621

Environmental Technology, Vol. 28. pp 621-628 Selper Ltd., 2007

PERFORMANCE AND COST COMPARISON OF A FWSAND A VSF CONSTRUCTED WETLAND SYSTEM

V. A. TSIHRINTZIS1*, C. S. AKRATOS1, G. D. GIKAS1, D. KARAMOUZIS2 AND A. N. ANGELAKIS3

1Laboratory of Ecological Engineering and Technology, Department of Environmental Engineering,Democritus University of Thrace, 67100 Xanthi, Greece

2Hydraulics, Soil Science and Agricultural Engineering Division, Department of Agriculture,Aristotle University of Thessaloniki, 54124 Thessaloniki, Greece

3Institute of Iraklio, National Foundation for Agricultural Research, P.O. Box 2229,71307 Iraklio, Greece

(Received 1 February 2006; Accepted 10 January 2007)

ABSTRACT

Two constructed wetland systems, treating domestic wastewater, are compared in terms of performance and costs. One is afree water surface (FWS) wetland system located in Pompia, Crete, south Greece, and the other one is a vertical subsurfaceflow (VSF) wetland system located in Gomati, Chalkidiki, north Greece. The FWS system is designed for 1200 p.e. Itsconstruction cost was 305,000, and the capital, operation and maintenance cost was 22.07 p.e.-1 yr-1 or 0.50 m-3 ofinfluent. The VSF system is designed for 1000 p.e. Its construction cost was 410,850, and the capital, operation andmaintenance cost was 36.81 p.e.-1 yr-1 or 0.56 m-3 of influent. Both systems achieved high removal rates for BOD5, COD,TSS, TKN, phosphorus, TC, and FC, which makes them ideal for small communities in the Mediterranean region.

Keywords: Free-water surface constructed wetland, vertical subsurface flow constructed wetland, treatment performance,

construction cost; operation cost

INTRODUCTION

Constructed wetland (CW) wastewater treatment

systems are considered more reliable compared to

conventional systems [1], and are ideal technologies for small

communities, due to their low construction, operation and

maintenance costs, easy adaptation to the environment and

limited generation of by-products [2,3].

One question however, is which is the optimum CW

type (i.e., free-water surface (FWS), horizontal subsurface flow

(HSF) or vertical subsurface flow (VSF) system) to use in a

specific region, in terms of performance, costs, area

requirements, and other factors. Most studies in the literature

emphasize specific systems in terms of general performance

[4-9]. Other studies examine the effect of various design

parameters [10-13]. Comparisons of various CW types in the

same region are limited (e.g., [14,15]). Construction and other

cost data for CW systems are also limited (e.g., [16]). The

necessity of pretreatment is an issue for discussion, since

modified VSF designs in France operate successfully without

pretreatment [17, 18]. Finally, small-scale on-site CW systems

are now installed for single family use (e.g., [19]).

The aim of this paper is to provide a perspective for

applying constructed wetland technology in the

Mediterranean regions and specifically in Greece.

Descriptions, design considerations, construction cost,

constituent removal performance, and operation and

maintenance (O&M) costs of two constructed wetland

systems (a FWS and a VSF) are presented. Both systems treat

domestic wastewater and were designed for comparable

treatment capacities.

METHODS AND MATERIALS

System Description

Two constructed wetland systems treating domestic

wastewater are compared in terms of costs and performance.

One is a FWS wetland system located in Pompia, Crete, South

Greece, and the other is a VSF wetland system located in

Gomati, Chalkidiki, Macedonia, North Greece.

622

FWS system

The main components of the FWS system are [20]: (a) a

septic tank (up-flow reactor simulation) equipped with three

screen vault filters [3]; (b) a FWS constructed wetland

consisting of two basins in series, with surface areas of 4300

m2 and 1200 m2 (the inflow is uniformly distributed across at

the inlet of each basin using manifolds); (c) two chambers, one

in each basin, for regulating the water level; (d) small pumps

and a pipeline for the recirculation of the effluent back to the

inlet of the first basin; (e) a compost filter for odor control in

the septic tank.

Vegetation selection included two species of reeds

(Phragmites australis and Arundo donax). The facility was

constructed in the early months of 1999. The vegetation was

planted late in the winter of the same year, and due to the

favorable climatic conditions prevailing in the area, it

established well very rapidly. By the end of the year, the

vegetation was very dense and more than two meters in

height.

The basic parameters used in the design of the FWS

facility are [20]: population served 1200 p.e.; mean daily flow

rate 144 m3 d-1; maximum daily flow rate 216 m3 d-1;

maximum hourly flow rate 27.7 m3 h-1; influent biochemical

oxygen demand (BOD5) 400 mg l-1; septic tank effluent BOD5250 mg l-1; wetland effluent BOD5 10 mg l-1 and (COD)

623

with Phragmites australis. The HSF cell was not planted at the

time of the study and was out of the wastewater stream. The

VSF system was planted in May 2003 and was immediately

put to operation. The plants were dense and nearly fully-

grown by October 2003.

The route of the wastewater through the system is the

following: from the inflow structure to the rotating disk

screen and to the settling tanks. The sludge is collected at the

bottom of the settling tank (estimated volume 4.8 m3 week-1),

it is then pumped to the sludge digestion-stabilization tanks,

and then to the VSF sludge basins. The leachate collected at

the bottom of the VSF sludge basins is pumped back to the

siphon, and together with the wastewater from the settling

basins feed the stage I VSF basin (2 cells operate at a time).

Then, the wastewater is led to the stage II VSF basins (2 cells

operate at a time). In the future, the flow will continue to the

stage III HSF basin (not in operation now). The effluent of the

basin discharges into a nearby stream, approximately 5 km

from the coast.

The main design parameters for this system are

summarized as follows: the design population is 1000 p.e. The

design mean daily flow of the system is 180 m3 d-1. The

maximum hourly flow is 28.5 m3 h-1. The design hydraulic

loading rate is about 36 m yr-1, and the organic loading rate

196 kg ha-1 d-1. Design influent and effluent concentrations are

as follows: for BOD, influent 330 mg l-1, settling tank effluent

196 mg l-1, VSF effluent 12 mg l-1, HSF effluent < 10 mg l-1. For

total suspended solids(TSS), influent 380 mg l-1, settling tank

effluent 80 mg l-1, VSF effluent 12 mg l-1, HSF effluent < 10 mg

l-1.

System Monitoring, Sample Analysis and Statistics

Grab samples were collected regularly at various points

along the FWS system (i.e., inlet, settling tank outflow and

system outflow) over a 3-year monitoring period from August

1999 to August 2003, and along the VSF system (i.e., inlet,

settling tank outflow, siphon, stage I VSF outflow and stage II

VSF outflow over a monitoring period from July 2003 to

August 2004. The samples were analyzed in the laboratory

following APHA standard methods for BOD5, COD, TSS,

Total Kjeldhal Nitrogen (TKN), Total Phosphorus (TP) or PO4,

Total Coliforms (TC) and Faecal Colifoms (FC). Calculations

of statistical parameters of measured constituent

concentrations were performed in a spreadsheet.

Cost Evaluation Method

The actual construction cost of the two systems was

used. Since the FWS system was constructed in 1999 and the

VSF in 2003, for comparison the cost of the FWS was

expressed in 2003 prices using a reported inflation rate of

approximately 3.1%. An economical life of 30 years and a

capital discount factor of 6% were assumed to calculate net-

present-value cost [21]. Operation and maintenance costs

were obtained for the operation time periods from the records

of the local authorities operating the two facilities, and were

expressed in 2003 prices for the FWS system.

RESULTS

System Performance

FWS system

The results during the 3-year period of the FWS facility

monitoring could be summarized as follows: mean BOD5,

COD and TSS removals about 95%, mean TKN and TP

removals about 53%, and TC and FC removals >97% (without

any disinfection). Removal efficiencies of BOD5, COD, TSS,

TKN and TP in the final effluent for the monitoring period are

presented in Table 1. Very high removal rates of BOD5, COD

and TSS (94.4%, 96.1%, and 95.6%, respectively) have been

observed in the septic tank. On the other hand, low removal

rates of TKN and TP of 52.5% and 53.1%, respectively, have

Table 1. Measured concentrations of BOD (mg l-1), COD (mg l-1), TSS (mg l-1), TKN (mg l-1), and TP (mg l-1) in the influent

(IN), the septic tank effluent (SE) and the final effluent (FE), and overall efficiency (TE, %) for the FWS system in

Pompia, Crete, Greece.

Parameter BOD COD TSS TKN TP

IN SE FE TE IN SE FE TE IN SE FE TE IN SE FE TE IN SE FE TE

mg l-1 % mg l-1 % mg l-1 % mg l-1 % mg l-1 %

Average 165 39 7.7 94.4 455 100 18 96.1 191 36 5.6 95.5 38 25 18 52.5 13 9.1 6.2 53.1

Std. Error 31 4.0 1.3 1.0 31 9.8 2.7 0.5 40 5.4 0.8 0.9 3.4 1.7 1.7 4.8 1.5 1.3 1.1 4.7

Min 52 11 2.0 86.5 280 44 2.0 92.7 38 4.0 1.0 86.8 17 8.0 4.0 23.1 4.8 2.3 1.6 10.6

Max 540 60 16 99.1 798 180 40 99.6 720 90 12 99.3 62 36 27 83.1 24 22 21 78.5

# of Data 14 14 15 14 17 18 18 17 17 18 18 17 17 18 18 17 17 18 18 17

624

been obtained in the tank. TN and TP removed in the septic

tank were probably in organic form as particulate organic

matter. Lower removal rates of various constituents in similar

septic tanks have been reported [3].

VSF system

The monitoring results of the VSF system during the 13-

month period of operation could be summarized as follows:

BOD5 removals >92%, TKN removals >89%, and TC removal

>99% (without any disinfection). Removal efficiencies of

BOD5, COD, TSS, TKN and TP in the final effluent for the

monitoring period are shown in Table 2. Removals are

satisfactory, considering that the facility was still new and the

plant root system was probably not fully developed yet.

Relatively, high removal rates of BOD5, COD and TSS have

been measured in the settling tanks. On the other hand, lower

removal rates of TKN (77%) were observed, while TP removal

showed fluctuation and some times increased along the

hydraulic path of the system. It is obvious that to improve

nitrogen and phosphorus removals the last stage of the

system should also be planted and be put soon in operation.

System Costs

FWS system

System cost calculations are presented in Table 3. The

actual capital cost for the FWS system was 305,000 (prices of

1999, including 18% VAT). To compare this cost with that of

the VSF system, it was expressed as 344,615 in 2003 prices

(287.18 p.e.-1) using the 3.1 % inflation rate. This cost also

includes 115,000 for access road and administration room

construction and other works. Some of this work was not

actually necessary, such as extra roads outside of the facility.

In addition, the soil used for planting in the treatment cells

was transported from a distance of more than 10 km with a

relatively high cost. This work was also unnecessary. The net-

Table 2. Measured concentrations of BOD (mg l-1), COD (mg l-1), TSS (mg l-1), TKN (mg l-1) and TP(mg l-1) in the influent (IN),

the settling tank effluent (STE), the VSF effluent (VSF) and overall efficiency (TE, %) for the VSF system in Gomati,

Chalkidiki, Greece.

Parameter BOD COD TSS TKN TP

IN STE VSF TE IN STE VSF TE IN STE VSF TE IN STE VSF TE IN STE VSF TE

mg l-1 % mg l-1 % mg l-1 % mg l-1 % mg l-1 %

Average 485 193 39 92 626 243 62 89 1077 208 9 95 77 51 14 77 17.5 8.2 5.6 62

Std. Error 246 111 29 6 260 119 31 6 1784 474 13 8 47 50 6 20 9.0 3.9 3.1 22

Min 62 10 4 78 238 96 0 81 26 23 0 75 0 8 0 31 7.5 4.3 2.4 24

Max 819 355 92 100 1171 465 106 100 7060 2158 47 100 187 251 27 100 29.3 14.9 11.9 89

# of Data 20 20 19 20 20 20 19 20 20 20 19 20 20 20 19 20 8 8 8 8

Table 3. Capital and operating costs () for the two facilities.

Cost ()

Cost category FWS System VSF System

Capital, including VAT (construction cost) 344,615 410,850

Construction cost per p.e. 287.18 410.85

Net-present-value cost 25,036 29,848

Annual average O&M cost 1,445 6,960

O&M cost per p.e. per year 1.20 6.96

O&M cost per m3 per year 0.03 0.11

Total annual cost (capital and O & M) 26,481 36,808

Total annual cost per p.e. 22.07 36.81

Total annual cost per m3 of influent 0.50 0.56

625

present-value cost was estimated at 25,036 yr-1 using a 6%

discount factor. The total mean operation and maintenance

(O&M) cost of the FWS system in the first three years of

operation was estimated at 1,445 yr-1 (i.e., 1045 for energy

used, 300 for works, and 100, for miscellaneous expenses)

or 1.20 p.e.-1 yr-1 or 0.03 m-3 of influent. Net-present-value

cost and O&M cost are added to a total annual cost of 26,481,

and the mean figures become 22.07 p.e.-1 yr-1 or 0.50 m-3 of

influent.

VSF system

The total construction cost of the VSF system was

410,850 (prices of 2003, including 18% VAT) or 410.85 p.e.-1

This construction cost also included costs for access road (250

m paved road), construction of 550 m sewer line to bring

wastewater to the facility and 400 m sewer line for effluent

disposal to the final receiver, fencing, landscaping and other

works. These extra works are estimated at about 100,000.

The net-present-value cost was estimated at 29,848 yr-1. The

operation cost for the VSF system (for the first 10 months of

operation) comprises electricity (lighting and operation of 9

pumps, estimated at 67.50 month-1 on the average), salaries

for the operator and maintenance works (500 month-1 on the

average) and miscellaneous other expenses (12.50 month-1).

Therefore, the total operation cost is approximately 580

month-1 or 6,960 yr-1 or 6.96 p.e.-1 yr-1 or 0.11 m-3

of influent. Net-present-value cost and O&M cost are added to

a total cost of 36,808 and the mean figures become 36.81

p.e.-1yr-1 or 0.56 m-3 of influent.

Design, Construction and Operation Problems

No major problems were observed in the FWS

constructed wetland. It seems that this CW has been

designed, constructed and is operated very successfully. The

VSF constructed wetland achieves a high removal efficiency

for all pollutants. Nevertheless, it is believed that its

performance could be even better if some design, construction

and operation problems were resolved. These can be

summarized as follows.

Design problems

A major design problem is the sizing of the siphon that

feeds the first stage of the VSF cells (Fig. 1). The dimensions of

this siphon are 10x4.2x0.8m or 3.2m3 of flooding volume. The

siphon floods two cells at a time, i.e., 320 m2, therefore, the

average flooding depth is 1.0 cm. If one considers surface

irregularities of the planted cells, it is obvious that this depth

is small. Usually, 4 to 5 cm of flooding depth are

recommended. This problem was obvious in this facility. The

flooding was limited to about a 1 m wide area around the

perforated feeding pipes (Fig. 2), something seen by denser

plant growth in this area. Therefore, a major part of the

available facility area was not used, reducing active treatment

area and performance. To fix this problem, it is recommended

that the siphon is replaced to one of a larger size that would

provide at least 4 cm of flooding.

Figure 2. View of the stage I distribution pipe.

626

Constructions problems

A construction problem was the proper placing of the

porous media and the installation of filtering material. Some

porous media was washed out from the drainage pipe at the

bottom of the wetland cells. Obviously, this was a result of not

placing proper filtering material. As a result, seepage holes

developed at areas of the cells from where wastewater could

seep out untreated. Another construction problem was in the

last stage (HSF), which was not perfectly level in the lateral to

the flow direction, resulting in preferential flow (and plant

growth) on one side (Fig. 3). Again, this resulted in reduction

in total active treatment area and perfomance.

Operation and maintenance problems

Operation and maintenance problems were also

observed in some of our visits. For example, plants were

grown (and not removed) inside the outlet overflow pipe of

the last HSF stage, obstructing outflow and resulting in

flooding of the system.

It is recommended that these problems are addressed to

improve the system treatment performance.

DISCUSSION AND CONCLUSIONS

In general, selection of the appropriate constructed

wetland system depends on wastewater characteristics,

experience gained, local conditions and site constraints. FWS

systems are less expensive to construct, to operate and to

maintain, are less sensitive and susceptible to problems, and

have greater potential for wildlife support. VSF systems

generally require less land area, are less susceptible to

freezing, mosquitoes and odor problems, and do not have

wastewater exposed at the surface, thus providing minimal

human contact and health risks. These systems are considered

more susceptible to clogging of the media. However,

neither odor nor clogging problem in either system has

been observed so far. It is noted that a possible problem

of mosquitoes in the FWS project was faced effectively by

Figure 3. View of the HSF constructed wetland cell.

627

planting from the start Gambusia spp. fish.

In terms of performance, the organic loading rates were

slightly higher in the VSF than in the FWS system.

Furthermore, ambient temperatures in the VSF system,

located in Northern Greece, are 5 to 10oC lower. Nevertheless,

the high efficiency of both systems has been observed. The

cost analysis, incorporating both capital and operation and

maintenance costs, also suggested a low cost for both systems.

The FWS system was less expensive to construct and to

operate. However, the VSF system required considerably less

land area (in this economic analysis the price of land was not

considered). In terms of construction and O&M problems, the

VSF system, which is more complex in design, construction

and operation, showed most problems, which, however, could

have been predicted and avoided from the beginning. Finally,

the FWS system may freeze for a few days in the winter, if

installed in areas where the temperature drops below 0oC

(e.g., North Greece).

More specifically, the following can be drawn from the

comparison of the FWS and VSF systems:

a. Constructed wetlands are considered appropriate

wastewater treatment systems for the Mediterranean

environment, generating excellent quality of effluent at

the secondary treatment level. In this comparative

study, BOD5, COD and TSS reductions of about 95%

were observed for the FWS CW. BOD5 and TSS

reductions were similar for the VSF system, while COD

reduction was about 89% for this system. In addition,

for the FWS system reductions of TKN and TP of about

53% were measured, and removal rates of TC and FC of

98.7% and 97.1%, respectively. For the VSF system, TKN

removal was 77% on average, while mean phosphorus

removal efficiency was 62%.

b. Reasons for the lower efficiency of the VSF system in

COD removal may be that it was new at the time of the

study and the plant roots were probably not fully

developed. Furthermore, the HSF basin of the system

was not in operation during the study. Nevertheless, the

two systems seem to be very promising in producing a

high effluent quality.

c. The total wetland area (after pre-treatment) of the FWS

system is 5500 m2 (4.58 m2 p.e.-1), while that of the VSF

system is 2040 m2 (including 240 m2 for sludge drying

and storage) or 2.04 m2 p.e.-1 Thus, as expected, the VSF

system provides comparable treatment at significantly

less area (less than half), for slightly higher average

design flow rate (180 m3 d-1 vs. 144 m3 d-1), and at lower

operational temperatures (north vs. south Greece).

d . When comparing the construction costs of the two

systems, it seems that the VSF is slightly more

expensive, probably due to the fact that this system

contains more concrete and several pumps. Generally,

the FWS system construction is much simpler. In terms

of the capital and operation cost, it also seems that the

FWS system is less expensive. Both systems are

considered less expensive, both in construction and

operation, when compared to equivalent conventional

treatment systems operating in the same areas.

e. When comparing design, construction operation and

maintenance problems it seems that, the VSF was more

susceptible to problems since it is a more complex

system. For both systems, careful design and

construction, and proper maintenance are very

important.

In conclusion, the treatment efficiencies of the two

systems are comparable (except for TKN and TP where the

VSF system had higher removal efficiencies), costs seem to be

less for the FWS system, and land requirements are quite

lower for the VSF system. Thus, one can select either system

in terms of treatment efficiency. When land is available, the

FWS system would be preferable because of its simplicity, less

expensive construction, and more reliable and problem-free

operation. If land availability is a problem or land value is

high, then the VSF system would be more preferable. A

careful design and construction, and proper maintenance are

necessary in any case to avoid operational problems.

ACKNOWLEDGEMENTS

We thank G. Dialynas, N. Kefalakis and K. Tsagarakis

for providing information on the study. Sample collection and

analyses for the VSF constructed wetland system were

performed by A. Paltsoglou, K. Vragalas and J.N.E.

Papaspyros. The evaluation of the VSF system was co-funded

by the European Social Fund & National Resources EPEAEK

II PYTHAGORAS II.

REFERENCES

1. Ansola G., Gonzlez J.M., Cortijo R. and de Luis E., Experimental and full-scale pilot plant constructed wetlands for

municipal wastewaters treatment. Ecol. Eng., 21, 43-52 (2003).

2. Kadlec R. and Knight R., Treatment Wetlands, CRC Press, USA, Boca Raton, Fl., (1996).

3. Crites R. and Tchobanoglous G., Small and Decentralized Wastewater Management Systems, WCB and McGraw-Hill. New

York. USA (1998).

4. Andersson J.L., Kallner Bastviken S. and Tonderski K.S., Free water surface wetlands for wastewater treatment in Sweden

nitrogen and phosphorus removal. In: Proc. of the 9th International Conference on Wetland Systems for Water Pollution Control,

628

IWA, Avignon, France, September 26-30, Vol. 1, pp. 39-47 (2004).

5. Koottatep T., Surinkul N., Polprasert C., Kamal A.S. and Srauss M., Treatment of septage in constructed wetlands in

tropical climate Lessons learnt after seven years of operation. In: Proc. of the 9th International Conference on Wetland Systems

for Water Pollution Control, IWA, Avignon, France, September 26-30, Vol. 1, pp. 249-257 (2004).

6. Cooper D., Griffin P. and Cooper P., Factors affecting the longevity of sub-surface horizontal flow systems operating as

tertiary treatment for sewage effluent. In: Proc. of the 9th International Conference on Wetland Systems for Water Pollution

Control, IWA, Avignon, France, September 26-30, Vol. 1, pp. 259-267 (2004).

7. Solano M.L., Soriano P. and Ciria M.P., Constructed wetlands as a sustainable solution for wastewater treatment in small

villages. Biosystems Eng., 87, 109-118 (2004).

8. Cameron K., Madramootoo C., Crolla A. and Kinsley C., Pollutant removal from municipal sewage lagoon effluents with a

free-surface wetland. Water Res., 37, 2803-2812 (2003).

9. Ran N., Agami M. and Oron G., A pilot study of constructed wetlands using duckweed (Lemna gibba L.) for treatment of

domestic primary effluent in Israel. Water Res., 38, 2241-2248 (2004).

10. He Q. and Mankin K., Performance variations of COD and nitrogen removal by vegetated submerged bed wetlands. J. Am.

Water Resour. Assoc., 38, 1679-1689 (2002).

11. Mayo A. W. and Mutamba J., Effect of HRT on nitrogen removal in a coupled HRP and unplanted subsurface flow gravel

bed constructed wetland. Phys. Chem. Earth, 29, 1253-1257 (2004).

12. Korkusuz E. A., Beklioglu M. and Demirer G.N., Comparison of gravel and slag based vertical flow reed bed performance

in Turkey. In: Proc. of the 9th International Conference on Wetland Systems for Water Pollution Control, IWA, Avignon, France,

September 26-30, Vol. 1, pp. 173-181 (2004).

13. Akratos C. and Tsihrintzis V.A., Pilot-scale constructed wetlands for wastewater treatment studies. In: Proc. of the

International Conference on Protection and Restoration of the Environment VII, Technical University of Athens, Greece and

Sterens Institute of Technology, USA. June 28-July 1, 2004, Mykonos, Greece, CD-Rom Section IV (Ecological Engineering

Applications), #1 (2004).

14. Rousseau D.P.L., Vanrolleghem P.A. and De Pauw N., Constructed wetlands in Flanders: a performance analysis. Ecol.

Eng., 23, 151-163 (2004).

15. Luederitz V., Eckert E., Lange-Weber M., Lange A. and Gersberg R.M., Nutrient removal efficiency and resource

economics of vertical and horizontal flow constructed wetlands. Ecol. Eng., 18, 157-171 (2001).

16. Sderqvist T., Constructed wetlands as nitrogen sinks in southern Sweden: An empirical analysis of cost determinants.

Ecol. Eng., 19, 161-173 (2002).

17. Molle P., Linard A., Boutin C., Merlin G. and Iwema A., How to treat raw sewage with constructed wetlands: An

overview of the French systems. In: Proc. of the 9th International Conference on Wetland Systems for Water Pollution Control,

IWA, Avignon, France, September 26-30, Vol. 1, pp. 11-19 (2004).

18. Paing J. and Voisin J., Vertical flow constructed wetlands for municipal wastewater and septage treatment in French rural

area. In: Proc. of the 9th International Conference on Wetland Systems for Water Pollution Control, IWA, Avignon, France,

September 26-30, Vol. 1, pp. 315-323 (2004).

19. Brix H., Danish guidelines for small-scale constructed wetland systems for onsite treatment of domestic sewage. In: Proc. of

the 9th International Conference on Wetland Systems for Water Pollution Control, IWA, Avignon, France, September 26-30, Vol.

1, pp. 1-9 (2004).

20. Dialynas G.E., Kefalakis N., Dialynas M.G. and Angelakis A.N., Performance of the first free-water surface constructed

wetland system in Crete, Greece. Water Sci. Technol., 46, 355-360 (2002).

21. Tsagarakis K.P., Mara D.D. and Angelakis A.N., Application of cost criteria for selection of municipal wastewater

treatment systems. Water, Air Soil Pollut., 142, 187-210 (2003).