Removal and Distribution of Iron, Manganese, Cobalt, and Nickel

within a Pennsylvania Constructed Wetland Treating Coal

Combustion By-Product Leachate
Z. H. Ye, S. N. Whiting, Z.-Q. Lin, C. M. Lytle, J. H. Qian, and N. Terry*
ABSTRACT

Constructed wetland treatment systems offer a practical alternative to the chemical treatment of drainage
from certain coal-related sources (Brodie et al., 1989;
Nu-Hoai et al., 1998). The advantages include low costs
of construction, operation, and maintenance, with effective removal of many contaminants such as heavy metals
and metalloids (National Rivers Authority, 1992; NuHoai et al., 1998). In addition, the systems are usually
better able to cope with fluctuating water flow rate and
variable contaminant concentrations than conventional
treatment systems (Bastian and Hammer, 1993). The
success of wetlands has resulted in the construction of
more than 400 wetlands in the USA for the purpose of
treating drainage water from coal mines alone (Perry
and Kleinmann, 1991).
Wetlands are capable of removing large quantities of
trace elements from wastewater passing through them.
However, there is considerable variation among metals,
and also between wetlands, in the degree to which each
metal is removed by the wetland vegetation. For example, in a review of constructed wetlands in the USA,
Skousen et al. (1994) reported that the removal of iron
(Fe) from acid mine drainage varied widely between 28
and 99%. In the case of manganese (Mn), where the
range of removal rates was between 8 and 98%, the
mean removal rate was actually less than 30%. The
mechanisms of removal for each contaminant from
wastewater, and the interrelatedness of these different
mechanisms, are not well understood (Faulkner and
Richardson, 1989; National Rivers Authority, 1992). It
is thought that the cleanup of water is accomplished
through a variety of physical, chemical, and biological
processes operating independently in some situations,
and interactively in others (Dunbabin and Bowmer,
1992; Hammer, 1993).
Physicochemical mechanisms of metal retention in
the sediments include immobilization via oxidative precipitation, sedimentation of metal-laden particulates,
adsorption on cation exchange sites, complexation with
organic matter, and sulfate reduction (Wieder, 1988;
Perry and Kleinmann, 1991; Brodie et al., 1993). Precipitation of metal oxides, following microbially mediated
oxidation, is thought to be one of the most important
removal mechanisms in wetlands (National Rivers Authority, 1992; Skousen et al., 1994). Indeed, 40 to 70%
of the total iron removed from acid mine drainage by
some wetlands was found as ferric hydroxides from the
hydrolysis of ferric iron or from microbial oxidation of
ferrous iron (Henrot and Wieder, 1990).

A flow-through wetland treatment system was constructed to treat

coal combustion by-product leachate from an electrical power station
at Springdale, Pennsylvania. In a nine-compartment treatment system,
four cattail (Typha latifolia L.) wetland cells (designated Cells 1
through 4) successfully removed iron (Fe) and manganese (Mn) from
the inlet water; Fe and Mn concentrations were decreased by an
average of 91% in the first year (May 1996May 1997), and by 94
and 98% in the second year (July 1997June 1998), respectively.
Cobalt (Co) and nickel (Ni) were decreased by an average of 39 and
47% in the first year, and 98 and 63% in the second year, respectively.
Most of the metal removed by the wetland cells was accumulated in
sediments, which constituted the largest sink. Except for Fe, metal
concentrations in the sediments tended to be greater in the top 5 cm
of sediment than in the 5- to 10- or 10- to 15-cm layers, and in Cell
1 than in Cells 2, 3, and 4. Plants constituted a much smaller sink for
metals; only 0.91, 4.18, 0.19, and 0.38% of the Fe, Mn, Co, and Ni
were accumulated annually in the aboveground tissues of cattail, respectively. A greater proportion of each metal (except Mn) was accumulated in cattail fallen litter and submerged Chara (a macroalga)
tissues, that is, 2.81, 2.75, and 1.05% for Fe, Co, and Ni, respectively.
Considerably higher concentrations of metals were associated with
cattail roots than shoots, although Mn was a notable exception.

oal-burning electric power generation facilities

can encounter compliance difficulties when disposing of metal-contaminated wastewater discharges (Electric Power Research Institute, 1998). Such discharges
are produced during the storage and processing of coal,
and the storage and land disposal of coal ash residues
produced during the combustion process. These discharges contain a variety of contaminants including metals, sulfur, and acid (low pH) (Kadlec and Knight, 1996),
which seriously degrade water quality and biodiversity
when released into streams and rivers. To comply with
the National Pollution Discharge Elimination System
(NPDES) effluent limitations, the contaminant load of
the wastewater must be reduced. Conventional techniques of controlling metal-contaminated water (for example, chemical treatment) are costly in terms of equipment and chemicals. Moreover, they require intensive
management and long-term maintenance (Brodie et al.,
1993; Skousen et al., 1994). Estimated costs for the
chemical treatment of waste solids and water produced
by the U.S. coal industry exceed $1 million per day
(Perry and Kleinmann, 1991).

Department of Plant and Microbial Biology, Univ. of California at

Berkeley, 111 Koshland Hall, Berkeley, CA 94720. Z.H. Ye, present
address: Institute for Resource and Environmental Management,
Hong Kong Baptist University, Hong Kong. Received 31 May 2000.
*Corresponding author (nterry@nature.berkeley.edu).

Abbreviations: APS, Allegheny Power Services; EC, electrical conductivity; DO, dissolved oxygen; NPDES, National Pollution Discharge Elimination System.

Published in J. Environ. Qual. 30:14641473 (2001).

1464

YE ET AL.: REMOVAL OF Fe, Mn, Co, AND Ni BY A CONSTRUCTED WETLAND

Plants and microbes are important in many different

ways for contaminant retention by wetlands. However,
the direct uptake of trace elements into the shoots of
plants represents only a minor proportion of the total
removed by the wetland; for example, the bioaccumulation of Fe by plants, even those with a large biomass
such as cattail, removed less than 1% of the Fe input to
some constructed wetlands (Sencindiver and Bhumbla,
1988; Faulkner and Richardson, 1989; Mitsch and Wise,
1998; Whiting and Terry, 1999). Nevertheless, plants
play a critical role in metal removal via filtration, adsorption, and cation exchange, and through plant-induced
chemical changes in the rhizosphere (Dunbabin and
Bowmer, 1992). Plants also provide habitat and energy
sources to maintain and stimulate a diverse microbial
population in the sediments (National Rivers Authority,
1992; Brodie, 1993; Skousen et al., 1994). These microbes drive the immobilization of contaminants in the
sediments through both oxidative and reductive processes (Johnson, 1998).
In 1995, the Electric Power Research Institute (EPRI),
in collaboration with Allegheny Power Services (APS),
constructed a wetland treatment system to treat the
leachate from a coal ash pile near the Springdale Electrical Power Station, Springdale, PA. The treatment
system included an equalization basin, four vegetated
wetland cells, two rock drains, an anaerobic (organic
reduction) cell, and an algal cell (Electric Power Research Institute, 1998). The project was implemented
to test the efficiency and economic feasibility of constructed wetlands for the treatment of coal combustion
by-product leachate. The treatment system was de-

1465

signed particularly to achieve regulatory compliance for

different trace metal contaminants (e.g., with the
NPDES limits of 1.6 mg L1 for Fe and 1.1 mg L1 for
Mn) before the leachate was discharged to a tributary
of the Allegheny River. The present study focuses on
the four vegetated wetland cells to determine (i) how
effectively the wetland cells removed iron (Fe), manganese (Mn), cobalt (Co), and nickel (Ni) from coal combustion by-product leachate and (ii) the compartmentalization of these four elements in different sinks (e.g.,
sediments, plant tissues, etc.) within the vegetated wetland cells.

MATERIALS AND METHODS

Study Site
The Springdale coal combustion by-product (mainly fly and
bottom ashes) disposal site was established in the 1950s near
the Allegheny Power Service in western Pennsylvania. As a
valley-fill operation, the disposal area (approximately 20 ha)
was subsequently taken off active status and revegetated in
the 1970s. To treat metal-contaminated leachate discharged
from the site underdrain, the Springdale surface-flow wetland
system was constructed in 1995. It comprises an aerobic equalization basin and eight lined cells, each of approximately 9
36 m (Fig. 1). The wastewater flowed sequentially from the
first treatment phase (equalization basin), through four fully
vegetated wetland cells, two rock drains, an anaerobic (organic
reduction) cell, and finally an algal uptake cell (Electric Power
Research Institute, 1998). An early study commissioned by
Electric Power Research Institute (1998) showed that the water influent to the equalization basin upstream of the wetland

Fig. 1. The Allegheny Power Services (APS) treatment system for coal combustion by-product leachate.

1466

J. ENVIRON. QUAL., VOL. 30, JULYAUGUST 2001

cells contained an average of 12.46 mg Fe L1, 2.71 mg Mn L1,

and 0.063 mg Ni L1 (Co was not measured). The oxygenated
equalization basin was effective for Fe removal (about 76%
of the total Fe in the leachate), but was less effective for Mn
(14%) and Ni (24%) (Electric Power Research Institute,
1998).
The present study focuses on metal removal by the four
vegetated wetland cells. These cells (Cells 14) are of approximately equal size, and have a total combined area of approximately 1300 m2. After construction, the cells were filled with
a blended compost substrate, comprising one part spent-mushroom compost to two parts native soil material. The native soil
was classified as Hazleton (loamy-skeletal, siliceous, subactive,
mesic Typic Dystrudrept). The cells were filled with 45 to 60
cm of this substrate. The wetland cells were planted with
commercially grown cattail rhizomes. Additional bulrush
(Scirpus spp.) and sedges (Carex spp.) were added as seed.
At maturity, cattail was by far the most dominant species,
and constituted the majority of the biomass produced in the
vegetated cells. A variety of invasive plant species, for example, arrowhead (Sagittaria latifolia Willd.), a native grass (Panicum dichotomiflorum Michx.), and smartweed (Polygonum
pensylvanicum L.), and also algae, for example, muskgrass
(Chara spp.) and Spirogyra spp. were found in the cells.
The wetland monitoring program by the University of California Berkeley began in May 1996 and continued to June
1998. Since the residence time of each vegetated wetland cell
was very short (i.e., about 5 h) (Electric Power Research
Institute, 1998), most of the other trace elements we studied,
such as boron (B), aluminum (Al), copper (Cu), or zinc (Zn),
were poorly removed by the vegetated wetland cells. In this
paper, we present data for Fe, Mn, Co, and Ni, that is, those
elements that were shown to be effectively removed from the
wastewater by the vegetated wetland cells.

Water Sampling
Influent and effluent waters were sampled quarterly from
May 1996 through May 1997, and then monthly from July
1997 to June 1998. The pH, temperature, electrical conductivity (EC), and dissolved oxygen (DO) content of the surface
water of each cell were measured in situ at the wetland using
a Corning Checkmate Modular System (Fisher Scientific,
Pittsburgh, PA). Water samples for trace element analysis
were collected from the inlet of Cell 1 and from the outlet of
Cell 4 of the vegetated cells. In March, April, and June 1998,
water samples were also collected from the outlet of Cells 1,
2, and 3. In each case, three replicate 250-mL grab samples
were collected using the bottle submersion method (Byrnes,
1994), and acidified immediately with 1 mL of concentrated
nitric acid. All samples were transported to the laboratory at
4C for chemical analysis.
Samples of pore water in the 0- to 10-cm sediment layer
were collected at the same time as the influent and effluent
waters. Pore water was collected from the same quadrats as
the vegetation sample, using 0.1-m Rhizon soil moisture filters (Ben Meadows, Canton, GA) connected to 20-mL Vacutainer tubes (Becton Dickinson, Franklin Lakes, NJ), which
were allowed to fill to capacity. This method provided a sample
from the composite 0- to 10-cm sediment profile. The sample
was transported to the laboratory at 4C for chemical analysis.

Plant Sampling
Plant samples were collected on a monthly basis from May
1996 through June 1998. Plant species composition and plant
species density were sampled in three randomly selected 1-m2
quadrats in each vegetated cell. Three replicate whole-plant

samples (shoot and root) were collected from the quadrats,

to give estimates of total plant biomass. Samples of algae
were also collected whenever available. All plant samples were
placed in polyethylene Ziploc bags and transported to the
laboratory at 4C. Plant samples were washed with tap water
and rinsed with distilled water. Shoot and root tissues were
separated, and the samples were dried at 70C for 1 wk. The
dried tissues were weighed and ground to pass a 40-mesh
screen using a Wiley mill in preparation for chemical analysis.

Sediment Sampling
Sediments were sampled quarterly over the first year (May
1996 to May 1997), and an additional collection was made at
the end of the study (November 1998). Three replicate sediment samples were collected from the top 15 cm of sediment
in each cell using an AMS (American Falls, ID) sediment
recovery probe with replaceable 5- 25-cm butyrate liners.
The sediments were transported to the laboratory in the
capped liners at 4C. Sediments were then removed, cut into
three 5-cm profile sections, air-dried, and ground to a fine
powder.

Chemical Analysis
All samples were analyzed for total trace elements. Water
samples were digested with HNO3, H2O2, and HCl according
to USEPA Method 3010A (USEPA, 1992). Trace element
concentrations were determined by inductively coupled
plasma atomic emission spectrometry (ICPAES). Elemental
analysis of plants was carried out by acid digestion of dry
samples (Zarcinas et al., 1987), followed by measurement of
total concentrations of all elements of interest in the acid
digest using ICPAES. Sediment samples were acid digested
with HNO3, H2O2, and HCl using USEPA Method 3050B
(USEPA, 1996) before chemical analysis by ICPAES.

Statistical Analysis
Sample means and standard errors were determined from
the replicate water, plant, and sediment samples. The data
were analyzed statistically using one-way analysis of variance
(ANOVA); significant differences between individual means
were determined using the Tukey multiple comparisons test
(Zar, 1984).

RESULTS
Physical Characteristics of Surface Water
There were no significant differences (P 0.05) between the four cells in the pH, EC, DO, and temperature
of the surface water. The values are therefore averaged
over the four cells (Fig. 2). Over the 26-mo study period,
the pH of the surface water averaged 7.2, and ranged
from 6.2 to 7.7. The average EC was 2.2 mS cm1, ranging from 1.9 to 3.0 mS cm1. There was some seasonal
variation in the DO concentration in the surface water,
with concentrations ranging from 5 to 14 mg L1. The
lowest values were in the summer (July through September), and the highest in the winter. The water temperature varied greatly from 0 to 30C during a year. The
DO and temperature were inversely proportional; high
temperatures resulted in the lowest concentrations of
DO (Fig. 2). The lowest DO values were most likely
due to the lower solubility of oxygen in summer and to
the greatest microbial activities when water temperature
was high.

YE ET AL.: REMOVAL OF Fe, Mn, Co, AND Ni BY A CONSTRUCTED WETLAND

1467

Trace Element Concentrations in Inlet and

Outlet Waters of the Vegetated Cells
The four vegetated wetland cells removed significant
quantities of Fe, Mn, Co, and Ni from the wastewater;
the best removal rates were seen for Fe and Mn. Greater
reductions in the concentrations of the four elements
were seen in the second year (i.e., after February 1997)
compared with the first year. During the study period,
the wastewater entering Cell 1 from the equalization
basin contained as much as 4.69 mg Fe L1, 3.3 mg Mn
L1, 0.022 mg Co L1, and 0.055 mg Ni L1 (Fig. 3).
However, in the outlet water from Cell 4 the concentrations of Fe and Mn were low, ranging from 0.002 to
0.49 mg L1 and 0.0005 to 0.44 mg L1, respectively.
Concentrations of Co and Ni in the outlet from Cell 4
were also low compared with their concentrations in the
inlet water. The monthly reduction (%) in concentration
shows that, in some months, almost 100% of these elements had been stripped from the wastewater by the
vegetated wetland cells. When these percentages were
averaged for the first year of the study, the concentrations of Fe, Mn, Co, and Ni were reduced by 90.8, 91,
38.5, and 47.3%, respectively. In the second year, the
metal concentration was reduced by 94% for Fe, 98%
for Mn, 96% for Co, and by 62.6% for Ni. Particularly,
Cell 1 appeared to remove most of the elements from
the wastewater by comparison with the removal by the
other three cells (Fig. 4).

Trace Element Distribution in Vegetated

Wetland Cells
Accumulation in Sediments
The average concentrations of Fe and Mn accumulated in the top 15 cm of sediment in the four vegetated
cells were 32 300 and 1200 mg kg1, respectively. In
contrast, only 18 mg kg1 Co and 41 mg kg1 Ni were
in the top 15 cm of sediments. Higher concentrations
of all four metals were observed in the surface layers
(05 cm) of sediment compared with the lower layers
(510 and 1015 cm) in Cell 1 (Table 1). Because the
sediment was mixed soil that was deposited into each
wetland cell when the wetland was constructed, the
background metal distribution in sediments should be
generally uniform prior to the beginning of this study.
It is evident therefore that metals (Mn, Co, and Ni, less
for Fe) from the inlet were being deposited in the surface
layer of sediments. In Cells 2, 3 and 4, this stratification
was not seen; all three layers had similar concentrations
of each metal (Table 1).
The concentrations of Mn, Co, and Ni in the top
sediment were less in Cell 2 than in Cell 1, but changed
little with increasing distance from Cell 2 to Cell 4. There
was a significant increase in the metal concentration in
the sediment of the four vegetated cells during the first
year of operation (Fig. 5). The average concentrations
of Fe in the sediment samples (015 cm) were typically
more than threefold higher than the concentration of Fe
in the substrate from which the wetland was constructed
(e.g., 32 500 mg Fe kg1 in May 1996 compared with
9000 mg Fe kg1 in November 1995). However, during

Fig. 2. Temporal changes in the surface water pH, electrical conductivity (EC), dissolved oxygen (DO), and temperature. Values are
averaged over the four vegetated cells for each month. Error bars
indicate standard errors (SE). Dashed horizontal line denotes overall average.

the second study period (May 1996 through November

1998) there were no significant changes in the concentration of Fe in top 15 cm of sediment. The concentrations
of Mn, Co, and Ni in the sediment (015 cm) increased
with time from November 1995 to November 1997
(Fig. 5).
Sediment Pore Water
The average concentrations of the metals, except for
Mn, in sediment pore water during the period of May
1996 to June 1998 were similar to their concentrations
recorded in the inlet water (Fig. 3 and 6). The metal
concentration of the pore water varied among the four
vegetated cells. The concentrations of Mn, Co, and Ni
(but not Fe) in pore water tended to decrease with
increasing distance from Cell 1 through to Cell 4 (Fig. 6).
Accumulation in Cattail, Fallen Litter, and Chara
The average concentrations of Fe, Mn, Co, and Ni in
cattail shoots and belowground tissues varied considerably for each element during the study period (May
1996June 1998) (Table 2). The greatest metal concentration in the shoots of cattail was Mn, followed by Fe
Ni Co. In the roots, however, the highest concentration was seen for Fe, followed by Mn Ni Co. The

1468

J. ENVIRON. QUAL., VOL. 30, JULYAUGUST 2001

Fig. 3. Temporal changes in the concentration of Fe, Mn, Co, and Ni in the inlet water to Cell 1 (hatched bars), and the outlet water from Cell
4 (filled bars). Values are means plus SE, n 3.

concentration of Co in the shoots and roots of cattail

from Cell 1 were higher than those in the plants growing
in the other three cells. The Fe, Mn, Co, and Ni contents
of cattail fallen litter were much greater than for the
living shoots of cattail; in the case of Mn, Co, and Ni
they contained higher concentrations than even the belowground tissues (Tables 2 and 3). The metal concentrations in the fallen litter decreased from Cell 1 to
Cell 4.
The average concentrations of all four metals on or in
the living tissues of Chara were very high, considerably
higher than for both the above- and belowground tissues
of cattail (Tables 2 and 3). Indeed, the concentrations
of Co and Ni in Chara were about 238 and 173 times
higher than that in the aboveground tissues of cattail,
respectively. In addition, the concentrations of Fe, Mn,

Co, and Ni in Chara growing in Cell 1 were dramatically

higher than those of plants grown in Cell 4 (Table 3).

Loading and Retention of Trace Elements

in Vegetated Cells
When the difference between the total metal loading
at the inlet of Cell 1 and the unloading of metals at the
outlet of Cell 4 was calculated, it showed that most of
the metals had been retained within the cells (Table 4).
On a per unit area basis, approximately 68 g of Fe, 149 g
of Mn, 0.58 g of Co, and 1.5 g of Ni per m2 were retained
by the vegetated cells during the period of July 1997
through June 1998; these amounts account for 94.6, 98.8,
96.1, and 62.1% of the annual total Fe, Mn, Co, and Ni
mass input to the vegetated wetland cells, respectively.
Most of the metals measured were retained in the sedi-

YE ET AL.: REMOVAL OF Fe, Mn, Co, AND Ni BY A CONSTRUCTED WETLAND

1469

Fig. 4. Concentrations of Fe, Mn, Co, and Ni in the inlet water to Cell 1 and the outlet water from Cells 1, 2, 3, and 4 in 1998. Solid line denotes
the percentage reduction in concentration. Values are means plus SE, n 3.

ments; only small amounts were contained in cattail

shoots, fallen litter, and Chara (Table 4). Cattail shoots
contained much higher amounts of Mn (4.18% of the
total Mn removed by the four cells) than the other
three elements.

DISCUSSION
Metal Removal Efficiency of the Vegetated Cells
The water flowing into Cell 1 of the vegetated part
of the wetland contained Fe and Mn concentrations
above the 1996 NPDES permissible limits. Previous
studies showed that constructed wetlands are effective
in removing Fe from metal-polluted waters, but are less
successful in removing Mn (Stillings et al., 1988; Fennessy and Mitsch, 1989; Hedin, 1989; Skousen et al.,

1994; Stark et al., 1994). Hedin (1989) noted that only 2

sites of 41 studied with inflow Mn concentrations greater
than 2.0 mg L1 achieved compliance with the standard.
Our study demonstrates that the four vegetated wetland
cells successfully removed a large proportion of both
the Fe and Mn loads from the inlet water; the concentrations of Fe and Mn were decreased by an average of
91% in the first year (May 1996May 1997), and by 94
and 98% in the second year (July 1997June 1998),
respectively. The vegetated cells also decreased the concentrations of Co and Ni but to a lesser extent (i.e., by
an average of 38.5 and 47% in the first year, and 97.5
and 62.6% in the second year, respectively).
The solubility of Fe and Mn in water is governed
mainly by their oxidation state, which in turn is affected
by pH. Skousen et al. (1994) suggested that ferric hy-

1470

J. ENVIRON. QUAL., VOL. 30, JULYAUGUST 2001

Table 1. Metal concentrations (mg kg1 ) at different depths in the sediment profile. Data are means (SE, n 6) of samples collected
from all four cells between May 1996 and November 1998. Differences between depths were determined by analysis of variance
(ANOVA); letter codes indicate significant differences (Tukey multiple comparisons test, P 0.05).
Metal
Fe

Mn

Co

Ni

Depth
cm
05
510
1015
05
510
1015
05
510
1015
05
510
1015

Cell 1

Cell 2

Cell 3

Cell 4

mg kg1
35 200 1 900
31 300 1 500
31 300 1 100
3 310 1 080a
1 390 270b
1 080 80b
31 5a
18 1b
16 1b
59 11a
37 3b
36 3b

droxides precipitate as pH increases above 3.5, while

Mn hydroxides require a pH of at least 7 to precipitate.
In our study, the surface water in the wetland cells had
a circumneutral pH (average 7.2, range 6.27.7). This
neutral pH may therefore have been a significant factor
contributing to the effective removal of Mn by the wetland cells. In an investigation of 14 constructed wetlands,
Brodie (1993) highlighted the apparent correlation between Mn removal and the alkalinity and/or acidity of
the influent water; it was found that Mn loading into
the 14 wetlands ranged from 0.17 to 2.00 g m2 d1, and
that the Mn unloading ranged from 0.15 to 1.87 g m2
d1, corresponding to a 1 to 96% Mn removal. Of the

31 300
32 000
31 600
1 350
960
890
21
16
16
45
38
37

1 100
1 500
800
170
100
40
2
1
1
4
3
2

32 600 1 000
31 800 1 400
32 500 1 000
830 71
832 51
859 42
17 1
16 1
16 1
40 2
39 2
39 2

32 600
33 300
32 100
872
856
892
17
17
16
42
41
37

1 200
900
1 200
75
74
76
1
1
1
2.5
3.0
1.4

systems with low removal rates, all were associated with

low pH in the wastewater (2.9 to 3.9). Hedin and Nairn
(1993) also indicated that the alkalinity of mine water
is a major determinant in the kinetics of contaminant
removal by wetlands. Whereas Fe removal rates increased with increasing pH, the removal of Mn occurred
only under alkaline conditions. Therefore, the superiority of our vegetated wetland cells in removing Mn was
almost certainly associated with a high water pH value
(i.e., pH 7.2).
Metal removal by the APS wetland was sustained
throughout the whole year. This is in contrast to some
other studies that found distinct seasonal variation in

Fig. 5. Concentrations of Fe, Mn, Co, and Ni in the top 15 cm of sediment in four vegetated wetland cells. The gray bars show the concentration
of each element in the substrate before the water treatment began (data from Electric Power Research Institute, 1998). Values are monthly
averages of all four cells plus SE, n 4.

1471

YE ET AL.: REMOVAL OF Fe, Mn, Co, AND Ni BY A CONSTRUCTED WETLAND

Table 3. Metal concentrations (mg kg1 ) in the fallen litter of

cattail and in Chara, a macroalga. Fallen litter samples were
collected from the four cells, during the study period of November 1996 to March 1997, and November 1997 to March 1998;
Chara samples were collected from Cells 1 and 4 during the
period of May 1996 to February 1998. Values shown are means
and one standard error. Means with different letters indicate
significant differences (Tukey multiple comparison test, P
0.05).
Metal

Cell 1

Cell 2

Cell 3

Cell 4

mg kg1
Cattail fallen litter (n 10)

Fig. 6. Concentrations of Fe, Mn, Co, and Ni in sediment pore water

(010 cm depth) in each vegetated wetland cell (May 1996 to June
1998). Values are means plus SE, n 16.

removal rates. Faulkner and Skousen (1994) showed

that climate, season, and nutrient availability dramatically affect the amount of sulfide generation and subsequent metal reduction. Henrot and Wieder (1990) noted
that Fe removal efficiency by peat increases linearly
with temperature in the 4 to 25C range; they suggested
that the performance of constructed wetlands in treating
trace elementcontaminated water varies seasonally,
with a higher efficiency in the summer. The results of
our study, however, showed that the vegetated cells
decreased the concentrations of metals by similar
amounts throughout the entire year (July 1997 through
June 1998) (Fig. 3), despite low temperatures in winter.
This might be attributable to the relatively constant pH
value with season and to the higher dissolved oxygen
level in winter, both of which favor oxidation and the
formation of insoluble hydroxides of metal elements.
The fact that the removal rates at the APS wetland were
not linked to seasonal changes is very beneficial because
it ensures that discharges are in compliance with regulatory limits throughout the year.
Table 2. Metal concentrations (mg kg1 ) in the aboveground and
belowground tissues of cattail collected from the four cells.
Values are means (SE, n 26) of samples collected between
May 1996 and June 1998. Differences between cells determined
by analysis of variance (ANOVA); letter codes indicate significant differences (Tukey multiple comparisons test, P 0.05).
Metal

Cell 1

Cell 2

Cell 3

Cell 4

kg1

Fe
Mn
Co
Ni

268
2 010
0.76
2.0

47
170
0.18a
0.2

mg
Aboveground tissues
178 26
222 58
2 350 230
2 440 230
0.33 0.07b
0.32 0.06b
1.9 0.2
2.4 0.4

Fe
Mn
Co
Ni

7 440
1 650
11
16

Belowground tissues
6 430 520
7 320 760
915 84
1 300 180
4.0 0.7b
3.9 0.6b
9.6 1.3
12 2

805
170
1.7a
2

218 48
2 060 180
0.28 0.04b
2.2 0.3
8 000
940
3.1
11

1 470
180
0.5b
2

Fe
Mn
Co
Ni

7 010
11 800
94.6
94.3

1 120a
1 800a
27.2a
19.7a

Fe
Mn
Co
Ni

36 200 13 700a
23 900 3 560a
325 92a
299 92a

1 690
6 800
24.3
47.7

470b 1 840 460b

1 300b 4 990 610bc
7.3b
7.2 2.8c
9.7b
48.9 8.5b

Chara (n 8)

1 310
4 050
3.3
38.7

180b
530c
1.1c
7.2b

9 280 1 760b
11 900 2 170b
14 5b
66 26b

Distribution of Metals within Sediments

and Pore Water
The concentrations of Fe, Mn, and Ni in the sediments
of the APS wetland (Table 1) were comparable with
those of other cattail-dominated treatment wetlands, for
example, 180 to 84 700 mg Fe kg1, 31 to 2620 mg Mn
kg1, and 1 to 24 300 mg Ni kg1 (Babcock et al., 1983;
Taylor and Crowder, 1983; Sencindiver and Bhumbla,
1988; Meiorin, 1989; Fernandes and Henriques, 1990).
However, the concentrations of Mn, Co, and Ni in the
top layer (05 cm) of the sediments in Cell 1 were higher
than their concentrations in the lower layers (510 and
1015 cm) of the same cell (Table 1). The surface layers
of sediment are generally aerobic, facilitating metal removal via the formation of hydrated metal oxides, and
have high microbial activity, important for the chemical
transformation of metals in sediment pore water (Hedberg and Wahlberg, 1998).
Table 4. Annual metal budget for the vegetated cells of the wetland (July 1997June 1998). Estimated metal loading and unloading rates calculated for the wetland cells. The retention
and distribution of the metals in the cells are estimated on a
unit area basis. The accumulation of metals in the plants is
also shown as a percentage of the metals removed by the
vegetated cells.
Fe

Mn

Co

Ni

Metal budget (g yr1 )

Loading - Inlet Cell 1
93 400
196 000
785
3 140
Unloading - Outlet Cell 4 5 500
3 590
32
1 200
Retention (g m2 yr1 )
68
149
0.58
1.5
Accumulation (g m2 )
Cattail shoots
0.62
6.21
0.001
0.006
Fallen litter
0.53
1.23
0.006
0.010
Chara
1.38
1.46
0.01
0.012
Metal uptake as a percentage
of total removed by the
vegetated cells (%)
Cattail shoots
0.91
4.18
0.19
0.38
Fallen litter
0.78
0.83
1.03
0.67
Chara
2.03
0.98
1.72
0.80
Mean inflow, 2.49 L s1; mean outflow, 2.53 L s1.
Peak aboveground biomass of cattail, 2800 g m2; fallen litter, 178 g
m2; peak biomass of Chara, 107 g m2.

1472

J. ENVIRON. QUAL., VOL. 30, JULYAUGUST 2001

There was a decrease in the concentration of metals

(except Fe) in sediment with increasing distance from
the inlet. The concentrations of the metals were greatest
in Cell 1, and least in Cell 4. Similar findings have been
reported for Fe, Mn, and other trace elements such as
As and Zn by several investigators (e.g., Beining and
Otte, 1996; Keller et al., 1998; Mitsch and Wise, 1998).
Except for Fe, the concentrations of Mn, Co, and Ni in
the sediment pore water also tended to decrease from
Cell 1 to Cell 4 (Fig. 6). Beining and Otte (1996) found
a similar reduction in the metal (Zn) load of the pore
water with increasing distance from the inlet.

Accumulation of Metals by Plants and Fallen

Litter: Role of Vegetation
To effectively use constructed wetlands for metal removal, it is important that the wetland vegetation can
tolerate the levels of potentially toxic metals present in
the wastewater. In our study, the growth of cattail was
not detrimentally affected by the metals present, despite
the high concentrations of metals in the water and sediment. The maximum aboveground biomass was 2800 g
dry weight m2, comparable with 2900 g m2 for healthy
cattail growing in a noncontaminated Texas pond
(Hill, 1987).
Although plants play a critical role in the removal of
metals by constructed wetlands through physicochemical processes and by supporting microbial activity (Kadlec and Knight, 1996), the uptake of metals into living
shoots represented only a minor component of the metal
removed by the APS wetland. In the present study, for
example, only 0.91, 4.18, 0.19, and 0.38% of Fe, Mn, Co,
and Ni were accumulated annually in the aboveground
tissues of cattail (the dominant species of the APS wetland) per m2, respectively (Table 4).
Considerably higher metal concentrations were measured in the fallen litter of cattail than were present in
the living cattail shoots (i.e., 3, 13, 85, and 28 times
higher for Fe, Mn, Co, and Ni, respectively). High concentrations of metals were also accumulated by the living tissues of Chara, a macroalga growing primarily in
Cells 1 and 4. Other researchers have noted that submerged plant tissues, including Chara, may constitute a
considerable sink for metal removal in wetlands (Reimer and Toth, 1968; Kepler, 1988; Sparling and Lowe,
1998). It is possible that the high concentrations of metals associated with both fallen litter and with Chara
were due partly to the fact that dead plant material and
living Chara tissues provide adsorption sites and sites
of precipitation for metal accumulation (Skousen et al.,
1994; Horne, 2000). The concentrations of all four metals in cattail fallen litter and Chara collected from Cell
1 were higher than those collected from Cell 4, respectively, probably as a direct consequence of the higher
concentration of metals in the surface water of Cell 1
compared with Cell 4. The accumulation of Fe, Co, and
Ni in fallen litter and submerged Chara tissues together
accounted for a significantly greater proportion of the
metals removed by the wetland (i.e., 2.81, 2.75, and
1.05% for Fe, Co, and Ni, respectively) than the propor-

tion removed by living cattail shoots. Interestingly, living

cattail shoots accumulated greater amounts of Mn than
fallen litter and Chara, that is, 4.2% compared with
1.8% of Mn removed (Table 4).
Cattail plants play an important role in metal retention by virtue of the immobilization of metals in the
oxygenated rhizosphere (Otte et al., 1995; Cacador et
al., 1996). Doyle and Otte (1997) reported that oxidation
processes and Fe concentrations were consistently
higher in the rhizosphere than in the bulk soil for all the
plant species they studied. Assuming that the biomass of
cattail roots and rhizomes are equal to those of the
shoots (Sencindiver and Bhumbla, 1988; Vymazal,
1995), we estimate that the belowground biomass of
cattail in the APS wetland could have taken up as much
as 20.2 g Fe m2. This represents 29.7% of the total
amount of Fe retained by the vegetated cells compared
with 0.91% for the shoots in the same area. Part of this
Fe was almost certainly associated with iron plaque,
which may be up to 15 to 17 m thick on cattail root
surfaces (Taylor et al., 1984). Our research has shown
that there may be approximately ninefold more Fe on
the surfaces of cattail roots than is accumulated within
the root tissues (Ye, unpublished data, 1999). Manganese may also be accumulated to high concentrations
in the iron plaque (Ye et al., 1997).
Our conclusions are in clear agreement with previous
studies (e.g., Sencindiver and Bhumbla, 1988; Dunbabin
and Bowmer, 1992; Mitsch and Wise, 1998) that indicate
that the substratum (sediment) is the primary sink for
metals in treatment wetlands. Metal accumulation in
the APS wetland, Pennsylvania, tended to be greater in
the surface layers of sediments as well as in the rhizomes
of cattail. The accumulation of metals in living shoot
tissues of cattail (the dominant plant species in the wetland), and the accumulation in cattail fallen litter and
submerged living Chara tissues, were relatively minor
sinks in comparison with the sediments. We conclude
that the vegetated constructed wetland at Springdale is
a highly effective, low-maintenance treatment system
for cleaning Fe and Mn in coal combustion by-product
leachate to comply with the NPDES permits.
ACKNOWLEDGMENTS
Funding for this study provided by the Electric Power Research Institute (EPRI) (4163-01), Palo Alto, CA. The authors
thank Dr. M. de Souza for reviewing this manuscript.

REFERENCES
Babcock, M.F., D.W. Evans, and J.J. Alberts. 1983. Comparative
uptake and translocation of trace elements from coal ash by Typha
latifolia. Sci. Total Environ. 28:203214.
Bastian, R.K., and D.A. Hammer. 1993. The use of constructed wetlands for wastewater treatment and recycling. p. 5968. In G.A.
Moshiri (ed.) Constructed wetlands for water quality improvement.
Lewis Publ., Boca Raton, FL.
Beining, B.A., and M.L. Otte. 1996. Retention of metals originating
from an abandoned lead-zinc mine by a wetland at Glendalough,
Co. Wicklow. Biol. Environ. 96:117126.
Brodie, G.A. 1993. Aerobic constructed wetlands to treat acid drainage: Case history of Fabius Impoundment 1 and overview of the
Tennessee Valley Authoritys program. p. 157165. In G.A. Moshiri
(ed.) Constructed wetlands for water quality improvement. Lewis
Publ., Boca Raton, FL.

YE ET AL.: REMOVAL OF Fe, Mn, Co, AND Ni BY A CONSTRUCTED WETLAND

Brodie, G.A., C.R. Britt, T.M. Tomaszewski, and H.N. Taylor. 1993.
Anoxic limestone drains to enhance performance of aerobic acid
drainage treatment wetlands: Experiences of the Tennessee Valley
Authority. p. 922. In G.A. Moshiri (ed.) Constructed wetlands
for water quality improvement. Lewis Publ., Boca Raton, FL.
Brodie, G.A., D.A. Hammer, and D.A. Tomljanovich. 1989. Treatment of acid drainage from coal facilities with man-made wetlands.
p. 903921. In D.A. Hammer (ed.) Constructed wetlands for wastewater treatment: Municipal, industrial and agricultural. Proc. from
the 1st Int. Conf. on Constructed Wetlands for Wastewater Treatment, Chattanooga, TN. 1317 June 1988. Lewis Publ., Chelsea, MI.
Byrnes, M.E. 1994. Field sampling methods for remedial investigations. Lewis Publ., Boca Raton, FL.
Cacador, I., C. Vale, and F. Catarino. 1996. Accumulation of Zn, Pb,
Cu, Cr and Ni in sediments between roots of the Tagus estuary
salt marshes, Portugal. Estuarine Coastal Shelf Sci. 42:393403.
Doyle, M.O., and M.L. Otte. 1997. Organism-induced accumulation
of iron, zinc and arsenic in wetland soils. Environ. Pollut. 96:111.
Dunbabin, J.S., and K.H. Bowmer. 1992. Potential use of constructed
wetlands for treatment of industrial wastewaters containing metals.
Sci. Total Environ. 111:151168.
Electric Power Research Institute. 1998. The Springdale Project:
Applying constructed wetland treatment to coal combustion byproduct leachate. Project 9065-01. Gannett Fleming, EES Division,
Hyde, PA.
Faulkner, S.P., and C.J. Richardson. 1989. Physical and chemical characteristics of freshwater wetland soils. p. 4172. In D.A. Hammer
(ed.) Constructed wetlands for wastewater treatment: Municipal,
industrial and agricultural. Proc. from the 1st Int. Conf. on Constructed Wetlands for Wastewater Treatment, Chattanooga, TN.
1317 June 1988. Lewis Publ., Chelsea, MI.
Faulkner, B., and J. Skousen. 1994. Treatment of acid mine drainage
by passive treatment systems. p. 250257. In Proc. Int. Land Reclamation and Mine Drainage Conf. 2429 Apr. 1994. Vol. 2. Bureau
of Mines SP 063-94. U.S. Dep. of Interior, Pittsburgh, PA.
Fennessy, M.S., and W.J. Mitsch. 1989. Treating coalmine drainage
with an artificial wetland. Res. J. Water Pollut. Control Fed.
61:16911701.
Fernandes, J.C., and F.S. Henriques. 1990. Metal levels in soils and
cattail (Typha latifolia L.) plants in a pyrite mine area at Lousal,
Portugal. Int. J. Environ. Studies 36:205210.
Hammer, D.A. 1993. Designing constructed wetlands systems to treat
agricultural nonpoint source pollution. p. 71111. In R.K. Olson
(ed.) Created and natural wetlands for controlling nonpoint source
pollution. C.K. Smoley, Boca Raton, FL.
Hedberg, T., and T.A. Wahlberg. 1998. Upgrading of waterworks with
a new biooxidation process for removal of manganese and iron.
Water Sci. Technol. 37:121126.
Hedin, R.S. 1989. Treatment of coal mine drainage with constructed
wetland. p. 349362. In S.K. Majumdar et al. (ed.) Wetlands ecology
and conservation: Emphasis in Pennsylvania. The Pennsylvania
Academy of Sci., Easton, PA.
Hedin, R.S., and R.W. Nairn. 1993. Contaminant removal capabilities
of wetlands constructed to treat coalmine drainage. p. 187195. In
G.A. Moshiri (ed.) Constructed wetlands for water quality improvement. Lewis Publ., Boca Raton, FL.
Henrot, J., and R.K. Wieder. 1990. Processes of iron and manganese
retention in laboratory peat microcosms subjected to acid mine
drainage. J. Environ. Quality 19:312320.
Hill, B.H. 1987. Typha productivity in a Texas pond: Implications
for energy and nutrient dynamics in freshwater wetlands. Aquatic
Bot. 27:385394.
Horne, A.J. 2000. Phytoremediation by constructed wetlands. p. 1339.
In N. Terry and G. Banuelos (ed.) Phytoremediation of contaminated soil and water. Lewis Publ., Boca Raton, FL.
Johnson, D.B. 1998. Biodiversity and ecology of acidophilic microorganisms. FEMS Microbiol. Ecol. 27:307317.
Kadlec, R.H., and R.L. Knight. 1996. Treatment wetlands. CRC Press,
Boca Raton, FL.
Keller, B.E.M., K. Lajtha, and S. Cristofor. 1998. Trace metal concentrations in the sediments and plants of the Danube Delta, Romania.
Wetlands 18:4250.
Kepler, D.A. 1988. An overview of the algae in the treatment of acid
mine drainage. p. 286290. In Mine drainage and surface mine
reclamation. Info. Circ. 9183. U.S. Bureau of Mines, Pittsburgh, PA.

1473

Meiorin, E.C. 1989. Urban runoff treatment in a fresh/brackish water

marsh in Fremont, California. p. 677685. In D.A. Hammer (ed.)
Constructed wetlands for wastewater treatment: Municipal, industrial and agricultural. Proc. from the 1st Int. Conf. on Constructed
Wetlands for Wastewater Treatment, Chattanooga, TN. 1317 June
1988. Lewis Publ., Chelsea, MI.
Mitsch, W.J., and K.M. Wise. 1998. Water quality, fate of metals, and
predictive model validation of a constructed wetland treating acid
mine drainage. Water Res. 32:18881900.
National Rivers Authority. 1992. Constructed wetlands to ameliorate
metal-rich mine waters. R&D Note 102. NRA, Bristol, UK.
Nu-Hoai, V.N., H.E. Farrah, G.A. Lawrance, and G.L. Orr. 1998.
Efficiency of a small wetland with an industrial urban catchment.
Sci. Total Environ. 214:221237.
Otte, M.L., S.C. Kearns, and M.O. Doyle. 1995. Accumulation of
arsenic and zinc in the rhizosphere of wetland plants. Bull. Environ.
Contam. Toxicol. 55:154161.
Perry, A., and R.L.P. Kleinmann. 1991. The use of constructed wetlands in the treatment of acid mine drainage. Nat. Resour. Forum 15(3):178184.
Reimer, D.N., and S.J. Toth. 1968. A survey of the chemical composition of aquatic plants in New Jersey. Bulletin 820. New Jersey
Agric. Exp. Stn., Rutgers University, New Brunswick, NJ.
Sencindiver, J.C., and D.K. Bhumbla. 1988. Effect of cattail (Typha )
on metal removal from mine drainage. p. 359366. In Mine drainage
and surface mine reclamation. Info. Circ. 9183. U.S. Bureau of
Mines, Pittsburgh, PA.
Skousen, J., A. Sexstone, K. Garbutt, and J. Sencindiver. 1994. Acid
mine drainage treatment with wetlands and anoxic limestone
drains. p. 263281. In D.M. Kent (ed.) Applied wetland science
and technology. Lewis Publ., Boca Raton, FL.
Sparling, D.W., and T.P. Lowe. 1998. Metal concentrations in aquatic
macrophytes as influenced by soil and acidification. Water Air Soil
Pollut. 108:203221.
Stark, L.R., F.M. Williams, S.E. Stevens, and D.P. Eddy. 1994. Metal
removal in wetland treatment systems. p. 8998. In Proc. Int. Land
Reclamation and Mine Drainage Conf. 2429 Apr. 1994. Vol. 2.
Bureau of Mines SP 063-94. U.S. Dep. of Interior, Pittsburgh, PA.
Stillings, L.L., J.J. Gryta, and E.A. Ronning. 1988. Iron and manganese
removal in a Typha-dominated wetland during ten months following wetland construction. p. 317324. In Mine drainage and surface
mine reclamation. Info. Circ. 9183. U.S. Bureau of Mines, Pittsburgh, PA.
Taylor, G.J., and A.A. Crowder. 1983. Uptake and accumulation of
heavy metals by Typha latifolia in wetlands of the Sudbury, Ontario
region. Can. J. Bot. 61:6373.
Taylor, G.J., A.A. Crowder, and R. Rodden. 1984. Formation and
morphology of an iron plaque on the roots of Typha latifolia L.
grown in solution culture. Am. J. Bot. 71:666675.
USEPA. 1992. Acid digestion of aqueous sample for metalsIAA/
ICP. Method 3010A. USEPA, Washington, DC.
USEPA. 1996. Acid digestion of sediments, sludges and soils. Method
3050B. USEPA, Washington, DC.
Vymazal, J. 1995. Algae and element cycling in wetland. Lewis Publ.,
CRC Press, Boca Raton, FL.
Whiting, S.N., and N. Terry. 1999. The Tennessee Valley Authority
constructed wetland at Widows Creek: Role of vegetation in the
removal of trace elements. TR-114220. Electric Power Res. Inst.,
Palo Alto, CA.
Wieder, R.K. 1988. Determining the capacity for metal retention in
man-made wetlands constructed for treatment of coalmine drainage. p. 375381. In Mine drainage and surface mine reclamation.
Info. Circ. 9183. U.S. Bureau of Mines, Pittsburgh, PA.
Ye, Z.H., A.J.M. Baker, M.H. Wong, and A.J. Willis. 1997. Copper
and nickel uptake, accumulation and tolerance in Typha latifolia
with and without iron plaque on the root surface. New Phytol.
136:481488.
Zar, J.H. 1984. Biostatistical analysis. 2nd ed. PrenticeHall, Englewood Cliffs, NJ.
Zarcinas, B.A., B. Cartwright, and L.R. Spouncer. 1987. Nitric acid
digestion and multi-element analysis of plant material by inductively coupled plasma spectrometry. Commun. Soil Sci. Plant
Anal. 18:131146.