Bioresource Technology 94 (2004) 185–192

A test of four plant species to reduce total nitrogen and
total phosphorus from soil leachate in subsurface
wetland microcosms
Lauchlan H. Fraser a,*, Spring M. Carty a, David Steer b
a
b

Department of Biology, The University of Akron, Akron, OH 44325-3908, USA
Department of Geology, The University of Akron, Akron, OH 44325-4101, USA

Received 12 September 2003; received in revised form 24 November 2003; accepted 24 November 2003
Available online 12 February 2004

Abstract
Four wetland plant species (Scirpus validus, Carex lacustris, Phalaris arundinacea, and Typha latifolia) were grown in monoculture and as a four-species mixture to compare eﬀectiveness of nutrient removal in controlled 18.93-l outdoor subsurface treatment
wetland microcosms. A nutrient treatment that mimicked single-resident domestic eﬄuent consisted of two levels of nitrogen (N)
and phosphorus (P) [low (56 mg/l N and 31 mg/l P) and high (112 mg/l N and 62 mg/l P)] of nutrient solution applied three times
weekly. The plants were established and maintained for one year before the nutrient treatment and monthly water sampling
commenced; water sampling began July 31, 2001 and ended October 23, 2001. We tested four hypotheses: (1) vegetated microcosms
are more eﬀective at reducing concentrations of total N and total P from soil leachate than unvegetated, (2) there is a diﬀerential
species eﬀect on the potential to reduce N and P, (3) plant mixtures are more eﬀective than monocultures at reducing N and P, and
(4) the microcosms will be least eﬀective at reducing N and P concentrations in October compared to August. We found support for
hypotheses 1, 2, and 4, but our results are inconclusive for the third hypothesis. Total N and total P in the soil leachate were
signiﬁcantly higher from unvegetated microcosms compared to vegetated. S. validus was most eﬀective and P. arundinacea was
generally least eﬀective at reducing N and P in monocultures, with treatment capabilities similar to unvegetated microcosms. The
four-species mixture was generally highly eﬀective at nutrient removal, however the results were not signiﬁcantly diﬀerent from the
monocultures. At the end of the growing season (October) treatment eﬃciency was signiﬁcantly less than earlier months, especially
for the unvegetated treatment.
 2003 Elsevier Ltd. All rights reserved.
Keywords: Treatment wetland; Phytoremediation; Nitrogen; Phosphorus

1. Introduction
The characteristic properties of wetlands make them
unique ecosystems; they contain anoxic soils, have
varying hydrology, distinct nutrient cycling and are
composed of plants tolerant of ﬂooded conditions. Since
wetlands are often found as a transitional zone between
aquatic and terrestrial ecosystems, they can receive a
wide array of dissolved substances through storm water
runoﬀ as well as through rivers, streams and water
channels. Wetlands are able to transform and reduce
*
Corresponding author. Tel.: +1-330-972-6141; fax: +1-330-9728445.
E-mail address: lfraser@uakron.edu (L.H. Fraser).

0960-8524/$ - see front matter  2003 Elsevier Ltd. All rights reserved.
doi:10.1016/j.biortech.2003.11.023

these compounds, so they have been utilized for water
treatment (Kadlec and Knight, 1996).
Widespread degradation of aquatic environments
including lakes, streams and even the oceans can be
attributed to the increase of agricultural practices, such
as fertilization, tilling, pesticides and herbicides (Ratcliﬀe, 1984; Hooda et al., 2000). An increasing human
population generates a greater agricultural demand and
a concomitant increase in agricultural runoﬀ. Increases
in wastewater, whether agricultural, municipal or
industrial create challenges for those seeking cost eﬀective treatment methods to process this wastewater
(Fraser et al., 2003; Steer et al., 2003). Wetlands have
been constructed in order to engineer secondary water
treatment facilities, and scientists have explored the

186

L.H. Fraser et al. / Bioresource Technology 94 (2004) 185–192

fundamental biogeochemical processes that enable wetlands to treat this wastewater, and the utility of phytoremediation––the use of plants to treat wastewater
(Brix, 1997; Carmen and Crossman, 2001).
As the numbers of constructed wetlands grow, an
increasing amount of research throughout the world is
being conducted in order to evaluate the eﬃciency of
treatment wetlands (e.g. Steer et al., 2002 [Ohio, USA];
Nerella et al., 2000 [Texas, USA]; Badkoubi et al., 1998
[Iran]; Haberl et al., 1995 [Europe]; Greenway and
Simpson, 1996 [Australia]). In Ohio, USA, 21 subsurface
ﬂow single-family treatment wetlands have been monitored throughout the northern half of the state for both
organics and nutrient removal (Steer et al., 2002). The
domestic treatment wetlands can reduce output of fecal
coliform 88 ± 27%, total suspended solids 56 ± 53%,
biochemical oxygen demand 70 ± 48%, ammonia
56 ± 31% and phosphorus 80 ± 20% (Steer et al., 2002).
Another study in Texas monitored eight household
constructed subsurface ﬂow wetlands (Nerella et al.,
2000). After one year, the BOD removal eﬃciency was
between 80% and 90%, however nutrient reduction was
relatively low, below 40% (Nerella et al., 2000).
In all these published constructed wetland experiments there is a very wide range in treatment eﬃciency
between wetlands and many of the experiments failed to
meet all governmental standards, especially for total
nitrogen and phosphorus (e.g. Rogers et al., 1991; Steer
et al., 2002). In order to discover how to more eﬃciently
construct treatment wetlands smaller scale mesocosm
and microcosm experiments have been initiated to test
the plant species composition (Gersberg et al., 1986;
Tanner, 1996; Coleman et al., 2001).
The purpose of this research was to determine the
relative eﬀect speciﬁc plant species, as well as a fourspecies plant mixture, have on the reduction of total N
(nitrogen) and total P (phosphorus) from the soil
leachate in subsurface wetland microcosms. The four
plants selected for our experiment were Carex lacustris
Willd. (lake sedge), Scirpus validus Vahl. (softstem bulrush), Phalaris arundinacea L. (reed canarygrass) and
Typha latifolia L. (broad-leaved cattail) (Gleason and
Cronquist, 1991). We chose these species because they
have been used in previously published wastewater
treatment experiments (Gersberg et al., 1986; Coleman
et al., 2001), and used in constructed wetlands (Cronk
and Fennessy, 2001; Steer et al., 2002). These wetland
plants are fast-growing, tall-stature, ‘‘clonal-dominants’’
(sensu Boutin and Keddy, 1993) that establish quickly,
process a lot of energy, and are therefore considered
suitable plants for treatment wetlands.
A microcosm experiment was designed to answer four
questions: (1) Are vegetated microcosms better than
non-vegetated microcosms at reducing N and P in the
soil leachate? (2) Are there diﬀerential responses between the four wetland plants in their eﬀectiveness at

reducing N and P? (3) Does a four-species mixture relatively enhance the eﬀectiveness of the removal of N and
P compared to monoculture treatments? and (4) is there
a diﬀerence between sampling dates with regard to N
and P in the soil leachate? Our hypotheses, based on the
research presented in the introduction, are (1) vegetated
microcosms are more eﬀective at reducing concentrations of total N and total P than unvegetated, (2) there is
a diﬀerential species eﬀect on the potential to reduce N
and P, (3) plant mixtures are more eﬀective than
monocultures at reducing N and P, and (4) the microcosms will be least eﬀective at reducing N and P concentrations in October compared to August.

2. Methods
2.1. Study site
The study was located at the Bath Nature Preserve in
Bath, Ohio. Beginning in the summer of 2000, a 4-m tall,
20 · 20 m fence was constructed at the site location. The
fencing was covered completely with bird netting and
secured with a padlocked gated entry. An on-site weather
station (Spectrume WatchDog Model 900ET) recorded
temperature and rainfall throughout the study period.
2.2. Experimental design
The experiment was established in the fall of 2000.
The experimental design was a six (plant type) by two
(nutrient addition) factorial combination. The plant
treatment had six levels: no plants, four monocultures,
(C. lacustris, S. validus, P. arundinacea and T. latifolia),
and a four-species mixture of these plants. The nutrient
treatment had two levels (low and high). Each type of
microcosm treatment had six replicates for a total
microcosm count of 72. The microcosms were arranged
in six rows. Two rows were placed back-to-back with an
approximate 1-m section in between. The placement of
microcosms was completely randomized.
2.3. Microcosms
The outdoor microcosms used to test the eﬀectiveness
of plants to treat wastewater were 18.93-l white buckets
ﬁlled with soil to 34 full, accounting for approximately
0.014 m3 of soil. The microcosms were set up to mimic
subsurface treatment wetlands; that is, wetlands where
the water level is below the soil surface (Kadlec and
Knight, 1996). The soil (Carlisle Muck) was taken from
the Panzner Wetland Restoration Site in Copley, Ohio.
This soil was used because it is a peaty wetland soil with
high carbon and nutrient levels, which aids in the
establishment of the plants. Three soil samples were
taken at random, before the microcosms were ﬁlled, and

L.H. Fraser et al. / Bioresource Technology 94 (2004) 185–192

sent to Spectrum Analytic Inc. (Washington Court
House, OH) to be analyzed for total nitrogen (TN), total
phosphorus (TP), pH, and percent organic matter. Each
microcosm had a hole in the bottom with a diameter of
1.9 cm that was plugged with a rubber stopper. A mesh
net covering the hole was fastened on the inside of the
microcosm in order to keep soil inside during the release
of water for testing.

187

Company, Loveland, CO) and a Spec 20 mass spectrometer.
2.7. Harvesting
During the month of November the above-ground
plant biomass was harvested. The plants growing in
mixture were separated to species. All plants were ovendried for approximately 48 h at 80 C and weighed.

2.4. Plant establishment
2.8. Statistical analysis
In May of 2001, rhizomatous cuttings of each species
from the ﬁeld were collected. Each microcosm was
planted with eight cuttings (approximately 10 cm in
length) of each species in monocultures and two of each
species in the four-species mixture microcosms. To
control for the disturbance experienced by the vegetated
microcosms during the planting of the rhizomes, the
unvegetated microcosms were similarly disturbed but
with no planting. The microcosms were monitored three
times per week, and invasive seedlings detected were
immediately removed.
The microcosms were watered three times weekly or
as needed depending on the weather. The water level
was kept constant at 5 cm below the level of the soil
surface. Nutrient addition began in August 2001, at
which time the plants were well established.

All statistical tests were performed using Systat Version 8 by SPSS Inc. In all cases, signiﬁcance was deﬁned
by p < 0:05. Six microcosms were removed from the
analyses because vegetation failed to establish. A twoway ANOVA was used to determine signiﬁcance of
species and nutrient eﬀect on biomass. Tukey’s LSD was
applied to test for signiﬁcance between treatment means.
Three-way ANOVAs were used to examine the eﬀects of
the three sampling periods (DATE), the two nutrient
treatments (NUTR), and the species treatment (SPECIES) on total nitrogen and total phosphorus in the
eﬄuent water. Linear regressions were performed within
each plant species to determine the possible eﬀect of
biomass on nitrogen and phosphorus reduction, but the
results were non-signiﬁcant and therefore we have not
included these results.

2.5. Nutrient addition
Rorison nutrient solution (Hendry and Grime, 1993)
was used to simulate the high nutrient input levels typically found in single-family domestic treatment wetlands (Steer et al., 2002). The standard solution
contained 56 mg/l of nitrogen (as (NH4 )6 Mo7 O24 Æ 4H2 O)
and 31 mg/l of phosphorus (as K2 HPO4 Æ 3H2 O). For
those microcosms receiving the lower level of nutrients
1 l of Rorison solution was added. For microcosms
receiving the higher level of nutrients the concentration
of the solution was doubled, but only 1 l was added (i.e.
112 mg/l of nitrogen and 62 mg/l of phosphorus). The
nutrient addition was applied three times weekly.
2.6. Water sampling
A baseline measurement of water was taken on July
30, 2001, before the nutrient addition commenced, followed by monthly samples until October 23, 2001.
Seventy-two 100 ml water samples were taken monthly,
one from each of the microcosms. Soil leachate was
drained from the bottom of the microcosms, through a
pre-drilled hole. Water samples were immediately placed
in a cooler, brought back to the lab and kept cool at 4
C until analyzed, which was within 48 h of sampling.
Total nitrogen (mg/l) and total phosphorus (mg/l) were
run using the HACH Test’N Tubee tests (HACH

3. Results
A number of baseline measurements were made before and during the experiment. Both the temperature
and the total amount of rainfall were recorded during
the study period by the weather station. The highest
recorded temperatures occurred at the end of July and
the beginning of August (37 C). One night during the
last week of August the temperature fell below freezing
()1 C), but daytime temperatures were still high (27
C). The lowest temperatures were recorded at the end
of the study in October when nighttime temperatures
were regularly below freezing. Rainfall was lowest during both July and September (6 cm/month). Rainfall in
the remaining months during the study period ranged
from 9.5 to 11.5 cm/month. Preliminary tests were run
on the soil at the beginning of the experiment (Table 1).
There was very little variation among the microcosms as
noted by the small standard deviations.
The harvested plant dry biomass was analyzed comparatively between nutrient treatments (Fig. 1, Table 2).
A two-way ANOVA (Table 2) showed that the biomass
of the species was signiﬁcantly diﬀerent, however nutrients had no signiﬁcant aﬀect on biomass. C. lacustris
had the greatest dry biomass at both nutrient levels,
and post-hoc analysis determined that the diﬀerence

188

L.H. Fraser et al. / Bioresource Technology 94 (2004) 185–192

was statistically signiﬁcant in comparison with all other
species as well as the mixture.
Results from the three-way ANOVA for total nitrogen to detect eﬀects of sampling date, nutrients and
species factors are shown in Table 3. This analysis
produced a signiﬁcant eﬀect of sampling date, nutrient
and species factors. In addition, the interactions between
date and nutrient, as well as date and species was signiﬁcant. The signiﬁcant date by nutrient combination
result indicates that the high nutrient treatment result in
higher nitrogen levels in the eﬄuent. The signiﬁcant date
by species combination illustrates a trend of higher
nitrogen levels measured later in the growing season
when temperature has declined.
A comparison of total nitrogen by species treatment,
and across sampling dates, is shown in Fig. 2a (low
nutrients) and Fig. 2b (high nutrient). The 7/31/01
sampling date is a pre-treatment baseline measurement.
Unvegetated microcosms show consistently and signiﬁcantly higher nitrogen levels at all three post-treatment
dates. At low nutrient (Fig. 2a), P. arundinacea has
consistently higher nitrogen values than the other vegetated microcosms. The high nutrient treatment shows
less consistent responses by date (Fig. 2b), but the
mixture has the highest values of vegetated microcosms
on 9/25/01, while T. latifolia has the highest values of
vegetated microcosms on 10/23/01. The general trend is
for increasing total nitrogen levels over time.
Table 4 shows the results from the three-way ANOVA for total phosphorus to detect eﬀects of sampling
date, nutrient and species factors. This analysis produced a signiﬁcant eﬀect of sampling date, nutrient and
species factors. In addition, the interactions between
date and species, as well as nutrient and species was
signiﬁcant. The nutrient by species response showed that
P. arundinacea had signiﬁcantly lower phosphorus levels
at the high nutrient treatment.
A comparison of total phosphorus by species treatment, and across sampling dates, is shown in Fig. 3a
(low nutrient) and Fig. 3b (high nutrient). The 7/31/01
sampling date is a pre-treatment baseline measurement.
Unvegetated microcosms show the highest total phos-

Table 1
Soil analysis of the Carlisle Muck used in the microcosms
Total nitrogen (mg/l)
Total phosphorus (mg/l)
pH
Organic matter (%)

Mean

Standard deviation

2.4
0.73
5.83
16.93

0.06
0.04
0.06
0.551

160

Above-ground dry biomass (g)

c

140
c

120
100

b

80
60

b

b

b

b

ab
ab

a

40
20
0
Scirpus

Carex

Phalaris

Typha

Mix

Fig. 1. Mean dry biomass (g) of total above-ground plants per
microcosm between plant treatments and nutrient levels. Open bars are
low nutrient and hatched bars are high nutrient treatments. Error bars
represent standard deviation. Bars sharing the same letter are not
signiﬁcantly diﬀerent based on Tukey’s HSD.

Table 2
Two-way ANOVA to determine signiﬁcance of species and nutrient
eﬀect on biomass
Source

Sum-of-squares

df

F -ratio

P

NUTR
SPECIES
NUTR * SPECIES
ERROR

593.198
33679.492
691.802
9896.132

1
4
4
43

2.578
36.586
0.753

0.116
<0.001
0.563

In the table, NUTR refers to nutrient treatment, and SPECIES refers
to the plant species treatment. ‘‘df’’ is degrees of freedom.

Table 3
Results of three-way ANOVA examining the eﬀects of the three sampling periods (DATE), the two nutrient treatments (NUTR), and the species
treatment (SPECIES) on total nitrogen in the outlet water
Source

Sum-of-squares

df

F -ratio

P

DATE
NUTR
SPECIES
DATE * NUTR
DATE * SPECIES
NUTR * SPECIES
DATE * NUTR * SPECIES
ERROR

7453.415
2065.655
13915.313
1425.715
2230.425
747.720
1496.974
18999.044

2
1
5
2
10
5
10
164

32.169
17.831
24.023
6.153
1.925
1.291
1.292

<0.001
<0.001
<0.001
0.003
0.045
0.270
0.239

‘‘df’’ is degrees of freedom.

L.H. Fraser et al. / Bioresource Technology 94 (2004) 185–192
80

189

14
12
10

T P (m g/L)

T N (m g/L)

60

NONE

40

SCIRPUS
CAREX
PHALARIS
TYPHA
MIX

20

8

NONE

6

SCIRPUS
CAREX
PHALARIS

4

TYPHA
MIX

2

0
7/31/01

0
7/31/01

8/29/01

(a)

9/25/01

8/29/01

9/25/01

10/23/01

9/25/01

10/23/01

10/23/01

(a)

Date

Date
14

80
12
10

T P (m g/L )

T N (m g/L)

60

40

8
6
4

20

0
7/31/01

2
0
7/31/01

8/29/01

(b)

9/25/01

8/29/01

(b)

10/23/01

Date

Date

Fig. 3. The mean total phosphorus (mg/l) in soil leachate for (a) low
nutrient, and (b) high nutrient treatments over the study period. Error
bars represent one standard error.

Fig. 2. The mean total nitrogen (mg/l) in soil leachate for (a) low
nutrient, and (b) high nutrient treatments over the study period. Error
bars represent one standard error.

Table 4
Results of three-way ANOVA examining the eﬀects of the three sampling periods (DATE), the two nutrient treatments (NUTR), and the species
treatment (SPECIES) on total phosphorus in the outlet water
Source

Sum-of-squares

df

F -ratio

P

DATE
NUTR
SPECIES
DATE * NUTR
DATE * SPECIES
NUTR * SPECIES
DATE * NUTR * SPECIES
ERROR

48.970
47.178
1365.292
2.531
189.173
191.539
39.489
1186.223

2
1
5
2
10
5
10
164

3.385
6.523
37.751
0.175
2.615
5.296
0.546

0.036
0.012
<0.001
0.840
0.006
<0.001
0.855

‘‘df’’ is degrees of freedom.

phorus levels at low and high nutrient levels, and they
do not change over time. Over time, vegetated microcosms show a general trend to increase. P. arundinacea
has high phosphorus levels at low nutrient, but reduced
levels at high nutrient on the last two sampling dates.

4. Discussion
Microcosms are extremely useful for controlled,
mechanistic investigations (Fraser and Keddy, 1997),
and have previously been used to test plants’ ability to

treat wastewater (Gersberg et al., 1986; Coleman et al.,
2001). Microcosms, however, have limitations. For
example, the spatial scale does not generally reﬂect what
occurs in nature, and abiotic conditions may be aﬀected
by the experimental conditions. In addition, research
has shown that in constructed systems more than one
year may be needed to reach ‘‘natural wetland conditions’’ (Sistani et al., 1996). Hence, our microcosms were
being tested during their second growing season.
Our purpose was to determine if plants can eﬀectively
reduce the levels of total N and total P in the soil water
leachate, and thereby be potentially useful in treating

190

L.H. Fraser et al. / Bioresource Technology 94 (2004) 185–192

single-family domestic wastewater in subsurface constructed wetlands. Biomass of the microcosms was not
signiﬁcantly diﬀerent among the high and low nutrient
treatments (Table 2, Fig. 1), which suggests that plant
growth was not limited by nutrients. Bachand and
Horne (1999) showed that an increase in biomass, both
living and dead, enhanced rates of denitriﬁcation and
improved removal eﬃciencies of nitrogen. We did not
ﬁnd a correlation between biomass and treatment eﬃciency, but perhaps given more time this would be
detectable.
Our results support the ﬁrst hypothesis that states
plants can reduce total N and total P at a level signiﬁcantly lower than unvegetated systems (Tables 3 and 4).
In some cases, the reduction of N and P in the soil
leachate was 10 times greater in vegetated microcosms
than unvegetated microcosms (e.g. Fig. 2b, 10/23/01 for
N and Fig. 3a, 8/29/01 for P). Our second hypothesis
tested whether four wetland plant species were diﬀerentially eﬀective at reducing N and P. We found that
plant species do have a diﬀerential response to reducing
N and P (Tables 3 and 4; Figs. 2 and 3). At low nutrients, both S. validus and the four-species mixture had
signiﬁcantly lower N and P in their soil leachate than the
other plant treatments. The least eﬀective species was P.
arundinacea, which tended to have signiﬁcantly higher N
and P in the soil leachate than the other wetland plant
treatments, and similar N and P concentrations as the
unvegetated microcosms. However, the high nutrient
treatment resulted in slightly diﬀerent results for some of
the species, which indicates that nutrient load might
aﬀect the treatment potential of plant species. For
example, the Typha microcosms performed quite well at
reducing N at low nutrient conditions (9.2 mg/l on 10/
23/01, Fig. 2a), but performed poorly under high
nutrient additions (31.5 mg/l on 10/23/01, Fig. 2b).
Phalaris was consistently poor at reducing P at low
nutrients across all dates (Fig. 3a), but under high
nutrient conditions, Phalaris was one of the more
eﬀective plants at reducing P (Fig. 3b).
Gersberg et al. (1986) investigated constructed wetlands that were vegetated by S. validus, Phragmites
communis, or T. latifolia or that were unvegetated in
order to determine the eﬃciency among the treatments
at reducing nitrogen. In all cases unplanted wetlands
were less eﬀective than any of the planted (Gersberg
et al., 1986). However, amongst species, S. validus was
the most eﬀective, which supports our ﬁndings. T. latifolia was the least eﬀective (Gersberg et al., 1986). Presumably due to the high ammonia loading, the T.
latifolia plants began to yellow after three months and
nearly all were dead after six months (Gersberg et al.,
1986). Gersberg hypothesized that T. latifolia was least
eﬀective due to its shallow rooting zone and the inability

to create an eﬀective environment for various microbial
communities (Gersberg et al., 1986). Contrary to the
ﬁnding of Gersberg et al. (1986), a study conducted by
Coleman et al. (2001) found that T. latifolia was very
eﬃcient at removal of nutrients.
In our experiment, T. latifolia was the slowest to
establish. Contrary to published reports (Clarke and
Baldwin, 2002; Svengosouk and Mitsch, 2001), T. latifolia did not produce more biomass with high nutrient
conditions. High nutrient levels appeared to stunt and
even kill the T. latifolia plants; very few transplanted
seedlings grew within the microcosms. Similar to results
reported by Gersberg et al. (1986), the T. latifolia plants
began to yellow after 2.5 months and were the ﬁrst
species to die oﬀ at the end of the season.
The third hypothesis was that a four-species plant
mixture is more eﬀective than a monoculture at reducing
nitrogen and phosphorus. At low nutrients, the mixed
microcosms consistently had among the lowest N and P
concentrations in the soil leachate (Figs. 2a and 3a), but
the only signiﬁcant diﬀerence between a monoculture
treatment was Phalaris for total P (Fig. 3a). At high
nutrients, the mixed microcosms did not have the lowest
N and P concentrations, and in fact had signiﬁcantly
higher P than Phalaris, Scirpus, and Carex on 10/23/01
(Fig. 3b). Therefore, our results do not support the
hypothesis that mixtures have the potential to reduce N
and P any more than monocultures. However, mixtures
may provide other beneﬁts over monocultures, such as
enhanced tolerance to abiotic stress or enhanced treatment eﬃciency of other toxins or nutrients not measured
in this study.
Coleman et al. (2001) tested unplanted versus planted
monocultures and three-species mixture mesocosms (400
l troughs) with two levels of water depth. The plants
used were Juncus eﬀusus, T. latifolia, and Scirpus cyperinus. Their mixed treatment was quite eﬀective at
reducing nutrient level, possibly due to root partitioning
of the soil. However, their results showed no signiﬁcant
diﬀerence between T. latifolia monocultures and the
mixed system (Coleman et al., 2001).
The results of our experiment supported the fourth
hypothesis; we found a general increase in N and P
concentration within the leachate over time, with the
highest levels in October. Temperature has been shown
to play a role in the removal of nitrogen through enhanced denitriﬁcation processes (Bachand and Horne,
1999). Some of the decline in treatment at the end of the
study might be attributed to lower temperatures and
thus lower microbial activity as well as a reduction in
macrophytic growth rates.
A review by Brix (1997) details the role macrophytes
play within a constructed wetland system and how they
are an integral part in nutrient cycling. Macrophytic

L.H. Fraser et al. / Bioresource Technology 94 (2004) 185–192

plants encourage the assimilation and breakdown of
nutrients within a wetland system. They have the ability
not only to bind high amounts of nutrients within their
system, but also to create an environment conducive to
decreasing nutrients. For example, (1) they provide surface area on their stems and leaves, which is necessary for
microbial growth; (2) their roots provide a structure for
microorganisms to adhere and perform the processes
necessary for transformation of nutrients; (3) roots not
only provide a place for microbes, but also serve to decrease erosion and increase the levels of oxygen, which
provides for the oxidation of toxic substances like
ammonia (NH3 ) and nitrites (NO2 ); (4) macrophytes are
not only beneﬁcial in the removal and transformation of
nutrients, but are also aesthetically pleasing to homeowners where many of these systems are located (see
Brix, 1997; Carmen and Crossman, 2001).

5. Conclusion
This study demonstrates the importance of macrophytes to reduce nutrient concentrations encountered by
single-family domestic subsurface constructed wetlands,
however the applicability of the results to actual constructed wetlands has yet to be determined. Vegetated
microcosms had signiﬁcantly lower total N and total P
in their soil leachate than unvegetated microcosms. S.
validus was shown to be eﬀective overall at treating
nutrient loaded water and should therefore be considered for the future design of single-resident treatment
wetlands. P. arundinacea was generally least eﬀective in
monocultures, with treatment capabilities similar to
unvegetated microcosms, but the performance of Phalaris at reducing phosphorus at high nutrient levels was
relatively high. The four-species mixture microcosms
were generally highly eﬀective at nutrient removal and
should also be considered for future design of singleresident treatment wetlands, however they did not show
any signiﬁcant diﬀerence between Scirpus, Typha and
Carex in monoculture.

Acknowledgements
We thank Adam Landaw and Jason Karnezis for
their help in the set up and maintenance of the experiment. Joel Duﬀ’s comments on earlier versions of the
manuscript are much appreciated. This publication was
ﬁnanced in part through a grant from the Ohio Environmental Protection Agency and the United States
Environmental Protection Agency, under the provisions
of Section 319(h) of the Clean Water Act.

191

References
Bachand, P.A.M., Horne, A.J., 1999. Denitriﬁcation in constructed
free-water surface wetlands: II. Eﬀects of vegetation and temperature. Ecol. Eng. 14, 17–32.
Badkoubi, A., Ganjidoust, H., Ghaderi, A., Rajabi, A., 1998.
Performance of a subsurface wetland in Iran. Water Sci. Technol.
38, 345–350.
Brix, H., 1997. Do macrophytes play a role in constructed treatment
wetlands. Water Sci. Technol. 35, 11–17.
Boutin, C., Keddy, P.A., 1993. A functional classiﬁcation of wetland
plants. J. Veg. Sci. 4, 591–600.
Carmen, E.P., Crossman, T.L., 2001. Phytoremediation. In: Nyers,
E.K. (Ed.), In situ Treatment Technology. Lewis Publishers, New
York, pp. 391–435.
Clarke, E., Baldwin, A.H., 2002. Responses of wetland plants to
ammonia and water level. Ecol. Eng. 18, 257–264.
Coleman, J., Hench, K., Garbutt, K., Sextone, A., Bissonnette, G.,
Skousen, J., 2001. Treatment of domestic wastewater by three
wetland plant species in constructed wetlands. Water, Air Soil Poll.
128, 283–295.
Cronk, J.K., Fennessy, M.S., 2001. Wetland Plants: Biology and
Ecology. Lewis Publishers, New York.
Fraser, L.H., Keddy, P., 1997. The role of experimental microcosms in
ecological research. TREE 12, 478–481.
Fraser, L.H., Bradford, M.E., Steer, D.N., 2003. Human appropriation and treatment of fresh water: a global hydrology model
incorporating treatment wetlands. Int. J. Environ. Sust. Dev. 2 (2),
174–183.
Gersberg, R.M., Elkins, B.V., Lyons, S.R., Goldman, C.R., 1986. Role
of aquatic plants in wastewater treatment by artiﬁcial wetlands.
Water Res. 20, 363–368.
Gleason, H.A., Cronquist, A., 1991. Manual of vascular plants of
northeastern United States and Adjacent Canada, second ed. The
New York Botanical Garden, Bronx, NY.
Greenway, M., Simpson, J.S., 1996. Artiﬁcial wetlands for wastewater
treatment, water reuse and wildlife in Queensland, Australia. Water
Sci. Technol. 33, 221–229.
Haberl, R., Perﬂer, R., Mayer, H., 1995. Constructed wetlands in
Europe. Water Sci. Technol. 32, 305–315.
Hendry, G.A.F., Grime, J.P., 1993. Methods in Comparative Plant
Ecology: A Lab Manual. Chapman & Hall, London.
Hooda, P.S., Edwards, A.C., Anderson, H.A., Miller, A., 2000. A
review of water quality concerns in livestock farming areas. Sci.
Total Environ. 250, 143–167.
Kadlec, R.H., Knight, R.L., 1996. Treatment Wetlands. Lewis
Publishers, New York.
Nerella, S., Weaver, R., Lesikar, B.J., Persyn, R.A., 2000. Improvement of domestic wastewater quality by subsurface ﬂow constructed wetlands. Bioresour. Technol. 75, 19–25.
Ratcliﬀe, D.A., 1984. The ﬂora and vegetation of Britain: origins and
changes––the facts and their interpretations: post medieval and
recent changes in British vegetation the culmination of human
inﬂuence. New Phytol. 98, 73–100.
Rogers, K.H., Breen, P.F., Chick, A.J., 1991. Nitrogen removal in
experimental wetland treatment systems: evidence for the role of
aquatic plants. Res. J. Water Poll. Control Fed. 63, 934–941.
Sistani, K.R., Mays, D.A., Taylor, R.W., 1996. Development of
natural conditions in constructed wetlands: biological and chemical
changes. Ecol. Eng. 12, 125–131.
SPSS, 1998. SYSTAT 8.0. SPSS Inc., Chicago, IL.
Steer, D., Fraser, L.H., Boddy, J., Seibert, B., 2002. Eﬃciency of small
constructed wetlands for subsurface treatment of single family
domestic eﬄuent. Ecol. Eng. 18, 429–440.

192

L.H. Fraser et al. / Bioresource Technology 94 (2004) 185–192

Steer, D., Aseltyne, T., Fraser, L.H., 2003. Life-cycle economic model
of small treatment wetlands for domestic wastewater disposal.
Ecol. Econ. 44, 359–369.
Svengosouk, L.S., Mitsch, W.J., 2001. Dynamics of mixtures of
Typha latifolia and Schoenoplectus tabernaemontani in nutrient-

enrichment wetland experiments. Am. Mid. Nat. 145, 309–
324.
Tanner, C.C., 1996. Plants for constructed wetland treatment systems––a comparison of the growth and nutrient uptake of eight
emergent species. Ecol. Eng. 7, 59–83.