Ecological Engineering 44 (2012) 337–343

Contents lists available at SciVerse ScienceDirect

Ecological Engineering
journal homepage: www.elsevier.com/locate/ecoleng

Short communication

A small-size vertical flow constructed wetland for on-site treatment of

household wastewater
Georgios D. Gikas, Vassilios A. Tsihrintzis ∗
Laboratory of Ecological Engineering and Technology, Department of Environmental Engineering, School of Engineering, Democritus University of Thrace, 67100 Xanthi, Greece

a r t i c l e i n f o a b s t r a c t

Article history: The design and performance of a small-scale vertical flow constructed wetland (VFCW) system, for on-site
Received 11 October 2011 treatment of domestic wastewater, are presented. The system serves a two-story, two-family (8 persons)
Received in revised form 31 March 2012 building, and comprises three treatment stages: two settling tanks in series, a VFCW and a zeolite tank.
Accepted 8 April 2012
The treatment performance of the system was monitored on a weekly basis for about forty months.
Available online 10 May 2012
Results show a satisfactory performance with the following mean removal efficiencies: 96.4% for BOD,
94.4% for COD, 90.8% for TKN, 92.8% for ammonia, 61.6% for OP and 69.8% for TP. The zeolite was found to
Keywords:
offer additional removal of nitrogen, total phosphorus and organic matter. The zeolite saturation time is
On-site wastewater treatment
VF constructed wetlands
estimated.
HSF constructed wetlands © 2012 Elsevier B.V. All rights reserved.
Zeolite
Organic matter
Nitrogen
Phosphorus
k–C* model

1. Introduction
Constructed wetland (CW) technology for wastewater treat-
ment is a relatively new technology, characterized by efficiency in
pollutant removal, simplicity, and low construction and operation
costs (Kadlec and Wallace, 2009; Vymazal et al., 1998). Constructed
wetland systems are used in many countries, serving relatively
small settlements, i.e., of the order of 1000 person equivalent (p.e.).
A small number of full-scale CW systems operate in Greece (Gikas
et al., 2007, 2011; Tsihrintzis et al., 2007; Tsihrintzis and Gikas,
2010). Several studies were completed showing the effectiveness
of CW systems in wastewater treatment under Mediterranean cli-
mate conditions (Akratos and Tsihrintzis, 2007; Akratos et al., 2008,
2009a,b; Kotti et al., 2010; Stefanakis and Tsihrintzis, 2012a).
Recent CW applications for domestic wastewater treatment
also include the service of isolated houses (4–10 people), mostly
located in agricultural and ecologically sensitive regions, or where
there is no possibility of house connection to a public wastewater
treatment system (Brix et al., 2003; Gikas and Tsihrintzis, 2010;
Zakova and Zak, 2003). Such systems are usually of the horizontal
subsurface flow (HSF) or vertical flow (VF) CW type, varying in
∗ Corresponding author. Tel.: +30 25410 79393; fax: +30 25410 79393;
mobile: +30 6974 993867.
E-mail addresses: tsihrin@env.duth.gr, tsihrin@otenet.gr (V.A. Tsihrintzis).
surface areas from 3 to 6 m2 /p.e. and planted with Phragmites aus-
tralis (Brix and Arias, 2005). Usually, two or three settling tanks are
placed before the CW for pretreatment of influent (Börner et al.,
1998; Vymazal, 2001; Zakova and Zak, 2003).
This paper describes the design and construction details of the
on-site wastewater treatment system for a two-family house in
Avdira, Xanthi, Northeast Greece, and presents performance results
from its 40-month monitoring under Mediterranean climate con-
ditions.
2. Materials and methods
2.1. Facility description
The on-site CW was built to treat the sewage from a two-family
residence of 8 people in Avdira, a village in Xanthi district, North-
east Greece (40◦ 58 49.6 N, 24◦ 56 52.8 E; Elevation 74 m). This
system consists of two settling tanks in series (ST1 and ST2), a vault
with a pump (ST3), a vertical flow CW, a zeolite tank (ZT), and an
effluent collection tank (Fig. 1). From the residence, the wastewater
inflows to the first settling tank (ST1), which overflows to the sec-
ond one (ST2) and then to the pump vault (ST3). The dimensions of
ST1 and ST2 are 1.3 m by 1.3 m in plan view. The operation depth is
1.5 m. The dimensions of ST3 are 0.7 m by 0.7 m in plan view. The
operation depth is 1.5 m (Fig. 1). In the vault (ST3), a pump is used
to discharge approximately every 3 h 150 L of wastewater onto the

0925-8574/$ – see front matter © 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.ecoleng.2012.04.016
338 G.D. Gikas, V.A. Tsihrintzis / Ecological Engineering 44 (2012) 337–343
Fig. 1. Constructed wetland system in Avdira, Xanthi district: (a) plan view; (b) aeration pipes; (c) porous media; (d) septic tanks; and (e) constructed wetland cells and
zeolite tank.
CW bed. The hydraulic residence time in the two septic tanks is
about 4.5 days.
The constructed wetland is of vertical subsurface flow type
(area 24.5 m2 ). The bed is made of reinforced concrete with ver-
tical walls and is separated in two equally sized cells CW1 and
CW2. The dimensions of CW1 and CW2 are 3.5 m by 3.5 m each
in plan view, and 1.3 m in depth (Fig. 1). The first cell (CW1) was
retained unplanted (for experimental reasons) and the second one
(CW2) was planted with common reed (Phragmites australis). The
CW contains porous media obtained from a river bed and placed
in two layers. The bottom layer is 30 cm thick well-washed coarse
gravel (D50 = 25 mm; range 20–40 mm). A layer of 70 cm fine gravel
(D50 = 6 mm; range 2–10 mm) was placed on top of the coarse gravel
layer. The CW contains networks of drainage and aeration pipes
(Fig. 1). The two beds CW1 and CW2 operate simultaneously, with-
out a resting period, and the total influent volume of 150 L is divided
in two equal parts of about 75 L in each CW bed.
The effluent from the bed is guided into the zeolite tank for
further removal of ammonia and phosphorus. The dimensions of
the zeolite tank are 1.8 m by 1.3 m in plan view, and 1.0 m in depth
(Fig. 1). The hydraulic residence time in the zeolite tank is about 1
day. After passing the zeolite tank, the treated wastewater enters
the effluent tank where it is temporarily stored and pumped for
irrigation.
2.2. Wastewater quality monitoring – statistical analyses
The CW was put in operation in January 2007. Its treatment
performance was evaluated in 82 sampling campaigns, which took
place in the 40-month period from January 2007 until May 2010.
Wastewater samples were collected at various locations along the
facility, i.e., at the influent and the effluent points of the facility, and
at various intermediate points. More specifically: one sample from
each settling tank and the pump vault (ST1–ST3), two from each
wetland cell effluent (CW1 and CW2), one from the inlet of the zeo-
lite tank and one from the effluent of the facility (Fig. 1). Collected
wastewater samples were properly stored in a cool box, trans-
ported to the laboratory and analyzed immediately for BOD, COD,
TKN, ammonia nitrogen (NH4 -N), nitrate (NO3 -N) and nitrite nitro-
gen (NO2 -N), ortho-phosphates (OP) and total phosphorus (TP),
following standard methods (APHA, 1998). In addition, tempera-
ture (T), electrical conductivity (EC), pH and dissolved oxygen (DO)
were measured in situ, using appropriate equipment (WTW 197-
series), at the same points where water samples were collected.
SPSS 15.0 for Windows was used in performing statistical anal-
yses, which included: (1) one-way between groups ANOVA and the
Tukey honestly significant difference (HSD) test, to determine sta-
tistically significant differences at 95% confidence interval (p < 0.05)
between the various treatment stages (i.e., influence of stage) of the
mean values of pollutant removals; and (2) the t-test, to determine
statistically significant differences between the mean values of pol-
lutant removals at temperatures above and below 15 ◦ C, and also
between the planted and unplanted CW cells.
3. Results and discussion
3.1. Mean physicochemical parameter variation
Measured physicochemical parameters and pollutant concen-
tration statistics are presented in Tables 1 and 2, respectively.
Figs. SM-1 and SM-2 [see Supplemental Online Material (SM)]
present the seasonal variation charts of these parameters. The
wastewater temperature did not show any significant spatial vari-
ations along the unit, and generally ranged between 8.5 ◦ C and
31.3 ◦ C, depending on the season (Table 1; Fig. SM-1a). Practi-
cally, temperature variation along the facility was minor. Since the
wastewater was of domestic origin, the mean pH value was in the
neutral area (Kadlec and Wallace, 2009) at every point of the facil-
ity, ranging between 7.4 and 7.7 (Table 1; Fig. SM-1b). Mean DO
concentrations increased along the facility, ranging from the inlet
to the outlet from 0.16 mg/L to 1.21 mg/L (Table 1; Fig. SM-1c). This
 G.D. Gikas, V.A. Tsihrintzis / Ecological Engineering 44 (2012) 337–343 339

Table 1
Physicochemical parameter statistics.

Parameter Sample location

ST1 ST2 ST3a Effluent CW1 Effluent CW2 Influentb ZT Final Effluent

Temperature (◦ C)
Mean 18.9 18.7 18.6 17.9 17.9 18.0 18.9
SD (n) 5.0 (82) 5.2 (82) 5.3 (82) 5.5 (82) 5.5 (82) 5.5 (82) 5.6 (82)
Min 11.0 10.4 10.1 8.5 8.8 8.7 9.4
Max 29.1 30.4 31.3 29.5 29.5 30.1 29.2
pH
Mean 7.4 7.5 7.7 7.6 7.6 7.6 7.5
SD (n) 0.6 (82) 0.6 (82) 0.4 (82) 0.4 (82) 0.4 (82) 0.4 (82) 0.4 (82)
Min 6.5 6.1 6.5 6.7 6.3 6.5 6.5
Max 8.9 8.9 8.9 8.8 8.7 8.7 8.1
DO (mg/L)
Mean 0.16 0.16 0.24 0.85 0.87 0.87 1.21
SD (n) 0.19 (67) 0.20 (67) 0.26 (67) 0.95 (67) 0.93 (67) 0.92 (67) 0.78 (67)
Min 0.00 0.00 0.00 0.00 0.00 0.00 0.06
Max 0.90 0.92 1.20 3.23 3.27 3.80 3.36
EC (␮S/cm)
Mean 1662 1674 1583 1445 1447 1447 1192
SD (n) 273 (82) 240 (82) 209 (82) 206 (82) 195 (82) 194 (82) 195 (82)
Min 950 863 1050 983 1020 1000 627
Max 2460 2140 1980 1893 1889 1889 1640

SD: standard deviation; n: number of data.

a
ST3 = influent CW (CW1, CW2).
b
Influent ZT = effluent from CW1 and CW2.

is a result of oxygenation by the aeration pipes and the plants. The
mean electrical conductivity decreased along the facility; mean
values ranged from 1662 ␮S/cm in the influent to 1192 ␮S/cm
in the effluent (Table 1). This is expected due to the removal of
dissolved nitrogen and phosphorus salts. There are cases, how-
ever, in constructed wetlands, where the final EC values may
be increased due to evapotranspiration and interactions between
porous media and wastewater. This may happen mostly during
the summer period when the wastewater temperature is high
(Fig. SM-1d).
There were observed variations of influent BOD and COD
concentrations (Fig. SM-2a and b), with observed mean values
during the monitoring period 475.9 mg/L and 960.3 mg/L, respec-
tively (Table 2). BOD and COD concentrations decreased along
the facility with mean effluent concentrations of 20.2 mg/L and
48.9 mg/L, respectively (Table 2). This indicates significant removal
of organic matter at all facility stages. Mean BOD and COD efflu-
ent concentrations were also lower than the limits set by EU
(EU directive 1991/271/EEC: effluent limits for BOD 25 mg/L, COD
125 mg/L, TN 15 mg/L, TP 2 mg/L; or, alternatively, minimum per-
centages of removal for BOD 70–90%, COD 75%, TN 70–80%, TP
80%).
The mean influent concentrations of TKN and NH4 -N were
95.0 mg/L and 82.4 mg/L, respectively, and mean effluent concen-
trations were 10.2 mg/L and 7.7 mg/L (Table 2 and Fig. SM-2c and d).
There were observed seasonal variations of TKN and NH4 -N con-
centrations at all facility stages, with lower effluent values observed
in the summer (mean values: TKN 8.8 mg/L; NH4 -N 5.9 mg/L). In the
winter, the effluent concentrations of TKN and ammonia (mean val-
ues 11.1 mg/L and 8.6 mg/L, respectively) were higher, indicating
that nitrification is insufficient, mainly due to low temperatures.
During the monitoring period, mean NO3 -N and NO2 -N influ-
ent concentrations were 1.11 mg/L and 0.19 mg/L, respectively,
and mean effluent concentrations were 7.37 mg/L and 0.96 mg/L,
respectively, indicating that nitrification takes place along the
facility stages. The mean effluent values for total nitrogen (TN)
(18.5 mg/L) were higher than those set by EU (TN: 15 mg/L), but
the mean% removal (about 80%) was within the range set by EU
(70–80%).
Phosphorus concentration also decreased along the facility,
without showing a seasonal variation (Fig. SM-2e and f). TP and
OP mean effluent concentrations were 2.9 mg/L and 1.9 mg/L,
respectively, during the monitoring period (Table 2). The mean
effluent values for TP (2.9 mg/L) were higher than those set by
EU (TP: 2 mg/L). TSS concentrations decreased along the facility
and the mean effluent concentrations were 14.9 mg/L (Table 2;
Fig. SM-2g). Total coliform (TC) effluent concentration was rela-
tively high (mean: 7000 N/100 mL), although it decreased along the
unit (Table 2; Fig. SM-2h).
3.2. Pollutant removal at the various stages of the facility
Fig. SM-3 presents the variation of mean removal efficiency
along the various parts of the unit for all the pollutants. Overall
removal for each constituent was calculated based on its concen-
trations at the inlet and outlet of the facility. Absolute removal for
each pollutant at the intermediate stages of the facility was cal-
culated based on its concentration at the inlet and outlet of the
stage. Relative removal at intermediate stages of the facility was
calculated based on concentrations of the constituent at the inlet
of the facility and the outlet of the particular stage, and subtract-
ing the removal in the previous stage. Relative removal indicates
the contribution of each stage in total performance of the facil-
ity. Cumulative removal results from the accumulation of relative
removals at various stages along the facility.
BOD and COD mean overall removal values were 95.8% and
94.9%, respectively (Fig. SM-3a and b). Organic matter absolute
removal was greater in the ZT stage than in the other stages. BOD
and COD absolute retentions were 82.9% and 76.2% in the ZT, 55.6%
and 60.2% in the septic tanks and 44.1% and 46.3% in the CW,
respectively (Fig. SM-3a and b). The removal of organic matter
by zeolite should be attributed to its relatively large pores that
allow for the adsorption of solid particles and dissolved organics
(Stefanakis and Tsihrintzis, 2012b). The stage where most rela-
tive removal of organic matter occurred was the septic tanks, with
mean removals of 55.6% and 60.2% for BOD and COD, respectively.
The stage with the lower relative removal values of organic matter
was the ZT, with mean relative removals for BOD and COD 20.0%
340 G.D. Gikas, V.A. Tsihrintzis / Ecological Engineering 44 (2012) 337–343

Table 2
Pollutant concentration statistics.

Parameter Sample location

ST1 ST2 ST3a Effluent CW1 Effluent CW2 Influentb ZT Final Effluent

BOD (mg/L)
Mean 475.9 314.1 211.3 154.9 109.5 118.2 20.2
SD (n) 155.3 (82) 111.8 (82) 77.8 (82) 70.4 (82) 48.6 (82) 48.7 (82) 14.3 (82)
Min 248.0 113.0 36.6 23.0 17.8 20.8 5.0
Max 815.0 653.0 400.0 385.0 220.0 238.0 69.0
COD (mg/L)
Mean 960.3 615.0 382.6 232.1 198.2 205.5 48.9
SD (n) 409.8 (82) 261.8 (82) 157.2 (82) 97.0 (82) 96.4 (82) 85.9 (82) 29.6 (82)
Min 101.4 96.0 71.3 55.4 34.8 43.4 14.9
Max 1776.0 1593.6 725.8 503.2 556.3 522.9 155.5
TKN (mg/L)
Mean 95.0 87.6 66.8 43.6 36.6 38.7 10.2
SD (n) 31.0 (71) 28.9 (71) 23.7(71) 18.5 (71) 15.1 (71) 15.4 (61) 3.8 (71)
Min 36.4 25.3 20.7 12.5 10.8 12.3 3.0
Max 149.5 131.6 116.5 106.7 82.9 98.3 23.8
NH4 -N (mg/L)
Mean 82.4 79.0 62.5 40.2 31.6 32.7 7.7
SD (n) 24.8 (71) 23.5 (71) 16.7 (71) 14.1 (71) 11.4 (71) 11.9 (71) 3.3 (71)
Min 15.1 7.0 19.7 17.1 13.8 15.0 2.5
Max 131.0 121.5 99.7 82.0 67.8 72.2 15.1
NO3 -N (mg/L)
Mean 1.11 1.44 2.05 5.84 6.23 6.17 7.37
SD (n) 1.75 (62) 2.68 (62) 3.28 (62) 7.05 (62) 8.15 (62) 7.98 (62) 9.11 (62)
Min 0.00 0.00 0.00 0.00 0.00 0.25 0.48
Max 9.96 15.74 17.03 30.66 39.43 32.61 38.14
NO2 -N (mg/L)
Mean 0.19 0.36 0.73 1.65 1.56 1.62 0.96
SD (n) 0.27 (62) 0.91 (62) 1.08 (62) 1.50 (62) 1.36 (62) 1.74 (62) 0.84 (62)
Min 0.00 0.01 0.01 0.00 0.00 0.00 0.01
Max 1.23 6.60 6.14 8.22 6.29 7.43 4.05
TP (mg/L)
Mean 8.8 7.1 6.0 5.3 4.7 4.8 2.9
SD (n) 3.4 (82) 3.2 (82) 2.1 (82) 2.1 (82) 2.2 (82) 2.0 (82) 1.1 (82)
Min 3.4 2.0 3.1 1.7 1.2 1.6 1.1
Max 19.9 16.6 12.0 12.0 11.5 11.5 6.5
OP (mg/L)
Mean 4.4 3.8 3.5 3.2 2.6 2.8 1.9
SD (n) 2.4 (82) 2.4 (82) 2.1 (82) 1.9 (82) 1.9 (82) 1.9 (82) 1.0 (82)
Min 1.0 0.0 0.4 0.5 0.0 0.1 0.2
Max 12.7 9.8 9.1 9.8 8.5 8.9 5.0
TSS (mg/L)
Mean 374.3 128.8 124.1 77.9 71.0 72.3 14.9
SD (n) 202.4 (78) 45.4 (78) 43.6 (78) 42.7 (78) 40.3 (78) 31.4 (78) 11.6 (78)
Min 31.8 19.4 20.0 10.8 6.5 8.0 0.7
Max 941.3 247.8 246.2 210.5 223.5 214.4 67.0
TC 105 N/100 mL
Mean 198.3 167.9 72.4 0.50 0.41 0.43 0.07
SD (n) 550.8 (30) 380.8 (29) 201.7 (30) 0.7 (29) 0.66 (28) 0.60 (29) 0.09 (22)
Min 4.0 2.0 0.8 0.01 0.01 0.01 0.01
Max 3000.0 2000.0 890.0 3.0 3.0 3.0 0.40

SD: standard deviation; n: number of data.

a
ST3 = influent CW (CW1, CW2).
b
Influent ZT = effluent from CW1 and CW2.

and 16.3%, respectively (Fig. SM-3a and b), but the zeolite-filter
contributed substantially to the improved performance. One way
between groups ANOVA analysis (BOD: F2,224 = 131.46, p = 0.001;
COD: F2,228 = 269.39, p = 0.001; Table SM-1) and post hoc compari-
son using the Tukey HSD test indicated for BOD and COD that the
mean score of the septic tanks was significantly different from that
of the CW and from that of the ZT. For COD, the mean score of the
CW was significantly different from that of the ZT (Table SM-1).
Nitrogen removal was also satisfactory; mean overall removal
values for TKN and NH4 -N were 89.3% and 90.7%, respectively
(Fig. SM-3c and d). Nitrogen absolute removal was greater in the ZT
than in the other stages. TKN and NH4 -N absolute retention were
29.7% and 24.2% in the STs, 42.0% and 47.5% in the CW and 73.7%
and 76.8% in the ZT, respectively (Fig. SM-3c and d). High relative
removal was achieved in the zeolite tank (30.0% for TKN and 30.6%
for NH4 -N; Figs. SM-3c and d). There was no statistically signifi-
cant difference of TKN relative removal values between treatment
stages (ANOVA: F2,195 = 1.09, p = 0.337; Table SM-1). However, there
was statistically significant difference of NH4 -N relative removal
values between treatment stages (ANOVA: F2,197 = 9.48, p = 0.001;
Table SM-1). Post hoc comparison using Tukey HSD test indicated
for NH4 -N that the mean score of the septic tanks was significantly
lower than that of the CW and than that of ZT (Table SM-1).
TP and OP overall removal values were 67.3% and 57.9%, respec-
tively (Fig. SM-3e and f). Highest absolute removal of TP and OP
occurred in the ZT (mean removals of TP and OP were 40.7% and
32.7%, respectively). In the other two stages, TP and OP absolute
removals were in the STs 31.8% and 21.4%, respectively, and in the
CW 19.1% and 20.3%, respectively. Septic tanks showed the high-
est contribution to TP and OP removal (mean relative removal for
 G.D. Gikas, V.A. Tsihrintzis / Ecological Engineering 44 (2012) 337–343
TP 31.8% and for OP 21.4%). The reason for this is that phosphorus
particles are attached to suspended solids, which are removed due
to settlement. The respective rates for TP relative removal in CW
and ZT stages were 13.0% and 22.4%, respectively, and for OP rel-
ative removal 16.0% and 20.5% (Fig. SM-3e and f). ANOVA analysis
for TP (F2,220 = 15.82, p = 0.001; Table SM-1) and post hoc compar-
ison using the Tukey HSD test indicated that the mean score for
the STs (mean = 31.8%) was significantly different from that of the
other treatment stages (i.e., CW: mean = 13.0%; ZT: mean = 22.4%),
and the mean score for the CW was significantly different from
that of the ZT. There was not significant difference of OP removal
values between treatment stages (ANOVA: F2,213 = 0.53, p = 0.591;
Table SM-1).
Total suspended solids were removed to a high degree with
mean removal 96.0%. TSS absolute removals were: 66.8% in the
STs, 41.7% in the CW and 79.4% in the ZT (Fig. SM-3g). The treat-
ment stage with the main contributions to TSS removal was the
septic tanks (66.8%). The CW and ZT stages showed compara-
ble contribution to TSS removal of 13.8% and 15.3%, respectively
(Fig. SM-3g). There was statistically significant difference of TSS
removal values between treatment stages (ANOVA: F2,222 = 354.91,
p = 0.001; Table SM-1). Post hoc comparison using the Tukey HSD
test indicated for TSS that the mean score of the septic tanks
was significantly greater than that of the CW and that of the ZT
(Table SM-1).
Total coliforms (TC) were removed in the septic tanks through
sedimentation and in the CW mainly through filtering in the porous
media; the total TC removal efficiency was 99.97% (Fig. SM-3h). The
highest absolute and relative removal of TC occurred in the CW
(mean absolute removal 99.6%) and in the STs (relative removal
63.5%), respectively. The lower relative removal of TC occurred in
the ZT (Fig. SM-3h). There was statistically significant difference of
TC removal values between treatment stages (ANOVA: F2,72 = 74.35,
p = 0.001). Post hoc comparison using Tukey HSD test indicated for
TC that the mean score of all stages were significantly different
(Table SM-1).
3.3. Effect of temperature
The influence of temperature on pollutant removal, in each stage
separately and overall, was also investigated. Mean values and t-
test results of pollutant removals (combined effluent from CW1
and CW2) for temperatures above and below 15 ◦ C are presented in
Table SM-2. This temperature value was selected based on previous
studies, because below it plant growth is halted and also bacteria
that contribute to nitrogen removal do not work efficiently (Akratos
and Tsihrintzis, 2007; Gikas and Tsihrintzis, 2010; Kotti et al., 2010).
Furthermore, evapotranspiration starts being significant at tem-
peratures above 15 ◦ C (Papaevangelou et al., 2012).
Based on t-test analysis (Table SM-2), the mean BOD and COD
removals (in each stage separately and overall) for temperatures
above and below 15 ◦ C were not statistically different, as also
observed in other cases of CWs (Gikas and Tsihrintzis, 2010; Gikas
et al., 2011). This implies that the removal of the organic matter
is mostly due to microbial activity of aerobic and anaerobic bacte-
ria, which function even at temperatures as low as 5 ◦ C (Vymazal,
2002).
The t-test (Table SM-2) showed that the mean values of NH4 -N
and TKN removals for temperatures above and below 15 ◦ C were
not statistically significantly different, based on data in septic tanks,
CW, ZT and all stages for TKN. The dependence of NH4 -N and TKN
removal efficiency on temperature was not generally found sig-
nificant in previous studies in VFCWs (Stefanakis and Tsihrintzis,
2012a; Gikas et al., 2011).
341
The t-test also showed that the mean OP and TP removals for
temperatures above and below 15 ◦ C were not statistically sig-
nificantly different except for the case of TP in the ZT stage and
all stages (Table SM-2). The explanation for this is that the main
phosphorus removal mechanisms are porous media adsorption and
microbial consumption, processes not affected directly by tem-
perature (Gikas and Tsihrintzis, 2010; Kadlec and Wallace, 2009;
Stefanakis et al., 2009). Finally, the t-test showed that the mean
TSS and TC removals for temperatures above and below 15 ◦ C were
not statistically significantly different (Table SM-2). The reason for
this is that coliform decline in CWs is due to the lethal effect of
solar radiation (however, not important in VFCW systems), natural
cell die-off, sedimentation, filtration, and adsorption (Kadlec and
Wallace, 2009).
3.4. Effect of planting
As mentioned before, the CW1 cell was unplanted and the
CW2 was planted with common reed. By comparing the removal
efficiencies of the two CWs, conclusions can be drawn about
the effect of vegetation. Table SM-2 also presents mean values
and t-test results of pollutant removals for CW1 (unplanted) and
CW2 (planted). The BOD and COD removal efficiencies were 35.8%
and 38.7% for CW1 and 49.4% and 47.7% for CW2, respectively
(Table SM-2). Concerning the removal efficiencies of organic mater,
the mean values of CW2 (planted) were found statistically signifi-
cantly higher than those of CW1 (unplanted; Table SM-2). The main
mechanism responsible for organic matter removal is the micro-
bial activity of aerobic and anaerobic bacteria (Vymazal, 2002);
it appears that plants play a role on removal efficiencies as they
transport oxygen to the porous media and offer their extensive root
system as media for microbial development.
Plants were found to also affect nitrogen removal. The mean
values of NH4 -N and TKN removal efficiencies of CW2 (49.4% and
45.3%; Table SM-2) were statistically significantly higher than those
of CW1 (35.5% and 34.8%; Table SM-2); this was also observed
by Akratos and Tsihrintzis (2007) and Stefanakis and Tsihrintzis
(2012a), and could be explained by the fact that mechanisms for
nitrogen removal, in addition to microbial activity (ammonifica-
tion, nitrification and denitrification), also include plant uptake
(Vymazal, 2002).
OP and TP also showed a significant dependence of removal effi-
ciency on the presence of plants. Specifically, the mean removal
efficiency of OP and TP (19.3% and 21.5%; Table SM-2) of CW2 were
statistically significantly higher than those of CW1 (8.6% and 10.6%;
Table SM-2). Although the main mechanism of phosphorus removal
is adsorption and precipitation, plant uptake plays apparently a sig-
nificant role. Finally, as Table SM-2 shows, there is no significant
effect of plants on TSS and TC removal. All the above findings show
that the presence of common reed in CW2 unit plays an important
role in organic matter, nitrogen and phosphorus removal. Also, the
presence of the reeds may help preventing clogging through their
root growth and stem movement by the wind (Brix and Arias, 2005).
3.5. Effect of zeolite
Enhanced CW wastewater treatment using special materials has
been assessed in the laboratory (Arias et al., 2001), in pilot-scale
VFCWs (Prochaska and Zouboulis, 2006; Prochaska et al., 2007;
Stefanakis and Tsihrintzis, 2012a), and in filters treating the efflu-
ent of CWs (Arias et al., 2003; Stefanakis et al., 2009; Stefanakis and
Tsihrintzis, 2012b).
In the case of the on-site CW system in Avdira, zeolite was used
as filter medium; no clogging was observed during the monitor-
ing period (40 months). As mentioned before, the zeolite filter was
342 G.D. Gikas, V.A. Tsihrintzis / Ecological Engineering 44 (2012) 337–343
Fig. 2. Temporal variation in the zeolite tank of the CW system in Avdira: (a) temperature; and (b) TP and NH4 -N removal.
designed for one-day retention time of wastewater. Fig. 2b presents
the temporal variation of mean monthly removal of TP and NH4 -N
in the zeolite tank. The temporal variation of TP removal showed
a pattern following the mean monthly wastewater temperature
(Fig. 2a). Removal performance is shown to increase with temper-
ature. The reason for this is increased ET in the CW during high
temperatures, which results in a reduced effluent quantity enter-
ing the zeolite tank, thus, resulting in increased residence time. The
negative slope of the trendline (−0.65% per month) indicates that
the zeolite filter looses capacity and the TP removal is reduced on
the average by approximately 7.8% per year. Nevertheless, adsorp-
tion sites were still available after 40 months of operation. The
temporal variation of NH4 -N removal showed a similar pattern
to TP, also following the mean monthly wastewater temperature
(Fig. 2a and b). In this case, the negative slope of the trendline
(−0.88% per month) indicates that the NH4 -N removal is reduced
by about 10.5% per year. Thus, the expected life of the filter is about
6.3 and 8.4 years for TP and NH4 -N, respectively.
The k–C* model (Kadlec and Wallace, 2009; Rousseau et al.,
2004) was validated with the experimental data and is presented
in Supplemental Online Material, Section SM-1. A comparison with
performance of an on-site HSF CW system (Gikas and Tsihrintzis,
2010) in the same geographic area is presented in Section SM-2.
4. Conclusions
In rural and ecologically sensitive areas, where there is no
possibility for sewer connection, small-size on-site constructed
wetlands provide an excellent solution for wastewater treatment
of one- or two-family residences. From the 40-month period of
operation and monitoring, it seems that the on-site VFCW facility
in Avdira operates quite satisfactorily, showing high removals for
BOD, COD, ammonia, TKN, phosphorus, TSS and TC. The zeolite tank
offers further CW effluent polishing.
Acknowledgement
This study was funded through the EU INTERREG IIIB-
MEDOCC project “Réseau Durable d’ Aménagement des Ressources
Hydrauliques (HYDRANET)” (2005–2007).
Appendix A. Supplementary data
Supplementary data associated with this article can be
found, in the online version, at http://dx.doi.org/10.1016/j.
ecoleng.2012.04.016.
References
Akratos, C.S., Tsihrintzis, V.A., 2007. Effect of temperature, HRT, vegetation and
porous media on removal efficiency of pilot-scale horizontal subsurface flow
constructed wetlands. Ecol. Eng. 29, 173–191.
Akratos, C.S., Papaspyros, J.N.E., Tsihrintzis, V.A., 2008. An artificial neural network
model and design equations for BOD and COD removal prediction in horizontal
subsurface flow constructed wetlands. Chem. Eng. J. 143 (1–3), 96–110.
Akratos, C.S., Papaspyros, J.N.E., Tsihrintzis, V.A., 2009a. Total nitrogen and ammo-
nia removal prediction in horizontal subsurface flow constructed wetlands: use
of artificial neural networks and development of a design equation. Bioresour.
Technol. 100 (2), 586–596.
Akratos, C.S., Papaspyros, J.N.E., Tsihrintzis, V.A., 2009b. Artificial neural network
use in ortho-phosphate and total phosphorus removal prediction in horizontal
subsurface flow constructed wetlands. Biosys. Eng. 102 (2), 190–201.
APHA, 1998. AWWA. In: Standard Methods for the Examination of Water and
Wastewater, 20th ed. American Public Health Association, American Water
Works Association, Washington, D.C.
Arias, C.A., Brix, H., Johansen, N.H., 2003. Phosphorus removal from municipal
wastewater in an experimental two-stage vertical flow constructed wetland
system equipred with a calcite filter. Water Sci. Technol. 48 (5), 51–58.
Arias, C.A., Del Bubba, M., Brix, H., 2001. Phosphorus removal by sands for use as
media in subsurface flow constructed reed beds. Water Res. 35, 1159–1168.
 G.D. Gikas, V.A. Tsihrintzis / Ecological Engineering 44 (2012) 337–343
Börner, T., Felde, K., Gschlössl, T.E., Kunst, S., Wissing, F.W., 1998. Germany. In:
Vymazal, J., Brix, H., Cooper, P.E., Green, M.B., Haberl, R. (Eds.), Constructed
Wetlands for Wastewater Treatment in Europe. Bachhuys Publishers, Leiden,
The Netherlands, pp. 123–152.
Brix, H., Arias, C.A., 2005. The use of vertical flow constructed wetlands for on-
site treatment of domestic wastewater: New Danish guidelines. Ecol. Eng. 25,
491–500.
Brix, H., Arias, A.C., Johansen, N.H., 2003. Experiments in a two-stage con-
structed wetland system: nitrification capacity and effects of recycling
on nitrogen removal. In: Vymazal, J. (Ed.), Wetlands-nutrients, Met-
als and Mass Cycling. Backhuys Publishers, Leiden, The Netherlands,
pp. 169–177.
Gikas, G.D., Tsihrintzis, V.A., 2010. On-site treatment of domestic wastewater using a
small-scale horizontal subsurface flow constructed wetland. Water Sci. Technol.
62 (3), 603–614.
Gikas, G.D., Akratos, C.S., Tsihrintzis, V.A., 2007. Performance monitoring of a vertical
flow constructed wetland treating municipal wastewater. Glob. NEST J. 9 (3),
277–285.
Gikas, G.D., Tsihrintzis, V.A., Akratos, C.S., 2011. Performance and modeling of a ver-
tical flow constructed wetland–maturation pond system. J. Environ. Sci. Health:
Part A 46 (7), 692–708.
Kadlec, R.H., Wallace, S.D., 2009. Treatment Wetlands, second ed. Taylor and Francis
Group, Boca Raton, USA. ISBN 978-1-56670-526-4.
Kotti, I.P., Gikas, G.D., Tsihrintzis, V.A., 2010. Effect of operational and design
parameters on removal efficiency of pilot-scale FWS constructed wetlands and
comparison with HSF systems. Ecol. Eng. 36, 862–875.
Papaevangelou, V.A., Gikas, G.D., Tsihrintzis, V.A., 2012. Evaluation of evapotranspi-
ration in small on-site HSF constructed wetlands. J. Environ. Sci. Health: Part A
47 (5), 766–785.
Prochaska, C.A., Zouboulis, A.I., 2006. Removal of phosphates by pilot-scale vertical-
flow constructed wetlands using a mixture of sand and dolomite as substrate.
Ecol. Eng. 26, 293–303.
343
Prochaska, C.A., Zouboulis, A.I., Eskridge, K.M., 2007. Performance of pilot-scale ver-
tical flow constructed wetlands, as affected by season, substrate, hydraulic load
and frequency of application of simulated urban sewage. Ecol. Eng. 31, 57–66.
Rousseau, D.P.L., Vanrolleghem, P.A., Pauw, N.D., 2004. Model-based design of hor-
izontal subsurface flow constructed treatment wetlands: a review. Water Res.
38, 1484–1493.
Stefanakis, A.I., Tsihrintzis, V.A., 2012a. Effect of loading, resting period, temper-
ature, porous media, vegetation and aeration on performance of vertical flow
constructed wetlands. Chem. Eng. J. 181–182, 416–430.
Stefanakis, A.I., Tsihrintzis, V.A., 2012b. Use of zeolite and bauxite as filter media
treating the effluent of vertical flow constructed wetlands. Micropor. Mesopor.
Mater. 155, 106–116.
Stefanakis, A.I., Akratos, C.S., Gikas, G.D., Tsihrintzis, V.A., 2009. Performance of nat-
ural zeolite (clinoptilolite) as filter medium providing an additional treatment
stage of the effluent of two pilot-scale horizontal subsurface flow constructed
wetlands. Micropor. Mesopor. Mater. 124 (1–3), 131–143.
Tsihrintzis, V.A., Gikas, G.D., 2010. Constructed wetlands for wastewater and acti-
vated sludge treatment in North Greece: a review. Water Sci. Technol. 61 (10),
2653–2672.
Tsihrintzis, V.A., Akratos, C.S., Gikas, G.D., Karamouzis, D., Angelakis, A.N., 2007. Per-
formance and cost comparison of a FWS and a VSF constructed wetland systems.
Environ. Technol. 28 (6), 621–628.
Vymazal, J., 2001. Constructed wetlands for wastewater treatment in the Czech
Republic. Water Sci. Technol. 44 (11–12), 369–374.
Vymazal, J., 2002. The use of sub-surface constructed wetlands for wastewater treat-
ment in the Czech Republic: 10 years experience. Ecol. Eng. 18, 633–646.
Vymazal, J., Brix, H., Cooper, P.F., Green, M.B., Haberl, R., 1998. Constructed Wetlands
for Wastewater Treatment in Europe. Backhuys Publishers, Leiden.
Zakova, Z., Zak, P., 2003. Small on-site constructed wetlands for wastewater
treatment in sensitive areas with scattered settlements. In: Vymazal, J. (Ed.),
Wetlands-nutrients, Metals and Mass Cycling. Backhuys Publishers, Leiden, The
Netherlands, pp. 169–177.