Ecological Engineering 75 (2015) 291302

Contents lists available at ScienceDirect

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

Temperature inuence on nitrogen removal in a hybrid constructed

wetland system in Northern Italy
Anna Mietto, Marco Politeo, Simone Breschigliaro, Maurizio Borin *
Department of Agronomy, Food, Natural Resources, Animals and EnvironmentDAFNAE, Agripolis, University of Padova, Viale DellUniversit 16, 35020
Legnaro, Padova, Italy

A R T I C L E I N F O

A B S T R A C T

Article history:
Received 7 March 2014
Received in revised form 16 September 2014
Accepted 9 November 2014
Available online xxx

The objective of this research was to investigate the efciency and seasonal performance of a full-scale
hybrid constructed wetland system (HCW) in reducing total nitrogen (TN), ammonia nitrogen (NH4-N)
and nitrate nitrogen (NO3-N). HCW with a total area of about 130 m2 and hydraulic load of 2 m3/day was
composed of three subsurface ow vertical systems (VF), working in parallel and one horizontal (HF)
connected in series. The system was loaded daily with synthetic wastewater having an average
concentration of TN of 250 mg/L (about 125 mg/L of NH4-N and 125 mg/L of NO3-N). Water samples were
collected and analyzed from May to July 2011 and from January 2012 to July 2012. Variations were
observed in nutrient removal performance related to temperature.
During the whole monitoring period median reduction efciency (RE) in the HCW was TN 95%, NH4-N
95% and NO3-N 93%, although three sub-periods characterized by different performances have been
observed. During the rst period (from May to July 2011) the RE was positive for the three nitrogen forms
considered, whereas from January to the end of March 2012 the RE was lower, particularly for TN and
NO3-N. From April 2012, when the temperature rose above 14.8
C, there was an increase in the
performance that reached the 2011 values.
Internal production of NO3-N was observed, mainly in the VF systems between January and March
2012. The median removals of mass pollutants per m2 of HCW per day were TN 3.1 g/m2/d, NH4-N 1.5 g/
m2/d, NO3-N 1.5 g/m2/d. Segmented regression analysis identied a breakpoint at 14.2
C for wastewater
temperature that caused variations in TN and NO3-N concentration reduction performances. According to
this approach the abatement was always positively correlated with temperature, but different regression
slopes were obtained below and above the breakpoint. In particular, with lower temperature the
abatement of NO3-N and TN increased by 1.7 and 2.0% per
C of temperature increase; with temperature
higher than 14.2
C the increase in abatement due to increased temperature was sharper, especially for
NO3-N.
2014 Elsevier B.V. All rights reserved.

Keywords:
Vertical ow (VF)
Horizontal ow (HF)
Temperature
VF operation mode
Phragmites australis
Canna indica

1. Introduction
Wetland technology emerged in the 1950s and the use of
controlled wetland environments for wastewater treatment has
since been developed. The major nitrogen removal mechanism is
achieved by biological processes that convert the organic and
ammonia nitrogen to nitrate in an aerobic environment (nitrication) and then reduce the nitrate to nitrogen gas in an anoxic
environment (denitrication) (Leverenz et al., 2010). Volatilization, absorption and plant uptake play a much less important role

* Corresponding author. Tel.: +39 0498272838; fax: +39 0498272839.

E-mail address: maurizio.borin@unipd.it (M. Borin).
http://dx.doi.org/10.1016/j.ecoleng.2014.11.027
0925-8574/ 2014 Elsevier B.V. All rights reserved.

in CWs (Kadlec and Wallace, 2009). The use of vertical-subsurface

ow constructed wetland (VF) systems became very popular in
Europe in the 1990s compared to the horizontal system (HF) due to
their enhanced ability to oxidize ammonia nitrogen (Stefanakis
and Tsihrintzis, 2012). Single-stage CWs cannot achieve high
removal of total nitrogen because of their inability to provide both
aerobic and anaerobic conditions at the same time.
The design of hybrid constructed wetland systems (HCW)
(combination of vertical and horizontal ow systems) has been
proposed to exploit the anoxic areas within the horizontal bed for
denitrication (Cooper et al., 1999; Kadlec and Wallace, 2009;
Molle et al., 2008). HCW systems have been used to treat domestic
or municipal sewage (Brix et al., 2003; Canga et al., 2011), and more
recently for many other types of wastewater including agro-

292

A. Mietto et al. / Ecological Engineering 75 (2015) 291302

industrial (Comino et al., 2011), agricultural (Borin et al., 2013;

Hunt and Poach, 2001; Kantawanichkul et al., 2003) and landll
leachate (Mhlum et al., 1999).
As reported in several previous studies (Akratos and Tsihrintzis,
2007; Kadlec and Wallace, 2009; Kotti et al., 2010; Kuschk et al.,
2003; Vymazal, 1999) temperature is one of the principal variables
that mainly inuences biological activity and so the seasonal
performances of constructed wetlands. Hill and Payton (1998)
reported that the efciency of treatment in a constructed wetland
decreases at low temperature primarily due to reduced biotic
activity. Kadlec and Reddy (2001) studied the temperature
dependence in surface ow wetlands on removal of contaminants.
They concluded that the performance of wetlands in treating
wastewater is seasonally cyclic and the biotic reactions are reduced
at temperatures lower than the optimal range (20 to 35
C). Kadlec
(2006) pointed out three reasons for the importance of water
temperature in treatment wetlands: (1) temperature modies the
rates of several key biological processes, (2) temperature is
sometimes a regulated water quality parameter, and (3) water
temperature is a prime determinant of evaporative water loss
processes. Several biogeochemical processes that regulate the
removal of nutrients in wetlands are affected by temperature, thus
inuencing the overall treatment efciency (Kadlec and Reddy,
2001).
The goals of this study are (1) to investigate the efciency of a
full-scale hybrid constructed wetland system (HCW) in reducing
total nitrogen (TN), ammonia nitrogen (NH4-N) and nitrate
nitrogen (NO3-N); (2) study the effects of temperature on nitrogen
forms abatement; (3) conrm or identify a wastewater temperature breakpoint that causes variation in nutrient removal performances; (4) study the effects on vertical ow cells performance in
relation to different vegetation, medium and operational mode.
2. Materials and methods
2.1. Hybrid constructed wetland conguration and characteristics
The HCW was located at a private pig farm in Carmignano di
Brenta, Padova, in Veneto Region, NE Italy, (E: 11419.58 ; N:
453745.16; 46 m a.s.l).
It was built in 2008 and designed to provide tertiary treatment
of 2 m3/d of pre-treated liquid fraction of pig slurry efuent. The
design guidelines provided by APAT were based on municipal
wastewater treatment wetlands (APAT, 2005).

The HCW system is composed of three vertical-subsurface ow

wetlands (VF1VF2VF3) in parallel with a total area of 21 m2,
followed by one horizontal-subsurface ow wetland (HF) connected in series (105 m2) (Fig. 1).
The entire system was designed for a hydraulic retention time
(HRT) of 7 days as minimum. Each VF unit was built in concrete
(length: 10 m; width: 1 m; depth: 0.7 m). Three different plastic
sheet liners were placed inside each cell to prevent leakage and
contact of wastewater with groundwater. The layers from bottom
to top were: nonwoven geotextile sheet with a basic weight of
400800 g/m2 and thickness 1 mm; interlayer EPDM geo-membrane; nonwoven geotextile sheet with a basic weight of 400
800 g/m2 and thickness 1 mm. The rst two cells were lled with
washed gravel: grain size 1020 mm (d10 = 8.5 mm; d60 = 9.7 mm)
with porosity of 40%. The rst one (VF1) was vegetated with Canna
indica L., the second (VF2) with Phragmites australis (Cav.) Trin. Ex
Steud (common reed). The third (VF3), planted as VF2 cell, was
lled with a 0.10 m deep gravel layer (grain size 1020 mm)
overlying a 0.10 m deep coarse sand (grain size 35 mm) and
zeolite (grain size 510 mm) transition layer and a 0.30 m deep
gravel drainage layer (1020 mm in size). The main components of
the zeolite were: chabasite 60%, K-feldspar 13%, phillipsite 5%, mica
5% and augite 2%. The synthetic wastewater was distributed evenly
over the surface of the VF beds by a pressurized PVC distribution
pipe 75 mm in diameter that ran along the VF wetland units. Three
lters with interchangeable cartridges were placed in series at the
inlet of VF system. The lters were used to remove particles larger
than 1 mm from the feeding tank efuent, and were installed to
minimize the accumulation of solids in the inuent distribution
pipe. At the inlet of each VF wetland units a water meter with ve
digit mechanical counter was attached at the distribution pipe to
measure the incoming wastewater quantity delivered to each cell.
A drainage pipe (diameter 75 mm and length 10 m) was located on
the bottom of each VF cell in order to facilitate efuent collection.
The drainage pipe was connected on one side to a 100 mm
diameter collection pipe that discharged the efuent from the bed
to a manhole that had a water level control structure equipped
with a siphon pipe where a timer-controlled pump was placed
(Fig. 2).
The siphon maintained water level at 0.30 m from the surface in
each VF cell. A 200 watt power submersible pump installed in the
same manhole was used to drain the porous media and transport
the leachate to separate sumps (OUT VF1, OUT VF2, OUT VF3). The
wastewater discharged from each VF sump was collected in a

Fig. 1. HCW system dimensions in overhead view. Sampling points: (1) inuent, (2) VF1 efuent, (3) VF2 efuent, (4) VF3 efuent, (5) inuent to HF, (6) HF efuent (nal
efuent).

A. Mietto et al. / Ecological Engineering 75 (2015) 291302

Fig. 2. Schematic representation (not to scale) of water level control structure of

each VF cell.

common sump (OUT VF) (length: 1.2 m; width: 0.8 m; depth:

0.66 m). The VF efuent was then transferred with a submersible
sump pump with an integrated oat switch to the horizontal
subsurface ow wetland. A water meter with ve digit mechanical
counter attached to the submersible pump outlet measured the HF
incoming wastewater volume.
The horizontal-subsurface ow wetland (HF) was a basin 25 m
long, 4 m wide and depth 0.7 m with a bottom slope of 1%. The
bottom and sides of the basin were waterproofed with the same
plastic sheets used for VF cells. The inlet and the outlet sections
were lled using two strips of coarse-rock material (grain size 50
100 mm) along two opposite edges of the basin, with washed
gravel: grain size 1020 mm (d10 = 8.5 mm; d60 = 9.7 mm) with
porosity of 40% placed in the central area. At the HF inlet a
distributor pipe was buried immediately below the surface
(diameter 100 mm), placed horizontally and perpendicular to
the direction of ow. At the outlet, a similar collector pipe was
buried at the bottom. The HF efuent volumes were measured by a
water meter, collected in an interred tank and recycled for cleaning
the piggery. The wetland was planted in April 2008 with
Phragmites australis (Cav.) Trin. Ex Steud. (common reed).
2.2. System mode operation
From May 2011 the HCW system was loaded daily with
synthetic wastewater. The wastewater was prepared daily in the
feeding tanks just before the feeding by dissolving ammonium

Table 1
Daily operation modes of the vertical-ow system. DOM1: from May to July
2011 and from January to 11th July 2012; DOM2: from 11th to 25th July 2012.
DOM 1
Phase

VF unit cycle

loading cycle 2

II

wet period

III

unloading
cycle
dry period

IV

Duration
(h)

14

DOM 2
Schedule

Duration
(h)

Schedule

10 a.m.12 a. 2
m.
12 a.m.6 p. 14
m.
6 p.m.8 p.m. 2

10 a.m.12 a.
m.
12 a.m.2 a.
m.
2 a.m.4 a.m.

8 p.m.10 a.
m.

4 a.m.10 a.
m.

293

nitrate fertilizer 26% N (13% nitrate and 13% ammonia) in 1.7 m3 of

fresh water. The average concentration was 250.3 (3.7) mg/L of
TN, 124.5 (2.5) mg/L of NO3-N and 124.9 (3.2) mg/L of NH4-N.
The feeding tank (pump chamber) consisted of a 5 m  0.7 m
concrete tank 0.7 m deep, with a submersible water pump inside
used to load the wastewater to each VF unit. The pump was
controlled by two programmable timers in series. One timer
dictated the time of loading cycles and was set to work once per
day, with one loading cycle in the morning from 10 am to 12 am.
The other timer dictated the loading time per cycle (one minute)
and the time between pumping events within each loading cycle
(every 4 min). Daily wastewater ow rate of 1.7 m3 was evenly
distributed to all three VF cells with average ow rate of 565 L/day
per cell.
Two programmable timers in series controlled the submersible
water pump placed inside the water level control structure of each
VF. The rst timer dictated the time of discharging cycles and was
set to work once per day, the other timer dictated the discharging
time per cycle (one minute) and the time between pumping events
within each discharging cycle (every 4 min).
To provide sufcient oxygen transfer for nitrication, load and
discharge cycles of VF unit were set with anoxic/oxic (A/O) steps.
The A/O stages were generated with rapid water ow through the
lter media, the phenomenon called passive air-pump (Green
et al., 1998). During the investigation period two different daily
operation modes (DOM) of the vertical-ow system were tested
(Table 1). We chosen this loading scheme to manage the VF cells
with alternation of period of saturation and unsaturation as
describe below. The rst (DOM 1) was applied from May to July
2011 and from January to 11th July 2012 with a feeding strategy
consisting of 2 h of wastewater inow, followed by 6 h of
completely full cell, 2 h of discharging and 14 h of completely
empty cell in order to assist the oxidation. This intermittent
feeding mode was chosen to provide good oxygen transfer to the
water phase. DOM 2 was applied from 11th to 25th July 2012,
programming 14 h of fully lled cells and 6 h of empty cells.
2.3. Sampling, chemical analysis and data elaboration
Water samples were collected and analyzed from May to July
2011 and from January to July 2012 from inow (1; IN), outow of
VF1 (2; OUT VF1), outow of VF2 (3; OUT VF2), outow of VF3 (4;
OUT VF3), inow of HF (5; IN HF), outow of HF and nal efuent
(6; OUT HF) (Fig. 1). IN sample was taken at the beginning of each
cycle, after synthetic feed preparation, OUT VF of each cell and IN
HF were taken at the end of the daily feeding period of the cycle
(after 6 h from inlet), and OUT HF sample was taken according to
the retention time (after 7 days from the start of the cycle).
Monitoring consisted of thirty seven weekly samplings during the
whole investigation period, collecting 211 samples.
In situ, measurements of pH, electrical conductivity (EC),
dissolved oxygen (DO), wastewater temperatures (T) and redoxpotential (Eh), were taken with a Hach Lange HQD 40d multiparameter meter with interchangeable probes according to
standard methods (APHA, 1998). Before testing, each probe was
carefully calibrated according to the manufacturer's procedures.
Samples were collected, preserved at 4
C and then analyzed
within a short time.
Total nitrogen (TN), ammonia-nitrogen (NH4-N) and nitricnitrogen (NO3-N), were determined photometrically using a HachLange DR-2800 spectrophotometer and adequate cuvette test kits
(cuvette-tests LCK 338, 302, 340), (Hach-Lange, 1989), according to
DIN (1985). Adequate sample dilutions were made with a stock
supply of deionized water.
Air temperature, humidity, solar radiation, rainfall volumes,
wind speed and direction were recorded every day from the

294

A. Mietto et al. / Ecological Engineering 75 (2015) 291302

beginning of the experiment by a meteorological station installed

at the site. Evapotranspiration (ETO) was calculated with the FAO
56 Penmam-Monteith equation (Allen et al., 1998) for short
reference crop ETos (Allen et al., 2005).
The data series of the parameters did not follow normal
distribution even after transformations. Thus, statistical analyses
were carried out with the KruskalWallis nonparametric test to
compare the six sampling positions (IN, OUT VF1, OUT VF2, OUT
VF3, IN HF, OUT HF) and Box and Whiskers were used to display the
variability.
In this study, the performance comparison of each HCW unit
was based on three different approaches:
(a) concentration percentage abatement (A), calculated on the
rst quartile (Q1), second quartile (Q2; median) and third quartile
(Q3) concentration values as (Eq. (1)):


Cin  Cout
A%
(1)
 100
Cin
where Cin is inow concentration (mg/L) and Cout is outow
concentration (mg/L);
(b) Reduction efciency (RE) calculated as (Eq. (2)):


Cin  Vin  Cout  Vout
RE%
(2)
 100
Cin  Vin
where Cin is inow concentration (mg/L), Vin is average inow
volume of synthetic wastewater applied (m3/d) with daily rainfall
volume (mm/d) included, Cout is outow concentration (mg/L),
Vout is outow volume detected at the outlet of the unit (m3/d);
(c) Areal load reduction (ALR), expressing the removed
pollutants mass per m2 of CW and time (g/m2/d). ALR is a useful
parameter to assess system efciency (Stefanakis and Tsihrintzis,
2012).
In addition, segmented linear regression analysis (or brokenline regression) with a non-parametric approach developed by
Pettitt (1979) was used to identify a change-point of wastewater
temperature that caused variation in nutrient percentage abatement. For this purpose, the Flat Steps method (Bai and Perron,
2003) was used, implemented in the Strucchange library of R
software (Zeileis et al., 2003). Partial F-test in one-way analysis of
variance was used to determine any signicant differences at
p < 0.05.

3. Results and discussion

3.1. Wastewater, air temperature and evapotranspiration
During the rst monitoring period (MayJuly 2011), slight
differences in air and wastewater temperatures were found at
different HCW sampling points (Fig. 3). Inlet temperature values
ranged between 20.5 and 22.3
C. These values result as being more
efcient for biological N removal (range of 2025
C) as reported by
Sutton et al. (1975) since ambient temperature positively
inuences microbial activity and diffusion rates (Phipps and
Crumpton, 1994).
During the second period (JanuaryJuly 2012) air temperature
vs inlet wastewater showed two different trends: in cold months
(JanuaryMarch) air temperature (average 6.3
C) was lower than
the inlet wastewater (average 11.9
C) whereas in the warm period
(from April to July) an opposite tendency was observed, especially
in JuneJuly. The freshwater temperature used to prepare synthetic
wastewater (average 14.2
C) showed less variation than the other
sampling points during the second part of the monitoring period,
ranged between 9.2 and 17.7
C. Until the beginning of March
2012 wastewater temperature measured from VF unit (average
8.1
C) was lower than inlet (average 11.1
C), whereas higher
values (average 19.8
C) were observed in VF with respect to the
inlet (average 15.3
C) till the end of the experimental period. HF
temperature followed the same tendency observed for VF
wastewater (Fig. 3).
The daily ETO results were in accordance with the data found in
the literature for similar conditions (size, latitude and measurement method) (Borin et al., 2011). The time pattern of the
cumulative ETos in the rst period, from May to September 2011,
showed an average daily ETO of 5.90 mm; from October 2011 to
March 2012, plants consumed 1.97 mm/day on average. In the
winter-spring period, daily ETO was 5.04 mm.
3.2. On site parameters
During rst monitoring period (MayJuly 2011), electrical
conductivity (EC) of the inuent wastewater was higher than
1600 mS/cm and increased after passage through VF cells. HF
slightly increased this to a median value of 1840 mS/cm (Fig. 4).

Fig. 3. Mean air and wastewater temperature at the sampling points of the HCW system during the monitored period.

A. Mietto et al. / Ecological Engineering 75 (2015) 291302

Fig. 4. Variation of EC, pH, DO, Eh at the sampling points of the HCW system during the monitored period.

295

296

A. Mietto et al. / Ecological Engineering 75 (2015) 291302

Fig. 5. Nitrogen forms concentration at the sampling points of the HCW system during the monitoring period.

A. Mietto et al. / Ecological Engineering 75 (2015) 291302

297

Fig. 6. Box-plot diagrams of nitrogen forms concentration (mg/L), at the sampling points of the HCW system during the monitoring period. Different letters indicate
signicant differences at p = 0.05 by KruskalWallis test.

Table 2
Slope and R2 for the linear relationships among wastewater temperature and
nitrogen forms abatements and for values below and above breakpoint.
Nitrogen form Linear
regression

TN
NH4-N
NO3-N

Values below
breakpoint

Values above
breakpoint

Slope

R2

Slope

R2

Slope

R2

3.63
0.62
6.5

0.88
0.66
0.83

2.01
0.36
1.7

0.38
0.4
0.1

2.9
0.44
6.6

0.7
0.23
0.62

During the second monitoring period (JanuaryJuly 2012), the

incoming EC did not exceed 1600 mS/cm until 28th March, then
rose to higher values until the end of monitoring: this is probably
related to the increasing water temperature that promoted
dissolving ammonium nitrate in water. The same trends as the
previous monitoring period were observed for VF and HF, mainly
due to substrate biolm interaction that may result in soluble salt
release from the media to the water (Stefanakis and Tsihrintzis,
2012). EC seasonal variations were observed during summer each
year probably due to increased evapotranspiration (MaySeptember 2011 and AprilJuly 2012) and plant growth, as reported by
Hench et al. (2003).
The median inuent pH during the rst and second monitoring
period was neutral or slightly alkaline 7.64  0.04. HF unit
exhibited alkaline pH values from 21st March 2012 till the end
of the monitoring period (Fig. 4). As nitrication proceeds
optimally at pH between 7.5 and 8.5 (Platzer, 1996), the pH values
were optimal in all three VF beds. However pH values measured in
the HF unit were not optimal for denitriers that operate best in
the range between 6.5 and 7.5 (Paul and Clark, 1996). Furthermore
ammonia nitrogen loss through volatilization was negligible since
it generally requires a pH of 9.3 (Jing and Lin, 2004).

Median dissolved oxygen concentration value at the inlet was

constant (4.17  0.05 mg/L) during the whole investigation period.
DO concentration above 1.5 mg/L is essential for nitrication to
occur (Ye and Li, 2009). During rst monitoring period DO
concentration increased after passage through VF cells to the
median value of 4.57 mg/L probably due to aerobic conditions
provided by the oxic stage during DOM 1 of the vertical-ow
system. The DO detected at the efuent of HF cell varied from
0.14 to 0.28 mg/L. The same trends were observed during the
second monitoring period in VF unit: from the beginning of March
to 11th July 2012 a slight increase appeared in efuent DO
concentration, possibly due to plant growth and enhanced oxygen
transfer to the plant roots. From 11th July 2012 to the end of
monitoring average DO concentration from VF unit efuent
decreased from 4.18 to 2.74 mg/L due to the longer anoxic stage
promoted by DOM 2. HF efuent DO concentration was very low
during end winter-early spring, from May to 4th July 2012 it
slightly rose with the common reed re-activation (Fig. 4).
Redox potential (Eh) variations within a range of a few hundred
mV from reduced (0 mV) to moderately oxidized redox conditions
(+270 mV) were observed (Fig. 4). During the rst period, median
wastewater redox potential value of 269 decreased to 206 mV after
VF cells passage. This suggests that the VF system provided
conditions suitable for nitrication, but not for denitrication. The
HF outow Eh decreased drastically in the range from +10 to 0 mV
probably inducing favorable (anaerobic) conditions suitable for
microbial denitrication. The second period followed the same
trend as the rst. From 11th July 2012 to the end of monitoring Eh
from VF unit efuent decreased drastically, possibly due to the
prolonged period of saturation of the cell promoted by DOM 2, as
demonstrated by DO values. In the HF cell Eh patterns differed
between the warm and cold periods probably affected by
wastewater temperature. Increase of temperature during the
warm period accelerated biochemical processes including bacteria
activity (Kadlec and Reddy, 2001). Moreover, oxygen solubility in

Fig. 7. Wastewater temperature and percentage abatement correlation charts with breakpoints for nitrogen forms.

298

A. Mietto et al. / Ecological Engineering 75 (2015) 291302

Table 3
Nitrogen pollutant percentage abatement (A) statistics for wastewater temperatures above and below identied breakpoint.
Temperature below break point
VF1

VF2

VF3

Temperature above break point

HF

HCW

VF1

VF2

VF3

HF

HCW

TN (A%)
Minimum
First quartile (Q1)
Median (Q2)
Third quartile (Q3)
Maximum

3.5
1.2
5.3
12.4
17.7

6.1
3.6
3.8
8
26

4.4
3.8
7.1
10.8
35

8.9
8.7
4.9
2.3
13.1

5.3
0.5
5.1
8.1
41.8

16.2
18.8
21
24.3
27.1

7.8
29.5
34.4
36.4
42.1

22.9
37.8
42.6
46.7
50.4

7.4
14.7
19.9
29.2
39

22.7
53.8
57.5
60.9
63.7

NO3-N (A%)
Minimum
First quartile (Q1)
Median (Q2)
Third quartile (Q3)
Maximum

54.4
50.6
47.8
44.3
3.2

49.8
46.7
41.8
39.9
3.3

61.6
51.2
47.1
39.9
2.6

2.7
2.1
0.2
1.2
2.9

55.7
51.3
47.3
45.6
1.8

28.5
10.4
13.4
14.4
22.1

22.3
16.7
26
28.7
33.8

20.7
18.2
26.5
30.4
46.2

1.8
14.6
25.3
30.6
52

27
35.6
45.4
59.3
68.4

water decreases with increasing temperature, so as a result, Eh

decreased, as also reported by Duek et al. (2008) (Fig. 4).

particular, the activity of denitrifying bacteria in CWs is generally

more robust in springsummer seasons than in autumnwinter
(van Oostrom and Russel, 1994).

3.3. Nitrogen forms concentrations

3.4. Effect of wastewater temperature
From May to July 2011, TN concentration decrease was clear and
constant accomplished by VF unit with median value of 173.4 mg/L
and HF unit with 105.6 mg/L. With low temperatures measured
during the rst part of the second period (JanuaryMarch 2012,
average temperature 10.4
C), TN concentration abatement was
negligible. Then, from the end of March to the rst ten days of May
following an increase in temperature from 15.5 to 20.4
C the
concentrations at the outlets of VF and HF progressively decreased.
Finally, when mean wastewater temperature of 20.2
C was
reached, TN outlet concentrations showed similar values as the
year before in the same period (Fig. 5).
The VF cells provided outlet values statistically lower than the
inlet ones; differences were also found among cells, with lower
values in VF3 with respect to VF1, probably related to the combine
effect of different plant and porous media types. The performance
of VF2 and VF3 units, both planted with P. australis, were similar
and the statistical analysis showed no signicant differences. The
HF system lowered the inlet median value from VF unit from
146 mg/L to a nal discharge concentration of 114 mg/L and slightly
increased the variability (Fig. 6).
Ammonia median concentration measured from VF and HF unit
showed a decreasing trend in both monitored periods. From May to
July 2011 median NH4-N concentration at VF outlet was 68.3 and
48.6 mg/L in HF. From January to July 2012 it was 68.4 and 50 mg/L,
respectively. The different wastewater temperature did not
inuence ammonia values at the outlet of wetland cells. No
statistically signicant differences were found between VF1 and
VF2 even though different plant species were used. Signicant
differences were found between VF3 and the rst two VF cells
(VF1 and VF2) probably due to composition of the medium layer
which comprised zeolite stones that could absorb NH4+ (Nguyen
and Tanner, 1998) (Fig. 6).
During the rst period, NO3-N trend concentrations followed
the same temporal pattern as the other nitrogen forms, with
decreasing values passing through VF and then HF. During the
second period, in the cold season efuent concentrations from VF
and HF unit exceeded inuent with median values of 179.7 and
182.7 mg/L respectively. With increasing wastewater temperature
during the warm season, measured concentration decreased
reaching values found in the rst period (Fig. 5). Statistical
differences were found only between inlet and outlet of HCW
system (Fig. 6). As reported by Kuschk et al. (2003), the microbial
activities related to nitrication and denitrication can decrease
considerably at water temperature below 15 or above 30
C. In

In our study, correlation analysis was performed between

percentage abatement (A%) for nitrogen forms calculated on a
weekly basis for all 37 weeks of monitoring and the corresponding
average wastewater temperatures during the whole monitoring
period. We focus on the results obtained for the whole HCW.
First, linear regressions between A% and temperature were
calculated for the nitrogen forms, obtaining signicant relationships, with higher R2 for TN and NO3-N (Table 2). The regression
slope for NO3-N suggested a sharp inuence of temperature on
abatement, which increased by 6.5% with every 1
C. On the
contrary, the effect of temperature on NH4-N was less marked, as
suggested by the lower R2 and regression slope.
In a second step a segmented regression analysis was conducted
according to Bai and Perron (2003) to identify any presence of a
breakpoint of wastewater temperature that caused variations in
nutrient abatement performances. The segmented analysis model
was signicant with respect to the linear regression for TN and
NO3-N and evidenced breakpoints at 14.2
C for both nitrogen
forms (Fig. 7).
According to this approach the abatement was always
positively correlated with temperature, but different regression
slopes were obtained below and above the breakpoint. In
particular, with lower temperature the abatement of NO3-N and
TN increased by 1.7 and 2.0% per
C of temperature increase; with
temperature higher than 14.2
C the increase in abatement due to
increased temperature was sharper, especially for NO3-N (6.6%
every 1
C) (Table 2).
Nitrogen pollutants abatement statistics were calculated for
wastewater temperatures below and above the breakpoint
identied with segmented regression analysis (Table 3). As
previously mentioned, higher performance was observed at
temperature above breakpoint for both TN as NO3-N. In particular,
NO3-N median abatement that was negative below 14.2
C, became
consistent above this critical temperature.
3.5. Nitrogen RE and ALR
Median TN RE during the rst period (from May to July 2011)
was 63%, 83% and 96% for VF, HF unit and SIF respectively (Fig. 8).
During the second period from January to the end of March
2012 the RE was lower, with values between 11% and 40% for VF
unit and the entire system. Considering the same period negative
reduction values were found for HF unit (ranging from 0.7 and

A. Mietto et al. / Ecological Engineering 75 (2015) 291302

Fig. 8. Nitrogen forms RE (%) at the sampling points of the HCW system during the monitoring period.

299

300

A. Mietto et al. / Ecological Engineering 75 (2015) 291302

Fig. 9. Nitrogen forms ALR of the HCW during the monitored period.

6%). From April 2012, when the temperature rose above14.8
C,

there was an increase of TN reduction in the system that reached
the 2011 values (Fig. 8).

VF and HF system provided a NH4-N reduction pattern similar to

that found for TN during 2011, with values ranging from 57% to 94%
in the entire system during the whole monitoring period (Fig. 8).

A. Mietto et al. / Ecological Engineering 75 (2015) 291302

Table 4
Mean values of nitrogen pollutant reduction efciency (RE) and areal load reduction
(ALR) measured during the two last two weeks of DOM1 (from 27th June to 11th July
2012) and DOM 2 (from 11th to 25th July 2012).
DOM2

DOM1
VF1

VF2

VF3

VF1

VF2

VF3

RE (%)
TN
NH4-N
NO3-N

63
73.7
57.1

70.3
76.3
65.4

70.5
77.3
58.8

64.1
68.6
59.4

68.3
72.5
63.7

65.5
73.1
58.4

ALR (%)
TN
NH4-N
NO3-N

13.3
7.7
6

15.3
8.2
7.1

12.7
6.9
5.3

13.4
7.2
6.2

14.7
7.8
6.8

12.1
6.7
5.4

VF reduction was slightly higher during the warm period of 2012

(median value 71%) than cold period of the same year (median
value 43%). The HF system performed better during the last
monitoring period, with median reduction ranging between 35
and 87%.
Median NO3-N reduction efciency during 2011 was higher for
HF (85%) than VF (57%) (Fig. 8). During the cold period of
2012 negative reduction values were observed for VF unit and in
the SIF. The negative RE value (30%) rose from the end of March
until reaching, at the end of the investigation period, a value higher
than 60% in VF.
The median removals of mass pollutants per m2 of HCW per day
were TN 3.1 g/m2/d, NH4-N 1.5 g/m2/d and NO3-N 1.5 g/m2/d
(Fig. 9). ALR removals of nitrogen forms showed a different trend
during the rst monitoring period of 2012. In particular, from
January to March 2012 the data reveal a low treatment capability,
probably related to temperature (Stefanakis and Tsihrintzis, 2012).
From April to the end of the monitoring period an increment was
observed: from the initially low performance, it reached the same
results observed during 2011 (Fig. 9). The combination of vertical
and horizontal stages ensures the right conditions to obtain a
higher nitrogen removal performance. Moreover, when considering a year of operation, we can suppose that the HCW would
remove approximately 1.1 Kg TN/m2.

301

4. Conclusions
The system showed higher performance in terms of TN, NO3-N
and NH4-N reduction. Nonetheless, seasonal variations appear to
affect the HCW system performance. Lower efuent concentrations were observed during the warm period (higher temperature),
especially for TN and NO3-N, whereas the performances ameliorated with the increase in wastewater temperature. NH4-N
reduction efciency was inuenced by seasonality to a lesser
extent in the VF than the HF.
Higher RE and ALR for TN and NH4-N were obtained with the
rst daily mode of operation (DOM 1) probably related to the dry
period of 14 h that seemed to promote the medium layer oxidation
and nitrication process, whilst similar ndings were obtained in
DOM 2 for NO3-N reduction efciency. VF systems planted with P.
australis (VF2 and VF3) showed slightly higher performance in TN
reduction efciency compared to VF1, planted with C. indica.
Segmented regression analysis identied a breakpoint at 14.2
C
for wastewater temperature that caused variations in TN and NO3-N
concentration reduction performances. According to this analysis
the abatement of NO3-N and TN increased by 1.7 and 2.0% per
C
when temperature was below breakpoint; with a temperature
higher than 14.2
C the increase of abatement due to increased
temperature was sharper, especially for NO3-N (6.6% every 1
C)
In the VF cells the mode of operation of loading/discharging
cycles induced a sharp variation in DO and Eh values; both
decreased passing from 6 h of completely full cells and 14 h of
empty cells to 14 h of full cells and 6 h of empty cells. As a
consequence, the NH4-N concentration at the outlet increased and
the reduction efciency decreased. Addition of zeolite to the
porous medium reduced the NH4-N concentration and P. australis
gave better results than C. indica.
It is important to highlight that during this study synthetic
wastewater was used and that temperature in the inuent was
affected by the environmental climate conditions. Slurry temperature is usually quite constant during the year, at around 1517
C
(Politeo, 2013) and this can explain the lower performance
observed during winter compared to summer season. This has
to be taken into account when considering the effective potential
of using wetland hybrid systems to treat slurry efuents as a higher
performance can be expected.

3.6. Comparison among vertical systems

Acknowledgment

Nitrogen pollutant reduction efciency (RE) and areal load

reduction (ALR) of VF system measured during DOM1 and
DOM2 are presented in Table 4. The performance comparison
was made between two loading cycles of each DOM with the same
conditions (air and wastewater temperatures). In particular, we
considered the subset data from the last two weeks of DOM1 (from
27th June to 11th July 2012) and the data of the two weeks of DOM
2 (from 11th to 25th July 2012). DOM 1 allowed higher RE and ALR
to be obtained, probably due to the longer empty period that
promoted better oxidative conditions. Lower average RE values for
TN (varying between 64.1 to 68.3%) and NH4-N (varying between
68.6 and 73.1%) were found with DOM 2. However RE values for
NO3-N were quite similar to those observed during DOM 1.
VF systems planted with P. australis (VF2 and VF3) showed
better performance in TN reduction efciency (70.3 and 70.5%,
respectively) than VF1 (63%), planted with C. indica. Average NH4-N
reduction efciency was 76.3 and 77.3% in VF2 and
VF3 respectively with higher ALR of 8.2 g/m2/d in VF2. NO3-N
values were lower than other nitrogen forms, best RE and ALR
performances were measured in VF2 with 65.4% and 7.1 g/m2/d.

Research supported by Progetto AGER, grant no. 2010-2220. We

thank Luigi Guarnieri, the owner of the pig farm and the hybrid
constructed wetland, for its kind hospitality.
References
APAT, 2005. Linee guida per la progettazione e gestione di zone umide articiali per
la depurazione dei reui civili. Agenzia per la protezione dell'ambiente e per i
servizi tecnici, Firenze, Italy.
American Public Health Association (APHA), 1998. Standard Methods for the
Examination of Water and Wastewater, 20th ed. American Water Works
Association (AWWA), and Water Environment Federation (WPCF), Washington,
DC.
Allen, R.G., Pereira, L.S., Raes, D., Smith, M., 1998. Crop Evapotranspiration:
Guidelines for Computing Crop RequirementsFAO Irrigation and Drainage
Paper 56. FAO, Rome, Italy.
Allen, R.G., Walter, I.A., Elliott, R.L., Howell, T.A., Itensu, D., Jensen, M.E., Snyder, R.L.,
2005. The ASCE standardized reference evapotranspiration equation. Am. Soc.
Civil Eng. Reston, VA 59 (with supplemental appendices).
Akratos, C.S., Tsihrintzis, V.A., 2007. Effect of temperature, HRT, vegetation and
porous media on removal efciency of pilot-scale horizontal subsurface ow
constructed wetlands. Ecol. Eng. 29 (2), 173191.
Bai, J., Perron, P., 2003. Computation and analysis of multiple structural change
models. J. Appl. Econ. 18, 122.

302

A. Mietto et al. / Ecological Engineering 75 (2015) 291302

Borin, M., Milani, M., Salvato, M., Toscano, A., 2011. Evaluation of Phragmites australis
(Cav:) Trin. evapotranspiration in Northern and Southern Italy. Ecol. Eng. 37,
721728.
Borin, M., Politeo, M., De Stefani, G., 2013. Performance of a hybrid constructed
wetland treating piggery wastewater. Ecol. Eng. 51, 229236.
Brix, H., Arias, C.A., Johansen, N.H., 2003. Experiments in a two-stage constructed
wetland system: nitrication capacity and effects of recycling on nitrogen
removal. In: Vymazal, J. (Ed.), Wetlands Nutrients, Metals and Mass Cycling.
Backhuys Publishers, Leiden, The Netherlands, pp. 237258.
Canga, E., Dal Santo, S., Pressl, A., Borin, M., Langergraber, G., 2011. Comparison of
nitrogen elimination rates of different constructed wetland designs. Water Sci.
Technol. 64 (5), 11221129.
Comino, E., Riggio, V., Rosso, M., 2011. Mountain cheese factory wastewater
treatment with the use of a hybrid constructed wetland. Ecol. Eng. 37, 1673
1680.
Cooper, P., Grifn, P., Humphries, S., Pound, A., 1999. Design of a hybrid reed bed
system to achieve complete nitrication and denitrication of domestic
sewage. Water Sci. Technol. 40 (3), 283289.
DIN (Deutsches Institut fr Normung), 1985. German Standard Methods for the
Examination of Water, Waste Water and Sludge. Deutches Institut fr Normung,
Berlin.

Duek, J., Picek, T., Ckov,
H., 2008. Redox potential dynamics in a horizontal
subsurface ow constructed wetland for wastewater treatment: diel, seasonal
and spatial uctuations. Ecol. Eng. 34 (3), 223232.
Green, M., Friedler, E., Safrai, I., 1998. Enhancing nitrication in vertical ow
constructed wetland utilizing a passive air pump. Water Res. 32 (12), 3513
3520.
Hach-Lange, 1989. Water Analysis Handbook. HACH, Company, Loveland, CO, USA.
Hench, K.R., Bissonnette, G.K., Sexstone, A.J., Coleman, J.G., Garbutt k, Skousen, J.G.,
2003. Fate of physical, chemical and microbial contaminants in domestic
wastewater following treatment by small constructed wetland. Water Res. 37
(4), 921927.
Hill, D.T., Payton, J.D., 1998. Inuence of temperature on treatment efciency of
constructed wetlands. Trans. ASAE 41 (2), 393396.
Hunt, P.G., Poach, M.E., 2001. State of the art for animal wastewater treatment in
constructed wetlands. Water Sci. Technol. 44 (1112), 1925.
Jing, S.R., Lin, Y.F., 2004. Seasonal effects on ammonia nitrogen removal by
constructed wetlands treating polluted river water in southern Taiwan. Environ.
Pollut. 127 (2), 291301.
Kadlec, R.H., 2006. Water temperature and evapotranspiration in surface ow
wetlands in hot arid climate. Ecol. Eng. 26 (4), 328340.
Kadlec, R.H., Reddy, K.R., 2001. Temperature effects in treatment wetlands. Water
Environ. Res. 73 (5), 543557.
Kadlec, R.H., Wallace, S.D., 2009. Treatment Wetlands, second ed. CRC Press/Lewis
Publishers, Boca Raton, Florida, pp. 1016.
Kantawanichkul, S., Somprasert, S., Aekasin, U., Shutes, R.B.E., 2003. Treatment of
agricultural wastewater in two experimental combined constructed wetland
systems in a tropical climate. Water Sci. Technol. 48 (5), 199205.

Kotti, I.P., Gikas, G.D., Tsihrintzis, V.A., 2010. Effect of operational and design
parameters on removal efciency of pilot-scale FWS constructed wetlands and
comparison with HSF systems. Ecol. Eng. 36 (7), 862875.
Kuschk, P., Wiebner, A., Kappelmeyer, U., Weibbrodt, E., Kstner, M., Stottmeister, U.,
2003. Annual cycle of nitrogen removal by a pilot-scale subsurface horizontal
ow in a constructed wetland under moderate climate. Water Res. 37 (17),
42364242.
Leverenz, H.L., Haunschild, K., Hopes, G., Tchobanoglous, G., Darby, J.L., 2010. Anoxic
treatment wetlands for denitrication. Ecol. Eng. 36 (11),
15441551.
Mhlum, T., Warner, W.S., Stlnacke, P., Jenssen, P.D., 1999. Leachate treatment in
extended aeration lagoons and constructed wetlands in Norway. In:
Mulamoottil, G., McBean, E.A., Revers, F. (Eds.), Constructed Wetlands
References 499 for the Treatment of Landll Leachates. Lewis Publisher/CRC
Press, Boca Raton, pp. 151163.
Molle, P., Prost-Boucle, S., Lienard, A., 2008. Potential for total nitrogen removal by
combining vertical ow and horizontal ow constructed wetlands: a full-scale
experiment study. Ecol. Eng. 34 (1), 2329.
Nguyen, M.L., Tanner, C.C., 1998. Ammonium removal from wastewaters using
natural New Zealand zeolites. N. Z. J. Agric. Res. 41 (3), 427446.
Paul, E.A., Clark, F.E., 1996. Soil Microbiology and Biochemistry, second ed. Academic
Press, San Diego, CA, pp. 340.
Pettitt, A.N., 1979. A non-parametric approach to the change-point problem. Appl.
Stat. 28 (2), 126135.
Phipps, R.G., Crumpton, W.G., 1994. Factors affecting nitrogen loss in experimental
wetlands with different hydrologic loads. Ecol. Eng. 3 (4), 399408.
Platzer, C., 1996. Enhanced nitrogen elimination in subsurface ow articial
wetlands - a multi stage concept. Proceedings of the Fifth International
Conference on Wetland Systems for Water Pollution Control, Vienna, Austria
September 1519, I/71I/79.
Politeo, M., 2013. Performance of hybrid constructed wetland for piggery
wastewater treatment. PhD Thesis. University of Padova, pp. 1134.
Stefanakis, A.I., Tsihrintzis, V.A., 2012. Effects of loading, resting period,
temperature, porous media, vegetation and aeration on performance of pilotscale vertical ow constructed wetlands. Chem. Eng. J. 181182 416430.
Sutton, P.M., Murphy, K.L., Dawson, R.N., 1975. Low temperature biological
denitrication of wastewater. J. Water Pollut. Control Fed. 47 (1), 122134.
van Oostrom, A.J., Russel, J.M., 1994. Denitrication in constructed wastewater
wetlands receiving high concentrations of nitrate. Water Sci. Technol. 29 (4), 714.
Vymazal, J., 1999. Nitrogen removal in constructed wetlands with horizontal subsurface ow can we determine the key process? In: Vymazal, J. (Ed.), Nutrient
Cycling and Retention in Natural and Constructed Wetlands. Backhuys
Publishers, Leiden, pp. 117.
Ye, F.X., Li, Y., 2009. Enhancement of nitrogen removal in towery hybrid constructed
wetland to treat domestic wastewater for small rural communities. Ecol. Eng. 35
(7), 10431050.
Zeileis, A., Kleiber, C., Krmer, W., Hornik, K., 2003. Testing and dating of structural
changes in practice. Comput. Stat. Data Anal. 44 (12), 109123.