517
13
Organic Chemicals
Organic compounds are major constituents of biochemical
oxygen demand (BOD) or chemical oxygen demand (COD)
in municipal wastewater, as discussed in Chapter 8. However,
there is an increasing number of treatment wetlands that tar-
get specic individual or groups of organic chemicals. These
chemicals pose a new and somewhat more difcult set of
problems because of their possible toxicity to plants and the
limitations of aerobic and anaerobic degradation. The major
routes for removal of hydrocarbons from wetland waters
include: volatilization, photochemical oxidation, sedimenta-
tion, sorption, biological (microbial) degradation, and plant
uptake. Three types of microbial processes can contribute:
fermentation, aerobic respiration, and anaerobic respiration.
The general principles and chemistry of processes affecting
carbon compounds are discussed in Chapter 8; see Equations
8.10 through 8.18.
Wetlands manufacture and contain a wide spectrum of
organic compounds. These compounds range from small
molecules such as methane to humic acids of very high
molecular weight. Many wetland soils are organic in nature
and possess an afnity for introduced organics, via sorption
and other binding mechanisms. Aliphatic hydrocarbons, both
straight-chain and branched, are present as natural waxes. As
a result, trace (background) amounts of some hydrocarbons
are present in all wetlands, whether constructed or natural.
13.1 PETROLEUM HYDROCARBONS
There is considerable information on the use of treatment
wetlands in the petroleum industry. Some of the general prin-

ciples and available data have been summarized in a 1998
industry report (Knight et al., 1997; Knight et al., 1999).
However, that compilation did not focus on specic hydro-
carbon classes, such as BTEX and its constituent components
(benzene, toluene, ethyl benzene, and xylenes). Two other
hydrocarbon designations of interest include Gasoline Range
Organics (GRO) and Diesel Range Organics (DRO). There is
some overlap and ambiguity in these designations. Typically,
GRO consists of hydrocarbons with 6–9 carbon atoms, while
DRO contains 10–40 carbon atoms (Chapple et al., 2002).
Total Petroleum Hydrocarbons (TPH) is a measure of the
sum of parafnic and aromatic constituents.
BTEX
Biodegradation
Biodegradation of BTEX in a wetland environment is compli-
cated by the existence of biolms on submerged plant parts,
plant litter, and gravel. Although small in terms of mass per
unit volume, these biolms are very active in biodegrada-
tion, and consequently serve as important sinks for organics.
In this aspect, treatment wetlands resemble conventional
attached-growth treatment processes.
Chang et al. (2001) established that all BTEX constituents
degrade rapidly (half-lives of one to two days) when inocu-
lated with a microbial consortium conditioned with toluene.
Their study showed that benzene, toluene, and ethylbenzene
were directly consumed as carbon sources, while xylene was
removed by co-metabolism.
Moore et al. (2002) measured both aerobic and anaerobic
biodegradation of BTEX in peat. Aerobic degradation was
tracked via oxygen consumption and carbon dioxide produc-

tion, and averaged 56 mg/kg·d. Anaerobic degradation was
inferred from the consumption of other electron acceptors,
including nitrate, sulfate, iron, manganese, and methane.
About one third of the BTEX loss could have been attributed
to anaerobic degradation.
Volatilization
BTEX constituents are volatile and may be easily lost from
water, especially shallow water bodies, such as FWS wet-
lands. Lee et al. (2004a) have reported that rst-order loss
rate constants for BTEX constituents exhibit a xed ratio to
the loss rate constant for benzene, independent of the water-
mixing regime (Table 13.1). Further, the presence of more
than one BTEX compound, or of surfactants, had only minor
effects on this ratio.
Of direct interest are the estimates, based on wetland
data, of Keefe et al. (2004a) for an FWS wetland in Arizona.
Values were determined of K
w
 0.015 m/hr (130 m/yr) for
toluene. This is a relatively high rate, compared to other
pollutants treated in FWS wetlands.
Moore et al. (2002) evaluated volatilization losses from
eld vapor collection in a peatland contaminated with BTEX.
Losses averaged 2,500 mg/m
2
·d over all seasons. This high
rate may have been in part due to the existence of a Light
Non-Aqueous Phase Layer (LNAPL) in the peat.
Plant Uptake
Willows (Salix babylonica) were shown to materially contrib-

ute to the removal of benzene from water under hydroponic
conditions (Corseuil and Moreno, 2001). Volatilization was
effectively suppressed in the experiments, as demonstrated
by control mesocosms. About 80% reduction in benzene was
found with an HRT of one day. Corseuil and Moreno sug-
gested that benzene was initially sorbed onto root biomass,
followed by plant uptake and biological degradation.
© 2009 by Taylor & Francis Group, LLC
518 Treatment Wetlands
Sorption
BTEX components are expected to partition strongly to
organic wetland substrates, although a denitive wetland
study is lacking. Wetland sediments typically have high
organic content, and therefore sorption may be an important
rst step in overall removal. Benzene sorption onto peat was
found to be strong, with a linear sorption coefcient of K
OC

5.6 L/kg (Moore et al., 2002), whereas binding to clay was
much weaker, with K
OC
 0.12 L/kg.
Wetland System Studies
In all existing petrochemical plant applications, wetlands
have been accompanied by pretreatment. However, other
applications, such as those involving remediation, do not
involve pretreatment. Hydrocarbon contamination of ground-
water includes closed and operating sanitary landlls, army
ammunition plants, and former oil renery sites. These facil-
ities were often the recipient of solvents and other organics,

which in the course of time have contaminated nearby
waters. Concentrations of petrochemicals in landll leachate
are typically lower than those associated with reneries and
terminals.
Gulf Strachan, Rocky Mountain House, Alberta
HSSF wetlands were tested for the ability to reduce hydro-
carbons, notably BTEX and TPH (Moore et al., 2000a).
The wetlands were planted with Phragmites australis and
Typha latifolia. Reductions of 40–60% were achieved within
14 days of detention time with no aeration, while aeration
produced 100% removal in the same detention time. Inuent
BTEX concentrations ranged from 4.5–12.1 mg/L, and inu-
ent ows ranged from 7–33 L/min.
At the same site, Moore et al. (2002) reported on the
natural attenuation of BTEX in a natural peatland, which
received both aqueous and nonaqueous phase hydrocarbons
for more than 15 years. No BTEX was detected leaving the
peatland. Companion studies elucidated some of the candi-
date removal mechanisms, including sorption, aerobic deg-
radation, volatilization, and anaerobic degradation. Aerobic
degradation, which was stimulated by air injection, was the
dominant removal mechanism.
Former Oil Refinery—United Kingdom
HSSF wetlands planted with Phragmites were tested for the
ability to reduce hydrocarbons, notably DRO (Chapple et al.,
2002). Reductions of 40–64% were achieved in less than one
day’s detention time. Gravel-based beds performed better (k
270 m/yr) than soil-based beds (k  137 m/yr) for DRO removal.
Mobil Oil AG Terminal at Bremen, Germany
The information here is taken from Altman et al. (1989),

which summarizes three years of research results at Bremen.
Highly contaminated wastewater (COD up to 14,000 mg/L)
was brought through an API separator, a parallel plate sepa-
rator, and a percolating reactor to a two-train pilot wetland
system patterned after the concepts of Seidel (1966; 1973) at
the Max Planck Institute. Each train had ve subsurface ow
wetlands in series, each being 2.5 m wide, 4.0 m long, and
0.8 m deep. Beds 1 and 2 were vertical ow, with passive
aeration to the underdrains. Beds 3, 4, and 5 were horizontal
ow. Each bed was lled with a stratied media, from the bot-
tom: 5 cm of 8/16 mm (pea) gravel, 20 cm of 36/72 mm stone,
25 cm of 8/16 mm g ravel , 10 cm of shar p sand, a nd toppe d wit h
2 cm of organic soil. Water is intermittently dosed to the sur-
face, where it spreads and inltrates, ultimately reaching the
(10 cm) under-drains in the second layer. The passive air is
admitted to the rst layer via perforated plastic pipe. Details
of this type of system are given in Vymazal et al. (1998).
The plants that proved best for Bremen were cattails (Typha
angustifolia) and bulrushes (Scirpus lacustris).
Hydrocarbon removal performance of this system was
primarily measured as total hydrocarbon, but BTEX and its
constituents were also measured less frequently. The inuent
BTEX in August was 3.8 mg/L (1.5 mg/L benzene), and the
outow contained no detectable BTEX, with all constituents
being below 0.01 mg/L. The ow of 5 m
3
/d corresponded to
about 5 cm/d hydraulic loading, or about ve days’ detention.
Although these results are encouraging, they do not answer
the question of how much of the ve-bed system was needed

to achieve a given reduction.
However, inferences about BTEX removal may be made
from the removal of total hydrocarbon (TH). Data were
taken at the outlet from each bed of the trains, and therefore
a removal model may be calibrated. A rst-order areal TH
model yields k  66 m/yr, with no evidence of temperature
dependence. This is consistent with the benzene removal rate
coefcient, which cannot be less than 90 m/yr for the data in
the preceding paragraph.
Williams Pipeline, Watertown, South Dakota
An aerated HSSF wetland was operated to reduce BTEX
from petroleum contact waters. Complete removal of
TABLE 13.1
Volatility of BTEX and Chlorinated
Benzenes Compared to Benzene
Compound Volatilization Ratio
Benzene 1.000
Toluene 0.966
Ethylbenzene 0.922
m-Xylene 0.930
1,3 Dichlorobenzene 0.882
1,3,5 Trichlorobenzene 0.569
1,2,3,4-Tetrachlorobenzene 0.530
Pentachlorobenzene 0.493
Note: The volatilization ratio is the ratio of rst-order loss
rate coefcients.
Source: Data from Lee et al. (2004a) Water Research
38(2): 365–374.
© 2009 by Taylor & Francis Group, LLC
Organic Chemicals 519

hydrocarbons was reported (Wallace, 2002), despite high
inuent concentrations (CBOD
5
 16,000 mg/L, NH
4
-N 
200 mg/L, and BTEX  1 mg/L), but this facility operated
at an extremely low hydraulic loading (1 mm/d), and conse-
quently had no water discharge.
ESSO, Chilliwack, British Columbia
A cardlock facility produced stormwaters contaminated by
diesel fuel. FWS wetlands were found successfully reduc-
ing water-phase diesel range organics to permit levels (Nix,
1995). Removal rate constants were about 10 m/yr.
Marathon-Pitchfork, Wyoming Produced Water Project
The Colorado School of Mines tested a combined pilot sys-
tem for hydrocarbon removal (Caswell et al., 1992). Overland
ow gravel beds (shallow) were loaded at 20–80 cm/d, and
reduced benzene from 15 to 2 µg/L. Further removal in SSF
wetlands reduced benzene to below detection.
Isanti-Chisago, Minnesota Leachate Treatment System
The Isanti-Chisago Sanitary Landll, an unlined municipal
solid waste facility located near Cambridge, Minnesota, was
closed in 1992 (Loer et al., 1999; Kadlec, 2003c). Leach-
ing of soluble wastes had contaminated the surcial and
increasingly deeper aquifers with toxic organic compounds
and heavy metals. Extraction wells permit pumping of leach-
ate to the top surface of the landll mound, about eight
meters above the surrounding landscape. Volatile Organic
Compounds (VOCs), including BTEX, are largely removed

by cascading the water down the side of the landll into a
sedimentation basin which serves to settle and store iron
precipitates. The next component of the treatment train is a
0.6-ha FWS wetland. During ve seasons of treatment, oper-
ating results indicated that the system efciency ranges from
85–100% for VOCs and 98% for iron.
Only low levels of BTEX enter the system: (benzene
3.7 µg/L; toluene 0.6 µg/L; ethylbenzene 1.1 µg/L, xylene
1.8 µg/L). The cascade and settling basin do all the BTEX
removal, although the FWS wetland does polish out any
remaining traces of BTEX. Benzene, toluene, ethyl benzene,
and xylenes are all individually below detection (0.1 µg/L) in
the system outow.
Saginaw Township, Michigan, Leachate Treatment System
Saginaw Charter Township’s Center Road Landll was
closed in the early 1980s (Kadlec, 2003c). Finger drains were
installed in the seepage zones, which connect to a collection
pipe. Leachate is then pumped to an aerator, which provides
some ammonia and BTEX stripping. The water is then held in
a sedimentation basin, to promote removal of solids. Water is
discharged periodically to one of two intermittent sand/gravel
lters, which provide ltration and nitrication. Underdrains
then convey the water to two parallel free water surface wet-
lands. BTEX and its constituents were monitored for ve years
after startup. A mean total BTEX of 39 µg/L was removed,
with only one detection in weekly samples (5.9 µg/L).
Phoenix, Arizona Wastewater Polishing
A demonstration wetland project was studied for VOC
removal (Keefe et al., 2004a). The wetland was a 1.4-ha
free water surface system of depth about 60 cm. The deten-

tion time was 3.9 days. An 80% reduction of toluene from
inlet to outlet was measured. Keefe et al. (2004a) attrib-
uted the reduction to volatilization, but concluded that theo-
retical predictions were only good for order-of-magnitude
estimation.
Alcoa, Tennessee Groundwater Remediation
A pilot wetland project was initiated and operated for a
period of over one year, in a moderate north temperate con-
tinental climate. DRO and GRO were among the targeted
substances (Gessner et al., 2005). GRO entered at monthly
average values of 0.04–0.37 mg/L, and never exceeded the
detection limit of 0.01 mg/L at the outlet of the FWS wet-
land. DRO entered at monthly average values of 0.29–1.08
mg/L, and exited at 0.11–0.44 mg/L. GRO removal rate
constants could not have been less than 100 m/yr. In con-
trast, the DRO removal rate constants were 12 m/yr annual
average. The rst wetlands at the site were SSF wetlands,
which were abandoned in favor of FWS because of continual
operational difculties.
Casper, Wyoming, Groundwater Remediation
A pilot scale subsurface vertical ow wetland system was con-
structed at the former British Petroleum Renery in Casper,
Wyoming, in order to determine benzene, toluene, ethylben-
zene, and xylene (BTEX) degradation rates in a cold-climate
application (Ferro et al., 2002). The pilot system, consisting
of four cells, each dosed at a nominal ow rate of 5.4 m
3
/d,
was operated between August and December 2002. The pilot
tested the effects of wetland mulch and aeration on system

performance. Areal rate constants (k
A
) were calculated based
on an assumed three tanks in series (3TIS). The presence of
both wetland sod and aeration improved treatment perfor-
mance. Mean k
A
values were 244 m/yr for cells without sod
or aeration, and improved to 356 m/yr for cells with sod and
aeration (Table 13.2).
Based on the results of the pilot system, a full-scale wet-
land system (capable of operating at 6,000 m
3
/d) was started
up in May 2003 (Wallace and Kadlec, 2005). The full-scale
system achieved permit compliance within one week of
start-up.
University of Edinburgh, Scotland
Bench scale vertical ow wetlands were operated to demon-
strate benzene removal (Eke and Scholz, 2006). The systems
were operated in a ll-and-drain batch mode, cycling twice
per week. Inuent concentrations of 1,000 mg/L benzene
were removed to 37–87 mg/L in the outdoor environment.
The presence of wetland media (gravel), fertilizer, and warm
temperatures were noted in improving treatment.
© 2009 by Taylor & Francis Group, LLC
520 Treatment Wetlands
ALKANES
Omari et al. (2001) determined removal efciencies in both
vegetated and unvegetated horizontal-ow gravel beds for

straight-chain alkanes of 10 to 26 carbon atoms. Removals
were typically above 80% for a vegetated top-fed wetland
mesocosm, and above 70% for a top-fed unvegetated meso-
cosm, in eight hours’ detention. Removals were somewhat
less for the lighter hydrocarbons, up to C15, but did not dif-
fer for the heavier alkanes. Omari et al. concluded that con-
structed wetlands planted with Typha latifolia were capable
of treating oil-polluted water.
Similarly, Hoffman (2003) found no pronounced effects
of chain length on alkane removals in willow mesocosms,
over the range C10–C32. Willows enhanced the removal of
TPH, with the efuent alkane concentrations in vegetated
mesocosms about half of that in unvegetated mesocosms.
Mechanisms of removal were not elucidated by these studies.
Salmon et al. (1998) evaluated removal of total hydro-
carbons in FWS mesocosms in France. The bed was planted
with Typha latifolia, and during the experiment, natural
development of Lemna minor occurred. Constructed wet-
lands removed 90% of total hydrocarbons.
Descriptions of other petroleum industry treatment wet-
land projects may be found in Knight et al. (1997). These
include the earliest such project at Mandan, North Dakota
(Litcheld and Schatz, 1989; Litcheld, 1990; Litcheld,
1993), as well as well-studied projects such as that at Bulwer
Island, Australia (Simi and Mitchell, 1999; Simi, 2000).
POLYCYCLIC AROMATIC HYDROCARBONS
Polycyclic aromatic hydrocarbons (PAHs) are fused ring
aromatic compounds formed during the incomplete combus-
tion of almost any organic material, and are ubiquitous in the
environment (Figure 13.1). Some of them are considered as

dangerous substances as a function of their toxic and muta-
genic or carcinogenic potentialities. The presence of PAHs
in contaminated soils and sediments may pose a risk to the
environment and human health. PAHs are hydrophobic com-
pounds, whose persistence within ecosystems is chiey due
to their low aqueous solubility. These materials are not vola-
tile, and no loss to the atmosphere is anticipated under wet-
land conditions.
Creosote consists of a mixture of organic compounds,
dominated by PAHs, many of which are individually desig-
nated as hazardous wastes. Wood may be treated with creo-
sote to enhance its resistance to decomposition. For instance,
railroad ties are typically treated with 128 kg/m
3
. Many of
the industrial sites that are or were used to treat ties are now
TABLE 13.2
Casper, Wyoming, Mean Pilot System Areal Rate Constants k
A
, in m/yr
Aeration No Aeration Aeration
Compound Wetland Mulch No Mulch Wetland Mulch No Mulch Full Scale
Benzene 518 (237) 456 (414) 317 (273) 226 (164) 240
BTEX 356 (218) 311 (136) 257 (151) 244 (151) 350
TPH 1,058 (1,141) 965 (722) 725 (558) 579 (506) 325
Note: Based on 3TIS; standard deviation in parenthesis.
Source: Data from Wallace and Kadlec (2005) Water Science and Technology 51(9): 165–171.
Naphthalene Anthracene Phenanthrene Pyrene Fluorene
Benzo(a)pyreneChrysene Fluoranthene Acenaphthene
FIGURE 13.1 Example polynuclear aromatic hydrocarbons.

© 2009 by Taylor & Francis Group, LLC
Organic Chemicals 521
classied as hazardous waste sites, due to spillage and the
associated soil contamination. Conversely, there is little or
no risk to aquatic and wetland environments from ties as
they are normally placed (Brooks, 2000b), nor from bridge
or dock pilings (Brooks, 2000a).
Biodegradation
Aromatics follow a pattern, with polyaromatic hydrocarbons
(PAHs) degrading more slowly than benzene; those with more
than three rings may not support microbial growth (Zander,
1980). The aerobic microbial degradation of PAHs having
two and three rings is well documented. A large number of
bacteria that metabolize PAHs have been isolated (Alcali-
genes denitricans, Rhodococcus spp., Pseudomonas spp.,
Mycobacterium spp.; Giraud et al., 2001). A variety of bac-
teria can degrade certain PAHs completely to CO
2
and meta-
bolic intermediates. As the number of fused rings increases,
the degree of degradation decreases (Cookson, 1995, as
referenced in Walsh, 1999). Degradation of PAHs by anaero-
bic organisms has not been very successful. However, some
degradation has been achieved under denitrifying, sulfate
reducing, and methanogenic conditions. Naphthalene and
anthracene have been found to be slightly degraded anaero-
bically under denitrifying conditions (Walsh, 1999).
PAH-degrading bacteria were isolated from contami-
nated estuarine sediment and salt marsh rhizosphere by
enrichment using naphthalene, phenanthrene, or biphenyl as

the sole source of carbon and energy (Daane et al., 2001).
Identication of the isolates assigned them to three main bac-
terial groups: gram-negative pseudomonads; gram-positive,
non-spore-forming nocardioforms; and the gram-positive,
spore-forming group, Paenibacillus (Table 13.3). This study
indicated that the rhizosphere of salt marsh plants contains
a diverse population of PAH-degrading bacteria, and the use
of plant-associated microorganisms has the potential for bio-
remediation of contaminated sediments. Contaminated sedi-
ment was obtained from Newtown Creek in the New York
Harbor, Brooklyn, New York. Chemical analysis showed the
dredged sediment to contain 2 to 7 ppm of the PAHs naph-
thalene, anthracene, and phenanthrene. An induction period
of 15 days was observed for pyrene, after which degradation
was complete in 10 to 15 days. It should be noted that both
the pure culture and microbial slurry experiments were per-
formed under highly oxygenated conditions and that the lim-
ited diffusion of oxygen into organic-rich sediments has been
found to restrict PAH biodegradation in the natural environ-
ment (DeLaune et al., 1981).
So
r
ption
Peat soils adsorb PAH compounds quite effectively. These
organics partition very strongly to carbonaceous soils (IT
Corporation, 1997; Pardue and Shin, 2000), and are not read-
ily desorbed (Pardue and Shin, 2000; Shin and Pardue, 2002).
It is therefore expected that PAHs would be sorbed and stored
in wetland peats. Partition coefcients in the range of 10
3

–10
4
L/kg were found for phenanthrene (Shin and Pardue, 2002).
Plant Uptake
Polycyclic aromatic hydrocarbons are not taken up by wet-
land plants to any signicant extent. Studies at Duluth,
Minnesota, have shown very low concentrations in above-
ground plant parts, and only trace amounts associated with
roots (IT Corporation, 1997). PAHs were found in dogwood
(Cornus stolonifera) and cattail (Typha spp.) shoots at the
Duluth site. Willows (Salix spp.) and bulrushes (Scirpus spp.)
have been reported to access tightly bound phenanthrene
(Gomez-Hermosillo et al., 2000; Gomez and Pardue, 2002).
Most of the PAH uptake was to the roots, and was associated
TABLE 13.3
Abilities of Several Isolates to Utilize a Variety of PAHs
% PAH Remaining
Isolate Naphthalene Biphenyl Fluorene Phenanthrene Pyrene
Pseudomonads
PR-N7 0 85 100 85 90
PR-N9 0 69 78 71 90
PR-N10 40 90 97 89 100
PS-P2 44 51 79 28 100
G
r
am-Positive Non-spore Formers
PR-N15 0 84 80 68 100
PR-P3 0 84 42 1 100
S
pore former

PR-P1 0 0 29 8 100
Source: Data from Daane et al. (2001) Applied and Environmental Microbiology 67(6): 2683–2691.
© 2009 by Taylor & Francis Group, LLC
522 Treatment Wetlands
with sorption. Because PAHs are not required for plant
metabolism and growth, up take is dependent on concentra-
tions and supplies to the pore water in the root zone.
Wetland System Studies
Testing of wetlands for PAH removal has presented mixed
results. Some tests show minimal removal, whereas others
show excellent reductions. Anecdotal monitoring of a cypress
swamp (forested peatland) receiving landll leachate in
Florida indicated no PAH removal after about 15 years
(Schwartz et al., 1999). However, other landll leachate con-
structed wetland studies have shown no detectable PAHs in
the efuent, including naphthalene at a Minnesota site; and
naphthalene, uorene, and phenanthrene at a Michigan site
(Kadlec, 2003c). Boving (2002) found virtually no retention
or removal of ten frequently detected PAHs in a stormwater
system comprised of ponds and wetlands. Boving showed that
heavier PAHs were present in urban freeway runoff at concen-
trations in excess of U.S. EPA benchmarks for chronic toxicity.
In particular, benzo(a)pyrene was present at over 1.0 Mg/L.
Subsurface ow constructed wetlands were used to treat
coke plant wastewaters at Sollac, Fos sur Mer, France (Jardinier
et al., 2001). Removal of 45% was achieved in 11 days’ deten-
tion. Jardinier et al. concluded that reedbeds were a valid
method to remove PAHs.
In a German study, naphthalene was removed using
hydroponic cultures of Carex gracilis and Juncus effusus

and using sand-bed reactors planted with Carex gracilis and
Juncus effusus, respectively, under batch and ow through
conditions (Wand et al., 2002). Concentrations of about
30 mg/L naphthalene were efciently eliminated over peri-
ods of up to six months. Vegetated cultures were found to
achieve a better removal rate than systems without vegeta-
tion. In the systems investigated, naphthalene was thought to
be mainly degraded by bacteria in the rhizosphere.
The behavior of the three-ring PAH phenanthrene was
investigated in a VF wetland system in Munich, Germany
(Machate et al., 1997), with overall removals of more than
99% at a detention time of 6.5 days. Intermediate degrada-
tion involved formation of 1-hydroxy-2-naphthoic acid as a
bacterial metabolite, which subsequently was also removed
in the wetland. Phenanthrene-degrading bacteria were enu-
merated, and found to be highest in the inlet zone of the sys-
tem (10
4
per mL) in comparison to the outlet end (fewer than
10 per mL). Feed concentrations of 0.385 mg/L phenan-
threne were reduced to less than 0.003 mg/L. The lava rock
substrate had only a small partition coefcient, in the range
0.1 to 1.8 L/kg.
An experimental subsurface ow constructed wetland
was developed in Curienne (Savoie-France), and operated
with a feed of wastewater dosed with uoranthene (Giraud
et al., 2001). Two beds were operated in series, with a total
detention time of three days. The inlet uoranthene concen-
tration was set at 6,660 mg/L, and no PAHs were detected in
the wetland outows. A total of 40 fungal species (24 gen-

era) were isolated and identified from samples (gravel and
sediments) from the test wetland and a control wetland. Fluor-
anthene was degraded efciently by 33 species whereas only
two species were able to remove anthracene by over 70%.
Salmon et al. (1998) evaluated removal of total hydro-
carbons in FWS mesocosms in France. The bed was planted
with Typha latifolia, and during the experiment, natural
development of Lemna minor occurred. Constructed wet-
lands removed 90% of total hydrocarbons.
13.2 CHLORINATED HYDROCARBONS
C
HLORINATED BENZENES
Chlorobenzenes were and are used in the production of
phenol, aniline, and DDT. Mono-, di-, and trichloroben-
zenes are used as solvents (Grayson, 1985). Various benzene
hexachloride isomers are used as broad-spectrum insecti-
cides, including Lindane™. Mono-, di-, and trichloroben-
zenes are resistant to photo-oxidation, and to both aerobic
and anaerobic degradation in purely aquatic environments,
with estimated half-lives of up to 6 to 24 months (Howard et
al., 1991). However, wetland environments are quite differ-
ent, because of the close interactions with plants and soils.
The major routes for removal of chlorobenzenes from wet-
land waters are: biological (microbial) degradation, sorption,
volatilization, and plant uptake.
Biodegradation
In general, chlorinated hydrocarbons may be dechlorinated
under anaerobic conditions, and the responsible microbial
consortia have been identied (van Eekert and Schraa, 2001).
Reductive dechlorination of chlorobenzenes occurs via an

anaerobic sequential pathway in wetland soils and sediments.
Jackson and Pardue (2000) established that the dichloroben-
zene (DCB) formed monochlorobenzene (MCB), which sub-
sequently mineralized:
DCB MCB CO H O Cl
2
ll

2
(13.1)
Their study showed that MCB produced reached half of the
initial DCB concentration in 30 days, and was accompa-
nied by the formation of methane. In separate experiments,
Jackson and Pardue (2000) recovered 13% of
14
C-MCB as
14
CO
2
in a surface sediment modulated decomposition.
Sorption
Chlorobenzenes sorb strongly to both organic and inorganic
wetland substrates (Pardue et al., 1993). Shin and Pardue
(2000; 2002) showed that there are both reversible and irre-
versible portions of the overall sorption. For several sediment
and soil samples, the reversible part ranged 1.88 a log
10
K
OC
a 2.88, while the irreversible part ranged 3.75 a log

10
K
OC
a
5.55. Thus, partitioning to organics is very strong, and domi-
nated by irreversible binding.
Suspended particulate matter forms a mobile sub-
strate for partitioned chlorobenzenes (Shugui et al., 1994).
© 2009 by Taylor & Francis Group, LLC
Organic Chemicals 523
The FWS wetland biogeochemical cycle creates and pro-
cesses very large quantities of total suspended solids (TSS),
and thus can play an important role in organic chemical
cycling and removal in these systems. The wetland environ-
ment is further complicated by the presence of large mol-
ecules of humic substances, which comprise a good share
of dissolved organic carbon. Organic solutes can partition to
these dissolved humic substances, thus creating two soluble
forms with different chemical characteristics.
Volatilization
The chlorobenzenes are volatile, and therefore may be easily
lost from water, especially shallow water bodies, such as free
water surface wetlands. Of direct interest are the estimates,
based on wetland data, of Keefe et al. (2004a) for a FWS wet-
land in Arizona. Values were determined of K
w
 0.011 m/hr
(100 m/yr) for chlorobenzene, and K
w
 0.008 m/hr (70 m/yr)

for 1,4 dichlorobenzene. These are relatively high rates, com-
pared to other pollutants treated in FWS wetlands.
Plant Uptake
Plants are capable of taking up organic chemicals (Trapp
and Matthies, 1995), and processing them in a number of
ways. They may be carried through the plant into the atmo-
sphere via the transpiration ux, metabolized, or accumu-
lated in plant tissues. For example, Leppich et al. (2000)
found up to 100 mg/kg of various chlorobenzenes in black
willow (Salix nigra) bark, and up to 25 mg/kg dw in leaves.
The willows were growing in a contaminated swamp site,
with large concentrations of di-, tri-, penta-, and hexa-chlo-
robenzenes. Leppich et al. (2000) concluded that partition-
ing of these organics to the plants formed an important part
of the site model.
Ove
rall Removal Coefficients
Despite the investigations detailed above, there is no reported
treatment wetland that has specically been designed to tar-
get chlorobenzenes. Therefore, results from several treatment
technologies are examined here to gain some insight as to the
anticipated rates of removal to be expected in wetlands.
Wetlands
Keefe et al. (2004a) calculated removal rate coefcients from
a FWS dataset to be 135 m/yr for chlorobenzene, and 67 m/yr
for 1,4 dichlorobenzene. The wetland was a 1.4-ha free water
surface system of depth about two feet. The detention time
was 3.9 days. These results correspond to a 67% reduction of
1,4 dichlorobenzene from inlet to outlet. Keefe et al. (2004a)
attributed the reduction to volatilization, but concluded that

theoretical predictions were only good for order-of-magni-
tude estimation.
Braeckevelt et al. (2006) studied monochlorobenzene
reduction in a pilot-scale horizontal subsurface ow wetland
constructed of local soil materials. Contaminated groundwater
containing up to 22 mg/L of monocholorobenzene was added
to the system at a hydraulic loading rate of 2.3 cm/d. Choro-
benzene reductions of up to 77.1% were observed in the
system. Monochlorobenzene reductions were highest in the
upper soil layer, possibly due to volatilization, and decreased
to 37.1% at the bottom of the wetland bed (0.5 m). Enrich-
ment of
13
C and low dissolved oxygen concentrations suggest
that reductive dehalogenation under anaerobic conditions
was the dominant removal mechanism.
CHLORINATED ETHENES
Perchloroethylene (PCE) and trichloroethylene (TCE) are
solvents that saw extensive use for metal cleaning and other
applications in previous decades. At many locations, these
materials found their way into groundwater, creating hazard-
ous waste conditions, because of concern due to their carcino-
genic properties. Wetlands provide environments for several
mechanisms of removal of chlorinated aliphatic organics,
including sorption, volatilization, reductive dechlorination,
direct biological oxidation, co-metabolism, and plant uptake
and metabolism (Pardue et al., 1993; Parkin, 1999; Pardue
et al., 2000).
Kassenga (2002) conducted continuous vertical ow col-
umn and wetland microcosm studies to investigate the atten-

uation potential of chlorinated volatile organic compounds.
Calibrated simulations indicated that removal of TCE in con-
structed wetland columns was controlled by biodegradation
whereas both sorption and biodegradation were important
natural wetland columns. Kassenga et al. (2003) evaluated
the removal of cis-1,2-dichlorethene (cis-1,2-DCE) in upow
wetland mesocosms planted with Scirpus americanus. The
results conrmed that biodegradation was occurring in the
system, and sand, peat, and Bion soil mixture had greater
degradation rate than the sand and peat mixture.
Lorah et al. (1997; 2002) observed complete removal of
TCE and daughter products as contaminated water moved
upward through peat to the surface. Reducing conditions
were present in the peat, and both methanogenic and iron-
reducing zones were identied. This important study pro-
vided the impetus to examine the future role of natural and
constructed wetlands in the remediation of TCE.
A former TCE recycling plant is now the site of a city
park in New Brighton, Minnesota. A TCE plume in a sur-
cial aquifer discharges into the wetland of an adjacent lake
(Bankston et al., 2002). Transect studies showed TCE in
monitoring wells just upgradient from the wetland, and to
a lesser degree in the fringe of the wetland. The anaerobic
degradation product of TCE, cis-1,2-DCE, was detected in
the aquifer and the wetland. It was hypothesized that the
indigenous cattail (Typha latifolia) assisted in phytoreme-
diation. Accordingly, microcosm studies were performed to
determine the fate of the removed TCE. The recovery of
14
C

totaled 94.1%, of which 46.8% was volatilized, most likely as
[
14
C] TCE because it was added to the microcosms by surface
application. Plant tissue contained 38.2% of the
14
C; 5.3%
was present as [
14
C] CO
2
, and 3.7% was recovered from the
© 2009 by Taylor & Francis Group, LLC
524 Treatment Wetlands
soil and water. This data suggested that natural attenuation
is a potential bioremediative strategy for TCE-contaminated
wetlands.
Given that natural wetlands contribute to reduction of
TCE, the next logical step is the reconstruction of former
wetlands that are positioned in the ow path of a contamina-
tion plume. Richard et al. (2002; 2003) lled and planted a
dredged channel that was conveying TCE from a contami-
nated site to a lake. In this case, an aquatic feature was con-
verted to wetland to aid in treatment. This Minnesota project
was completed in 2000.
In other cases, terrestrial landforms may be converted to
constructed treatment wetlands to provide reductions in TCE.
The Schilling Farm Project at Hillsdale, Michigan, intercepts
a TCE plume for treatment in a constructed FWS wetland,
built in a former corn eld. The wetland system is made up

of
four treatment cells in parallel (see Figures 4.21 and 4.26).
The two large central cells were sited to intercept the plume,
and the two small anking cells were added to ensure the
full capture of that plume. Groundwater moves down a slight
incline into a 4-m deep capture trench, designed to intercept
approximately the top 2 m of the underground ow. This
trench is lled with coarse rock to eliminate safety hazards
and control rodent problems. The water ows upward and
out across the FWS wetland cells, carrying TCE and DCE
into the wetlands. During passage through the wetland, TCE
undergoes reductive dechlorination to DCE and then to vinyl
chloride (VC) in anoxic zones, and these are further degraded
to carbon dioxide and water in aerobic zones. There is also
volatilization of VC, and to a lesser extent DCE and TCE.
Water is collected in rock-lled trenches, approximately 1 m
deep, at the downstream end of all four cells. These four
outows are metered, and merged to form a single project
outow for compliance monitoring. Additionally, another
independent rock-lled trench is positioned across the entire
downstream end of the four cells, which is drained via perfo-
rated pipe into the compliance outow. The purpose of this
trench is to capture waters that may pass totally underneath
the treatment wetland.
Control of this wetland system is solely by means of set-
ting water levels in each of the four cells, by use of the weir
settings in the outlet structures. It is totally passive, with no
pumps, and runs year-round in a cold climate. Therefore,
water level control must compensate for ice formation. The
correct operating strategy is the subject of ongoing investiga-

tions. Herbivory and short-circuiting were created by musk-
rats (Ondatra zibethicus), necessitating removal of both the
animals and their habitat, by lling trenches with large rock,
and by fencing the wetland.
T
a
ble 13.4 shows the performance results for the sys-
tem for 88 of the 95 months of operation, October 1998
through September 2005. Figure 13.2 shows performance for
October 2003 through September 2005, respectively. The rst
seven months were a dormant period for vegetation planted in
autumn 1998, which remained sparse for that period. Efu-
ent standards for TCE (limit 150 µg/L) were met in 84 of
88 months after startup. Exceedances for VC (limit 13 µg/L)
tend to occur in late summer, when wetland surface ow con-
tributes little to the outow.
In summary, these projects all utilize the characteristics
of wetlands to reduce TCE. Research work continues at the
time of this writing at Wright–Patterson Air Force Base in
Ohio (Entingh, 2002; Blalock, 2003). The addition of recycle
in 2007 has eliminated VC from the outow.
13.3 ORGANIC CHEMICALS
E
XPLOSIVES
It has been estimated that approximately 100 military bases
and explosives manufacturing facilities have soil and/or
groundwater contaminated with munitions (Medina et al.,
2000). Studies have shown that plants are capable of trans-
forming 2,4,6-trinitrotoluene (TNT) without microbial con-
tribution, but very little accumulation of TNT has been found

in plant material. Therefore, plant-enhanced degradation, or
phytoremediation, of TNT by aquatic macrophytes has been
proposed as a promising groundwater treatment process.
TABLE 13.4
Performance of the Schilling Farm Constructed FWS Wetlands
Annual Winter Spring Summer Autumn
HLR (cm/d) 1.00 1.14 1.39 0.85 0.60
T (nC)
11.9 1.0 12.4 23.6 12.7
TCE C
i
(µg/L) 726 726 726 726 726
TCE C
o
(µg/L) 37 87 25 18 21
cis-1,2-DCE C
i
(µg/L) 472 472 472 472 472
cis-1,2-DCE C
o
(µg/L) 44 65 21 46 46
VC C
o
(µg/L) 5.8 5.4 1.6 8.5 7.6
Note: Inlet concentrations are average undiluted plume values.
© 2009 by Taylor & Francis Group, LLC
Organic Chemicals 525
TNT is a reactive molecule that biotransforms readily
under both aerobic and anaerobic conditions to give amino-
dinitrotoluenes (Brannon and Myers, 1997; Hawari et al., 2000;

Xiang, 2001; Esteve-Nunez et al., 2001). The resulting amines
biotransform to give several other products, including azo,
azoxy, acetyl, and phenolic derivatives, leaving the aromatic
ring intact. Little or no mineralization is encountered dur-
ing bacterial or wetland treatment. The nonaromatic cyclic
nitramine explosives hexahydro-1,3,5-trinitro-1,3,5-triazine
(RDX) and octahydro-1,3,5,7-tetranitro-1,3,5,7-tetrazocine
(HMX) lack the electronic stability enjoyed by TNT or its
transformed products. Therefore, an enzymatic change on
one of the N-NO
2
or C-H bonds of the cyclic nitramine could
lead to a ring cleavage, and subsequent mineralization.
Medina et al. (2000) performed a series of batch-scale
experiments to assess the engineering kinetics of phytodeg-
radation of TNT under a variety of operational conditions.
Parrotfeather (Myriophyllum aquaticum) was hydroponically
grown in laboratory microcosms. TNT was degraded accord-
ing to a near-rst order model in vegetated microcosms, but
not in unvegetated systems (Figure 13.3). Measurements at
varying plant densities indicated that rate constants increased
with increasing plant abundance. Removal rate constants
also increased with increasing temperature from 2 to 34°C,
leveled off between 34 and 43°C, and at 54°C, no activity
was found. This pattern is also found for enzyme kinetics,
in which rates increase until the enzyme denatures. Plants
appeared healthy up to 34°C, but wilted at 43°C. Plants incu-
bated at 54°C were dead by the end of the experiment. The
modied Arrhenius temperature coefcient was 1.093 for
water temperatures between 2 and 30°C. Hughes et al. (1997)

found that Eurasian water milfoil (Myriophyllum spicatum)
degraded TNT to aminonitrotoluenes, whereas unvegetated
controls did not.
0
100
200
300
400
500
0 30 60 90 120 150 180 210 240 270 300 330 360
Julian Day
TCE or DCE (µg/L)
TCE Out
DCE Out
TCE In
DCE In
FIGURE 13.2 Annual pattern of treatment of TCE and DCE in the Schilling Farm wetland system. The inlet concentrations have been
adjusted for dilution with clean groundwater entering the ank cells. Winter conditions provide lesser treatment. There is a second period of
lesser treatment in late summer and autumn, which is occasioned by very low surface ows and dominance of underground ows.
0.0
1.0
2.0
3.0
4.0
5.0
6.0
7.0
0 5 10 15 20 25
Time (hr)
TNT (mg/L)

Parrotfeather
Unvegetated
FIGURE 13.3 Disappearance of TNT in batch microcosms. The vegetated system points represent the mean of ve replicates. The half-life
of TNT was about seven hours. (Data from Medina et al. (2000) Water Research 34(10): 2713–2722.)
© 2009 by Taylor & Francis Group, LLC
526 Treatment Wetlands
Zoh and Horne (2000) also performed a series of batch-
scale experiments. Utilizing straw, cattail, and bulrush lit-
ter in water, they found no degradation without litter, and
rst-order behavior in the presence of litter. Sorption and
14
C
studies indicated that removal was due to initial sorption fol-
lowed by degradation to aminonitrotoluenes. Volumetric rate
coefcients (k
V
) translate to areal rates (k
A
) on the order of
25 m/yr for TNT, regardless of litter type. A modied Arrhe-
nius temperature coefcient (Q) of 1.17 t the observed dif-
ference in rates at 10 and 20°C.
The ability of ten species of submerged aquatic to phy-
toremediate explosives-contaminated groundwater was
investigated by Best et al. (1997a). Phase I of this project
provided for laboratory-scale plant screenings to evaluate
locally adapted aquatic and wetland species for their dif-
ferential ability to diminish levels of TNT and RDX. These
were evaluated under hydroponic batch conditions. Analysis
of the data according to a batch rst-order areal model shows

remarkable similarity among species (Table 13.5).
Best et al. (Best et al., 1999a,b) reported that per unit
of mass, uptake of TNT was higher in submerged (Elodea
canadensis, Potamogeton pectinatus, Heteranthera dubia,
Myriophyllum aquaticum) rather than emergent species
(Acorus calamus, Phalaris arundinacea, Scirpus cyperinus)
and biotransformation of TNT had occurred in all plant treat-
ments after a seven-day incubation in 1.6 to 3.4 mg/L TN.
TNT declined less with substrates, and least in controls with-
out plants. Mineralization to CO
2
was very low, and evolu-
tion into C-volatile organics negligible. RDX disappeared less
rapidly than TNT from groundwater.
Cattails (Typha angustifolia) in FWS mesocosms were
used to test treatment of mono-, di-, and trinitrotoluene mix-
tures at the Volunteer Army Ammunition Plant, Tennessee
(Best et al., 2000; 2001). Rate coefcients (Table 13.6) ranged
from 16 to 45 m/yr. The potential contribution of photodeg-
radation was determined by shielding nonplanted mesocosms
from UV in sunlight. Radiation accounted for 30% of TNT,
60% of DNT, and 10% of NT removal in the absence of plants.
TABLE 13.5
Batch Kinetics of TNT Reduction for Several Submersed Species
with Data Including Two Different Plant Densities
Scientific Name Common Name k
A
(m/yr)
Myriophyllum aquaticum Parrotfeather 39
Myriophyllum spicatum Milfoil 48

Egeria densa Egeria 30
Eiodea canadensis Elodea 40
Vallisneria amerilcana Vallisneria 50
Potamogeton crispus Curly leaf pondweed 48
Potamogeton pectinatus Sago pondweed 40
Heteranthera dubia Star-grass 45
Eleocharis parvula Spikerush 49
Chara vulgaris Stonewort 49
Source: DatafromBest et al. (1997a) Screening aquatic and wetland plant species for phytore-
mediation of explosives-contaminated groundwater from the Iowa Army Ammunition Plant.
Technical Report EL-97-2, U.S. Army Corps of Engineers Waterways Experiment Station:
Vicksburg, Mississippi.
TABLE 13.6
Continuous-Flow Kinetics of Nitrotoluenes Reduction for Typha angustifolia Mesocosms
Contaminant
Inlet
(mg/L)
Outlet
(mg/L)
Percent Removal
(season)
Areal Rate Constant
(m/yr)
Trinitrotoluene 2.73 0.30 79% 31
2,4-dinitrotoluene 16.66 6.00 58% 16
2,6-dinitrotoluene 5.17 1.40 61% 17
2 nitrotoluene 42.64 2.20 — 45
Source: Data from Best et al. (2000) Explosives removal from groundwater at the Volunteer Army Ammunition Plant, Tennessee, in small-scale wetland mod-
ules. Means and Hinchee (Eds.), Wetlands and Remediation: An International Conference; Battelle Press: Columbus, Ohio, pp. 365–374; and Best et al.
(2001) Water Science and Technology 44(11–12): 515–521.

Note: Season refers to June through October.
© 2009 by Taylor & Francis Group, LLC
Organic Chemicals 527
However, planted mesocosms had higher removal rates than
unshielded unplanted systems (Best et al., 2001).
The U.S. Army Corps of Engineers constructed treat-
ment wetlands with the purpose of containing and treating
residual TNT (2,4,6-trinitrotoluene) and RDX (hexahydro-
1,3,5-trinitro-1,3,5-triazine) contamination at the Iowa Army
Ammunition Plant (Thompson et al., 2003). Screening stud-
ies established the relative capabilities of various plants to
reduce explosives found at the site (Best et al., 1997a,b).
Demonstration studies were then conducted to establish veg-
etative uptake, fate, and toxicity of TNT and RDX. Results
from two years of testing indicated that TNT concentrations
were undetectable in surface water or plant tissues. Concen-
trations of RDX in wetland surface waters were below the
U.S. EPA human health advisory level of 2 µg/L and were
undetectable in plant tissues during the summer months.
Based upon these results, and other wetland design con-
siderations, testing then proceeded to pilot scale. Behrends
et al. (1997) reported that a FWS wetland system of two cells,
each 9.4 × 24 meters, performed poorly for TNT and RDX
removal at Milan, Tennessee. This was attributed to an inabil-
ity to maintain the vegetation, parrotfeather (Myriophyllum
aquaticum), in the face of tadpole herbivory and hailstorms. In
contrast, a two-cell SSF system (473 m
2
total) performed well
for TNT, RDX, and HDX. Cold weather limited the gravel-

based system’s ability to reduce the total nitrobodies below the
50 ppb demonstration goal (Sikora et al., 1997).
Analysis indicated that the full-scale system could be
economically resized to overcome the winter performance
decit by increasing the retention time from 10–14.5 days
(ESTCP, 1999).
Xiang (2001) tested VF wetland mesocosms, in series, the
rst two planted with Typha spp., and the last three with Scir-
pus lacustris. Virtually all of the TNT was removed in the
rst two cells, at a combined detention time of 2.4 days. In
radiolabelled experiments, less than 0.1% of the radioactivity
added to the system as TNT was present in the metabolically
produced
14
CO
2
. Consequently, mineralization was not an
important pathway. Small amounts of mono- and dinitrotolu-
enes were detected, indicating some utilization of a pathway
that removed NO
2
groups. Sorption experiments showed neg-
ligible binding to the media. Planted cascades removed 80%
of the TNT, while unplanted removed only 23%. It was noted
that photodegradation might have contributed to the removals
in the planted tank. Only small amounts of TNT and amino-
dinitrotoluene were found in plants in the rst tank, about 1.0
mg/kg of TNT in shoots, and 0.25 mg/kg (dry weight unless
otherwise noted) in roots. Slightly greater amounts of amino-
dinitrotoluene were found, 1.8 mg/kg in stems and 0.8 mg/kg

in roots. Xiang (2001) found enhanced removal with the addi-
tion of a carbon source (molasses or sucrose).
DE-ICING COMPOUNDS
The waste generated by aircraft de-icing operations repre-
sents a threat to the environment arising from the high bio-
logical oxygen demand of commonly used propylene and
ethylene glycols (Castro et al., 2004). Other candidates for
de-icing include urea, which can also pose a pollution threat
to nearby receiving waters (Thoren et al., 2003). In temper-
ate climates, pollutant loadings are seasonal, and generally
occur from November to April.
Both glycol and urea are biodegradable. Microbial metab-
olism of glycol within a wetland system utilizes the material
as it would other forms of BOD
5
(see Chapters 8 and 13). Deg-
radation of urea follows patterns established for other species
of nitrogenous compounds (see Chapter 9). However, pollutant
loads occur at cold temperatures, which minimize some bio-
logical processes, notably those for nitrogen reduction. The
threat of bird–aircraft strikes indicates the use of SSF wet-
lands in many cases.
The applicability of treatment wetlands for reducing the
concentration of ethylene glycol is being tested or implemented
at several airports, including Pearson Airport in Toronto,
Edmonton International Airport in Alberta, Heathrow
Airport in London, and Westover Air Reserve Base in Mas-
sachusetts. Constructed wetlands are also in use for nitrogen
de-icer control at Kalmar Airport, southeast Sweden. Pilot
wetland testing has occurred at the Airborne Express airport

in Wilmington, Ohio (Naval Facilities Engineering Service
Center and Wetland Solutions, Inc., 2004b); Zurich, Switzer-
land (Richter et al., 2004); and Schonefeld Airport in Berlin,
Germany (Abydoz Environmental, Inc., 2005).
In 2000, the Greater Toronto Airports Authority (GTAA)
constructed the rst full-scale vertical ow treatment wet-
land in Canada for reduction of ethylene glycol in stormwa-
ter (Flindall et al., 2001). The facility consists of a two-cell
treatment wetland preceded by a sedimentation forebay.
The wetland facility collects water discharged from a drain-
age area of approximately 382 hectares, comprised of ter-
minal aprons, taxiways, and runways where aircraft receive
glycol de-icing compounds. The facility provides 24 to 48
hours of retention to attenuate storm ows and provide removal
of suspended sediment. Cell 1 of the wetland is a 0.42-ha VF
system, and Cell 2 is a 1.38-ha FWS system, and is primar-
ily intended for storage of runoff rather than treatment. Both
treatment cells were planted with Phragmites australis. A
key component of the facility is the sediment forebay, which
provides temporary storage of ows entering the facility. The
forebay allows removal of some of the sediment load in the
runoff, which is necessary to prevent blockage of the vertical
ow reed bed. Additionally, forebay storage allows intermit-
tent dosing of the vertical ow reed bed, which is an important
design feature of the vertical ow system.
Urea (chemical formula: NH
2
-CO-NH
2
) is used as a

de-icing agent at Kalmar Airport, southeast Sweden (Thoren
et al., 2003). To reduce stream transport of nitrogen from
airport, agricultural, and other diffuse sources to the Baltic
Sea, a FWS wetland system was constructed in 1996. The
18-ha wetland consists of four ponds (1.5 m deep), which are
dominated (75%) by the submerged plant Elodea canaden-
sis, separated by shallower water (0.5 m deep) dominated by
common reed (Phragmites australis). The theoretical water
retention times for the wetland are 5.3 and 2.4 days during
© 2009 by Taylor & Francis Group, LLC
528 Treatment Wetlands
average and high water discharges, respectively. Annual wet-
land retention of TN varied in the range of 6 to 36% dur-
ing 1998 to 2001. During airport de-icing, January through
March 2001, 660 kg urea-N out of 2,600 kg applied urea-N
reached the wetland, and approximately 40% of the incom-
ing urea-N was eliminated in the wetland system at air tem-
peratures around 0°C.
The Edmonton Airport in Canada has a subsurface wet-
lands treatment facility to treat stormwater contaminated
with ethylene and propylene glycol from the de-icing activi-
ties that occur at the airport (Higgins and Maclean, 2002;
Edmonton Airports, 2003). The HSSF facility is comprised
of six trains of two cells each. The 4.5-ha wetlands started
operating in 2001. Cells are vegetated with cattails (Typha
spp.), with the exception of one cell being planted with reed
canary grass (Phalaris arundinacea). The new system inte-
grates the stormwater collection system serving the apron
area by channeling the uids into a 300,000 cubic meter
detention facility, prior to the wetlands. Treated water ows

out of the area to surface receiving waters. The cost to con-
struct the wetlands was CDN $2 million.
A 0.24-ha HSSF wetland system was installed at the
Westover Air Reserve Base in Chicopee, Massachusetts to
demonstrate the treatment of stormwaters from on-site de-
icing operations (Karrh et al., 2002; ESTCP, 2004; Naval
Facilities Engineering Service Center and Wetland Solu-
tions Inc., 2004a). The apparent hydraulic residence time in
the bed was estimated at 1.3 to 1.9 days. Event mean ow
through the wetland was 278 m
3
/d for an average hydrau-
lic loading rate of 11.4 cm/d. Peak inow BOD concentra-
tions ranged from 974 to 15,098 mg/L during ten de-icing
events in 2002, and these were reduced by more than 50%
in 5 of the 10 events. Event ow-weighted mean concen-
trations for BOD were 1,183 and 937 mg/L at the wetland
inow and outow for an estimated mass reduction of 21%.
Overall averages for the season are given in Table 13.7. The
estimated mass removal rate was 286 kg/ha·d. The system
was overloaded by a factor of four due to extreme condi-
tions during the winter of testing.
London’s Heathrow airport began pilot testing of wet-
lands for glycol removal in 1994 (Chong et al., 1999; Revitt
et al., 2001). Surface and subsurface ow systems and oat-
ing rafts were operated, with detention times of less than
one day. Modest reductions (approximately 30%) of various
glycols were achieved (Worrall et al., 2002). Subsequently, a
1.2-ha oating raft wetland and a 2.08-ha HSSF were built,
and planted with Phragmites. The design detention times

were both just over one day (Revitt et al., 2001; Richter et
al., 2003). Evaluation of performance is in progress, with a
ve-year data acquisition period (Richter et al., 2003; Richter
et al., 2004). The system responds to pulse events by delaying
and reducing the incoming COD peak (Figure 13.4).
The Buffalo–Niagara International Airport (BNIA) has
studied the use of aerated vertical ow wetlands for glycol
removal. A pilot-scale system was constructed to deter-
mine necessary design parameters (Higgins et al., 2006a).
A volumetric rate coefcient of k
v
 3.3 d
1
was observed for
reduction of glycol-produced COD. This rate coefcient was
developed based on a hydraulic ow model of two tanks in
series (2TIS) and an operating temperature of 4°C.
PHENOLS
Phenol is a compound of considerable industrial importance
and a frequent pollutant in industrial efuents. It also has been
studied for over four decades in the wetland treatment context.
Early work at the Max Planck Institute in Krefeld, Germany,
focused on SSF microcosms with bulrushes (Schoenoplectus
(Scirpus) lacustris) and resulted in a series of publications on
removals and effects on plant physiology (for example, Seidel,
1966). That data showed degradation rates of 5 to 20 g/m
3
·d,
and in some instances, an induction period of one to three days
TABLE 13.7

Summary of Results from the 2002–2003 De-icing Season at Westover
Air Reserve Base
Constituent Inlet Outlet Percent Reduction
BOD (mg/L) 2,226 2,094 6%
COD (mg/L) 1,883 1,335 29%
MeBT (mg/L) 0.68 0.72
6%
DO (%) 52.2 47.7
9%
pH (S.U.) 7.58 9.54 —
Redox (mV) 391 172
56%
Turbidity (NTU) 5.22 4.61 12%
Note: MeBT is methyl-1H-benzotriazole, an additive to the de-icing uid.
Source. Data from ESTCP (2004) Enhanced biological attenuation of aircraft deicing uid runoff
using constructed wetlands. Environmental Safety Technology Certication Program (ESTCP)
Cost and Performance Report CP-0007, U.S. Department of Defense: Arlington, Virginia.
© 2009 by Taylor & Francis Group, LLC
Organic Chemicals 529
before reduction commenced. The Seidel data may be mod-
eled by a zero-order reaction. The zero-order rate constant for
the Krefeld data was k
0
 81 o 18 mg/d in summer, and 34
o 10 mg/d in winter. The corresponding modied Arrhenius
temperature coefcient is Q 1.083.
Wolverton and McDonald (1981) evaluated removal of phe-
nol in experimental mesocosms lled with 16 cm of 2.5–7.5 cm
diameter railroad rocks covered with 5 cm deep of 0.25–1.3 cm
diameter pea gravel. The mesocosms were planted with Phrag-

mites australis and Typha latifolia. After 24 hr the concentra-
tion of phenol dropped down from 104 to 7 mg/L in Phragmites
unit whereas in Typha unit the phenol concentration dropped
down from 101 to 17 mg/L during this time. After 48 hours the
respective phenol concentrations were 4 and 7 mg/L. Phenol
concentration in unplanted control unit dropped from 100 to 40
and 28 mg/L after 24 and 48 hours, respectively.
Microcosm research on FWS wetlands with cattails pro-
duced similar results (Kadlec and Srinivasan, 1995). The
average value of the zero-order rate constant was 76 o 33 mg/d
(loss  0.55 g/m
2
·d) for vegetated microcosms, and 108 o
26 mg/d (loss  0.78 g/m
2
·d) for unvegetated microcosms.
The sizes of these microcosms were comparable to those at
Krefeld. There was little selective loss of phenol from micro-
cosms containing only water and the phenol dose. This study
showed no effect of water depth, but a slight effect of soil
type, with more organic content promoting phenol removal.
There were higher evaporative losses in summer, with phenol
being lost with water at the bulk water concentration. There
was not signicant evaporative concentration or selective
stripping of phenol.
Polprasert et al. (1994; 1996) operated both FWS micro-
cosms and a FWS pilot wetland in Bangkok, Thailand, for
phenol reduction at high strength. Nearly complete removal
of phenol was measured for loading rates up to 300 kg/ha·d.
Some phenol was volatilized; some was found in the roots of

the cattails. Removal followed a rst-order model, with rate
constants on the order of 100 m/yr. About one third of the
removal was reported due to volatilization, and the remaining
two thirds due to biodegradation, plant uptake, and sorption.
Abira et al. (2005) studied duplicate pilot scale subsur-
face ow wetlands for phenol reduction, with two cells each
of unvegetated gravel, Cyperus papyrus, Phragmites mauri-
tanius, and Typha domingensis. Removals in the 15-month
study ranged from 50–75%, with higher reductions for ten
days detention compared to ve days. Unvegetated cells
performed nearly as well as vegetated cells. The paper mill
wastewater had high COD (300–600 mg/L), but only about 1
to 3 mg/L of phenol.
The Jinling Petrochemical Company, Beijing, China,
reported reductions in several parameters, including phenol, in
water hyacinth wetlands (Tang and Lu, 1993). Two wetlands,
25 m r 150 m r 1 m deep, received wastewater at detention
times ranging from 2 to 12 days. One channel was 80% cov-
ered by Eichhornia crassipes, the other was unvegetated. The
calculated rate constants for phenol removal were 150 m/yr
for the vegetated system, and 76 m/yr for the unvegetated
system.
Dong and Lin (1994a) report data and models for a
research facility polishing secondarily treated petrochemi-
cal wastewater in six wetlands and six ponds comprising a
1.5-ha test unit at Yanshan Petroleum, in Niukouyu, China.
This research facility tested a variety of plants and soil sub-
strates over the seasons. Phenol was reduced by 28–37%,
with higher values in warm months. Dong and Lin (1994b)
then described a full-scale facility, built in 1990, consist-

ing of 50 ha of wetlands and ponds treating 100,000 m
3
/d.
Average inow and outow phenol concentrations were
27 mg/L and 10 mg/L, respectively, with a mean removal
efciency of 63%. The full-scale facility had a phenol rate
constant removal rate coefcient of 100 m/yr.
The treatment system at Amoco Mandan, North Dakota,
is comprised of an API separator, lagoon, and a series-parallel
set of FWS wetlands. Approximately 5,700 m
3
/d of process
water is directed to an API separator for primary treatment,
and then passed through the oxidation lagoon for second-
ary treatment. Process wastewater and stormwater are then
directed through an 0.8-km earthen canal to 6 of the 11
cascading ponds (16.6 ha) before eventual discharge from
Dam 4 to the river. The remaining ve ponds (19.1 ha) are
reserved for wildlife. (Litcheld and Schatz, 1989; Litcheld,
1990; Litcheld, 1993). Removal of phenol amounted to 94%
with an average outow concentration of 5 mg/L.
Microcosms and pilot scales display removal rates that
are greater than those observed in some full-scale eld situ-
ations. For instance, the Mandan, North Dakota, wetlands
do not reduce phenols to zero, as they should, because of a
detention time on the order of months (Litcheld and Schatz,
1989; Litcheld, 1990; Litcheld, 1993). Further, the loss
rates for the Jinling project are also much smaller than would
be predicted from the microcosm data (Tang and Lu, 1993).
0

20
40
60
80
100
120
02468101214
Time (days)
COD (mg/L)
Inlet
Outlet
FIGURE 13.4 Time series of COD into and out of the Heathrow
reedbeds. (Data from Richter et al. (2004) Treatment performance
of Heathrow constructed wetlands. Liénard and Burnett (Eds.).
Proceedings of the 9th International Conference on Wetland Sys-
tems for Water Pollution Control, 26–30 September 2004; Asso-
ciation Scientique et Technique pour l’Eau et l’Environnement
(ASTEE), Cemagref, and IWA: Avignon, France, pp. 125–132.)
© 2009 by Taylor & Francis Group, LLC
530 Treatment Wetlands
Both of these systems receive phenol at concentrations much
lower, by a factor of more than 100, than the lab studies. It is
clear that microcosms are not adequate for determination of
design parameters for phenol. For instance, the areal mixing
efciency of the wetland may be an important determinant
of trace amounts of phenol in the efuents from pilot and
full-scale wetlands.
The Listowel studies included the monitoring of phenols
in ve wetlands for four years (Herskowitz, 1986). These
substances were present in the aeration cell and in the lagoon

that provided the wetland inuent, at concentrations on the
order of 10 µg/L. Winter levels were higher than summer,
and the wetlands did not provide removal at these low con-
centration levels.
SURFACTANTS
Linear alkylbenzene sulfonate (LAS) is a widely used syn-
thetic surfactant for domestic detergents. This group of
substances is amenable to phytoremediation using both
terrestrial and wetland plants (Schwitzguebel et al., 2001).
Wetland plant detritus is an important source of carbon
and energy and offers a large surface area for sorption of
dissolved organics. To examine the biodegradation of sur-
factants by detrital microorganisms, Federle and Ventullo
(1990) obtained submerged oak leaves from a laundromat
wastewater pond, which were incubated in water amended
with
14
C-labeled LAS. Sorption and evolution of
14
CO
2
were
followed with time. LAS that was sorbed to the detritus was
mineralized without a lag, with a half-life of 12.6 days. This
study showed that detritus represents a signicant site of sur-
factant removal in detritus-rich systems.
Barber et al. (2001) performed synoptic sampling of sev-
eral existing treatment wetlands, including analysis for LAS
(Table 13.8). Results were mixed, with removal occurring in
some, but not all, systems.

Adsorption and degradation were found to both be opera-
tive in the wetland and in companion laboratory research (Inaba,
1992). LAS was effectively reduced in a FWS Typha–Phrag-
mites wetland. The 474-m
2
system treated gray water from a
100-inhabitant community. The calculated plug ow volumet-
ric removal coefcients (k
V
) for the removal of these surfac-
tants were in the range of 30 to 80 m/yr, depending upon chain
length and isomeric form, with greater removal rates for the
longer chain lengths. Seasonal dependence was strong, with a
temperature factor of Q 1.087 o 0.005 averaged over all forms
of the detergent. The inlet concentration totaled about 5 mg/L,
and was reduced by over 90% in summer and over 40% in win-
ter at an average hydraulic loading of approximately 5 cm/d.
The surfactant partitioned strongly to wetland TSS, with a
partition coefcient of 30 L/kg for chain length 11, isomer 5.
Values for other forms of the surfactant were higher, ranging
up to 520 L/kg for chain length 13, isomer 2. Sorption was
not diminished at cold temperatures. Therefore, the sorption
capacity and TSS trapping efciency of the wetland partially
offset the diminished degradation during winter.
LAS was investigated in SSF pilot subsurface ow con-
structed wetlands comprised of eight beds of 55 m
2
with dif-
ferent aspect ratios, two water depths and two medium sizes
(Huang et al., 2004). The Phragmites wetlands treated urban

wastewater. Under the most favorable conditions, LAS was
biodegraded up to 71%. Shallower water depth (27 cm) was
considerably more efcient than deeper (47 cm) for LAS
removal. Water temperature has a signicant positive effect
on the LAS reduction. Biodegradation of LAS was found to
occur under both sulfate- and nitrate-reducing conditions.
C13 LAS homologues were generally removed in higher
extent than the shorter alkyl chain counterparts, thus rein-
forcing the ndings of Inaba (1992).
Del Bubba et al. (2000) used a pilot HSSF system planted
with Phragmites australis to study LAS removal in Florence,
Italy. With HLR of 3.7 cm/d the inow LAS concentra-
tions in unltered samples of 278 mg/L and 18.8 mg/L were
reduced to 0.18 mg/L and 0.06 mg/L. This resulted in respec-
tive removal efciencies of 99.9% and 99.7%. The authors
also pointed out that the biodegradation of LAS homologues
increased with increasing carbon atom number: C14  C13 
C12  C11.
Kantawanichkul and Wara-Aswapati (2005) reported that
LAS inow concentration of 40 mg/L was reduced by 89%
in an experimental HSSF constructed wetland planted with
Typha angustifolia in Thailand. The highest LAS removal
TABLE 13.8
Results of Synoptic Sampling of Linear Alkylbenzene Sulfonate (LAS) at Several Treatment Wetland Sites
Site Wetland Type Source Water Area
(ha)
T
n
(days)
C

i
(Kg/L)
C
o
(Kg/L)
Hemet FWS Municipal 10 14 6.2 4.0
Arcata treatment FWS Lagoon 6 1 0.9
0.1
Arcata enhancement FWS Wetland 12 20
0.1
1.9
Halsey HSSF Pulp mill 0.13 2–10 10.0 11.2
Halsey FWS Pulp mill 0.13 2–10 10.0 3.4
Halsey FWS Pulp mill 0.13 2–10 10.0 7.9
Corvallis FWS Dairy 0.02 7 92.6 1.9
Source: Data from Barber et al. (2001) Environmental Science and Technology 35(24): 4805–4816.
© 2009 by Taylor & Francis Group, LLC
Organic Chemicals 531
was achieved at HLR 15 cm/d while at HLR of 10 cm/d and
20 cm/d the respective removals amounted to 87% and 77%.
Billore et al. (2002) operated a 300-m
2
SSF wetland,
vegetated with Phragmites karka, and studied LAS reduc-
tion over a 12-month period. Three days’ nominal detention
provided 43, 61, 76, and 88% reduction of C10, C11, C12,
and C13 LAS, respectively, again reinforcing the ndings of
Inaba (1992).
LAS was monitored for one year at three SSF Phragmites
reedbed sites in the United Kingdom (Thomas et al., 2003).

Hydraulic loadings were 133 cm/d at Brynsiencyn, 14.5 cm/d
at Clutton, and 77 cm/d at Rosset; and had average reductions
of 57% at Clutton and 55% at the other two sites. The longer
alkyl chain homologues were removed to a greater extent.
MISCELLANEOUS HYDROCARBONS
The estrogenic hormones 17-estradiol (E2) and 17-ethinyl
estradiol (EE2) have been detected in municipal wastewa-
ter efuent and surface waters at concentrations sufcient
to cause feminization of male sh (Gray and Sedlak, 2004).
These hormones were added to an engineered treatment wet-
land, and 36% of E2 and 41% of EE2 were removed in 3.5
days’ nominal retention time. The observed attenuation was
attributed to sorption, followed by biotransformation. Sorp-
tion was indicated by the retardation of the hormones rela-
tive to a conservative (lithium) tracer. Biotransformation was
indicated by elevated concentrations of the E2 metabolite,
estrone.
A full-scale FWS treatment wetland has been in opera-
tion since 1995 at Quimigal S.A. in Estarreja, Portugal, for
reducing organics in the efuent from a sulfanilic acid plant
(Haberl et al., 2003) (Figure 13.5).
This wastewater contains nitrobenzene, aniline, and
sulfanilic acid. Four vertical ow wetland beds totaling one
hectare are operated at a hydraulic loading rate of 2.4 cm/d.
Removals were 6 g/m
2
·d (99%) for aniline, 2 g/m
2
·d (98%) for
nitrobenzene, and 4 g/m

2
·d (99%) for aniline. Phragmites has
survived well in this industrial wastewater (Figure 13.5).
Lemna minor batch mesocosms displayed half-lives of
one to three days for a variety of chlorinated phenols (Tront,
2004; Tront et al., 2004). Rapid sorption (less than ve min-
utes) was observed, followed by rst-order removal. There
was no clear pattern associated with chemical structure. The
microcosm research at Krefeld, Germany, found a surpris-
ingly large amount of pentachlorophenol was absorbed in
wetland systems (Seidel, 1976). This result was somewhat
unexpected, as pentachlorophenol is used as a wood preser-
vative and does not biodegrade readily.
Polar organics are also degraded in constructed wet-
lands. Oil sands tailings water contains toxic constituents
naphthenic and naphthoic acids (Wood et al., 1995). Cattail
mesocosms were used to examine the removal of naphthoic
acid in a Kadlec and Srinivasan (1995) project. A lag time
was observed before removal started, as observed in the
Krefeld experiments with phenol (Seidel, 1976). However,
that lag phase was followed by a steady but slow decline of
this hydrophilic hydrocarbon. The zero-order rate constant is
smaller than for phenol, k
0
 0.044 g/m
2
·d.
Three polar organic solvents, acetone, tetrahydrofuran
(THF), and 1-butanol, were tested in experimental micro-
cosm subsurface constructed vegetated with Juncus effusus,

Carex lurida, Iris pseudacorus, Pondeteria cordata, and
unplanted controls (Grove and Stein, 2005). Ninety percent
removal of 1-butanol typically took less than three days,
acetone required from 5 to 14, and THF required at least ten
days, but was frequently not achieved in 14 days. Removal
was typically better in planted replicates, and was slower in
winter. Herrera-Melian et al. (2004) found essentially com-
plete removal of methanol and formaldehyde in planted and
unplanted gravel mesocosms in less than one day. Small SSF
wetlands were examined for potential use in reduction of
benzoic acid (Zachritz et al., 1996). Planted (Schoenoplectus
(Scirpus) validus) and unplanted rock bed mesocosms were
fed 5 to 50 mg/L benzoic acid at hydraulic loading rates of
4.9 and 9.6 cm/d. Removals were 30 to 96% in single- and
double-tank systems.
Phthalate acid esters (PAEs) are important at many indus-
trial products, and occupy 80–85% of the global plasticizer
market (Eljertsson et al., 1997). Studies have shown that PAE
contamination in waters, sediments, soils, and biota (Staple et
al., 1997) may be toxic to mammals, and aquatic organisms
(Adams et al., 1995; Foster et al., 2000; Patyna and Coo-
per, 2000). In addition, the high bioaccumulation rate and
mutagenic action posed a potential threat to human health
(Zhou et al., 2005b). Di-n-butyl phthalate (DBP) is the most
frequently identied PAE in diverse environmnetal symplex,
and it has been listed as a priority polutant by the U.S. EPA.
Zhou et al. (2005b) evaluated the fate of di-n-butyl
phthalate in a downow/upow constructed wetland in China.
The results indicated that on the tenth day, 35% of the DBP
in the surface soil and 62% of the DBP in the subsurface soil

remained. After 30 days, 95.7% DBP was degraded in the
surface soil and 64.2% DBP was degraded in the subsurface
soil. The degradation t a rst-order model and the kinetic
equations gave rate constants for surface and subsurface soils
of 0.50 and 0.17 d
-1
and the half-life of DBP in the respective
soils was 1.4 and 4.0 days.
13.4 PESTICIDES
The list of pesticides is very long, and therefore wetland
fate and transport studies lag far behind their introduction
and use. A distinction may be drawn between the persistent
chemicals used prior to the 1950s and the more degradable
substances used since that time. Possible retention factors are
adsorption to soil particles and organic matter, sedimenta-
tion of particles, photodegradation, plant uptake, and biodeg-
radation. The U.S. Department of Agriculture, Agricultural
Research Service publishes a compendium of chemical and
physical properties of 334 widely used pesticides. Informa-
tion included in the database focuses on 16 of the most impor-
tant properties that affect pesticide transport and degradation
characteristics (U.S. Department of Agriculture, 2001).
© 2009 by Taylor & Francis Group, LLC
532 Treatment Wetlands
Many of the “old” pesticides, such as DDT, are very
persistent in the environment. These substances partition
strongly to particulate matter. It is doubtful that wetlands can
provide any very effective mechanism for degradation, but
wetlands can act as a trap for the particulates that carry most
of the load. Very little data is available for this class of com-

pounds, but Winter (as referenced in Gilges, 1991) reports
some success treating PCBs and Lindane™ in SSF systems.
Studies at the Des Plaines, Illinois, treatment wetlands
indicated the presence of some of the old “hard” pesticides.
The river and wetland sediment samples all contained quan-
tiable levels of PCBs and some of them also had one or more
of the chlorinated pesticides or metabolites DDT, DDE, and
dieldrin. All identied species were found in the low micro-
grams-per-kilogram (parts per billion) range. PCBs were
present at the 20–25 µg/kg level in river-borne sediments
throughout most of the year, and were found in the sediments
of wetland EW3 at less than 10 µg/kg. DDT, together with its
decomposition product DDE, were present at approximately
2 µg/kg. Wetland cell EW3 sediments contained similar low
levels of DDT and DDE. Dieldrin ranged from 0.1 to 3.0 µg/kg
in river sediments and 0.5 µg/kg in wetland sediments.
Modern pesticides degrade more readily, and wetlands
have been found to generally reduce levels of many of these
compounds. Some of these may also volatilize (Suntio et al.,
1988). However, the many types of wetlands and modes of
operation (continuous ow, event ow, batch) compound with
a very long list of pesticides to create very numerous combi-
nations. Examples of studies that partially dene treatment
e
x
pectations are shown in Table 13.9. Perhaps the most com-
monly used agricultural herbicide is atrazine, and it has there-
fore been the subject of many treatment wetland studies.
ATRAZINE
Atrazine (2-chloro-4-(ethylamino)-6-(isopropylamino)-s-tri-

azine) is a triazine herbicide used to control broadleaf weeds
in corn, sorghum, sugarcane, turfgrass, fruits, and vegetables
(Kao et al., 2001b). Agricultural practices produce runoff with
atrazine at concentrations which sometimes peak in excess
of the federal drinking water standard (Alvord and Kadlec,
1995; Anderson et al., 2002). The atrazine–wetland interac-
tion is very complex, including removal from the area by con-
vection in the water, loss of chemical identity by hydrolysis
to hydroxytriazine and dealkylation, and sorption on wetland
sediments and litter. Atrazine degrades to trihydroxytriazine
in ultraviolet light (Hequet et al., 2001), but this process is
not rapid in natural waters (Konstantinou et al., 2001). Deg-
radation half-lives in various natural waters range from 26 to
6
5
days (Table 13.10). In the absence of light, aqueous deg-
radation rates are much slower, half-lives greater than 1,000
days, even with added carbon sources (Chung et al., 1996;
Ghosh and Philip, 2003).
The studies of Seybold et al. (2001) illustrate that wet-
land soils and water, by themselves, are not efcient reducers
of atrazine. Soil samples for the study were collected from a
tidal freshwater wetland site along the James River in Prince
George County, Virginia. The soil was a Levy silt loam with
an organic matter content of 14.1%. The half-life was 38 days
for atrazine in the anaerobic sediments, and 86 days in the
overlying water. Metabolites detected in the aqueous phase
above the soil were hydroxyatrazine, deethylatrazine, and
deisopropylatrazine. The hydroxyatrazine accumulated in
the sediments, and was resistant to further degradation.

Anderson et al. (2002) studied the mineralization of
atrazine in sediments from a constructed wetland in cen-
tral Ohio. Atrazine mineralization potential was measured
by
14
CO
2
evolution from [U-ring-14C]-atrazine in biometers.
The constructed wetland showed 70 to 80% mineralization
of atrazine within one month (Figure 13.6). The highest lev-
els of mineralization were localized to the top 5 cm of the
wetland sediment.
Weaver et al. (2004) pointed out that atrazine dissipated
in saturated and ooded soils with a half-time of approxi-
mately 23 days, but only 10% of atrazine was mineralized to
CO
2
. The major route of atrazine dissipation was incorpora-
tion of atrazine residues into methanol-nonextractable (soil-
bound) components, with minimal extractable metabolite
accumulation. A mixed-mode extractant (potassium phos-
phate:aceto-nitrile) recovered greater amounts of
14
C-resi-
dues from atrazine-treated soils, suggesting that hydrolysis
of atrazine to hydroxylated metabolites was a major compo-
nent of the bound residues.
Atrazine transport, sorption, and identity loss were stud-
ied at the Des Plaines River constructed wetland site, and
in accompanying laboratory work. Sorption was effective for

soils and sediments, but the more organic materials, such as
litter, showed a stronger afnity for atrazine than the mineral
base soils of the wetland cells at Des Plaines, Illinois (Alvord
and Kadlec, 1995; 1996). Atrazine was found to degrade
on those sediments according to a rst-order rate law, with a
half-life of 40 to 90 days. However, degradation was faster on
cattail litter, precluding a rate measurement. The estimated
half-life on cattail detritus was on the order of ve days.
FIGURE 13.5 Reed bed treating sulfanilic acid at Quimigal S.A.
in Estarreja, Portugal.
© 2009 by Taylor & Francis Group, LLC
Organic Chemicals 533
TABLE 13.9
Removal of Pesticides in Treatment Wetlands
Pesticide
Wetland
Type Size
Flow
Regime Vegetation
HRT
(days)
Reduction
(%) Reference
Aldrin FWS Field Continuous/
Event
Freshwater marsh 2.1 86 Lopez-Flores et al. (2003)
Azinphos-methyl FWS Field Event Typha capensis 0. 5–2.9 100 Schulz and Peall (2001)
Chlorothalonil FWS Mesocosm Batch Scirpus cyperinus 0.5 98 Sherrard et al. (2004)
Chlorpyrifos FWS Field Event Typha capensis 0.5–2.9 100 Schulz and Peall (2000)
Chlorpyrifos FWS Mesocosm Batch Elodea densa 2.5 40 Karen et al. (1998)

Chlorpyrifos FWS Mesocosm Batch Scirpus cyperinus 3 98 Sherrard et al. (2004)
Chlorpyrifos FWS Field Dosed — 15–30 95 Darby (1995)
Cinosulfuron FWS Field Batch Oryza sativa 21 81 Ferrero et al. (2001)
DDT FWS Field Continuous/
Event
Freshwater marsh 2.1 100 Lopez-Flores et al. (2003)
Diazinon FWS Field Continuous Scirpus spp. 3.2 22 City of Phoenix
unpublished data
Diazinon FWS Field Continuous Scirpus spp. 6.3 12 City of Phoenix
unpublished data
Diazinon FWS Field Continuous Scirpus spp. 6.9 60 City of Phoenix
unpublished data
Diazinon FWS Field Continuous Scirpus spp. 6.8 28 City of Phoenix
unpublished data
Endosulfan FWS Field Event Typha capensis 0.5–2.9 100 Schulz and Peall (2000)
Fenpropimorph FWS Field Event Sparganium
erectum
0.25 23 Braskerud and Haarstad
(2003)
Lindane FWS Field Continuous/
Event
Freshwater marsh 2.1
19
Lopez-Flores et al. (2003)
Linuron FWS Field Event Sparganium
erectum
0.25 17 Braskerud and Haarstad
(2003)
Metalaxyl FWS Field Event Sparganium
erectum

0.25 15 Braskerud and Haarstad
(2003)
Metamitron FWS Field Event Sparganium
erectum
0.25 33 Braskerud and Haarstad
(2003)
Methyl Parathion FWS Mesocosm Batch Juncus effusus 4 100 Schulz et al. (2003c)
Metolachlor SSF Mesocosm Dosed Scirpus validus 2–19 34–97 George et al. (2003)
Stearman et al. (2004)
Metolachlor Pond Mesocosm Dosed Scirpus validus 12–20 50 Mazanti et al. (2003)
Metribuzin FWS Field Event Sparganium
erectum
0.25 30 Braskerud and Haarstad
(2003)
Permethrin FWS Field Continuous/
Event
Freshwater marsh 2.1 4 Lopez-Flores et al. (2003)
Propachlor FWS Field Event Sparganium
erectum
0.25 41 Braskerud and Haarstad
(2003)
Propiconazole FWS Field Event Sparganium
erectum
0.25 19 Braskerud and Haarstad
(2003)
Simazine SSF Mesocosm Dosed Scirpus validus 2–19 51–96 George et al. (2003)
Stearman et al. (2004)
Simazine FWS Field Continuous/
Event
Freshwater marsh 2.1 100 Lopez-Flores et al. (2003)

Simazine FWS Microcosm Batch Myriophyllum
aquaticum
7 51 Wilson et al. (2001)
© 2009 by Taylor & Francis Group, LLC
534 Treatment Wetlands
Outows from the Des Plaines wetland cells contained
reduced amounts of atrazine compared to the river water
inputs. During 1991, atrazine peaked in the river due to two
rain events. Only about 25% of the incoming atrazine was
removed in wetland cell EW3, but 95% was removed in wet-
land cell EW4, due to longer detention time. Either time aver-
aged or dynamic models, both embodying a nonideal mixing
model, t the data for wetland EW3 (Figure 13.7). A removal
rate coefcient (k) of approximately 14 m/yr calibrated the
models to the data.
McKinlay and Kasperek (1999) tested two-stage, recircu-
lating VF mesocosms. Clubrush (Schoenoplectus (Scirpus)
lacustris), cattails (Typha latifolia), iris (Iris pseudocorus),
and Phragmites australis provided complete removal of
atrazine, as did sterile unplanted systems, but rates were
slower in the sterile systems. These results conrm those of
Stearman et al. (2004) for a SSF system treating simazine,
another triazine herbicide (Table 13.11). There were not
large differences among plant species. Acclimatization was
observed in serial dosing, with later atrazine doses removed
at higher rates. The fastest removals showed a half-life of
about ve days.
Moore et al. (2000b) dosed atrazine into 0.1-ha FWS
demonstration wetlands at Oxford, Mississippi (Rodgers
and Dunn, 1992). Between one third and two thirds of

the atrazine was removed, with observed half-lives of 16
to 48 days. Atrazine was not detected in plant or sediment
samples.
Detenbeck et al. (1996) performed comprehensive stud-
ies on FWS wetlands, for the purpose of determining ecolog-
ical effects of atrazine on receiving wetlands. Experiments
were carried out in the downstream wetland portions of four
of the eight outdoor experimental wetland/stream meso-
cosms located at the U.S. EPA Ecological Research Station in
Monticello, Minnesota. Dominant emergent plants consist of
cattail (Typha spp.), reed canary grass (Phalaris arundina-
cea), and wild rice (Zizania aquatica). Dominant submersed
vegetation consists of waterweed (Elodea canadensis),
pondweed (Potamogeton spp.), and coontail (Ceratophyl-
lum demersum), and oating-leaved vegetation dominated by
duckweed (Lemna minor). Half-lives of 8 to 14 days were
found. Sorption measurements gave isotherms for sediments
that agreed with those found by Alvord and Kadlec (1995),
but sorption alone predicted half-lives of 36 days. Therefore,
other mechanisms were involved in the overall reductions.
Ecosystem effects included repression of periphyton, and
early senescence of the wild rice.
Runes et al. (2001) evaluated the microbial degradation
of atrazine in sediments taken from a constructed wetland
treating irrigation runoff from a container nursery near
Portland, Oregon. To stimulate atrazine-degrading ability
in wetland sediments, dried shoot material from cattail was
incorporated, and sediment soil was bioaugmented with an
atrazine spill-site soil known to contain atrazine-mineralizing
microbial populations. The population density of atrazine-

degrading microorganisms in unbioaugmented sediment
was increased from about 10
2
/g to 10
4
/g by bioaugmentation
(1:100 w/w) and increased by another 60-fold (6 r 10
5
/g) after
incubation with 10 μg/g of atrazine. A high population of
atrazine degraders (approximately 10
6
/g) and enhanced rates
of atrazine mineralization also developed in bioaugmented
sediment after incubation in ooded mesocosms planted with
Typha latifolia and supplemented with atrazine (3.2 mg/L,
1 μg/g sediment). In the absence of atrazine, neither the pop-
ulation of atrazine degraders nor the atrazine mineralizing
potential of bioaugmented sediment increased, regardless of
the presence or absence of cattails.
TABLE 13.10
Photolysis Half-Lives of Triazine Herbicides in Natural Waters (in days)
Lake Water River Water Groundwater Distilled Water
Atrazine 53 43 26 35
Propazine 65 53 29 33
Prometryne 51 52 28 32
Note: Dark blanks have been deducted.
Source: Data from Konstantinou et al. (2001) Journal of Environmental Quality 30(1): 121–130.
0
10

20
30
40
50
60
70
80
90
100
0 5 10 15 20 25 30 35
Time (days)
Percent Mineralization
0–5 cm
10–15 cm
FIGURE 13.6 The mineralization of atrazine to CO
2
in treat-
ment wetland sediments. (Data from Anderson et al. (2002) Water
Research 36(19): 4785–4794.)
© 2009 by Taylor & Francis Group, LLC
Organic Chemicals 535
Runes et al. (2003) then conducted eld experiments
in a constructed wetland consisting of ve sequential cells,
each approximately 3 m r 40 m lled with a silt loam,
and planted with Typha latifolia at greater than 75% cov-
erage. Water came from a container nursery facility, and
was delivered episodically, and dosed with atrazine. Flow
events were ranked by intensity (low: less than 250 L/min;
medium: 250 to 500 L/min; high: greater than 500 L/min)
and frequency of the events (low: fewer than 6 events per

week, medium: 6 to 9 events per week, high: greater than 9
events per week). Atrazine treatment efciency at the outlet
of the constructed wetland during a seven-day period ranged
from 76 to 82% in 1998 (experiments 1 through 3) and 83 to
84% in 1999 (Table 13.12), but zero treatment was found in
one monitoring period. For all experiments, deethylatrazine
and deisopropylatrazine accounted for 13 to 21% of the ini-
tial application. Hydroxyatrazine was rarely detected in the
water.
13.5 CYANIDE
Cyanide has an active cycle in nature, where it is present in
green plants as part of the methionine cycle (ethylene pro-
duction), and is therefore prevalent in autumn foliage and
ripening fruit (Neuhauser and Swallow, 2005). If not han-
dled properly, black and bing cherries (Prunus serotina and
Prunus avium) and cassava can cause toxicity in humans.
There are clear risks associated with cyanide release into the
environment because of toxicity. The most toxic species are
the free cyanides (HCN and CN
-
), which are infrequent in
uncontaminated soils and groundwater. Iron cyanide is less
toxic, and often predominates in environmental samples. The
U.S. Environmental Protection Agency and various state
agencies regulate cyanide in groundwater and surface water
differently. Agencies either choose to regulate the total or the
free cyanide form. Because of the potential for photolytic
production of free cyanide from iron cyanide, regulatory
agencies may consider total cyanide to be indirectly toxic.
U.S. EPA criteria for free cyanide include a drinking water

standard of 200 µg/L, a freshwater acute criterion of 22 Mg/L,
and a freshwater chronic criterion of 5.2 Mg/L.
Cyanide compounds are classied into those that are
readily converted into hydrogen cyanide under acidic condi-
tions, or into CN
under basic conditions, and those that are
strongly complexed. This Weak Acid-Dissociable cyanide
(WAD) is distinguished from strong complexes formed with
iron and cadmium, as examples. Total cyanide is the sum of
both forms (Kavanaugh et al., 2003).
Cyanide is a pollutant of concern in several situations.
Gold mine wastewaters may contain cyanide as the result of
the extraction process. Manufactured gas plants historically
left residues rich in cyanide and other constituents, which
may subsequently leach to natural water bodies. Aluminum
processing produces spent pot liners, which, when landlled,
cause similar leaching of cyanide.
0.00
0.05
0.10
0.15
0.20
0.25
0.30
0.35
0 5 10 15 20 25 30 35 40 45
Time (Days from May 28, 1991)
Atrazine Outflow Rate (g/hr)
Data
Model I

Model II
FIGURE 13.7 Des Plaines wetland EW3 atrazine mass outow as a function of time. Model I is a steady-ow, tracer-based RTD calculation.
Model II is a dynamic simulation which ts the hourly changes in ows of water and atrazine. Both models use the same postulated rst-
order homogeneous reaction rate constant (k  14.4 m/yr). (Data from Alvord and Kadlec (1996) Ecological Modelling 90(1): 97–107.)
TABLE 13.11
Percentage Simazine Removed in HSSF Wetland Cells
HRT
(days)
Vegetated
(%)
HRT
(days)
Unvegetated
(%)
2.3 65 2.1 64
3.5 67 3.2 64
5.1 78 4.1 57
7.9 81 6.4 65
13.3 96 8.3 78
20.6 93 8.3 64
20.7 95 12.9 67
Source: Data from George et al. (2003) Water Environment Research 75(2):
101–112.
© 2009 by Taylor & Francis Group, LLC
536 Treatment Wetlands
Both aerobic and anaerobic microorganisms are able
to destroy cyanide (Hinchee et al., 1995; Goncalves et al.,
1998; Alexander, 1999). Further, plants such as willows are
able to ingest cyanide, and metabolically degrade it (Larsen
et al., 2004). As a consequence, a natural treatment system

that provides maximum opportunity for contact between
cyanide-contaminated water and plants and microbes would
have strong promise for removal and degradation of cyanide.
Treatment wetlands are leading candidates to accomplish
these goals.
The processing of cyanide in wetlands is known to
involve microbial degradation, volatilization, plant uptake,
and complexation with iron (Young, 2003). Additionally, iron
cyanide decomposes photolytically to free cyanide (Kim et al.,
1998). The aerobic and nutrient-rich environment promotes
the growth of the microbial population capable of uptake,
conversion, sorption, and/or precipitation of thiocyanate,
cyanide, ammonia, nitrate, sulfate, and metals. The micro-
bial removal involves production of ammonia (Goncalves
et al., 1998; Akcil, 2003):
MM(CN) H O O HCO NH
222
2+
3
3
l 

422
(13.2)
SCN H O 2O SO NH HCO H
2244 3

l 3
2
(13.3)

where M represents a metal species.
Various Pseudomonas species are responsible for oxi-
dation of cyanide and thiocyanate. Some of the organisms
known to oxidize cyanide include species of the genera Acti-
nomyces, Alcaligenes, Arthrobacter, Bacillus, Micrococcus,
Neisseria, Paracoccus, Pseudomonas, and Thiobacillus
(Akcil and Mudder, 2003).
However, plant uptake is also potentially active (Ebbs et
al., 2003; Larsen et al., 2004; Samiotakis and Ebbs, 2004;
Larsen et al., 2005). Larsen et al. (2005) found Michaelis-
Menten kinetics with half-lives of less than a day (batch,
with leaves or roots). Samiotakis and Ebbs (2004) tracked the
transport and metabolism of two cyanide compounds, potas-
sium cyanide and potassium ferrocyanide, by willow (Salix
eriocephala), using a hydroponic system that preserved
cyanide speciation and solubility. The cyanide compounds
were labeled with
15
N to quantify transport. These analyses
revealed that while little free cyanide was detected in the
aerial tissues of plants exposed to either of these two cyanide
compounds, signicant enrichments in
15
N were observed,
suggesting transport and subsequent metabolism of free
cyanide as well as ferrocyanide. Yu et al. (2005) found tem-
perature to be a factor for uptake by willow (Salix baby-
lonica), from which Arrhenius factors of about 1.07 may be
calculated. Disappearance rates in solution, using detached
leaves, were characterized by half-lives of about eight hours.

These encouraging process studies have led to several
treatment wetland projects targeting cyanide reduction.
Garcia (2003) found more than 90% reduction in wetland
mesocosms (Typha). Alvarez et al. (2006a; 2006b) experi-
mented with wetland-based passive systems at laboratory and
eld scale at the site of a gold mine in northern Spain. The
pilot-scale system was comprised of aeration cascades and
aerobic and anaerobic cells. Results suggest that this technol-
ogy is able to treat wastewaters with about 25% reduction
for dissolved cyanide for nominal detention of six days. The
waste strengths in this study were quite high, with total cya-
nide of about 200 mg/L and dissolved cyanide of 75 mg/L.
The removal of WAD cyanide was 99%, from 15 mg/L to
0.14 mg/L. Vegetated cells (Typha latifolia) performed only
marginally better than gravel control cells.
A large-scale trial was conducted at the Rustler’s Roost
mine near Darwin, Australia. A 3.2-ha site was congured
as ponds and wetlands, covered with peat, and planted with
Typha orientalis (Halford, 1999). A light hydraulic loading
(1 cm/d) was applied during the dry season of 1996 (July–
September). The incoming water had total cyanide of about
2,000 µg/L, and the system discharge was 40–400 µg/L.
Weak acid dissociable cyanide (WAD) was reduced from
about 900 to 200 µg/L.
Wetland remediation of groundwaters contaminated with
cyanide from landlls associated with the aluminum indus-
try has been shown to be effective (Gessner et al., 2005). A
pilot project at Alcoa, Tennessee, consisted of two lined FWS
cells operated in series. Softstem bulrushes (Schoenoplectus
(Scirpus) tabernaemontani) were planted in Cell 1 (around

5 m r 15 m), and cattails (Typha latifolia) were planted in Cell
TABLE 13.12
Percentage Atrazine Removed in FWS Wetland Cells under Event-Driven Operation
Runoff Frequency Runoff Intensity Cell 1 Cell 3 Cell 5
High Medium 3 80 82
High Medium 90 69 76
Medium High 99 94 80
High High 27 76 83
Low Low 55 48 84
Source: Data from Runes et al. (2003) Water Research 37: 539–550.
© 2009 by Taylor & Francis Group, LLC
Organic Chemicals 537
2 (around 7 m r 12 m). Inow concentrations typically ranged
from 0.03 mg/L to 0.05 mg/L, while all outow concentrations
except one were below the lowest reporting level (LRL) of
0.01 mg/L. Total cyanide showed a statistically signicant
(p  0.00005, N  25) decrease in mean concentration (88%)
between the treatment wetland inow and outow. By the
time water reaches the midpoint of the treatment wetland,
total cyanide average concentrations are at, or below, the LRL.
A second pilot project was operated during 2004–2006, also
in Alcoa, Tennessee. This 2,300-m
2
system was one lined
FWS cell that received about 300 µg/L of total cyanide at
130 m
3
/d, resulting in a nominal detention time of about three
days. This system achieved well over 90% reduction in total
cyanide (Figure 13.8).

The early successes of wetland treatment have caused opti-
mism for continued application of wetlands to reduce cyanide.
The PIRAMID summary research report stated that “the safe,
passive destruction of residual cyanide leaching from gold
mine tailings has been achieved using compost-based wet-
land-type passive systems” (PIRAMID Consortium, 2003b).
This report goes on to suggest that WAD removal in such sys-
tems can be expected to average around 10.6 mg/m
3
·d, which
equates to an area-adjusted removal rate of 4.2 g/m
2
·d for a
substrate 0.5 m in depth. However, this recommendation is ten
times higher than the results of Alvarez et al. (2006b), who
found 0.49 g/m
2
·d for WAD cyanide.
0
50
100
150
200
250
300
350
400
450
0246810121416182022
Days

Total Cyanide (µg/L)
July 2004 In
March 2005 In
July 2004 Out
March 2005 Out
FIGURE 13.8 Input–output data for two intensive sampling periods for total cyanide at the Duck Spring FWS treatment wetland at Alcoa,
Tennessee. (Unpublished data from Alcoa Aluminum.)
SUMMARY
Wastewaters containing a wide variety of hydrocarbon con-
stituents are presently being treated in wetland systems, and
performance has been generally good. Rate constants are
typically high, especially for volatile materials. The surfac-
tant work has collectively shown that individual compounds
may behave quite differently, depending upon isomeric form
and upon small changes in molecular weight. This is some-
what daunting, because it means pre-project investigations
for specialty chemicals are a necessity; rate constants may
not be inferred from data on related compounds. The avail-
able data on phenol show that microcosms are only quali-
tative predictors of the performance of full-scale wetland
systems. This means that pilot-scale tests are a necessity for
generating design data.
However, there is some removal of nearly every organic
compound tested in wetland treatment systems, whether full
scale, pilot scale, or microcosm. This includes organics from
very important generators: pesticides, petroleum and petro-
chemicals, food wastes, and remediation. In each of these cat-
egories, there are multiple studies that have initiated the design
database. Current knowledge shows great promise for wetlands
technology for organics control, especially in situations where

passive systems can be t into the landscape and accumulated
organics will not pose a threat to wetlands biota.
© 2009 by Taylor & Francis Group, LLC