Bioresource Technology 96 (2005) 1039–1047

The interacting effects of temperature and plant community type

on nutrient removal in wetland microcosms
Christian R. Picard a, Lauchlan H. Fraser a,*,1
, David Steer b

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

Received 18 May 2004; received in revised form 16 September 2004; accepted 16 September 2004
Available online 13 November 2004

Abstract

Treatment wetlands can remove nutrients from inflow sources through biogeochemical processes. Plant composition and temper-
ature play important roles in the nutrient removal efficiency of these wetlands, but the interactions between these variables are not
well understood. We investigated the seasonal efficiency of wetland macrophytes to reduce soil leachate concentrations of total
nitrogen and total phosphorus in experimental microcosms. Each microcosm contained one of six vegetation treatments: unplanted,
planted with one of four species (Carex lacustris, Scirpus validus, Phalaris arundinacea and Typha latifolid) in monoculture or
planted with an equal abundance of all four species. Microcosms were also subjected to two temperature treatments: insulated
microcosms and microcosms exposed to environmental conditions. A constant nutrient solution containing 56 mg/l N and 31 mg/
l P was added to all microcosms three times a week. Water samples were analyzed monthly for total dissolved nitrogen and total
dissolved phosphorous. Microcosms exhibited a typical pattern of seasonal nutrient removal with higher removal rates in the grow-
ing season and lower rates in the winter months. In general, planted microcosms outperformed unplanted microcosms. Among the
plant treatments, Carex lacustris was the least efficient. The four remaining plant treatments removed an equivalent amount of nutri-
ents. Insulated microcosms were more efficient in the winter and early spring months. Although a seasonal pattern of nutrient
removal was observed, this variation can be minimized through planting and insulation of wetlands.

2004 Elsevier Ltd. All rights reserved.

Keywords: Treatment wetlands; Phytoremediation; Eutrophication; Nitrogen; Phosphorous; Seasonality; Microcosms

1. Introduction 1980). Nutrient enrichment, or eutrophication, of aqua-

tic ecosystems can cause an increase in algae and aquatic
Human activities have had a large effect on global plants, loss of component species and loss of ecosystem
biogeochemical cycles. Through agricultural practices, function (Smith et al., 1999). Eutrophication is the larg-
urbanization, industrialization and other alterations, est water quality problem throughout the world (Car-
humans have increased the input of nutrients into bio- penter et al., 1998). For example, 61% of 2048 water
geochemical cycles, especially nitrogen (Vitousek et al., bodies located in the United States failed to meet EPA
1997) and phosphorous (Reckhow and Simpson, standards with regard to total nitrogen and total phos-
phorous (Smith et al., 1997a,b). Wetlands have been
investigated as a possible solution to these global
*
Corresponding author. Tel.: +1 250 377 6135; fax: +1 250 828 eutrophication and water quality problems (Mitsch
5450. et al., 2001; Fraser et al., 2003).
E-mail address: lfraser@cariboo.bc.ca (L.H. Fraser).
1
Present address: Department of Natural Research Sciences,
Total nitrogen and total phosphorous removal in
Thompson Rivers University, Kamloops, British Columbia, Canada treatment wetlands can range from 3–98% to 31–99%
V2C 5N3. respectively (Spieles and Mitsch, 2000; Steer et al.,

0960-8524/$ - see front matter
2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.biortech.2004.09.007
1040 C.R. Picard et al. / Bioresource Technology 96 (2005) 1039–1047

2002). On average, the removal of nitrogen and phos- taining polycultures during all seasons. Throughout the
phorous from these systems is about 50% (Verhoeven experiment, planted microcosms should be more effi-
and Meuleman, 1999). Plants and seasonal temperature cient than unplanted microcosms. It was also predicted
change are two factors that greatly affect the nutrient re- that insulated microcosms should provide a more suita-
moval efficiency of these systems. Plants, such as species ble environment for nutrient removal.
from the genera Typha, Scirpus, and Phragmites play an
important role in the biogeochemical cycle of treatment
wetlands (Brix, 1997; Wood et al., 1999; Schutes, 2001; 2. Methods
Fraser et al., 2004).
Wetlands are affected by solar radiation and ambient 2.1. Study site
temperatures, which cycle on an annual and daily basis.
These abiotic factors mediate the temperature of the This experiment was carried out in an enclosed area
00
wetland environment causing cyclical patterns in evapo- at the Bath Nature Preserve in Bath, Ohio (4106 0 39
00
transpiration, photosynthesis and microbial activity N, 8151 0 42 W). The research area consisted of fencing
(Kadlec, 1999). Laboratory studies demonstrate the that was 2.4 m tall and enclosed a 15.2 · 15.2 m area of
optimum temperature for nutrient removal to be 30 C grassland within the preserve. The fencing was covered
(Wood et al., 1999). ‘‘Biological zero’’ or 5 C is where with bird netting and secured with a locked gate entry.
biological processes drastically slow down or cease The area was mowed and outfitted with three 189.3-l
(Mitsch and Gosselink, 1993). More specifically, nitrifi- water tanks, which provided water throughout this
cation rates in wetlands become inhibited at water tem- experiment. A weather station (Watchdog) was placed
peratures of about 10 C and rates drop rapidly at 6 C in the research area and recorded the ambient tempera-
(Werker et al., 2002). Phosphorous removal is affected ture at 15-min intervals throughout the study period.
less because it is dominated by sediment adsorption as
opposed to biological processes. Many treatment wet- 2.2. Microcosms
lands in temperate climates often operate at a much
lower level of nutrient removal efficiency in the colder The microcosms used in this experiment consisted of
months (Spieles and Mitsch, 2000). Wittgren and Maeh- 18.93-l buckets (28.58 cm diameter and 35.88 cm high)
lum (1997) state that nitrogen cycling is inhibited in that were a quarter filled with soil. The microcosms
colder months due to the decrease of oxygen availabil- were filled with approximately 0.02 m3 of highly organic
ity. Furthermore, extreme temperatures inhibit the soil (Carlisle muck), which was taken from the Panzner
growth rate of nitrogen reducing bacteria (Spieles and Wetland Restoration Site in Copley, Ohio. Each micro-
Mitsch, 2000). Conversely, constructed wetlands have cosm had a 2 cm hole in the bottom plugged with a
effectively run at atmospheric temperatures of
4 C in removable rubber stopper to allow drainage. A mesh
northern China and
20 C in Norway (Yin and Shen, lining was glued to the bottom of the bucket to prevent
1995; Werker et al., 2002). These wetlands are often soil loss during drainage. Four 2 cm holes were also
insulated by some natural or artificial means. Many drilled into the sides of the buckets at a point about
cold-climate wetlands are also specifically designed to 5 cm above the soil surface. These holes were evenly dis-
operate in cold conditions. tributed around the circumference of the bucket. The
Little research has been done on the interacting purpose of these holes was to ensure that the water lev-
effects of the role of plants in treatment wetlands and els in the buckets did not exceed 5 cm above the soil
seasonal temperature changes on nutrient removal effi- surface.
ciency. The objective of this experiment was to investi-
gate the seasonal changes in efficiency with respect to 2.3. Experimental design
plant communities using microcosms. This study also
explored the use of insulation as a method for increasing The experimental design for this study involved a
nutrient removal efficiency by mediating wetland tem- six by two factorial with six replicates. Hence, the ex-
perature. The major questions to be answered were: periment consisted of 72 microcosms. Each microcosm
(1) How does nutrient removal change throughout the contained one of six vegetation treatments. These treat-
seasons with respect to different plant communities? ments consisted of second year wetland plants in mono-
(2) Do plants exhibit species-specific rates of nutrient re- culture or an equal mixture of these species. There was
moval? (3) Does insulating microcosms improve their also a set of unplanted microcosms, which acted as a
effectiveness in extreme temperatures by mediating soil control. The four species of wetland plants used were
temperature? We expected that all microcosms would Carex lacustris, Scirpus validus, Phalaris arundinacea,
be more efficient in the growing season as compared to and Typha latifolia. Rhizomatous cuttings from wild
the winter months. Based on previous research, nutrient plants were planted in the microcosms in May 2001.
removal was expected to be greatest in microcosms con- Each microcosm was planted with six cuttings of each
 C.R. Picard et al. / Bioresource Technology 96 (2005) 1039–1047 1041

species in monoculture and two of each species in the 672 mg/l Ca(NO3)2 and 372 mg/l K2HPO4. These nutri-
combination treatments. Each cutting was approxi- ent loads are similar to those of actual treatment
mately 12 cm in length. The majority of the microcosms wetlands.
had well established plant communities by August of
2001. However, some plants did not survive the winter 2.5. Sampling regime and laboratory analysis
of 2001–2002. Hence, additional rhizomatous cutting
were added to the microcosms in April 2002 using the A 125 ml sample of filtered effluent (using a 50 lm
same procedure. By June 2002, all of the microcosms sieve) was taken from each of the microcosms during
housed a well-established plant community. Throughout the last few days of each month, with the exception of
the experiment, microcosms were hand weeded regularly February. Due to extremely low temperatures, the sam-
to remove any invading plants. The surrounding re- ples for February were taken during the first week of
search area was also kept mowed to ensure that the sur- March. At the time of sampling, each microcosm was
rounding, natural plant growth was not interfering with also drained of any remaining water in an attempt to
the plant communities within the microcosms. mimic natural systems. All water samples were collected
In addition to the vegetation treatment, half of the in Nalgene
bottles and stored at approximately 4 C
microcosms were also subjected to an insulation treat- within a laboratory refrigerator until further analysis.
ment. At the research site, holes of the approximate size The data logger information was also obtained during
of the microcosms were drilled into the ground using a each sampling event using Boxcar 3.7 software (Forestry
gas-powered auger. These holes were then fitted with a Suppliers).
sleeve of Reflectix
insulation to provide further insula- To ensure accurate results, all water samples were
tion, as well as to fortify the structure of the hole. Half analyzed for total dissolved nitrogen (TDN) and total
of the microcosms were placed in these holes so the soil dissolved phosphorous (TDP) within five days of sam-
levels of the bucket were slightly below the actual soil pling. This analysis involved a modified Kjeldahl tech-
level. The remaining microcosms were left at regular soil nique utilizing HACK Test N 0 Tube kits. Percent
level in order to be exposed to the environment. absorbance was obtained from the samples using a
Microcosms were arranged into six pair wise exposed HACH DR/4000 Spectrometer. The absorbance was
and insulated blocks for a total of 12 blocks. The order transformed into concentration values for TDN and
of the pair wise arrangement was randomly assigned. TDP using a standardized curve.
Within each block, the six vegetation treatments were
randomly distributed. Six Stowaway Tidbit Weather- 2.6. Calculations and statistical analysis
proof and Waterproof Temperature Loggers (Forestry
Suppliers Inc.) were placed in randomly designated insu- With the exception of June through August, biweekly
lated and exposed microcosms. These data loggers were means and standard error were calculated for each set of
buried about 15 cm below the soil surface. The locations data loggers. Only monthly averages were available for
of the data loggers were changed weekly to compensate June through August. Temporary logger malfunction
for the low quantity of loggers being utilized. However, caused gaps in the data set. Consequently, no further
this became difficult in the winter months when the top statistical analysis could be done on this data.
layer of soil was often frozen. Hence, the loggers were All statistical analyses were done using Systat Version
moved to new locations as the weather permitted during 8 (SPSS, 1998). Data were analyzed for homoscedastic-
these colder months. These loggers took and stored ity and normality of residuals. Monthly mean values
readings of soil temperature every 30 min. and standard errors for TDN and TDP were calculated
for each cross treatment. A general linear model was ap-
2.4. Nutrient additions plied to determine potential block effects for TDN and
TDP. TDN and TDP were analyzed with a repeated
The experiment officially began in June 2002 and ran measures two-way analysis of variance (ANOVA).
for one year. A modified Rorison solution of 56 mg/l These ANOVA were used to test the effects of the plant
Ca(NO3)2 and 31 mg/l K2HPO4 was added three times treatment and insulation treatment on effluent nutrient
a week to all microcosms (Hendry and Grime, 1993). concentrations, as well as the interacting effects of these
Microcosms were also watered at a frequency that kept treatments. The monthly sampling was set as the re-
the water level at approximately the soil surface. In peated measures.
order to avoid ice formation in the microcosms, a higher A set of one-way ANOVA tests was run to detect the
concentration Rorison solution was added once a week significance of the insulation treatment on each plant
in the colder months. Nutrient additions were added at treatment. A final set of ANOVA tests was run to detect
approximately midday throughout the duration of the the monthly significance of the insulation treatment. For
experiment. Nutrient additions were standardized so all aforementioned statistical tests, the significance level
that on a monthly basis each microcosm received was set at p < 0.05.
1042 C.R. Picard et al. / Bioresource Technology 96 (2005) 1039–1047

3. Results treatment in January (100.6 mg/l ± 4.2) and a minimum

value in July (24.7 mg/l ± 3.3) within the insulated
Mean data logger and ambient temperature readings microcosm group. Scirpus validus demonstrated a maxi-
are represented in Fig. 1. All temperature readings were mum and minimum TDN output within the exposed
highest in the month of July. These maximum tempera- treatment in January (87.1 mg/l ± 3.4) and August
ture readings were as follows: insulated = 23.0 C ± 0.1, (4.7 mg/l ± 0.5). Carex lacustris demonstrated a maxi-
exposed = 24.4 C ± 0.1 and ambient = 23.0 C ± 0.1. mum and minimum TDN output within the exposed
Temperature readings were the lowest from 1/19/03 to treatment in January (95.7 mg/l ± 1.8) and September
1/25/03. These minimum temperature readings were as (15.3 mg/l ± 3.6). Phalaris arundinacea demonstrated a
follows: insulated =
0.7 C ± 0.1, exposed =
8.9 C ± maximum and minimum TDN output within the ex-
0.1 and ambient =
11.1 C ± 0.1. There was little differ- posed treatment in February (88.7 mg/l ± 3.0) and July
ence between data logger readings within the tempera- (8.9 mg/l ± 2.3). Typha latifolia demonstrated a maxi-
ture treatments as demonstrated by the low standard mum and minimum TDN output within the exposed
errors. treatment in January (92.6 mg/l ± 1.4) and August
Monthly mean TDN readings for each plant (10.2 mg/l ± 2.7). The combination of plants demon-
treatment within the two insulation treatments are strated a maximum TDN value within the exposed treat-
represented in Fig. 2. The unplanted treatment demon- ment in January (75.7 mg/l ± 4.1) and a minimum value
strated a maximum TDN output within the exposed in July (9.8 mg/l ± 3.2) within the insulated microcosm
group. The results of a two-way repeated measures
ANOVA (Table 1) show that both the plant and insula-
tion treatments had a significant effect on TDN output
(p < 0.001). The interacting effect of the plant and insu-
lation treatments was also significant (p < 0.05). The ef-
fects of month, month and plant treatment interactions
and month and insulation treatment interactions were
all significant (p < 0.001).
Monthly mean TDP readings for each plant
treatment within the two insulation treatments are rep-
resented in Fig. 3. The unplanted treatment demon-
strated a maximum TDP output within the exposed
treatment in February (42.9 mg/l ± 2.1) and a minimum
value in July (8.4 mg/l ± 2.2) within the insulated micro-
cosm group. Scirpus validus demonstrated a maximum
and minimum TDP output within the exposed treatment
Fig. 1. Mean temperature readings from the microcosm data loggers in February (23.1 mg/l ± 1.9) and August (7.8 mg/
and the weather station. l ± 0.3). Carex lacustris demonstrated a maximum and
minimum TDP output within the exposed treatment in
February (37.2 mg/l ± 2.2) and September (4.5 mg/
l ± 0.8). Phalaris arundinacea demonstrated a maximum
and minimum TDP output within the exposed treatment
in February (24.6 mg/l ± 0.6) and July (4.9 mg/l ± 1.9).

Table 1
Two-way repeated measures ANOVA for total dissolved nitrogen
Source SS DF F-value P-value
Between subjects
Plant 141,167.241 5 115.988 0.001
Insulation 11,405.891 1 46.858 0.001
Plant · insulation 3327.458 5 2.734 0.027
Error 14,604.974 60
Within subjects
Month 560,457.878 11 420.326 0.001
Month · plant 45,526.841 55 6.829 0.001
Month · insulation 13,745.929 11 10.309 0.001
Fig. 2. Monthly mean total dissolved nitrogen values for each plant Month · plant · insulation 7245.579 55 1.087 0.316
treatment within the insulation treatments. Error bars represent ±1 SE. Error 80,003.3630 660
 C.R. Picard et al. / Bioresource Technology 96 (2005) 1039–1047 1043

Table 3
One-way ANOVA analysis of total nitrogen data to determine effects
of the insulation treatment on each plant treatment
Plant Source SS DF F-value P-value
Unplanted Insulation 252.082 1 0.258 0.613
Error 138,914.022 142
Scirpus validus Insulation 1351.628 1 1.597 0.208
Error 120,217.370 142
Carex lacustris Insulation 148.450 1 0.145 0.704
Error 145,345.465 142
Phalaris Insulation 5242.074 1 6.495 0.012
arundinacea Error 114,611.712 142
Typha latifolia Insulation 3678.887 1 4.394 0.038
Error 118,891.002 142
Combination Insulation 4002.239 1 6.798 0.010
Error 83,601.697 142

Fig. 3. Monthly mean total dissolved phosphorous values for each

plant treatment within the insulation treatments. Error bars represent Table 4
±1 SE. One-way ANOVA analysis of total phosphorous data to determine
effects of the insulation treatment on each plant treatment
Plant Source SS DF F-value P-value
Typha latifolia demonstrated a maximum and minimum
Unplanted Insulation 426.433 1 3.027 0.084
TDP output within the exposed treatment in February Error 20,004.755 142
(22.5 mg/l ± 2.2) and September (2.6 mg/l ± 0.4). The Scirpus validus Insulation 212.253 1 4.480 0.036
combination of plants demonstrated a maximum and Error 6728.052 142
minimum TDP value within the exposed treatment in Carex lacustris Insulation 2.268 1 0.022 0.883
February (18.6 mg/l ± 1.8) and August (3.5 mg/l ± 1.0). Error 14,896.370 142
Phalaris Insulation 392.278 1 10.116 0.002
The results of a two-way repeated measures ANOVA arundinacea Error 5506.212
(Table 2) show that both the plant and insulation treat- Typha latifolia Insulation 637.550 1 14.233 0.001
ments had a significant effect on TDP output Error 6360.899 142
(p < 0.001). The effects of month, month and plant treat- Combination Insulation 224.328 1 8.926 0.003
ment interactions and month and insulation treatment Error 3568.545
interactions were all significant (p < 0.001). The com-
bined interaction of month, plant and insulation was
also significant (p < 0.05). microcosms for Scirpus validus (p < 0.05), Phalaris arun-
Throughout the study period, the TDN output of dinacea (p < 0.05), Typha latifolia (p < 0.001) and the
insulated microcosms was significantly less than that plant combination (p < 0.05) (Table 4). For all plant
of exposed microcosms for Phalaris arundinacea treatments, the insulated microcosms exhibited signifi-
(p < 0.05), Typha latifolia (p < 0.05) and the plant com- cantly less TDN output than exposed microcosms for
bination (p < 0.05) (Table 3). Similarly, the TDP output the following months: December (p < 0.05), January
of insulated microcosms was less than that of exposed (p < 0.001), April (p < 0.05) and May (p < 0.001) (Table
5). For all plant treatments, the insulated microcosms
exhibited significantly less TDP output than exposed
microcosms for the following months: June (p < 0.05),
Table 2 December (p < 0.05), January (p < 0.001), February
Two-way repeated measures ANOVA for total dissolved phosphorous
(p < 0.05), March (p < 0.05), April (p < 0.05) and May
Source SS DF F-value P-value (p < 0.001) (Table 6).
Between subjects
Plant 24,218.410 5 96.941 0.001
Insulation 1569.975 1 31.421 0.001
Plant · insulation 356.052 5 1.425 0.228
4. Discussion
Error 2997.907 60
The objective of this study was to examine the effects
Within subjects
Month 27,480.241 11 119.057 0.001
of season on the nutrient removal efficiency of various
Month · plant 8017.503 55 6.947 0.001 types of plant communities using microcosms. The
Month · insulation 2958.042 11 12.816 0.001 experiment also set out to investigate if there are spe-
Month · plant · insulation 1692.846 55 1.467 0.018 cies-specific rates of nutrient removal and to explore
Error 13,848.940 660 the use of insulation for the increase of nutrient
1044 C.R. Picard et al. / Bioresource Technology 96 (2005) 1039–1047

Table 5 amount of nutrients. Also, insulated microcosms were

One-way ANOVA analysis of total nitrogen data to determine the predicted to be more effective in removing nutrients than
monthly effects of the insulation treatment
exposed microcosms.
Month Source SS DF F-value P-value During the year long study, all microcosms removed
June Insulation 199.341 1 1.634 0.205 large amounts of both nitrogen and phosphorous com-
Error 8541.470 70 pared to large-scale natural and treatment wetlands.
July Insulation 10.467 1 0.054 0.817
Error 13,508.004 70
These high rates of nutrient removal may be due to
August Insulation 28.743 1 0.070 0.792 the small treatment area of the microcosms (Tanner
Error 28,609.021 70 et al., 1995). The microcosms exhibited a definite pattern
September Insulation 136.891 1 0.440 0.509 in seasonal nutrient removal. In general, nutrient
Error 21,791.870 70 removal was highest during the growing season (June–
October Insulation 43.264 1 0.216 0.644
Error 14,030.095 70
October) and lowest in the cold months (November–
November Insulation 1979.463 1 2.644 0.108 March). Nutrient removal then increased again in April
Error 52,405.144 70 and May and was associated with an increase in temper-
December Insulation 2622.232 1 8.203 0.006 ature and plant growth. This pattern of nutrient removal
Error 22,376.118 70 is similar to that of many natural and treatment wet-
January Insulation 3863.952 1 16.502 0.001
Error 16,390.498 70
lands (Mitsch and Gosselink, 1993; Spieles and Mitsch,
February Insulation 43.984 1 0.359 0.551 2000; Tanner, 2001b). There was less variance in sea-
Error 8571.791 70 sonal phosphorous removal when compared to nitrogen.
March Insulation 437.335 1 0.996 0.322 This difference in variance is demonstrated in many
Error 30,731.821 70 treatment wetlands and may be due to year-round sedi-
April Insulation 7902.633 1 10.690 0.002
Error 51,748.989 70
mentary binding of phosphorous (Kadlec and Knight,
May Insulation 7883.515 1 23.817 0.001 1996; Wittgren and Maehlum, 1997).
Error 23,170.636 70 Planted microcosms almost always outperformed
unplanted microcosms in removing nutrients. This has
been observed in various microcosm and full-scale
Table 6 experiments (Hunter et al., 2001; Juwarkar et al., 1995;
One-way ANOVA analysis of total phosphorous data to determine the Zhu and Sikora, 1995; Fraser et al., 2004). This pattern
monthly effects of the insulation treatment
was also demonstrated in a natural wetland comparing
Month Source SS DF F-value P-value nutrient retention in unvegetated patches and patches
June Insulation 180.114 1 4.184 0.045 containing Juncus effusus (Mann and Wetzel, 2000). In
Error 3013.372 70 general, planted microcosms were also less vulnerable
July Insulation 96.216 1 2.592 0.112
to monthly fluctuations in nutrient removal efficiency.
Error 2598.307 70
August Insulation 92.725 1 2.971 0.089 The elevated nutrient uptake in planted systems may
Error 2184.923 70 be directly due to plant processes and more stable
September Insulation 11.849 1 0.449 0.505 year-round temperatures in planted systems (Hill and
Error 1847.646 70 Payton, 2000). The direct uptake of nutrients by macro-
October Insulation 30.695 1 0.622 0.433
phytes has been shown to be significant in various exper-
Error 3452.818 70
November Insulation 1.448 1 0.013 0.909 iments. For example, a microcosm experiment
Error 7647.915 70 demonstrated that uptake from Scirpus validus ac-
December Insulation 986.827 1 12.828 0.001 counted for over 90% of nitrogen removal and over
Error 5384.972 70 74% of phosphorous removal (Hunter et al., 2001). Sim-
January Insulation 976.739 1 14.635 0.001
ilar experiments concluded that aquatic plants were
Error 4671.755 70
February Insulation 674.571 1 6.191 0.015 responsible for approximately 90% of nitrogen removal
Error 7627.172 70 (Rogers et al., 1991).
March Insulation 538.976 1 4.805 0.032 Unplanted microcosms also consistently removed
Error 7851.257 70 large concentrations of nutrients. Some unplanted
April Insulation 720.430 1 12.875 0.001
microcosms were colonized by duckweed (Lemna
Error 3916.993 70
May Insulation 187.669 1 15.849 0.001 spp.), which have been shown to remove nutrients from
Error 828.885 70 wastewater in Czechoslovakian treatment wetlands
(Vymazal, 2002). The unplanted microcosms that re-
mained free of duckweed also reduced nutrient levels
removal. It was hypothesized that all microcosms would to a great extent. This suggests that microbial processes
experience a decrease in nutrient removal during the are the major pathways of nutrient removal in all the
winter months as compared to the growing season. Plant microcosms and plants provided supplemental nutrient
polycultures were predicted to remove the greatest removal. There is also a great deal of research support-
 C.R. Picard et al. / Bioresource Technology 96 (2005) 1039–1047 1045

ing microbial processes as the impetus behind nutrient in this experiment. Consequently, these characteristics
removal in treatment wetlands. The monitoring of four could have lead to the lower abundance or diversity of
wetlands used to treat dairy wastewater demonstrated soil microbes when compared to the other plant commu-
that plant nutrient uptake was variable with nitrogen nities. These results are different than those of a similar
and phosphorous removal ranging from 3–19% to 3– experiment involving the same wetland plant species.
60% respectively (Tanner et al., 1995). In addition, The previous experiment found that Carex lacustris
experiments on full-scale wetlands valued direct uptake community removed nutrients at an intermediate level,
by plants as only 7% for nitrogen and 8% for phospho- while the Phalaris arundinacea community was the least
rous (Tanner, 2001a). It is assumed that plant-mediated efficient of the plant treatments (Fraser et al., 2004). The
microbial processes were the major pathways of nutrient differences in results may be due to the relative matura-
removal in these systems. In this experiment, macro- tion of the microcosm communities. Fraser et al. (2004)
phytes may be removing some nutrients through direct experimented on microcosms that had established for
uptake and providing an environment for more intense one year, whereas our experiment was conducted on
microbial activity. Microbial communities are often 2–3 year-old planted microcosms. Studies demonstrate
more numerous and diverse in planted treatment wet- that both treatment and natural wetlands often are more
lands as compared to wetlands with little or no vegeta- efficient at removing nutrients after the wetland commu-
tion (Ottova et al., 1997; Werker et al., 2002). This nity is allowed to develop (Biesboer, 1984; Maehlum
could be due to the diverse microhabitats within the soil and Stalnacke, 1999; Werker et al., 2002). An important
structure created by plant root growth. Plants may also community characteristic that is directly related to nutri-
be supplying carbon to the microbes, which allows these ent removal is the development of the root zone (Wood
microbes to be more successful in surviving and remov- et al., 1999). The root zone of the microcosms contain-
ing nutrients (Lin et al., 2002). ing Phalaris arundinacea, a plant with high amounts of
Contrary to some studies, there was no clearly de- belowground biomass, may have developed greatly over
fined pattern of species-specific nutrient removal found the past year. In relation to this concept, the roots of
in this experiment (Bachand and Home, 2000; Gersberg Phalaris arundinacea have been known to release high
et al., 1986; Lin et al., 2002). This specificity is often due concentrations of organic carbon that facilitates micro-
to nutrient needs, plant physiology and plant morphol- bial nutrient removal pathways (Zhu and Sikora, 1995).
ogy. Scirpus validus, Phalaris arundinacea, Typha latifo- Temperature and the use of insulation was the other
lia and the plant combination communities all exhibited variable being investigated in this experiment. The insu-
near equal nutrient removal efficiency for most of the lation treatment in this experiment had a varied effect on
year. In other treatment wetland research, Typha spp. nutrient removal. Overall, insulation had no effect on
removed greater amounts of nutrients as compared to nutrient removal during the growing season. However,
Scirpus spp. (Coleman et al., 2001). Similar research insulated microcosms were often significantly more
showed Scirpus spp. to outperform Typha spp. in nutri- effective in the winter months, as well as the early spring
ent removal efficiency (Bachand and Home, 2000; Gers- months. During the warmer months both sets of micro-
berg et al., 1986). Phalaris spp. was shown to be of cosms must have experienced the same rates of heat
greater or equal nutrient removal efficiency when com- absorption. Insulated microcosms were probably able
pared to Typha spp. and Scirpus spp. depending on to better retain this heat in the winter. With respect to
the type of nitrogen species (Zhu and Sikora, 1995). plant treatment, insulation only affected planted micro-
These results exhibit the complexity of wetlands systems, cosms. As demonstrated by the data logger information,
as well as research involving these systems. soil temperature of insulated microcosms was only
It has been hypothesized that plant polycultures out- greater than that of exposed microcosms in the winter
perform plant monocultures in removing nutrients (Cole- months. Hence, this difference may have lead to higher
man et al., 2001; Kadlec and Knight, 1996; Karpiscak amounts of nutrient removal during the winter. The
et al., 1996). In this experiment, the polycultures and insulation may also have protected the belowground
their associated communities tended to outperform the biomass of the macrophytes allowing for more vigorous
other plant treatments in certain months. The growing growth in the early spring. This could be the reason for
season of wetland plants is species-specific with some differences in nutrient removal between the insulation
species having longer growing seasons than other spe- treatments during early spring.
cies. This phenomenon supports the theory of using Nutrient removal in treatment wetlands has been
polycultures to maximize nutrient removal (Scholes shown to be temperature dependent. Hence, nutrient
et al., 1999). removal is often a primary factor when designing cold-
The Carex lacustris community was the least efficient climate treatment wetlands (Werker et al., 2002). Low
of the plant treatments. This may be due to Carex lacus- technology methods of combating cold-climate effects
tris having lower amounts of below ground biomass or a include construction of subsurface flow wetlands and
different root morphology than the other plants utilized the allowance of high retention times (Maehlum and
1046 C.R. Picard et al. / Bioresource Technology 96 (2005) 1039–1047

Stalnacke, 1999; Werker et al., 2002). In addition, cold Finally, we acknowledge all past and present members
climate wetlands may also operate successfully due to of L FraserÕs lab for their aid in the field.
insulation by a natural vegetation layer, snow cover or
a thin ice layer (Smith et al., 1997a,b; Werker et al.,
2002; Wittgren and Maehlum, 1997). Incoming waste- References
water may also provide a heat source for many treat-
ment wetlands (Kadlec, 2001). More complex methods Bachand, P.M., Home, A.J., 2000. Denitrification in constructed free-
of insulation have also been shown to allow for year- water surface wetlands: effects of vegetation and temperature.
Ecological Engineering 14 (1–2), 17–32.
round operation of cold-climate wetlands. These meth- Biesboer, D.D., 1984. Nitrogen fixation associated with natural and
ods include the enclosure of treatment wetland systems, cultivated stands of Typha latifolia (Typhaceae). American Journal
as well as the addition of industrial byproducts to the of Botany 71 (4), 505–511.
wetland substrate (Merlin et al., 2002; Peterson and Brix, H., 1997. Do macrophytes play a role in constructed treatment
wetlands? Water Science and Technology 35 (5), 11–17.
Teal, 1996).
Carpenter, S.R., Caraco, N.F., Correll, D.F., Howarth, R.W.,
Sharpley, A.N., Smith, V.H., 1998. Nonpoint pollution of surface
waters with nitrogen and phosphorous. Ecological Applications 8
5. Conclusion (3), 559–568.
Coleman, J., Hench, K., Garbutt, K., Sextone, A., Bissonnette, G.,
This experiment demonstrated that there is a distinct Skonsen, J., 2001. Treatment of domestic wastewater by three plant
species in constructed wetlands. Water, Air and Soil Pollution 128,
seasonal pattern of nutrient removal within treatment 283–295.
wetland microcosms. Microcosms removed the greatest Fraser, L.H., Bradford, M.E., Steer, D.N., 2003. Global supply of
amount of nutrients during the growing season and lesser freshwater: the role of treatment wetlands. International Journal of
amounts in the colder months. However, seasonal differ- the Environment and Sustainable Development 2 (2), 174–
ences in nutrient removal can be minimized through the 183.
Fraser, L.H., Carty, S.M., Steer, D.N., 2004. A test of four plant
use of plant growth and insulation. Planted microcosms species to reduce total nitrogen and total phosphorus from soil
outperformed unplanted microcosms stressing the leachate in subsurface wetland microcosms. Bioresearch Technol-
importance of macrophytes in treatment wetland opera- ogy 94 (2), 185–192.
tion. Plants play a supplemental role in nutrient cycling Gersberg, R.M., Elkins, B.V., Lyon, S.R., Goldman, C.R., 1986. Role
to microbial processes, which seem to be the major path- of aquatic plants in wastewater treatment by artificial wetlands.
Water Resources 20 (3), 363–368.
way of nutrient removal in these microcosms. Within Hendry, G.F., Grime, J.P., 1993. Methods In Comparative Plant
planted microcosms, the Carex lacustris community per- Ecology: A Lab Manual. Chapman and Hall, London.
formed the poorest. Hence, monocultures of Carex Hill, D.T., Payton, J.D., 2000. Effect of plant fill ratio on water
lacustris and other sedges may not be the optimal plants temperature in constructed wetlands. Bioresource Technology 71
to utilize in treatment wetlands. More reliable results (3), 283–289.
Hunter, R.G., Combs, D.L., George, D.B., 2001. Nitrogen, phospho-
may have been attained with temperature and insulation rous and organic carbon removal in simulated wetland treatment
effects if this experiment was performed in a green house systems. Archives of Environmental Contamination and Toxicol-
or climate controlled structure. This study also focused ogy 41, 274–281.
on only a few variables that may affect treatment wet- Juwarkar, A.S., Oke, B., Juwarkur, A., Patnaik, S.M., 1995. Domestic
wastewater treatment through constructed wetlands in India.
lands. Studies of other variables and interactions are
Water Science and Technology 32 (3), 291–294.
needed to better understand these systems. These types Kadlec, R.H., 1999. Chemical, physical and biological cycles in
of microcosm studies should be an impetus for further treatment wetlands. Water Science and Technology 40 (3), 37–44.
large-scale experiments that deal with wetland design Kadlec, R.H., 2001. Thermal environments of subsurface treatment
and function. The coupling of these two research strate- wetlands. Water Science and Technology 44 (11), 251–258.
gies could lead to better management of these systems. Kadlec, R.H., Knight, R.L., 1996. Treatment Wetlands. Lewis
Publishers, Boca Raton.
Consequently, these wetlands could be operated more Karpiscak, M.M., Gerba, C.P., Watt, P.M., Foster, K.E., Falabi, J.A.,
successfully in temperate and cold climate regions. 1996. Multi species plant systems for wastewater quality improve-
ments and habitat enhancement. Water Science and Technology 33
(10–11), 231–236.
Acknowledgements Lin, Y.F., Jing, S.R., Wang, T.W., Lee, D.Y., 2002. Effects of
macrophytes and external carbon sources on nitrate removal from
groundwater in constructed wetlands. Environmental Pollution
The authors acknowledge the Ohio Environmental 119, 413–420.
Protection Agency and the United States Environmental Maehlum, T., Stalnacke, P., 1999. Removal efficiency of three cold-
Protection Agency, under the provisions of Section climate constructed wetlands treating domestic wastewater: effects
319(h) of the Clean Water Act for funding the majority of temperature, season, loading rates and input concentrations.
Water Science and Technology 40 (3), 273–281.
of this research. We thank Bath Township for providing Mann, C.J., Wetzel, R.G., 2000. Effects of emergent macrophyte
the space for this research. Also, thanks to Dr. Peter Juncus effusus L. on the chemical composition of interstitial water
Lavrentyev for his helpful comments on this experiment. and bacterial productivity. Biogeochemistry 48, 307–322.
 C.R. Picard et al. / Bioresource Technology 96 (2005) 1039–1047 1047

Merlin, G., Pajean, J.L., Lissolo, T., 2002. Performances of con- a comparison of low and high nutrient riverine systems. Ecological
structed wetlands for municipal wastewater treatment in a rural Engineering 14 (1–2), 77–91.
mountainous area. Hydrobiologia 469, 87–98. SPSS., 1998. Systat 8.0 for Windows. Prentice Hall, Chicago.
Mitsch, W.J., Gosselink, J.G., 1993. Wetlands, second ed. John Wiley Steer, D., Fraser, L.H., Boddy, J., Seibert, B., 2002. Efficiency of small
and Sons Inc., New York. constructed wetlands for subsurface treatment of single family
Mitsch, W.J., Day, J.W., Gilliam, J.W., Groffman, P.M., Hey, D.L., domestic effluent. Ecological Engineering 18, 429–440.
Randall, G.W., Wang, N., 2001. Reducing nitrogen loading to the Tanner, C.C., 2001a. Growth and nutrient dynamics of soft-stem
Gulf of Mexico from the Mississippi River Basin: Strategies to bulrush in constructed wetlands treating nutrient-rich wastewaters.
counter a persistent ecological problem. Bioscience 51 (5), 373–388. Wetlands Ecology and Management 9, 49–73.
Ottova, V., Balcarova, J., Vymazal, J., 1997. Microbial characteristics Tanner, C.C., 2001b. Plants as ecosystem engineers in subsurface flow
of constructed wetlands. Water Science and Technology 35 (5), treatment wetlands. Water Science and Technology 44 (11–12), 9–
117–123. 17.
Peterson, S.B., Teal, J.M., 1996. The role of plants in ecologically Tanner, C.C., Clayton, J.S., Upsdell, M.P., 1995. Effect of loading rate
engineered wastewater treatment systems. Ecological Engineering 6 and planting on treatment of dairy farm wastewater in constructed
(1–3), 137–148. wetlands—II. Removal of nitrogen and phosphorous. Water
Rogers, K.H., Breen, P.H., Chick, A.J., 1991. Nitrogen removal in Research 29 (1), 27–34.
experimental wetland treatment systems: evidence for the role of Verhoeven, J.A., Meuleman, A.M., 1999. Wetlands for wastewater
aquatic plants. Research Journal of Water Pollution Control treatment: opportunities and limitations. Ecological Engineering 12
Federation 63 (7), 934–941. (1–2), 5–12.
Reckhow, K.H., Simpson, J.T., 1980. A procedure using modeling and Vitousek, P.M., Aber, J.D., Howarth, R.W., Likens, G.E., Matson,
error analysis for the prediction of lake phosphorus concentration P.A., Schindler, D.W., Schlesinger, W.H., Tilman, G.D., 1997.
from land use information. Canadian Journal of Fisheries and Human alteration of the global nitrogen cycle: Sources and
Aquatic Science 37, 1439–1448. consequences. Ecological Applications 7, 737–750.
Scholes, L.L., Schutes, R.E., Revitt, D.M., Purchase, D., Forshaw, M., Vymazal, J., 2002. The use of subsurface constructed wetlands for
1999. The removal of urban pollutants by constructed wetlands wastewater treatment in the Czech Republic: 10 years experience.
during wet weather. Water Science and Technology 40 (3), 333– Ecological Engineering 18 (5), 633–646.
340. Werker, A.G., Dougherty, J.M., McHenry, J.L., Van Loon, W.A.,
Schutes, R.E., 2001. Artificial wetlands and water quality improve- 2002. Treatment variability for wetland wastewater treatment
ment. Environmental International 26 (5–6), 441–447. design in cold climates. Ecological Engineering 19 (1), 1–11.
Smith, I.D., Bis, G.N., Lemon, E.R., Rozema, L.R., 1997a. A thermal Wittgren, H.B., Maehlum, T., 1997. Wastewater treatment wetlands in
analysis of a subsurface, vertical flow constructed wetland. Water cold climates. Water Science and Technology 35 (5), 45–53.
Science and Technology 35 (5), 55–62. Wood, S.L., Wheeler, E.F., Berghage, R.D., Graves, R.E., 1999.
Smith, R.A., Schwartz, G.E., Alexander, R.B., 1997b. Regional Temperature effects on wastewater nitrate removal in laboratory-
interpretation of water quality monitoring data. Water Resources scaled constructed wetlands. Transactions of the ASAE 42 (1),
Research 33 (12), 2781–2798. 185–190.
Smith, V.H., Tilman, G.D., Nekola, J.C., 1999. Eutrophication: Yin, H., Shen, W., 1995. Using reed beds for winter operation of
Impacts of excess nutrient inputs on freshwater, marine, and wetland treatment systems for wastewater. Water Science and
terrestrial ecosystems. Environmental Pollution 100 (1–3), 179– Technology 32 (3), 111–117.
196. Zhu, T., Sikora, F.J., 1995. Ammonia and nitrate removal in vegetated
Spieles, D.J., Mitsch, W.J., 2000. The effects of season and hydrologic and unvegetated gravel bed microcosm wetlands. Water Science
and chemical loading on nitrate retention in constructed wetlands: and Technology 32 (3), 219–228.