF-HF 2-stage hybrid CWs. The removal efficiencies for BOD, COD, SS, T-N and T-P in the first and second stages on filter media at different depth of VF bed were in the order of 90 cm ≥ 70 cm > 50 cm in VF-HF 2-stage hybrid CWs.

Fig. 4 Removal efficiency of BOD, COD, SS, T-N and T-P in the water at different depth of VF and HF beds in VF-HF 2-stage hybrid constructed wetlands (error bar not displayed). CS: coarse sand; BS: Broken stone, SS: Steel slag; MF: Mixed filter media.

The removal efficiency for BOD and COD in the first and second stages on filter media increased as the depth increased to 70 cm. For a VF bed depth greater than 70 cm, the removal efficiency for BOD and COD in the first and second stages on filter media increased slightly with depth. It was shown that the removal efficiency of both stages (VF and HF) exhibited a similar trend.

The removal of BOD and COD in the VF bed was much higher than that in the HF bed. The BOD and COD consumption rate by microbes in the VF bed was higher in comparison with that in the HF bed likely because of the activity of the aerobic bacteria, which provided greater oxidation of organic matter than anaerobic bacteria.¹⁷ The removal efficiencies of BOD and COD in coarse sand and broken stone were higher than that measured using the other filter media. In contrast, removal by steel slag was lower than that for the other filter media. This would suggest that the steel slag in VF-HF systems do not provide conditions suitable for the microbes and plant growth because the measured pH of leachate from steel slag can exceed pH 12.

The removal efficiency for SS on filter media in the first stage increased rapidly as the depth increased to 50 cm, whereas, depth above 50 cm, showed similar removal efficiency. The removal efficiency of SS in the first stage was much higher than that in the second stage because most of the SS had filtered or settled out within the first few centimetres past the inlet. In general, suspended solids in constructed wetlands are effectively removed by filtration and settlement.^27,28

At the 90 cm depth, the removal efficiencies of T-N by the first stage of CS-CS, BS-BS, SS-SS, and MF-MF configurations were 23.2 ± 5.2, 26.1 ± 4.9, 36.8 ± 5.4 and 35.7 ± 4.6%, respectively. The removal efficiency of T-N in the VF bed was lower than that for BOD, COD and SS. This suggest that the VF bed (first stage) in VF-HF systems provided conditions suitable for nitrification but not for denitrification. In addition, the removal efficiency of T-N using steel slag and mixed filter media was slightly higher than that measured using the other filter media. This was likely because steel slag had a high pH level (pH > 11) which provided suitable condition for ammonia volatilization. Ammonia (NH₃) volatilization alone is an undesirable mechanism for the removal of nitrogen (N) from constructed wetlands because the steel slag in VF-HF CWs due to high pH (> pH 12) did not provide suitable conditions for microbial and plant growth which can also remove nitrogen. In the second stage, the removal efficiencies of T-N by CS-CS, BS-BS, SS-SS, and MF-MF configurations with 90 cm depth bed in VF-HF 2-stage hybrid CWs were 64.3 ± 4.7, 65.1 ± 4.2, 54.0 ± 5.2 and 56.7 ± 4.3%, respectively. The measured removal efficiency of T-N in the HF bed was higher than that in the VF bed. This would suggest that the HF bed (second stage) in VF-HF systems provided conditions suitable for denitrification. In VF-HF CWs, the first stage provided suitable conditions (aerobic) for nitrification while the second stage provided suitable conditions (anoxic/anaerobic) for denitrification to occur.^14,15,17

In the first stage, the removal efficiencies of T-P by CS-CS, BS-BS, SS-SS, and MF-MF configurations in 90 cm depth bed were 35.0 ± 2.8, 33.5 ± 2.3, 60.3 ± 0.8 and 57.2 ± 1.9%. In the second stage, the removal efficiencies of T-P by CS-CS, BS-BS, SS-SS, and MF-MF configurations in 90 cm depth bed were 65.5 ± 2.2, 66.8 ± 1.7, 85.0 ± 0.6 and 82.1 ± 1.2%. The removal efficiency of T-P in steel slag and mixed filter media was higher than that in the other filter media because steel slag has a high P adsorption capacity. This is attributed to Ca–P and Fe–P formation since steel slag contains 40.1% CaO and 21.9% FeO, and most extractable P would be bound to Ca and Fe in the steel slag. Calcium bound-P is the predominant form of P in alkaline soils while Fe–P and Al–P are the predominant forms in acidic soils.²⁹ Results suggest that steel slag could perhaps be a suitable filter media for enhancing phosphorus removal in some constructed wetlands.

Steel slag used as the filter medium for the establishment of constructed wetland (improving the accumulating capacity of phosphorus) would also extend the longevity of the constructed wetland. However, the efficiency for BOD, COD and T-N removal by steel slag would be low.

Based on the above results, the optimum depth of VF and HF beds for the effective removal of BOD, COD, SS, T-N and T-P in VF-HF CWs was 90 cm or greater. Vymazal¹⁴ reported that VF-HF systems at Colecott exhibit a high removal of organics (COD and BOD₅) and suspended solids. As compared to single HF systems, there was a much higher removal of total nitrogen as a result of high nitrification in the VF section. Nitrate produced in the VF section is successfully removed in the HF section. However, removal of phosphorus would be low.

Removal efficiency of pollutants at different volume ratio of VF and HF beds in VF-HF 2-stage hybrid constructed wetlands. Based on removal efficiency obtained in the 90 cm depth of VF and HF beds, the same depth was used VF-HF 2-stage hybrid CWs. The removal efficiency for BOD, COD, SS, T-N and T-P from wastewater was evaluated in order to identify the optimum volume ratio of VF and HF beds. Fig. 5 presents removal efficiencies of BOD, COD, SS, T-N, and T-P in the VF-HF 2-stage hybrid CWs.

Fig. 5 Removal efficiency of pollutants from the effluent at different volume ratio of VF and HF beds in VF-HF 2-stage hybrid constructed wetlands (error bars not displayed).

For coarse sand, the removal efficiencies of BOD, COD and SS from the effluent (using different volume ratio of VF and HF beds) were in the order of 1:3 ≈ 1:2 ≥ 1:1. The removal efficiencies of BOD, COD and SS were high 98.8–99.4, 95.7–97.2 and 99.4–99.9%, respectively, regardless of the ratio of aerobic and anaerobic beds. The removal efficiencies of T-N and T-P from the effluent for coarse sand at different volume ratio of VF and HF beds were in the order of 1:3 > 1:2 > 1:1. The removal efficiency of T-N and T-P from the effluent for coarse sand gradually increased as the volume ratio of VF and HF beds increased.

For broken stone, the removal efficiencies of BOD, COD, SS, T-N and T-P from the effluent using different volume ratios in the VF and HF beds were in the order of 1:3 ≈ 1:2 ≥ 1:1. The removal efficiency of BOD, COD, SS, T-N and T-P for broken stone from the effluent was slightly higher than that measured using coarse sand. At a volume ratio of 1 (VF bed): 2 (HF bed), the removal efficiencies of BOD, COD, SS, T-N and T-P for broken stone were 99.3, 98.2, 99.9, 68.2 and 71.8%, respectively.

Using steel slag, the removal efficiencies of BOD, COD, SS, T-N and T-P for all the volume ratio conditions were 89.2–89.3, 88.0–89.6, 87.8–90.1, 54.0–57.9 and 85.0–96.8%, respectively. The removal efficiency of pollutants by steel slag gradually increased as the volume ratio of VF and HF beds increased. For mixed filter media, the removal efficiency of BOD, COD, SS, T-N and T-P exhibited a similar trend as measured using steel slag.

The removal efficiency for BOD, COD, SS and T-N by steel slag and mixed filter media at all volume ratio conditions were lower than that for coarse sand and broken stone. In contrast, the removal efficiency for T-P was much higher. Results of the experiment showed that steel slag would be a suitable filter media to effectively treat phosphorus in constructed wetlands. However, cautious should be taken in use as a filter medium in constructed wetland because the VF or HF beds with steel slag in constructed wetlands do not provide conditions suitable for the microbes and plant growth. In general, steel slag is alkaline with a solution pH in the range of 8 to 10. However, the pH of leachate from steel slag can exceed a pH value of 12.³⁰ If steel slag was used for removal of phosphorus from wastewater in constructed wetlands, the pH of VF or HF beds with steel slag must be adjusted to a neutral pH (between pH 6–8).

Based on the above results, the optimum volume ratio of VF and HF beds for the effective removal of BOD, COD, SS, T-N and T-P in VF-HF 2-stage hybrid CWs was shown to be 1:2. Because constructed wetlands generally require a large area for construction, it is reasonable that the ratio of VF and HF beds is to be configured as 1:2. In addition, the optimum filter medium for use in the VF and HF beds should be broken stone in VF-HF 2 stage hybrid CWs. However, based on optimum depth and volume ratio of VF and HF beds, the efficiency for T-N and T-P removal by broken stone would be low. Therefore, the low removal efficiency of T-N and T-P by broken stone in 2-stage hybrid CWs must be improved in order to meet water quality standards. Also, in order to utilize the constructed wetlands system for treating wastewater from individual housing units, 2-stage hybrid CWs receiving domestic sewage must be downsized to reduce the total area required. Thus, in order to reduce the CWs area and improve the T-N and T-P removal efficiencies in VF-HF 2-stage CWs, VF-HF 2-stage hybrid CWs was modified for evaluation representing a smaller area VF/HF(I)-HF(II) and VF/HF(I)-HF(III) 3-stage hybrid CWs (Fig. 3).

2. 3-stage hybrid constructed wetlands experiment

Removal efficiency of pollutants at different configuration in 3-stage hybrid constructed wetlands. The optimum conditions obtained from the previously study (filter media depth of 90 cm, sewage loading of 0.6 m³/m²/day, volume ratio of 1:2 (VF: HF), and broken stone, which had the highest removal efficiency of pollutants) was used in system A (VF-HF 2-stage CWs; control), B (VF/HF(I)-HF(II) 3-stage hybrid CWs) and C (VF/HF(I)-HF(III) 3-stage hybrid CWs). The removal efficiency for BOD, COD, SS, T-N and T-P from wastewater was quantified in order to identify the optimum 3-stage hybrid CWs. Fig. 6 presents the results of the removal efficiency of the BOD, COD, SS, T-N, and T-P in system B and C.

Fig. 6 Removal efficiency of pollutants from the effluent for various configuration of 3-stage hybrid constructed wetlands. Data represent mean and standard deviation (error bars).

In system B, the removal efficiencies for BOD, COD and SS from the effluent were 99.3, 98.3 and 99.9%, respectively. The removal efficiency for BOD, COD and SS from the effluent in system B in comparison with system A (control) showed similar values regardless of the treatment methods. In system B in comparison with system A (control), the T-N and T-P treatment efficiency presented a especially high removal efficiency (83.2 and 75.2%, respectively).

In system B, it was evident that this configuration increased nitrogen and phosphorus treatment efficiency. The nitrogen treatment efficiency in system B rapidly increased by 15% in comparison with system A (control) because the HF(I) (second stage) bed provided suitable conditions (anoxic/anaerobic) for denitrification.

Denitrification is illustrated by the following equation:³¹

6(CH₂O) + 4NO₃⁻ → 6CO₂ + 2N₂ + 6H₂O

This reaction is irreversible, and occurs in the presence of available organic substrate only under anaerobic or anoxic conditions (Eh = + 350 to + 100 mV), where nitrogen is used as an electron acceptor in place of oxygen. Greater evidence is being provided from pure culture studies that nitrate reduction can occur in the presence of oxygen. Hence, in saturation soils nitrate reduction may also start before the oxygen is depleted.^32,33

In addition, in system C, the removal efficiencies for BOD, COD and SS from the effluent were 96.3, 95.9 and 99.5%, respectively. The removal efficiency for BOD and COD from the effluent in system C was slightly lower than that in system A (control). In system C in comparison with system A (control), the T-N treatment efficiency presented a significantly high removal efficiency (about 85.2%), whereas, the removal efficiency of T-P from effluent was lower than that in the control. The nitrogen treatment efficiency in system C rapidly increased by 17% in comparison with system A (control) because the HF(I) (second stage) and HF(III) (third stage) beds provided suitable conditions (anoxic/anaerobic) for denitrification to occur. Result clearly demonstrate that system C configurations (VF/HF(I)-HF(III) 3-stage CWs) increased nitrogen treatment efficiency, whereas BOD, COD, T-P treatment efficiency decreased in comparison with system A (VF-HF 2-stage CWs).

Based on the above results, in order to reduce the CWs area and increase the T-N removal obtained by VF-HF 2-stage CWs, the optimum 3-stage hybrid CWs would have the VF/HF(I)-HF(II) 3-stage CW configuration. As compared to VF-HF 2-stage systems, there is much higher removal of total nitrogen as a result of high denitrification in the HF(I) and HF(II) section. Especially, nitrogen which in the VF/HF(I)-HF(II) is successfully removed in the HF(I) section.

Removal velocity of pollutants in VF/HF(I)-HF(II) 3-stage hybrid constructed wetlands. The removal velocity of the various pollutants in system B (VF/HF(I)-HF(II) CWs), which had the highest removal efficiency of pollutants in 3-stage hybrid CWs, are shown in Fig. 7 and Table 2. The HRT in each bed was compared with the ln(C_o/C). The removal velocities of BOD, COD, SS, T-N and T-P were expressed as linear regression.

Table 2 Removal velocity constant and linear equation of BOD, COD, SS, T-N and T-P in VF/HF(I)-HF(II) 3 stage hybrid CWs, respectively

	VF (1^st stage) bed		HF(I) (2^nd stage) bed		HF(II) (3^rd stage) bed
	k^a	Equation	k^a	Equation	k^a	Equation
a k: pollutant's removal velocity constant. b (* and ** denote significance at 5.0 and 1.0% levels, respectively).
BOD	8.80	y = −8.80x + 0.11 (R^2 = 0.975^b)	2.09	y = −2.09x − 1.86 (R^2 = 0.996^b)	2.36	y = −2.36x − 1.54 (R^2 = 0.924^b)
COD	7.15	y = −7.15x + 0.06 (R^2 = 0.987^b)	1.63	y = −1.63x − 1.49 (R^2 = 0.963^b)	2.07	y = −2.07x − 1.14 (R^2 = 0.986^b)
SS	11.35	y = −11.35x + 0.16 (R^2 = 0.971^b)	4.53	y = −4.53x − 1.70 (R^2 = 0.923^b)	2.12	y = −2.12x − 3.87 (R^2 = 0.967^b)
T-N	1.10	y = −1.10x + 0.01 (R^2 = 0.990^b)	1.53	y = −1.53x + 0.10 (R^2 = 0.994^b)	1.11	y = −1.11x − 0.25 (R^2 = 0.994^b)
T-P	1.47	y = −1.47x + 0.01 (R^2 = 0.995^b)	0.88	y = −0.88x − 0.18 (R^2 = 0.993^b)	0.88	y = −0.88x + 0.17 (R^2 = 0.997^b)

---

Fig. 7 Removal velocity of BOD, COD, SS, T-N and T-P in VF/HF(I)-HF(II) 3-stage hybrid constructed wetlands, respectively. Data represent mean and standard deviation (error bars).

In the VF (first stage) bed of system B, the removal velocity constants (k) of BOD, COD, SS, T-N and T-P were 8.8, 7.2, 11.4, 1.1 and 1.5/day, respectively. The removal velocity of pollutants in the HF (first stage) bed of system B configuration was rapid on the order of SS > COD > BOD ≫ T-P ≥ T-N. Especially, the removal velocity constant (k) of BOD, COD and SS was greater than that for T-N and T-P. This would suggest that the VF bed (first stage) in system B provided suitable conditions (aerobic) for BOD, COD and SS removal. In addition, the removal velocity of T-N in the VF bed of system B was slightly slower than that for BOD, SS, SS and T-P. This suggests that the first stage (VF bed) provided suitable conditions (aerobic) for nitrification but not for denitrification to occur. Nitrification coupled with denitrification is the major removal process for nitrogen in many constructed wetlands.^14,15,17

In the HF(I) (second stage) bed of system B configuration, the removal velocity constants (k) of BOD, COD, SS, T-N and T-P were 2.1, 1.6, 4.5, 1.5 and 0.9/day, respectively. The removal velocity of pollutants in the HF(I) (second stage) bed of system B was rapid on the order of SS > BOD ≥ COD ≈ T-N > T-N. The removal velocity of T-N in the HF(I) bed of system B was rapid as compared to the VF bed of system B. Results suggested that the second stage (HF(I) bed) provides suitable conditions (anaerobic/anoxic) for denitrification.

In the HF(II) (third stage) bed of system B configuration, the removal velocity constants (k) of BOD, COD, SS, T-N and T-P were 2.4, 2.1, 2.1, 1.1 and 0.9/day, respectively. The removal velocity of pollutants in the HF(II) (third stage) bed of system B configuration was rapid in the order of BOD ≥ COD ≒ SS > T-N ≥ T-P.

In summary, the removal velocity of BOD, COD and SS in system B was rapid, being in the order of VF (1^st stage) ≫ HF(II) (3^rd stage) ≥ HF(I) (2^nd stage), VF (1^st stage) ≫ HF(II) (3^rd stage) > HF(I) (2^nd stage) and VF (1^st stage) ≫ HF(I) (2^nd stage) > HF(II) (3^rd stage), respectively. Results suggest that the optimum bed for BOD and COD removal from domestic sewage in system B was the VF bed. Seo et al.¹⁷ reported that the BOD and COD consumption rate by microbes in the VF bed was high in comparison with that in the HF bed likely because of the activity of the aerobic bacteria, which provided greater oxidation of organic matter than anaerobic bacteria. In addition, this would also suggest that removal velocity of SS from domestic sewage in system B decreased as treatment stage increased.

In the VF bed of system B, the removal velocity of pollutants from domestic sewage was greater than that for the HF(I) and HF(II) beds because the hydraulic retention time (HRT) of domestic sewage in the VF bed was short in comparison with that in the HF bed. Vymazal¹⁴ reported that most of the SS are filtered out and settled within the first few meters beyond the inlet zone regardless of the flow types. Also, the removal velocity of T-N and T-P in system B was rapid, being in the order of HF(I) (2^nd stage) > HF(II) (3^rd stage) ≥ VF (1^st stage) and VF (1^st stage) > HF(I) (2^nd stage) > HF(II) (3^rd stage), respectively. This result agrees with the observations by Seo et al.¹⁷ for the removal velocity of COD and T-P between HF and VF beds in HF-VF-HF CWs for treating greenhouse wastewater. However, for the removal velocity of T-N, a different result was reported by Seo et al.¹⁷ using greenhouse wastewater. They found that the removal velocity of T-N in HF-VF-HF CWs was rapid in the order of VF > HF. This would suggest that domestic sewage from individual housing units contained high concentration of ammonia in this study, whereas greenhouse wastewater from greenhouses contained high concentration of nitrate.

In the HF beds of system B, the removal velocity of T-N from domestic sewage was greater than that for the VF bed because the HF beds provided suitable conditions (anaerobic/anoxic) for denitrification. Results demonstrate that the removal velocity of pollutants is an important parameter for determining the optimum depth and hydraulic retention time (HRT) in system B.

Conclusions

In a hybrid constructed wetland (CW) designed for treating domestic sewage from individual housing unit surrounding agricultural villages, the optimum depth of filter media was 90 cm. The optimum volume ratio of vertical flow (VF) and Horizontal flow (HF) beds was 1:2. The optimum filter medium was broken stone in the VF-HF 2-stage hybrid CWs. However to utilize constructed wetlands for treating domestic sewage for individual housing units would require 2-stage hybrid CWs to be downsized in order to utilize a smaller area. In addition, the low removal efficiency of T-N and T-P in VF-HF 2-stage hybrid CWs required improvements in other respects to meet acceptable water quality standards. For reducing the CWs area and increasing the T-N and T-P removals in constructed wetlands, the 3-stage hybrid CWs is recommended. In the 3-stage hybrid CWs, Phragmites australis plant was transplanted in the VF bed, and Iris pseudoacorus plants was transplanted in the HF(II) bed. Utilizing the VF/HF(I)-HF(II) configuration of 3-stage hybrid CWs also increased T-N removal. In the VF/HF(I)-HF(II) 3-stage hybrid CWs, removal efficiencies of BOD, COD, SS, T-N, and T-P were 99, 98, 99, 83, and 75%, respectively. Using this system, removal velocity of BOD, COD and SS was rapid in the order of VF (1^st stage) ≫ HF(II) (3^rd stage) ≥ HF(I) (2^nd stage), VF (1^st stage) ≫ HF(II) (3^rd stage) > HF(I) (2^nd stage) and VF (1^st stage) ≫ HF(I) (2^nd stage) > HF(II) (3^rd stage), respectively. Also, the removal velocity of T-N and T-P in VF/HF(I)-HF(II) CWs was rapid in the order of HF(I) (2^nd stage) > HF(II) (3^rd stage) ≥ VF (1^st stage) and VF (1^st stage) > HF(I) (2^nd stage) > HF(II) (3^rd stage), respectively. Removal velocity was shown to be an important parameter in determining the optimum depth, volume ration, configuration and the hydraulic retention time (HRT) for the VF/HF(I)-HF(II) 3-stage hybrid CWs.

In summary, to effectively treat domestic sewage from individual housing unit surrounding agricultural villages, the optimum hybrid CWs would be the VF/HF(I)-HF(II) 3-stage constructed wetlands configuration.

Acknowledgements

This study was carried out with the support of “Specific Joint Agricultural Research-promoting Projects (Project No. 20060101033215)”, RDA, Republic of Korea. This work was also supported by the Brain Korea 21 Program in Ministry of Education in Korea and by Technology Development Program for Agriculture and Forestry, Ministry of Agriculture and Forestry, Korea.

References

1. H. J. Gijzen, Water Sci. Technol., 2002, 45(10), 321–328.
2. C. Bartone, J. Bernstein, J. Leitmann and J. Eigen, Towards environmental strategies for cities, Urban Environmental Management Paper, 18, World Bank, Washington, USA, 1994.
3. J. Niemczynowics, Wat. Qual. Int., 1997, 2, 9–11.
4. J. A. Elliot, An introduction to sustainable development, Routledge, London (UK) and New York (USA), 2nd edn, 1999.
5. W. Rodhe, Crystallization of eutrophication concepts in North Europe.in Eutrophication, Causes, Consequences, Correctives, National Academy of Sciences, Washington D.C., Standard Book Number 309-01700-9, 1969, pp. 50–64.
6. Ministry of Enviroment, White paper on the environment, Ministry of Environment Republic, Seoul (in Korea), 2005.
7. ILEC/Lake Biwa Research Institute(ed.), Survey of the State of the World's Lakes, International Lake Environment Committee, Otsu and United Nations Environment Programme, Nairobi, 1988–1993, Vol. I–IV.
8. J. Bartram, W. W. Carmichael, I. Chorus, G. Jones and O. M. Skulberg, in Toxic Cyanobacteria in Water: A guide to their public health consequences, monitoring and management, World Health Organization, 1999, Ch. 1.
9. S. R. Carpenter, N. F. Caraco and V. H. Smith, Ecol. Appl., 1998, 8, 559–568.
10. K. R. Reddy and W. H. Smith, Aquatic plants for water treatment and resource recovery, Magnolia Publishing Inc, Orlando FL, 1987, pp. 1032.
11. W. J. Mitsch and J. K. Cronk, Adv. Soil Sci., 1992, 17, 217–255.
12. J. Vymazal, Ecol. Eng., 2002, 18, 633–646.
13. C. Platzer, Water Sci. Technol., 1999, 40(3), 257–263.
14. J. Vymazal, Ecol. Eng., 2005, 25, 478–490.
15. J. Vymazal, Sci. Total Environ., 2007, 380, 48–65.
16. R. H. Kadlec and R. L. Knight, Treatment wetlands, Lewis Publishers, 1996, pp. 893.
17. D. C. Seo, S. H. Hwang, H. J. Kim, J. S. Cho, H. J. Lee, R. D. DeLaune, A. Jugsujinda, S. T. Lee, J. Y. Seo and J. S. Heo, Ecol. Eng., 2008, 32, 121–132.
18. D. C. Seo, J. S. Cho, H. J. Lee and J. S. Heo, Wat. Res., 2005, 39, 2445–2457.
19. E. H. Kim, H. G. Park and N. C. Sung, Kor. J. Environ. Agric, 1998, 17(4), 329–334.
20. Ministry of Enviroment, White paper on the environment, Ministry of Environment Republic, Seoul (in Korea), 2003.
21. APHA-AWWA-WEF, Standard methods for the examination of water and wastewater, American Public Health Association, Washington, DC, 21st edn, 2005.
22. RDA(Rural Development Administration, Korea), Methods of Soil Chemical Analysis.National Institute of Agricultural Science and Technology, RDA, Suwon (in Korea), 1988.
23. D. C. Seo, B. I. Jang, I. S. Jo, S. C. Lim, H. J. Lee, J. S. Cho, H. C. Kim and J. S. Heo, Kor. J. Environ. Agric, 2006, 25(1), 25–33.
24. T. Mæhlum and P. Stålnacke, Water Sci. Technol., 1999, 40(3), 273–281.
25. G. M. Mitterer-Reichmann, Data evaluation of constructed wetlands for treatment of domestic wastewater, in proceedings of the IWA 8th International conference on Wetland Systems for Water Pollution Control, Arusha, Tanzania, 2002, pp. 40–46.
26. S. O'Hogain, Water Sci. Technol., 2003, 48(5), 119–126.
27. P. Cooper and J. de Maeseneer, Hybrid systems-what is the best way to arrange the vertical and horizontal flow stages? IAWQ Specialist group on the use of macrophytes in water pollution control, 1996, 15, 8–13.
28. J. Vymazal, H. Brix, P. F. Cooper, R. Haberl, R. Perfler and J. Laber, Removal mechanisms and types of constructed wetlands.in Constructed Wetlands for Wasterwater Treatment in Europe, J. Vymazal, H. Brix, P. F. Cooper, M. B. Green and R. Haberl (eds.), Backhuys Publishers, Leiden, The Netherlands, 1998, pp. 17–66.
29. L. A. Dean, Adv. Agron., 1949, 1, 391–411.
30. W. H. Kang, I. S. Hwang and J. Y. Park, Chemosphere, 2006, 62, 285–293.
31. R. D. Hauck, Atmospheric nitrogen chemistry, nitrification, denitrification, and their relationships.in The handbook of environmental chemistry, O. Hutzinger, Editor, Springer-Verlag, Berlin, 1984, vol. 1, part C, pp. 105–127.
32. J. G. Kuenen and L. A. Robertson, Ecology of nitrification and denitrification.in The nitrogen and sulfur cycles, J. A. Cole and S. J. Ferguson, Editors, Cambridge University Press, Cambridge, UK, 1987, pp. 162–218.
33. H. J. Laanbroek, Aquat. Bot., 1990, 38, 109–125.

---