TREATMENT WETLANDS - CHAPTER 16 pptx

593
16
Design Basis
The preliminary step in the design of a treatment wetland
is to acquire a fundamental understanding of the site of the
wetland. Site conditions dictate the physical, chemical, and
biological environment of a wetland treatment system. Con-
ditions that should be evaluated during planning of a wetland
treatment system include climate, geography, groundwater
and its chemistry, soils and geology, rainfall and runoff water
chemistry, biology, and socioeconomic factors. The impor-
tance of each of these conditions may vary, but all should be
investigated to some extent. Detailed studies may be needed
to determine the importance of those site conditions that
affect technical feasibility.
This book primarily considers performance-based design
algorithms. The rst steps in the process require the assembly
of the basis of the design, which includes the following steps:
1. Determine inlet concentrations and ows.
2. Determine target concentrations (regulatory limits
and allowable exceedance factors).
3. Determine allowable inow and seepage rates.
4. Determine rain, ET, and temperature ranges for
the project site.
5. Select wetland type (FWS or SSF).
Often, the establishment of ows and concentrations will
require the acquisition of data on ows and concentrations,
at least to conrm estimates based on prior operations or
knowledge of the technology of the source.
There are, unfortunately, numerous examples of inappro-
priate selection of the design basis for treatment wetlands. The

difculty is often not the misunderstanding of current condi-
tions, but rather incorrect assumptions about future conditions.
This may involve actions outside the control of the designer.
For instance the “failed” Gustine, California, system
added wetlands to existing lagoons (Walker and Walker,
1990). The source water was a combination of municipal and
milk processing wastewaters, the latter having very high bio-
chemical oxygen demand (BOD). The design presumption
was that milk wastewater would be discontinued, but this did
not occur. Instead of the design inuent BOD of 60 mg/L,
the wetland actually experienced BOD of approximately
600 mg/L. The wetland could not meet design targets for the
unplanned tenfold-higher inlet concentrations. Another exam-
ple is the “successful” wetland system treating potato waste-
water at Connell, Washington (Kadlec et al., 1997; Burgoon
et al., 1999). The system was built according to a design
based on operating data from a xed-capacity processing
plant. However, coincident with wetland start-up, the plant
implemented water conservation. The loads of pollutants
remained the same, but concentrations went up considerably
as ows decreased. Fortunately, in this case the wetland sys-
tem was robust enough to accommodate the change and still
achieve goals. These anecdotes serve notice that the basis of
design must be carefully set forth, and reasonable changes
anticipated in inuent ows and loads.
16.1 PROJECT SETTING
S
PACE CONSIDERATIONS: LIMITED VERSUS UNLIMITED SPACE
Free water surface (FWS) treatment wetlands are in the cate-
gory of land-extensive technologies. At the end of this chapter,

it will be seen that horizontal subsurface ow (HSSF) wetlands
for the same purpose are not much different in size. However,
the site conditions—primarily property boundaries and topog-
raphy—can limit the potential size of a treatment wetland for
a particular source-water volume. This is particularly true of
urban stormwater wetlands, which need to be sited in built-up
areas and which often utilize high-value lands. Siting of wet-
lands inside the boundaries of major cities, such as Orlando,
Florida (Palmer and Hunt, 1989), or Toronto, Ontario (Heleld
and Diamond, 2004), means that the size of the wetland is dic-
tated by existing streets, highways, and buildings, not by the
size needed to achieve a particular performance goal.
Topography may also limit the potential size of a treat-
ment wetland. The presence of steep slopes adjacent to the
site can preclude construction beyond a certain limit, dened
by the practicality of earth moving (Figure 16.1). When the
site demands, the treatment wetland may be established in
terraces, with elevation drops occurring between the succes-
sive cells of the system (Navarra, 1992; Inman et al., 2003).
Land ownership can also constrain opportunities for wet-
land construction. Above and beyond questions of acquisition
costs, there is the issue of the willingness of the owner to sell
property. Building a treatment wetland rarely falls into the
category of eminent domain acquisition, although that has
happened in connection with the phosphorus removal wet-
lands of South Florida. For large wetlands, suitable parcels
are often already in agricultural use. Wetlands are frequently
viewed as valuable landforms across the regional landscape,
but aquatic and terrestrial landforms are also valuable. The
construction of a treatment wetland implies the removal of

other types of plant, animal, and human communities. There-
fore, competing uses may block the construction of a wetland
on a particular plot of land.
Perhaps the most serious potential constraint of preexist-
ing landform is the presence of naturally occurring wetlands
on property under consideration for a treatment wetland. In
© 2009 by Taylor & Francis Group, LLC
594 Treatment Wetlands
the United States, it is generally not allowed to build any proj-
ect that destroys existing wetlands. But what if the project is
a constructed wetland? That situation is obviously confusing
and unclear, and it is therefore not surprising that a variety of
rules and regulations apply in various states.
Sometimes the constructed treatment wetland may be
viewed as self-mitigating; that is, it inherently compensates for
the loss of preexisting wetlands. That situation has occurred
at the West Jackson County, Mississippi, constructed wetland
site. It is probably most acceptable when the preexisting wet-
land is degraded, and of low regional value. However, in many
other circumstances, construction in wetlands must be avoided.
For instance, HSSF and vertical ow (VF) wetlands do not
offer the same type of habitat that occurs in natural wetlands,
and construction of these systems in natural wetlands is often
blocked by regulatory constraints. Of course, the extreme cir-
cumstance is the use of natural wetlands for wastewater treat-
ment, which is outside the scope of this book.
For these area-constrained situations, the design methods
described herein are not used to select wetland area, but rather
are used to forecast performance of the available wetland area.
This predictive mode is readily accommodated in a rate coef-

cient approach, but is very awkward, if not impossible, for
a loading design approach due to the data scatter inherent in
loading charts.
SOILS AND GEOLOGY
For planning purposes, site soils in the United States can
be characterized by using USDA Soil Conservation Service
soil surveys, which are generally available for most coun-
ties within the United States. Other countries often have
similar mapping resources. Soil surveys typically include
maps of soil types as well as summaries of soil properties,
groundwater conditions, climatic information, and plant
community information.
Soils are classied by soil scientists based on a complex
array of physical and chemical characteristics. Soil informa-
tion that might be important during project planning includes
the presence of hydric soils, which occur in natural wetlands
(even if formerly drained) and could be a potential regula-
tory constraint for a constructed wetland site; soil texture and
composition as a suitable medium for berm construction or
for impeding leakage to the groundwater; depth to seasonal
high groundwater; and depth to conning layers of clays or
rock horizons. On-site soils are typically preferred for the
rooting media in FWS wetlands. In some cases, the sorption
potential of these rooting soils will be a design variable, such
as for metal removal.
The construction of wetlands entails the excavation of the
wetland basin, including any deep zones, possibly together
with conveyance and seepage interception canals. Therefore,
the soil thickness above bedrock is an important piece of
design information, because that material is movable without

blasting. The characteristics of the bedrock are important if
s
u
ch blasting is required (Figure 16.2). Construction in rock
is extremely expensive, and is to be avoided if possible.
At the other extreme, on-site materials may be unsuitable
for the construction of embankments, because they cannot
withstand exposure to the water (Figure 16.3).
GROUNDWATER
Inltration of wastewater to the groundwater is important
because inltration affects the wetland water balance and
could pose regulatory problems under some conditions. Soil
inltration rates published in soil surveys typically overesti-
mate the actual inltration rates under sustained, saturated
soil conditions and are not reliable for project planning or
design. Surface inltrometer tests or well slug tests provide
better estimates of the groundwater leakage that can be
FIGURE 16.1 The treatment wetlands in the Tucush valley of the
high Andes Mountains of Peru (4,100 masl) are constrained to a
xed area by extremely steep slopes. They treat the drainage com-
ing from the wasterock dump of a mine operation. (Photo courtesy
Compañía Minera Antamina. Reprinted with permission.)
FIGURE 16.2 Deep zones and canals for South Florida’s storm-
water treatment areas require blasting of the limestone bedrock. A
thin veneer (0.3–1.0 m) of peat overlies the limestone.
© 2009 by Taylor & Francis Group, LLC
Design Basis 595
expected from a full-scale wetland treatment system. Meth-
ods for measuring inltration rates are described by the Soil
Conservation Service (SCS) (Hansen, 1980; U.S. Bureau of

Reclamation, 1993). Field tests are the most reliable method
of estimating groundwater inltration rates. For constructed
wetlands, it may be necessary to construct pilot wetland
basins on a proposed site and then instrument inows and
outows to develop an accurate water balance.
Wetlands can be built on leaky soils as long as regulatory
requirements can be met and adequate hydroperiods can be
maintained with the wastewater addition and net rainfall. In
fact, wetlands have been designed with groundwater recharge
as a specic project goal (Ewel and Odum, 1984; Knight and
Ferda, 1989). Groundwater inltration can be eliminated as
a project concern for constructed wetlands by using a clay or
synthetic impervious liner. Although this approach may not
be necessary if the wastewater has received secondary pre-
treatment, it is recommended when wastewater is less than
secondary quality, or is known to contain contaminants of
concern for the regional groundwater and its intended uses.
Percolation tests are often used as the basis of sizing
inltration elds for septic tank efuent disposal, although
such tests are probably insufcient for ensuring adequate
performance of the eld (Crites and Tchobanoglous, 1998).
The allowable hydraulic loading for the inltration eld is set
according to a published table or curve, relating the allowable
loading to the time for the water level in a test pit to drop a
specied amount (usually 2.5 cm). Data collected from perco-
lation tests are then typically related to a prescriptive hydraulic
loading that is usually much less than the observed percolation
rate. The reduction in hydraulic loading is to account for the
long-term accumulation of microbial biomass and particulate
matter in the soil, which substantially reduces the inltration

rate (Tyler and Converse, 1994). Allowable hydraulic load-
ings are usually in the range of 1–5 cm/d. Additionally, there
must be a specied vertical travel distance to the groundwater
table, typically about 1 m of unsaturated soil (to allow for the
removal of pathogens). These requirements are commonly set
forth in local codes and rules, and are enforced as a condition
for acceptability of new on-site (septic) systems. These codes
are typically intended for single-home treatment systems, but
are often extrapolated to larger systems due to a lack of more
appropriate regulatory guidelines.
The focus of on-site (septic) system codes is the dis-
posal of primary efuent into the soil matrix. When water
is pretreated, organic and pathogen loads are substantially
reduced, and soil-based treatment is less critical for regula-
tory compliance.
Given this basis, constructed wetlands are frequently
viewed by the on-site regulatory community as a means for
justifying higher loadings or lesser unsaturated travel dis-
tances in the inltration bed, or both. For example, the state
of Indiana allows reduction in the size of the absorption eld
associated with a subsurface-constructed wetland based on
the soil loading rate (Indiana Department of Environmental
Management, 1997). For soil loading rates less than or equal
to 5 cm/d but greater than 2 cm/d, the allowable reduction in
eld size is 50%. For soil loading rates of less than 2 cm/d
but greater than or equal to 1 cm/d, the allowable reduction
in the eld is 33%. Similar reductions in inltration area are
allowed in other states.
In general, it is benecial to understand the directions
and ows of regional groundwater under the project site. Dif-

ferent levels of hydrogeological surveys may be performed,
depending on the requirements of the specic project. Con-
siderable detail is necessary for groundwater remediation
wetlands that intercept a plume of contamination, because
those studies provide the ows and concentrations needed
to determine wetland size or performance. For instance, the
design of the Hillsdale, Michigan, project involved multiple
monitoring wells, studied over several years, and three-
dimensional computational uid mechanics (Ecology and
Environment Engineering, 2004). Modeling at a similar level
was necessitated at the Columbia, Missouri, project, because
of proximity to the city’s potable water well eld (Brunner
and Kadlec, 1993). If the water leaving the system is trans-
ported by unsaturated ow, more complex models will be
required (Langergraber, 2001; Davis, 2007).
ALTITUDE
As the use of treatment wetland technology has grown across
the planet, the site conditions have broadened to include a
wider range of conditions, among which is the altitude of the
project. A few experiences have identied special issues, such
as the types of wetland plants that are adapted to high-altitude
conditions: Phragmites is not a mountain plant! (Navarra,
1992). Other concerns have yet to be explored. For instance,
treatment wetlands have now been built at up to 4,000 m above
sea level (see Figure 16.1), at which altitude the atmosphere is
approximately at half sea-level density. Therefore, the partial
pressure of oxygen is half that at sea-level, with potential con-
sequences on the ability of the wetlands to process reactions
that require dissolved oxygen, such as nitrication.
FIGURE 16.3 This collection canal in the Lakeland, Florida, FWS

system was built using unstable materials from on site. Despite the
attempt to reinforce the embankment with concrete matting, ero-
sion caused the discharge structure to drop into the water.
© 2009 by Taylor & Francis Group, LLC
596 Treatment Wetlands
BIOLOGICAL CONDITIONS
The addition of any type of water or wastewater will alter
biological conditions at a site. Constructed wetlands fre-
quently replace upland habitats with wetland vegetation. The
upland habitats that are lost might include plant communi-
ties such as grassland, forest, scrub, desert, or agriculture.
The environmental values of these upland habitats should be
assessed during project planning. Likewise, wastewater dis-
charge to natural wetlands can cause biological changes of
varying magnitudes (see Chapter 3). Existing plant and ani-
mal communities in natural wetlands will change depending
on the degree of changes to surface water quality and hydrol-
ogy. Construction-related impacts will result in replace-
ment of part of the existing vegetation by distribution pipes,
boardwalks, and monitoring structures. For most constructed
wetland projects, site-specic biological conditions do not
represent a major technical constraint.
16.2 CHARACTERIZATION OF DOMESTIC
AND MUNICIPAL WASTEWATER
Wastewater quality varies widely among domestic, municipal,
industrial, agricultural, and stormwater categories. Different
wastewater sources have unique mixtures of potential pollut-
ants, so that even a single wastewater source category, such as
municipal wastewater or urban runoff, may vary considerably
depending on local, site-specic circumstances. However, for

some chemical constituents, the qualitative and quantitative
composition of wastewaters from different sources varies less.
In general, any summary of “typical” wastewater concentra-
tions and loads must be considered cautiously.
Site-specic wastewater data showing historical ows
and mass loads provide the best information for wetland
treatment system design. However, because many treatment
systems are designed for new facilities or because historical
monitoring may be nonexistent or insufcient, it is useful
to know the typical concentrations of major constituents in
similar wastewaters. This section summarizes information
from a number of sources on the typical pollutant composi-
tion of wastewater applied to engineered wetlands. These
“typical” concentrations and loads should only be used
when site-specic information is not available.
The total municipal wastewater ows from municipal
sources undergoing treatment in the United States is 45 ×
10
9
m
3
per year, serving approximately 72% of the popula-
tion (U.S. EPA, 2007). In addition to industrial and munici-
pal wastewaters, nonpoint source pollution contributes about
two thirds of the total pollution load to U.S. inland surface
waters (U.S. EPA, 1989). Sources of nonpoint ows include
urban and suburban runoff, diffuse agricultural runoff,
forestry activities, runoff from concentrated agricultural
activities such as feedlots, mine drainage, and runoff from
undisturbed areas. However, in certain areas urban runoff

or other stormwater sources provide the greatest percentage
of uncontrolled pollutants. Wetlands are often used in con-
junction with other treatment devices, including septic tanks,
lagoons, and mechanical treatment plants (Figure 16.4). In
those circumstances, the water quality of interest for the wet-
land design is that exiting a pretreatment step.
The amount and timing of the water to be treated is the
rst and foremost item of the design basis. This informa-
tion should include the possible seasonality of ows and the
anticipated progression of ows over the life of the design.
This is more important for treatment wetland design than for
conventional concrete and steel treatment plants, because of
the implied life cycle of the process and the nature of urban
and industrial growth. It is customary to plan for a 20-year
life expectancy for conventional wastewater treatment plants,
because mechanical equipment often wears out during this
period. But wetlands clearly can continue to function for
far longer periods than two decades; for example, there are
receiving wetlands that have been in operation for periods of
70 years (Great Meadows; Yonika et al., 1979) and 90 years
Surface
Discharge
Infiltration
Bed
Subsurface
Discharge
Sludge
Reed Bed
VF
Wetland

HSSF
Wetland
FWS
Wetland
Settling
Basin
Lagoon
Oxidation
Pond
Activated
Sludge
Biofilm
(RBC)
Source
Septic
Tank
Sludge
Bed
Combination
Wetland
FIGURE 16.4 Simplied options for treatment trains involving treatment wetlands.
© 2009 by Taylor & Francis Group, LLC
Design Basis 597
(Brillion; Spangler et al., 1976). Projecting ow estimates far
into the future is risky, so it is necessary to be explicit about
ow capacity at the time of design.
Most of the pollutants that are common to many of these
wastewater sources can be effectively treated by wetland sys-
tems. The normal concentration range of these pollutants is
an important consideration in evaluating wetland treatment

system options. This section compares and contrasts these
wastewater sources to facilitate initial alternative evaluation.
WATER QUANTITY
The information on water quantities and timing is assembled
into the annual and monthly water budgets for the design,
including any seasonal or event storage that may be necessary.
Such water budgets are easily prepared within the framework
of a spreadsheet program on a personal computer. This infor-
mation is later linked to the computation of the expected reduc-
tions in pollutant concentrations. Interestingly, the addition of
a wetland to any of the several forms of pretreatment provides
dampening of ow pulses. Although it is necessary to account
for the diurnal cycles in the inows for hydraulic purposes,
the wetland will typically “hold” several such daily pulses,
because of the extended detention time used in the wetland.
SMALL DOMESTIC SYSTEMS
Most design information in engineering textbooks is based
on large-scale sewer networks that have a continuous base
ow. Small-scale wastewater treatment systems often do not
have a continuous base ow. On the contrary, low ows are
zero (no ow), and peak ows are many times larger than the
average ow. These differences in water use patterns raise
issues that are not encountered in the design of larger sewage
treatment works.
SMALL FLOWS
For design of single-family home treatment systems, the
accepted practice in the United States is to base the design
ow on the number of bedrooms within the home. These pre-
scriptive ow determinations (typically ranging from 455 to
568 L/d per bedroom) are used to provide a sufcient factor

of safety for soil inltration of septic tank efuent. They do
not represent actual water use. Prescriptive ow determina-
tions are commonly interpreted as representing the maximum
expected occupancy of the home (two occupants per bedroom)
and a corresponding peak ow rate. As a result, peaking fac-
tor determinations and inltration/inow allowances are typi-
cally not necessary when using prescriptive ows based on
a bedroom count. Special provisions may apply in some cir-
cumstances (Minnesota Pollution Control Agency, 1999).
Flow projections may be based on population for small
communities. A prescriptive criterion of 379 L/d per person
is commonly used in North America (Great Lakes UMRB,
1997). This per-person ow guideline is intended to repre-
sent an average dry weather ow from domestic wastewater
sources plus a “normal” amount of inltration for gravity
sewers built with modern construction techniques. If the
only available information is the number of homes, an aver-
age number of people per household may be used to approxi-
mate the total population. The average household size in the
United States is 2.7 people (American Housing Survey, 2003),
although this varies by geographic location. An appropriate
peaking factor must be applied to determine peak ows.
PATTERNS OF SMALL FLOWS
Wastewater ow from individual residences is delivered to a
small-scale treatment system via a series of discrete pulses
triggered by ush toilets, washing machines, dishwashers, etc.
Low ows in small systems will be zero (no ow). Most water
use occurs in the morning, evening, and at mealtimes. In the
United States, water use from single-family homes has been
idealized for design purposes, as indicated in Figure 16.5. As

more and more homes are added to the system, ow pulses
overlap. If there are enough homes in the collection network,
ow pulses overlap to form a continuous base ow, and ow
peaks start to attenuate.
10%
15%
5%
0
0:00 3:00 6:00 9:00 12:00 15:00 18:00 21:00 24:00
Percent of Daily Flow
Time of Day
FIGURE 16.5 Idealized water use pattern for an individual home. (Adapted from NSF International (2000) Residential wastewater treat-
ment systems. NSF/ANSI 4–2000, NSF International: Ann Arbor, Michigan. Reprinted with permission.)
© 2009 by Taylor & Francis Group, LLC
598 Treatment Wetlands
For single-family homes, the ratio of the peak ow to the
average ow (peaking factor) can be ve or higher. Larger
treatment systems will experience lower peaking factors due
to overlapping ow pulses and the presence of a continuous
base ow. In the United States, Recommended Standards for
Wastewater Facilities (Great Lakes UMRB, 1997) suggest a
formula based on population to determine the ratio of the
peak hourly ow to the average daily ow (Equation 16.1).
For small populations (less than 100 people), this relation-
ship results in a peaking factor of approximately 4.5.
Q
Q
P
P
peak hourly

average day



18
4
(16.1)
where
Flow rate, any units
Population in
Q
P

 tthousands
The pattern of peak ow events can be altered dramatically
if wastewater is collected and pumped into the treatment
system, as might be the case for septic tank pretreatment.
Sources of inow and inltration from homes (sump pumps,
footing drains, roof leaders, and furnace drains) can easily
produce much higher ows unless they are identied during
the design process with a home plumbing survey program
and subsequently separated from the wastewater collection
system.
ACTUAL WATER USE
Water use studies in the United States estimate an average
daily water use of 189 to 265 L/d per person for homes built
before 1994; implementation of standards for water-efcient
appliances since then has reduced water use in newer homes
to approximately 161 to 227 L/d per person (U.S. EPA,
2002c). Water use is strongly inuenced by cultural practices

and varies widely from country to country. Across Europe,
typical ow rates in small communities (less than 500 people)
range from 80 to 120 L/d per person (IWA Specialist Group
on Use of Macrophytes in Water Pollution Control, 2000). In
Germany, water use rates are much lower than in the United
States (Gesellschaft zur Förderung der Abwassertechnik d.V
(GFA), 1998), at 100–150 L/d per person. In urban areas of
developing countries, water use is approximately 60 L/d per
person (Nhapi et al., 2003).
Lagoons
There are many variants on the concept of aquatic units for
wastewater treatment, ranging from single-pond units (Water
Environment Federation, 2001) to complex arrays of multiple
units (Craggs, 2005). Often, other treatment process units are
added to complement the pond itself (Middlebrooks et al.,
2005). The combination of a pond followed by a wetland has
been explored at a number of locations (Horne, 1995; Stein-
mann et al., 2003; Tanner and Sukias, 2003; Kadlec, 2003d;
Polprasert et al., 2005; Wang et al., 2005; Kadlec, 2005e).
Because the wetland is often an add-on, the ow of the water
exiting the pond is often known from performance data.
The prescription for lagoon operation may be continuous
discharge, typical of warm climates, or episodic discharge,
typical of cold climates. Lagoon systems often discharge to
surface waters, for which the goal is to minimize water quality
impacts. Maximum dilution occurs at high ow of the recipi-
ent, which in turn occurs during freshets, i.e., the spring thaw
and the autumn wet season. Therefore, lagoon discharges are
traditionally scheduled for those times of maximum dilu-
tion. When a wetland is added to the system, there are more

options for scheduling the discharge. For instance, the system
may be designed for discharges that avoid ammonia toxicity
in the recipient (Kadlec and Pries, 2004). Winter storage may
be contemplated, provided capacity is present or designed.
Therefore, the designer has an added degree of freedom: the
total annual volume may be managed to optimize treatment,
perhaps at the expense of more pond volume. This design
feature is discussed in more detail subsequently.
Me
ch
anical Plants
Pretreatment systems, such as activated sludge plants, are
small-retention devices, which do not typically have much
capacity to dampen the incoming ow pulses. Hour-to-hour,
day-to-day, and month-to-month ow variations are likely to
be passed through the pretreatment system, and thus affect
what is entering a follow-on treatment wetland. These pulses
will then be partially evened out by an add-on wetland.
Flows, whether municipal or industrial, are often seasonal in
character. It is necessary to anticipate those patterns, because
the wetland must function appropriately under these variable
hydraulic conditions. Monthly ow estimates will be required
for most point-source projects.
INFILTRATION AND INFLOW
Inltration is dened as groundwater that seeps into a waste-
water collection system. It invariably introduces additional
ow into the collection network. Inltration is strongly
inuenced by groundwater elevation, workmanship of sewer
construction, quality of construction materials, and fraction
of the overall collection network that relies on gravity ow.

Typical sources of inltration include poorly installed service
laterals, leaking joints on sewer pipes, cracked sewer pipes,
and leaking manholes. Exltration (movement of water out of
the collection system) can also occur. Portions of collection
systems that are pumped (such as pressure sewers) have posi-
tive internal pressures and are often pressure-tested during
construction. As a result, pressure sewer collection systems
have a much lower potential for inltration.
Inow is dened as extraneous water that is directly dis-
charged to the wastewater collection system. In combined
sewer systems, stormwater is a major source of inow. It is
driven by rainfall intensity and amount of impervious sur-
face present within the catchment area. In newer collection
networks, stormwater is almost always excluded. In these
© 2009 by Taylor & Francis Group, LLC
Design Basis 599
situations, major sources of inow are generally limited to
roof leaders, sump pumps, and foundation drains. Because
most inow sources are driven by rainfall, these tend to be
high-ow, short-duration events. These “surge” events can
have major impacts on treatment systems. The combined
effect of inltration and inow depend on a number of fac-
tors, including the integrity of the sewer system, size of the
collection pipes, the presence of high groundwater, and other
factors. Typical allowances for combined inltration/inow
range from 0.09 to 0.9 m
3
/d/cm/km (Metcalf and Eddy,
1998).
WATER QUALITY

The concentrations of the pollutants in the water to be treated
are critical to the sizing process, and to the prediction of the
wetland performance in the face of unknown future varia-
tions. A clear denition of the incoming water quality, includ-
ing the anticipated temporal distribution of concentrations,
is essential. There are often seasonal uctuations for point
sources, as well as diurnal uctuations. Incoming patterns
of chemical composition propagate through the wetland and
undergo modication, resulting in a spectrum of output com-
positions. Some of this output variability may be predicted
by the design models, namely, those variations that represent
responses to moderately slow input changes (those which
occur on monthly or less). Faster events involve ecosystem
processes that are not included in the design models available
at the present time, and therefore will give the appearance of
generating stochastic variations.
In domestic and municipal wastewater collection sys-
tems, the following components contribute to sewage ow:
Human excreta (feces and urine)
Wastewater generated by personal use, including
washing, laundry, food preparation, etc. (graywa-
ter), and water used as the carrier media for human
bodily wastes (blackwater)
Water that inadvertently leaks into the collection
system (inltration and inow)
Wastewater from commercial or industrial sources
•
•
•
•

Component concentrations of sewage will depend on the
individual circumstances of the community. In communities
with hand-carried water supplies or limited water resources,
graywater is often looked at as an irrigation resource and is
not commingled with excreta. In industrial societies, mixing
graywater and blackwater to produce “combined sewage” is
the norm (Günther, 2000). As a result of these differences in
water usage, and hence composition, wetland designers need
to consider how the community is utilizing water.
S
m
allDomesticSystems
There is often no composition data to be used for the design
of treatment systems for small systems. It is necessary to
resort to estimating methods that consider water use and
population in the source community. Untreated human urine
and fecal material can introduce a variety of pollutants into
the environment. Typical per-person generation rates are
su
mmarized in Table 16.1 (Del Porto and Steinfeld, 2000).
Graywater includes spent water from bathtubs, showers,
washbasins, washing machines, laundry tubs, kitchen sinks,
and dishwashers. In developed countries, graywater accounts
for 50 to 82% of household water use and represents about
half of the organic waste solids produced in the home. When
conventional ush toilets are used in a waterborne sewer sys-
tem, graywater is often combined with blackwater. Relative
contributions of pollutants by source (for a combined sewer
system) are summarized in Table 16.2 (U.S. EPA, 2002c).
Typical constituent concentrations for residential septic tank

systems are given Table 16.3.
Lagoons
Another source of treatment wetland inuents arises from
pond treatment as the initial component of the treatment
train. One or more facultative, anaerobic or aerated ponds
or lagoons may be used (Shilton, 2005). Because the wet-
land is often an add-on, the quality of the water exiting the
pond is often known from performance data. If the entire
system is constructed at the same time, the lagoon elements
should be designed according to the currently accepted
TABLE 16.1
Typical Per-Person Waste Generation Rates
Parameter Urine Feces Combined
Volume (L per capita·day) 1.2 0.15 1.4
Dry solids (g per capita·day) 60 45 105
Moisture content (g per capita·day) 95% 70% —
Organic carbon (g per capita·day) 8.5 22 30
BOD
5
(g per capita·day) 7.5 11 18.5
Nitrogen (g per capita·day) 11 2 13
Phosphorus (g per capita·day) 1 0.6 1.6
Source: From Del Porto and Steinfeld (2000) The Composting Toilet System Book. Center
for Ecological Pollution Prevention, Concord, Massachusetts. Reprinted with permission.
© 2009 by Taylor & Francis Group, LLC
600 Treatment Wetlands
methods (Shilton, 2005). Lagoon systems are typically
designed for reduction of BOD and total suspended solids
(TSS), and occasionally ammonia. Data and older models
exist for pond pathogen reduction, but have not been recently

updated and synthesized (Davies-Colley, 2005). Phosphorus
data for lagoon systems are not voluminous, because it is not
frequently regulated in lagoon discharges. Some approxima-
tions of the efuent characteristics of several types of lagoons
are shown in Table 16.4.
Mechanical Plants
Table 16.5 summarizes the typical quality of medium-
strength, raw, municipal wastewater in the United States and
provides a range of values for commonly observed constitu-
ents. Municipal wastewater is composed of a variable array
of components characterized by the presence of biodegrad-
able organic matter (paper, feces, and food), particulate and
dissolved solids, and nutrients. Many municipal wastewaters
also receive some component of industrial waste. These ows
and residential sources may add trace metals and pesticides
to typical municipal wastewater.
Table 16.5 also provides a range of estimated treat-
ment efciencies for conventional primary and secondary
treatment processes, and summarizes the typical quality of
secondarily treated municipal wastewaters. These removal
efciencies vary widely depending on the types of treatment
processes. However, it is generally observed that at least 70%
of the BOD and TSS are removed from municipal wastewater
during primary and secondary treatment. Treatment require-
ments have generally increased over the past decades, and
many treatment plants now include at least partial nitrica-
tion, perhaps denitrication, and phosphorus removal. These
blur the terminology, because they range from “advanced
secondary” to “tertiary” and beyond. The follow-on treat-
ment wetland may therefore be termed “tertiary” or, as might

be supposed, “quaternary.”
This summary can be used as a rough estimate of the inu-
ent water quality to be applied to a wetland system designed
for primary, secondary, or advanced wastewater treatment.
TABLE 16.3
Typical Wastewater Component Concentrations Entering and Leaving a Residential Septic Tank
Parameter
Raw Waste
Central Estimate Range
Septic Tank Effluent
Range
Without Kitchen Solids
(Central Estimate)
With Kitchen Solids
(Central Estimate)
BOD
5
(mg/L) 450 210–530 180 190 140–200
TSS (mg/L) 500 230–600 80 85 50–90
NH
4
-N (mg/L) 40 7–40 40 45 40–60
TKN (mg/L) 70 50–90 70 75 50–90
NO
x
-N (mg/L) <1 0–1 <1 <1 0–1
Total phosphorus
(mg/L)
17 10–30 16 16 12–20
Fecal coliforms

(CFU/mL)
—10
6
–10
10
——10
3
–10
6
Viruses (PFU/mL) — — — — 10
5
–10
7
Source: Data from Metcalf and Eddy Inc. (1991) Wastewater Engineering, Treatment, Disposal, and Reuse. Tchobanoglous and
Burton (Eds.), Third Edition, McGraw-Hill, New York; Crites and Tchobanoglous (1998) Small and Decentralized Wastewater
Management Systems. McGraw-Hill, New York.
TABLE 16.2
Typical Per-Person Combined Sewage Generation Rates
Parameter (Mean Values)
BOD
5
(g per capita·day)
Suspended Solids
(g per capita·day)
Nitrogen
(g per capita·day)
Phosphorus
(g per capita·day)
Garbage disposal 18.0 26.5 0.6 0.1
Toilet 16.7 27.0 8.7 1.6

Bathtubs, sinks, appliances 28.5 17.2 1.9 1.0
Approximate total 63.2 70.7 11.2 2.7
Source: Adapted from U.S. EPA (2002c) Onsite wastewater treatment systems manual. EPA 625/R-00/008. U.S. EPA Ofce of Research
and Development: Washington, D.C.
© 2009 by Taylor & Francis Group, LLC
Design Basis 601
TABLE 16.4
Typical Composition of Lagoon Discharge Water and Percent Removals at Various Levels of Treatment
Parameter Primary Anaerobic Secondary Aerobic Facultative Aerated Facultative Aerated Partial Mix
BOD Reduction, % 50–85 — 75–95 — Up to 95
Median efuent, mg/L — — — — 25
Efuent range, mg/L — 20–40 15–35 1–45 15–40
TSS Reduction, % — — — — —
Median efuent, mg/L — — — — 40
Efuent range, mg/L 80–160 80–140 10–90 1–90 20–60
NH
4
-N Reduction, % — — 23–97 — —
Median efuent, mg/L — — 20 — 4
Efuent range, mg/L — — 0.2–25 0–12 —
TP Reduction, % — — 50 — 15–25
Median efuent, mg/L — — — — 5
Efuent range, mg/L — — — 3–4 —
FC Reduction (log
10
) — — 1–5 1–5 —
Median efuent, CFU/100 mL — — 200 100 100
Source: Shilton (2005) In Pond Treatment Technology. Shilton (Ed.), IWA Publishing, London; Metcalf and Eddy Inc. (1991) Wastewater Engineering, Treat-
ment, Disposal, and Reuse. Tchobanoglous and Burton (Eds.), Third Edition, McGraw-Hill, New York; Crites and Tchobanoglous (1998) Small and Decentral-
ized Wastewater Management Systems. McGraw-Hill, New York; Crites et al. (2006) Natural Wastewater Treatment Systems. Meyer (Ed.), CRC Press, Boca

Raton, Florida; U.S. EPA (1983a) Design manual: Municipal wastewater stabilization ponds. EPA 625/1-83/015, U.S. EPA Ofce of Water: Cincinnati, Ohio;
U.S. EPA (1983c) Wastewater stabilization ponds: Nitrogen removal. U.S. EPA Ofce of Water: Washington, D.C.; Rich (1999) High Performance Aerated
Lagoons. American Academy of Environmental Engineers, Annapolis, Maryland.
TABLE 16.5
Typical Composition of Municipal Wastewater and Percent Removals at Various Levels of Treatment
Constituent
Raw Wastewater (mg/L) Percent Removal Secondary Effluent (mg/L)
Typical Range Primary Secondary Typical Range
BOD
5
220 110–400 0–45 65–95 20 10–45
COD 500 250–1,000 0–40 60–85 75 35–75
TSS 220 100–350 0–65 60–90 30 15–60
VSS 165 80–275 — — — —
NH
4
-N 25 12–50 0–20 8–15 10 <1–20
NO
x
-N 0 0 — — 6 <1–20
Org-N 15 8–35 0–20 15–50 4 2–6
TKN 40 20–85 0–20 20–60 14 10–20
Total N 40 20–85 5–10 10–20 20 10–30
Inorg P 5 4–15 — — 4 2–8
Org P 3 2–5 — — 2 0–4
Total P 8 6–20 0–30 10–20 6 4–8
Arsenic 0.007 0.002–0.02 34 28 0.002 —
Cadmium 0.008 <0.005–0.02 38 33–54 0.01 <0.005–6.4
Chromium 0.2 <0.05–3.6 44 58–74 0.09 <0.05–6.8
Copper 0.1 <0.02–0.4 49 28–76 0.05 <0.02–5.9

Iron 0.9 0.10–1.9 43 47–72 0.36 0.10–4.3
Lead 0.1 <0.02–0.2 52 44–69 0.05 <0.02–6.0
Manganese 0.14 0–0.3 20 13–33 0.05 —
Mercury 0.001 <0.0001–0.0045 11 13–83 0.001 <0.0001–0.125
Nickel 0.2 — — 33 0.02 <0.02–5.4
Silver 0.022 0.004–0.044 55 79 0.002 —
Zinc 1.0 — 36 47–50 0.15 <0.02–20
Source: W
PCF (1983) Nutrient Control. Manual of Practice FD-7, Water Pollution Control Federation: Washington, D.C.; Metcalf and Eddy Inc. (1991)
Wastewater Engineering, Treatment, Disposal, and Reuse. Tchobanoglous and Burton (Eds.), Third Edition, McGraw-Hill, New York; Richardson and
Nichols (1985) In Ecological Considerations in Wetlands Treatment of Municipal Wastewaters. Godfrey (Ed.), Van Nostrand Reinhold Company, New York,
pp. 351–391; Krishnan and Smith (1987) In Aquatic Plants for Water Treatment and Resource Recovery. Reddy and Smith (Eds.), Magnolia
Publishing, Orlando, Florida, pp. 855–878; Williams (1982) In Water Reuse. Middlebrooks (Ed.), Ann Arbor Science, Ann Arbor, Michigan, pp. 87–136.
© 2009 by Taylor & Francis Group, LLC
602 Treatment Wetlands
16.3 CHARACTERIZATION OF
OTHER WASTEWATERS
I
NDUSTRIAL WASTEWATERS
Although industrial wastewater quality varies among indus-
tries, it has a fairly consistent intrasystem efuent quality.
Table 16.6 summarizes the typical quality of raw wastewater
from a number of industries that have used wetlands treatment
technology. Raw industrial wastewater usually receives some
level of pretreatment before discharge to a wetlands treatment
system. If total concentrations of BOD, suspended solids, and
ammonia nitrogen in untreated industrial wastewater are in
the concentration range of hundreds to thousands milligrams
per liter, it is generally not acceptable for wetlands discharge
without additional pretreatment.

LANDFILL LEACHATES
Treatment and disposal of liquid leachates is one of the most
difcult problems associated with the use of sanitary land-
lls for disposal of solid waste. Leachates are produced when
rainfall and percolated groundwater combine with inorganic
and organic degraded waste. In unlined landlls, leachates
TABLE 16.6
Typical Pollutant Concentrations in a Variety of Untreated Industrial Wastewaters
Constituent Units
Pulp and
Paper
a
Landfill
Leachate
b
Coal Mine
Drainage
d
Petroleum
Refinery
e
Electroplating
f
Breweries
g
BOD
5
mg/L 100–500 42–10,900 — 10–800 — 1,500–3,000
COD mg/L 600–1,000 40–90,000 — 50–600 — 800–1,400
TSS mg/L 500–1,200 100–700 — 10–300 4–600 100–500

VSS mg/L 100–250 60–280 — — 30–100 50–500
TDS mg/L — — — 1,500–3,000 800–5,800 50–350
NH
4
-N mg/L — 0.01–1,000 — 0.05–300 — —
TN mg/L — 70–1,900 — — 10–120 25–45
TP mg/L — <0.01–2.7 — 1–10 20–50 –
pH S.U. 6–8 3–7.9 3–5.5 8.5–9.5 4–10.5 5–7
Sulfate mg/L — 10–260 20–2,000 ND–400 30–120 —
Conductance µS — 1,200–16,000 — — — —
TOC mg/L — 11–8,700 — 10–500 — —
Aluminum mg/L — 0.5 50 — — —
Arsenic mg/L — 0.011–10,000 — — — —
Barium mg/L — 0.1–2,000 — — — —
Cadmium µg/L — 5–8,200 — — 10,000–50,000 —
Chromium mg/L — 0.001–208 — ND–3 10–120 —
Iron mg/L — 0.09–678 50–300 — 2–20 —
Lead µg/L — 1–19,000 — — — —
Manganese mg/L — 0.01–550 20–300 — — —
Selenium µg/L — 3–590 — — — —
Oil and Grease mg/L — — — 10–700 — —
Phenols mg/L — <0.003–17 — 0.5–100 — —
Cyanide mg/L — — — — 1–50 —
Note: ND  not detected.
a
From Jorgensen (1979) Studies in Environmental Science 5. Elsevier, New York.
b
From Staubitz et al. (1989) Constructed Wetlands for Wastewater Treatment: Municipal, Industrial, and Agricultural.
Hammer (Ed.), Lewis Publishers, Chelsea, Michigan, pp. 735–742; Lema et al. (1988) Water, Air, and Soil Pollution 40:
223–250; Bolton and Evans (1991) Water, Air, and Soil Pollution 60: 43–53.

c
From Wildeman and Laudon (1989) In Constructed Wetlands for Wastewater Treatment: Municipal, Industrial, and Agricul-
tural. Hammer (Ed.), Lewis Publishers, Chelsea, Michigan.
d
From Girts and Kleinmann (1986) National Symposium on Mining, Hydrology, Sedimentology, and Reclamation. University
of Kentucky Press, Louisville, Kentucky, pp. 165–171.
e
From Adams et al. (1981) Development of Design and Operational Criteria for Wastewater Treatment. Enviro Press, Nash-
ville, Tennessee; ANL (1990) Environmental consequences of, and control processes for, energy technologies. Argonne
National Laboratory (ANL) and Noyes Data Corporation: Park Ridge, New Jersey.
f
From OECD (1983) Emission Control Costs in the Metal Plating Industry. Organization for Economic Cooperation and
Development (OECD).
g
From Cooper (1978) The textile industry. Environmental control and energy conservation. Noyes Data Corporation. Park
Ridge, New Jersey; Wildeman et al. (1993a) Wetland Design for Mining Operations. Bitech Publishers, Vancouver, British
Columbia.
© 2009 by Taylor & Francis Group, LLC
Design Basis 603
frequently discharge to groundwater or appear as surcial
drainage around the base of the landll. In modern lined
landlls, leachates are collected from the lined cells and
routed to treatment units. The use of constructed wetlands to
treat these landll leachates is a well-developed technology,
currently undergoing rapid expansion in application. This
application is discussed further in Chapter 25.
The highly variable nature of solid waste, differences in
age and decomposition, and the diversity of chemical and
biological reactions that take place in landlls result in a
wide range of chemical quality of leachates (McBean and

Rovers, 1999). Reviews of “average” landll concentrations
of COD, volatile acids, and nitrogenous compounds show
increases during the rst few years of operation and then
decline over ten or more years. Table 16.6 provides typical
ranges encountered in landll leachates. Flows are generally
low, but vary depending on management and minimization
of percolation from rainfall. Clearly, the expected volume
and chemical quality of a landll leachate is highly site-spe-
cic, may change over time, and must be estimated on a case-
by-case basis for wetland treatment system design. A detailed
discussion of the leachate quality that may reach a treatment
wetland is found in McBean and Rovers (1999).
PULP AND PAPER WASTEWATER
The pulp and paper industry converts wood products includ-
ing pines, spruce, poplar, beech, birch, and aspen, as well as
recycled paper, into liqueed cellulose pulp and paper. Raw
wood and wood chips are converted to pulp (cellulose bers)
by mechanical grinding (ground wood) or through chemi-
cal degradation and leaching (sulte and Kraft processes).
At an increasing number of pulp and paper mills, this pulp
is bleached to delignify and decolorize the cellulose bers
before paper manufacture. About 29–34 m
3
of raw wastewa-
ter is produced for each metric ton of pulp and paper pro-
duced (Britt, 1970). Total wastewater ow for the U.S. pulp
and paper industry is about 20 × 10
6
m
3

/d (Greyson, 1990).
Table 16.6 summarizes the typical composition of this waste-
water, although different manufacturing processes result in
different wastewater qualities. This ow is equivalent to a
raw organic matter (BOD
5
) load of about 15 × 10
6
metric tons
per day (Greyson, 1990).
Raw wastewater from pulp and paper mills typically
receives primary treatment through settling, either in ponds
or in primary clariers. When required to meet discharge
limitations, secondary treatment at most pulp and paper
mills includes biological conversion of BOD
5
and additional
solids settling in aerated lagoons or in conventional acti-
vated sludge treatment systems. To meet reduced efuent
limitations, some pulp and paper mills are being required to
provide treatment beyond the secondary level. The goals of
additional treatment depend on site conditions, such as the
quality of the efuent after secondary treatment and water
quality permit limits in the receiving water. One goal may
be to further reduce BOD
5
, TSS, nitrogen, phosphorus, color,
chlorinated organics (such as adsorbable organic halides or
dioxin), and whole efuent toxicity. Constructed and natural
wetland treatment systems have been used at a number of

pulp and paper mills to provide this advanced secondary or
tertiary treatment (Knight et al., 1994; Knight, 2004).
MINE DRAINAGE
During and following mining operations, runoff and leachate
from tailings and from abandoned tunnels and shafts dissolve
trace metals, contaminating nearby surface waters. Leach-
ates from coal and metal mines contain residual trace met-
als, notably iron, manganese, and aluminum from coal mines
(Table 16.6). Seeps from abandoned mines typically received
no treatment in the past, but there is an increasing emphasis
on corrective measures (Younger et al., 2002). Constructed
wetlands are used as a technology to reduce metal concentra-
tions in mining wastewater (Younger et al., 2002; PIRAMID
Consortium, 2003a, 2003b). As discussed in Chapter 11, met-
als are precipitated and sequestered in sediments, and taken
up by plants in wetlands. Wetland design for metals removal is
sometimes limited by the need to avoid toxic concentrations in
tissues that could subsequently accumulate in the food chain.
Information necessary to evaluate the ability of wetlands to
provide treatment of these waste products is summarized
in Chapter 11 and in the references listed previously. Both
removal rate specication and rate constants have been used
in design, but the available level of detailed design data sup-
port is not as large as for municipal wastewater treatment.
During coal mining, iron pyrite and other metal-
bearing minerals are exposed to percolating water, which
leads to the release of acidic leachates to surface water. These
drainages typically have low pH and elevated concentrations
of dissolved iron, sulfate, calcium, and magnesium. In addi-
tion, the drainages have variable and somewhat elevated

concentrations of aluminum, copper, manganese, nickel, and
zinc (Table 16.6). Many of the streams and impoundments in
the Appalachian coal mining region of the United States are
affected by acid mine drainage. Conventional treatment of
leachates at these sites includes surface grading and recon-
touring to reduce or divert ows and chemical buffering and
precipitation with mechanical treatment plants to improve
water quality. Because these processes have relatively high
capital and lifetime costs, there has been considerable inter-
est in developing more cost-effective alternatives. Beginning
in the early 1980s, research focused on the potential of aero-
bic wetlands for precipitation of ferric sulfate to neutralize
pH and reduce dissolved ferrous iron concentrations. Con-
structed wetlands are now used at many sites in the United
States and Europe to increase the pH and reduce concentra-
tions of iron and manganese at coal mine sites (Kleinman
and Hedin, 1989; Younger et al., 2002; PIRAMID Consor-
tium, 2003a, 2003b) of coal mine drainage.
PETROLEUM INDUSTRY WASTEWATER
Because of the diverse processes at reneries and associ-
ated transportation facilities, and the storage of ammable
© 2009 by Taylor & Francis Group, LLC
604 Treatment Wetlands
liquids, land area requirements are large and include many
kilometers of piping and hundreds of tanks and storage areas.
Wastewater is generated by manufacturing processes, cool-
ing tower blowdown, water and sludge drainage from tanks,
and stormwater drainage and runoff (UNEP, 1987). Typical
wastewater pollutants at petroleum reneries include BOD
5

,
COD, oil and grease, TSS, NH
4
-N, phenolics, H
2
S, trace
organics, and heavy metals. Concentrations of many of these
pollutants are reduced through source control and prelimi-
nary treatments such as sour water stripping, oxidation and
neutralization of spent caustics, and cooling tower blowdown
treatment. Table 16.6 lists some examples of pollutant con-
centrations remaining in renery wastewater.
Raw wastewater from petroleum reneries typically
receives additional treatment including gravity separation of
oils and greases, primary clarication, dissolved air otation,
and secondary treatment, including oxidation ponds, aerated
lagoons, activated sludge, trickling lters, and activated car-
bon. The API separator process typically removes 60 to 99%
of the oil and grease, and smaller proportions of other pollut-
ants. Primary treatment removes 20 to 70% of the BOD
5
and
TSS and 10 to 60% of the COD. Secondary treatment will
reduce 40 to 99% of the BOD
5
, 30 to 95% of the COD, 40 to
90% of the TOC, 20 to 85% of the TSS, 60 to 99% of the oil
and grease, 60 to 99% of the phenol, 9 to 99% of the NH
4
-N,

and 70 to 100% of the sulde (ANL, 1990).
As described in Chapter 13, constructed wetlands are pro-
viding advanced secondary and tertiary treatment of process
water and stormwater at a large number of reneries (Knight
et al., 1999; API, 1999). Constructed wetlands typically will
reduce remaining concentrations of BOD
5
, COD, TSS, NH
4
-N,
oils and grease, phenols, and metals to advanced treatment
levels.
ANIMAL INDUSTRY WASTEWATERS
Animal industry wastewater contains high BOD
5
, COD, TSS,
and nutrients and is qualitatively similar to municipal waste-
water. Mass loadings from animal feed lots and other con-
centrated agricultural activities require intensive treatment
systems to provide environmental protection. Traditional
treatment methods such as anaerobic lagoons and spray irri-
gation are not always adequate to provide high-quality water
for off-site discharge. Constructed wetlands are being used
in a growing number of cases to receive pretreated dairy and
swine wastes (NADB database, 1998; Knight et al., 2000).
These wetland treatment systems must be designed with rea-
sonable organic loadings to prevent plant mortality, odors,
and poor treatment efciencies. Treatment wetlands are a
compatible component of on-farm, total waste management.
Their land intensiveness is not a serious limitation in most

instances. Farmers typically have the equipment and skills
necessary to build their own wetlands and operate them suc-
cessfully. Table 16.7 summarizes the composition of wastes
from animal operations, both entering and leaving treatment
wetlands.
STORMWATER RUNOFF
Concentrations of most parameters in stormwater are time
dependent. Stormwater concentrations and loads are cyclic
TABLE 16.7
Average Wetland Influent and Effluent Concentrations of Selected Animal Facilities (mg/L)
Wastewater Type: Dairy Poultry Cattle Swine
Number of Systems: 60 5 9 58
BOD
5
In 442 153 137 104
Out 141 115 24 44
% Reduction 68 25 83 58
TSS
In 1,111 — 291 128
Out 592 — 55 62
% Reduction 47 — 81 52
NH
4
-N
In 105 74 5.1 366
Out 42 59 2.2 221
% Reduction 60 20 57 40
TN
In 103 89 — 407
Out 51 70 — 248

% Reduction 51 22 — 39
Source: Data from NADB database (1998) North American Treatment Wetland Database (NADB), Version 2.0. Compiled by
CH2M Hill. Gainesville, Florida; and Knight et al. (2000) Ecological Engineering 15(1–2): 41–55.
© 2009 by Taylor & Francis Group, LLC
Design Basis 605
with periods of dry fall and deposition, then the rst ush
of runoff after rain, followed by exponential decreases in
runoff constituent concentrations as storages rinse from the
landscape, and nally dry conditions and deposition until
the next storm event. Chapter 14 provides a more complete
description of the expected ows and concentrations for such
event-driven systems.
Table 16.8 provides typical mean concentrations for con-
stituents. The averages are ow-weighted to provide realistic
estimates of the total constituent load that escapes during
multiple storm events. Instantaneous concentrations will
be considerably higher than these averages. Pollutant con-
centrations and loads generally range from low levels, from
undeveloped and park lands to low-density residential and
commercial, to agricultural, to higher-density residential and
commercial, and nally to high-density commercial, indus-
trial, and agricultural land uses. Mean concentrations per
event for BOD
5
vary from 1.45 mg/L for undeveloped lands
to 20 mg/L for high-density urban areas. TSS concentrations
vary from 11 mg/L for undeveloped areas to 150 mg/L for
high-density urban areas.
The mass loading rates provided in Table 16.8 represent
normalized pollutant loads that are somewhat independent

of local rainfall amounts. Because pollutant loads per area
per time are relatively constant between similar land use
areas, variable local rainfall washes these loads off the land
in a few large events or over many smaller events. Urban
pollutant loads increase with the imperviousness of the
watershed. Although 20 to 40% of the material on street
surfaces is organic, it does not biodegrade easily because
TABLE 16.8
Composition and Mass Loading Rates for Stormwaters
Constituent
Urban Industrial Residential/Commercial Agricultural
Concentration
(mg/L)
Load
(kg/ha·yr)
Concentration
(mg/L)
Load
(kg/ha·yr)
Concentration
(mg/L)
Load
(kg/ha·yr)
Concentration
(mg/L)
Load
(kg/ha·yr)
BOD
5
20 (7–56) 90 9.6 34–98 3.6–20 31.59–135.2 3.8 11.59

COD 75 (20–275) — — — — — — —
TSS 150 (20–2890) 360 93.9 672–954.5 18–140 84.28–797 55.3 24.14
VSS 88 (53–122) — — — — — — —
NH
3
N 0.582 — — — — — 0.33–0.48 —
TKN 1.4 (0.57–4.2) — — — — — 2.16–2.27 —
TN 2.0 (0.7–20) 11.2 1.79 7.8–18.06 1.1–2.8 9.144–32.18 2.32 10.61
Ortho-P 0.12 — 0.13 1.321 0.05–0.40 0.568–3.302 0.13–0.227 0.942
TP 0.36 (0.02–4.3) 3.4 0.31 2.2–3.151 0.14–0.51 1.412–4.85 0.344 1.362
Copper 0.05 (0.01–0.40) 0.049 — 0.077 — 0.045 — —
Lead 0.18 (0.01–1.20) 0.174 0.202 0.269–2.053 0.065–0.214 0.157–2.431 — —
Zinc 0.20 (0.01–2.9) 0.630 0.122 0.98–1.240 0.046–0.170 0.218–1.88 — —
Chromium — 0.28 — 0.044 — 0.026 — —
Cadmium 0.0015 0.16 — 0.024 — 0.013 — —
Iron 8.7 — — — — — — —
Mercury 0.00005 0.043 — 0.065 — 0.038 — —
Nickel 0.022 0.032 — 0.030 — 0.029 — —
Oil and Grease 2.6 — — — — — — —
Source: Dames and Moore (1990) Lakeland Comprehensive Stormwater Management and Lake Pollution Study, Volume I. Report to the City of
Lakeland, Florida (May 1990); U.S. EPA (1983b) Design principles for wetland treatment systems. EPA 600/2 83/026, Hammer and Kadlec (Eds.),
National Technical Information Service; Marsalek and Schroeter (1989) Water Pollution Research Journal of Canada 23: 360–378; Bastian (1986)
Potential Impacts on Receiving Water. Urbonas and Roesner (Eds.). Proceedings of the ASCE Engineering Foundation Conference: Urban Runoff
Quality—Impact and Quality Enhancement Technology, 23–27 June 1986. American Society of Civil Engineers: Henniker, New Hampshire, pp.
157–160; Lager et al. (1977) Urban stormwater management and technology: Update and user’s guide. EPA 600/8–77/014, U.S. EPA Ofce of
Research and Development, Municipal Environmental Research Laboratory: Cincinnati, Ohio; Marsalek (1990) Water Science and Technology 22:
23–30; Driscoll (1986) Lognormality of Point and Non-Point Source Pollutant Concentrations. Urbonas and Roesner (Eds.), Proceedings of the
ASCE Engineering Foundation Conference: Urban Runoff Quality—Impact and Quality Enhancement Technology, 23–27 June 1986. American
Society of Civil Engineers: Henniker, New Hampshire, pp. 438–458; Shelley and Gaboury (1986) Estimation of Pollution from Highway Runoff—
Initial Results. Urbonas and Roesner (Eds.). Proceedings of the ASCE Engineering Foundation Conference: Urban Runoff Quality—Impact and

Quality Enhancement Technology, 23–27 June 1986. American Society of Civil Engineers: Henniker, New Hampshire, pp. 459–473; Novotny
(1992) Water Environment Technology January: 40–43.
© 2009 by Taylor & Francis Group, LLC
606 Treatment Wetlands
it comes from leaf and wood litter, rubber, and road sur-
face material (Novotny, 1992). The high metal content of
highway solids comes from vehicle emissions. Novotny
(1992) reported that the average total nitrogen load from
urban lands is 5 kg/ha·yr (1 to 38.5 kg/ha·yr), and the total
phosphorus load averages 1 kg/ha·yr (0.5 to 6.25 kg/ha·yr).
Urban and residential runoff is being treated with wetland
detention basins (Kehoe, 1993) and constructed wetlands
(Carleton et al., 2001).
16.4 TREATMENT GOALS
Wastewater treatment and disposal are regulated by an ever-
increasing number of federal, state (provincial), and local
laws, rules, ordinances, and standards. In some cases, the
most challenging part of implementing a wetland treatment
project is complying with regulations through the permitting
process. A detailed knowledge of the pertinent regulations
is essential to evaluate the feasibility of a wetland treatment
project, and to design it properly. An up-to-date, detailed
survey of federal, state, and local ordinances should be con-
ducted to determine those that might be relevant to specic
projects.
Treated water may be destined for one of three primary
receivers: surface water, groundwater, or irrigation (reuse).
There are often stringent specications of quality that must
be met to allow discharges to these recipients, and they are
quite different. The intent of specications for surface water

discharges is the preservation or improvement of the des-
ignated uses of those waters. No matter what the receiving
ecosystem or post-wetland treatment element, proper design
requires a clear statement of the required water quality leav-
ing the treatment wetland.
RECEIVING WATER STANDARDS
In the United States, the Clean Water Act created the National
Pollutant Discharge Elimination System (NPDES) permit-
ting program. An NPDES permit is required for nearly all
point discharges of water or wastewater into waters of the
United States, including municipal and industrial wastewa-
ter. NPDES permits specify allowable ows and chemical
quality of discharges into waters of the United States based
on established water quality standards for those receiving
waters. The Clean Water Act guides water quality standards,
which are promulgated individually by the states. Water
quality standards vary among water bodies within a state
and among states, depending on specic receiving water
resources. Many wetland treatment systems discharge to
surface waters, and therefore must meet the conditions of a
discharge permit. The conditions of the permit dictate the
required performance of the wetland, and therefore govern
its sizing.
Traditionally, permits have been developed to control
both ows and loads of pollutants. There are typically annual
averages and monthly and weekly maxima, perhaps adjusted
seasonally. The relation between averages and maximum
allowable concentrations may have been determined from
other technologies, and may be inappropriate for a wetland
system. Similar procedures are in place for other countries.

Discharge in some circumstances is directed to down-
stream wetlands, which are often a combination of surface
and subsurface waters. These wetlands are often federally
regulated waters, and subject to appropriate regulations.
However, the upstream treatment wetland is a treatment
system, typically regulated according to a different set of
rules. The design goals for the treatment wetland therefore
become the water quality and quantity desired for the man-
agement of the downstream, jurisdictional wetland. These
are likely to be stricter than for discharge to a large river, for
example (due to lower dilution factors in the natural wetland
environment).
GROUNDWATER DISCHARGES
Groundwater discharges are regulated via either the codes
for single-home on-site (septic) drainelds or the rules for
inltration of treated wastewaters from municipal treatment
plants. The septic draineld codes are not based explicitly
on water quality, and typically use a prescriptive approach
based on a presumed reduction in pathogen counts. Larger
systems typically target specic water quality parameters,
such as nitrate, and hydrogeologic modeling is often required
to determine the fate and transport of these parameters in the
subsurface environment.
Groundwater discharges of treated water are feasible in
a number of circumstances. The problems of avoidance of
eutrophication of surface receiving waters are replaced by
problems of ensuring proper quality for the aquifer to be
recharged. If the groundwater beneath the wetland, or a follow-
on inltration bed, is a drinking water source, then atten-
tion must be paid to nitrate, pathogens, and metals, as well

as to trace organic chemicals. However, many aquifers are
not, and will not be, used for potable water supply. Wetlands,
therefore, have a role in pretreatment for conventional rapid
inltration basins (RIBs) and in posttreatment for nitrate
removal from underdrained RIBs.
The primary concepts of regulation of groundwater dis-
charges concern nitrates, pathogens, and perhaps salts. The
limit of 10 mg/L nitrate-nitrogen as a drinking water stan-
dard in the United States results in specications of nitrate,
or by implication, total nitrogen in such discharges. The need
to regulate pathogens if any drinking water use is present is
obvious. Salt content is of concern if it is a perceived threat
to potable water supplies. Groundwater discharges may be
a preferred alternative because of the phosphorus-binding
potential of many soils. Land application has a long tradition
as a means of wastewater disposal. But it is often plagued
by a surplus of nitrogen, which escapes crop utilization and
nitries during transport to the groundwater. Wetlands have
the potential to strip excess wastewater nitrogen before land
application. In this application, wetland design targets the
requirements of the subsequent treatment process (i.e., land
application).
© 2009 by Taylor & Francis Group, LLC
Design Basis 607
The design of the treatment wetland interfaces with the
design of the inltration system. Several sources discuss the
design of rapid inltration systems (Crites and Tchobano-
glous, 1998; Water Environment Federation, 2001; Crites
et al., 2006).
INTERFACING TO REUSE

In many parts of the globe, water is in short supply. As
treatment technologies improve, it has become possible to
consider the treated water as a resource, with a variety of
potential benecial uses. The irrigation of agricultural crops
is the leading consumer of treated water. Crops include trees,
pastures, and fodder crops in North America, but food crops
are irrigated with treated water in other parts of the world.
The key consideration is the potential for passing of patho-
gens to consumers of the crop. Treatment wetlands have been
used to help condition the water for these applications (Crites
and Tchobanoglous, 1998). Landscape irrigation is another
reuse candidate, with applications for ornamentals and golf
courses. Again, treatment wetlands have been employed as
conditioners in this application (Wallace and Kadlec, 2005).
The reuse water quality standards to be met vary from
state to state in the United States. U.S. EPA has guidelines
for various categories of reuse (U.S. EPA, 2004). State regu-
lations or guidelines are in place in virtually all states, and
typically apply to several categories: (a) unrestricted urban
use, (b) restricted urban use, (c) nonfood crops, and (d) food
crops. Some of these regulations contain extremely low
requirements for solids and pathogens. For instance, the
state of California requirement is for turbidity less than 2
NTU, and total coliforms less than 2.2 MPN/100 mL, for
the highest reuse category (which incidentally includes use
for ushing toilets) (California Code of Regulations, 2001).
Other states have less stringent requirements; for instance,
Colorado allows unrestricted irrigation of water with a
BOD less than 20 mg/L, fecal coliform bacteria less than
25 MPN/100 mL, and TSS less than 40 mg/L (Colorado

Department of Health Water Quality Control Division,
2005). States that are driven primarily by liability concerns
tend to have very stringent reuse limits, whereas states driven
by water scarcity often have less strict limits to reduce the
economic barriers to water reuse.
In water-scarce areas such as Mediterranean countries,
water scarcity may be a driving factor for reuse, and lower
levels of treatment may be acceptable (e.g., control of para-
sites but not bacterial or viral contaminants). In these cases,
alternate treatment guidelines will apply (Korkusuz, 2005).
Clearly, there is a lower limit to the treatment achiev-
able in constructed wetlands, because natural processes cre-
ate background concentrations that may be in excess of local
regulatory requirements for water reuse.
EXCURSION CONTAINMENT AND SAFETY FACTORS
All treatment technologies possess a spectrum of efuent
concentrations, which is predictable only in the probabilistic
sense. Therefore, in addition to the mean efuent concen-
tration (which may vary in a deterministic way with tem-
perature and loading), there is an associated bandwidth of
concentration. Regulations may constrain both the mean and
the maximum of the band, via specication of a limit on the
maximum daily, weekly, or monthly value, together with a
limit on the average annual value. In design, care must be
taken to accommodate the most restrictive of multiple aver-
aging tests given by the regulation.
The average system performance will depend on either
the season of the year or the water temperature, or both.
These are deterministic variations; that is, they may be pre-
dicted from the seasonality of the removal rate coefcient or,

more directly, from information on observed trends in treat-
ment wetland outlet concentrations. Theta factors and trend
properties (given in Part I) allow the designer to forecast sea-
sonal deterministic trends. There remains the variability not
predicted by such seasonality (see Chapter 9, Figure 9.48, for
example).
Probabilistic effects are important in the utilization of
design models for predicting removal performance of treat-
ment wetlands. Regulatory requirements often employ a
standard other than a long-term average. There may be a
maximum monthly concentration not to be exceeded, or a
specied concentration not be exceeded for more than a cer-
tain percentage of samples. To illustrate, consider the repre-
sentation of Figure 16.6, showing the hypothetical reduction
curve for a typical wetland. The information in Part I is
unequivocal; as wetland size increases, there is a downward
trend in pollutant concentration.
However, there is also a scatter in the individual measure-
ments that make up the trend. During any specied part of
the year, the concentration of a pollutant follows a decreasing
curve, with wetland size (detention time), and has an associ-
ated bandwidth of scatter in expected values (Figure 16.6). In
this hypothetical example, that scatter is shown as a uniform
distribution about the trend line, with a bandwidth propor-
tional to the trend value. It is supposed that the regulatory
limit is a concentration of 30 mg/L, as a maximum allow-
able. The trend model tells us what size wetland is needed
to meet 30 mg/L as a long-term average, which is a deten-
tion time of 6.1 days. However, the scatter is such that half
the time the measured values will be higher, up to 42 mg/L.

The exceedance frequency is expected to be 50% because
the design is based on the mean performance. As this level of
excursions is likely to be quite unacceptable from a regula-
tory standpoint, it would be necessary to increase the size of
the wetland.
At a detention time of 9.1 days (almost a 50% increase
in wetland size), a large majority of excursions are contained
(95% this hypothetical example) below the regulatory limit,
and the system would experience exceedances only 5% of
the time. In this case, the 50% increase in size is needed for
excursion containment of an expected, quantied scatter. At
the size that contains excursions (9.1 days’ detention), the
trend value is 21 mg/L, or 70% of the limit value for the
design. This fraction is called the coefcient of reliability
© 2009 by Taylor & Francis Group, LLC
608 Treatment Wetlands
(COR). Crites and Tchobanoglous (1998) present a method
for its estimation from the coefcient of variation (variance/
mean) of one or more datasets, adapted from activated sludge
technology. In this book, an exceedance multiplier is used,
which is just the reciprocal of the COR:
Exceedance multiplier  1
1
9
COR
(16.2)
where
= additive fraction of trend of a give9 nn exceedance
frequency, dimensionless
Values of such multipliers were determined for many pollut-

ants for many wetlands, and the average values are tabulated
in Part I for various exceedance frequencies. The relationship
between monthly trend average efuent concentration and the
90th percentile monthly concentration for typically regulated
constituents is given in Table 16.9, which is extracted from
the various results in Part I.
Exit concentrations uctuate with an amplitude of about
1.75 times the mean for the 90th percentile, meaning that
this percentile is about 75% higher than the mean. (Fecal
coliform bacteria are a separate case; the multiplier for FC
is on log
10
).
Beyond excursion containment, standard engineering
practice is to allow a margin for the unexpected. There are
no xed rules for the selection of a safety factor for wetlands.
Water Environment Federation (2001) suggests that an extra
15–25% area is required for treatment wetlands. Equiva-
lently, the designer may elect to design to a more stringent
standard than that specied in the regulation. For instance,
in the example in Figure 16.6, it could be decided to contain
excursions below 20 mg/L instead of 30 mg/L. That leads
to a yet larger area requirement, corresponding to 13.1 days’
detention in this example.
TABLE 16.9
Trend Multipliers Required to Contain the 90th Percentile
of Excursions around Trend Means for Various Pollutants
Parameter FWS HSSF VF
BOD 1.80 1.78 —
TSS 2.21 1.84 —

NH
4
-N 1.89 1.76 3.03
NO
x
-N 1.74 — —
TN 1.53 — —
TP 1.94 1.66 —
FC (log
10
) 1.23 1.48 —
Note: Other percentiles can be found in the pollutant chapters of Part I.
FIGURE 16.6 A hypothetical example of design for excursion containment. The trend represents P  3, C*  5 mg/L, and k  30 m/yr.
The scatter is a uniform distribution with a o50% bandwidth. On average, a goal of 30 mg/L can be met with 6.1 days’ detention. To avoid
exceedances at the 95% level (one time in 20), the detention time should be 9.1 days. More wetland area (more detention time) may be added
as a safety factor; i.e., 13.1 days should contain outlet concentrations to less than 20 mg/L most of the time.
20
0
40
60
80
100
120
140
024 86 10121416
HRT (days)
Concentration Out (mg/L)
Scatter
Trend
95th Percentile

Trend design
(6.1d)
Excursion containment
(9.1d)
Safety factor
(13.1d)
30 mg/L
© 2009 by Taylor & Francis Group, LLC
Design Basis 609
Historically, Kadlec and Knight (1996) determined mul-
tipliers corresponding to the 100th percentile of monthly
means from the NADB. These were relative to the long-term
mean value for a particular wetland, and therefore seasonal
variations, whether temperature driven or not, were included
in the multiplier. In this book, an annual trend is computed as
the basis of the multiplier, thus excluding seasonal phenom-
ena from this measure of random scatter. Further, wetland
data may sometimes contain sufcient intrasystem variabil-
ity to place the 100th percentile above the median inlet con-
centration. For wetlands with very low inlet concentrations,
it may not be possible to design a wetland to totally avoid
the possibility of monthly exceedances. Consequently, multi-
pliers for various frequencies of occurrence are tabulated in
Part I, and are used in Part II for design purposes.
Curiously, the use of a COR has been ascribed as an attri-
bute of areal rst-order models but not an attribute of volu-
metric rst-order models (Water Environment Federation,
2001; Crites et al., 2006). Incredibly, the absence of a COR
in a volumetric model has been described as an advantage,
because then it has “no limiting impact on the mathematical

results of design models” (Crites et al., 2006). Conversely,
Crites et al. (2006) portray the use of a COR to adjust the
design goal as a disadvantage of the areal model, because it
“may result in excessive wetland sizes to achieve low con-
centrations.” Of course, the use of a COR has nothing to do
with how one determines the trend values, as is apparent
from its use with the volumetric model (Crites and Tchob-
anoglous, 1998). And there is no doubt that the larger wetland
sizes needed to contain excursions are required to achieve an
acceptable level of regulatory compliance.
OTHER DESIGN PARAMETERS
Some of the specications of regulatory permits or licenses are
outside the commonly encountered groups of rational design
parameters, which include BOD, TSS, nitrogen compounds,
phosphorus pathogens, metals, organics, and temperature.
These do not have loading charts, nor do they have k-values.
pH
There is often a specied range of allowable pH for dis-
charges to surface waters, typically 6.0–9.0. Most treatment
wetland applications are not likely to exceed such ranges, as
detailed in Chapter 5. Exceptions are the acid mine drainage
wetlands, in which the design goal includes raising the pH of
the incoming water. Treatment wetlands are not particularly
effective at neutralizing strong acid, whereas they are quite
good at creating circumneutral pH for more benign inuents,
such as food wastewaters. Other exceptions include industrial
processes, for which a neutralization step is included as part
of pretreatment.
Tox
icity

The U.S. Clean Water Act prohibits the discharge of toxic sub-
stances to waters of the United States. For this reason, whole
efuent toxicity (WET) monitoring is included in the NPDES
permits for many municipal treatment plants. WET tests
are the standardized procedure to detect levels of acute and
chronic toxicity in municipal and industrial efuents. Individ-
ual pollutants that contribute to toxicity may be monitored in
some instances, such as for nitrite and ammonia, but biomoni-
toring is required to assess the overall net potential for acute
and chronic toxicity to receiving water biota or surrogates.
The short-term chronic toxicity tests that were eventu-
ally developed by the U.S. EPA are a relatively inexpensive
method of assessing WET (U.S. EPA, 1994). Freshwater
chronic toxicity tests utilize three organisms:
The water ea (Ceriodaphnia dubia)
The fathead minnow (Pimephales promelus)
The green alga (Selenastrum capricornutum)
Test methods require a seven-day (96 hours for the algal
test) testing period, with renewals of the testing solution
using the tested efuent three times during that seven-day
period. For the water ea, the test encompasses three repro-
ductive cycles, involving three broods of young (neonates).
The adult water eas are typically fed three times per day
during the testing period. Acute toxicity is assessed via mor-
tality. Chronic toxicity is assessed via the number of young
produced per female. Tests may be conducted using just con-
trol water and 100% efuent, or one or more diluted efuent
concentrations.
Fathead minnow testing utilizes larval sh, tested over
a seven-day growth period with test water renewals. Acute

toxicity is assessed through observed mortality of the sh.
Chronic toxicity is assessed by measurement of the nal dry
weight of the tiny sh at the end of the test period. Laboratory
controls are utilized and one or more efuent concentrations
are tested to assess the lethal and sublethal effects of the efu-
ent on the sh. Some states have specic protocols that vary
from the federal WET guidelines, because of the use of other
vertebrate and invertebrate species. Two alternate species are
rainbow trout (Salmo gairdneri) and a different water ea,
Daphnia magna. Fathead minnows may be replaced with
indigenous sh species such as the bluegill (Lepomis macro-
chirus) or the bannern shiner (Cyprinella leedsi).
TOXICITY REDUCTION IN FWS WETLANDS
A number of studies have shown that constructed wetlands
can be effective in reducing toxicity (Knight et al., 1997;
U.S. EPA, 1999; Wetland Solutions, Inc., 2003). U.S. EPA-
sponsored synoptic studies at six constructed treatment wet-
lands in the United States (McAllister, 1992; McAllister,
1993a; McAllister, 1993b), including standardized toxicity
tests. The Collins, Mississippi, wetland had signicant acute
and chronic toxicity at the wetland inow, probably due
to high unionized ammonia concentrations, but acute and
chronic toxicity were absent in the wetland outow. West
Jackson County, Mississippi, had slight acute and chronic
toxicity to the water ea at the wetland inow but no toxicity
•
•
•
© 2009 by Taylor & Francis Group, LLC
610 Treatment Wetlands

at the outow. Survival and reproduction of the water eas
were increased by the Lakeland, Florida, constructed wet-
land. No chronic toxicity was present at either the inow or
the outow of the Orlando Easterly, Florida, wetland.
No toxicity was detected at the inow of the Incline Vil-
lage, Nevada, wetland, but the water in the terminal cell of
this near-total evaporative wetland was acutely and chroni-
cally toxic to the water ea. A similar reduction in repro-
duction was observed with passage through the Show Low,
Nevada, constructed wetland, another near-zero discharge
(evaporative) system. It is likely that toxicity was increased
in these wetlands (Incline Village and Show Low) as a result
of evaporative concentration of salts and other constituents.
Toxicity test results are reported for four treatment wet-
lands in the NADB v. 2.0 (U.S. EPA, 1999): Santa Rosa,
California; Minot, North Dakota; American Crystal Sugar
Company in Hillsboro, North Dakota; and Arcata, California
(Wetland Solutions, Inc., 2003). An average of 9% mortality
in rainbow trout was observed in samples collected at three
points in the Santa Rosa wetland system, but there were no
inuent data. At the Minot constructed wetland outlet, no
signicant mortality was recorded for water eas and fat-
head minnows for test dilutions up to 87.5% efuent. The
Hillsboro sugar processing wastewater treatment wetland
produced efuents slightly toxic to fathead minnows (15%
mortality) and to water eas (16% mortality). No wetland
inuent data were available from this site. At the Arcata wet-
land, data indicated little change in survival between the wet-
land inlet and outlet (94 to 96%).
Whole water toxicity was measured with water eas and

fathead minnows in quarterly tests of operation in six wet-
lands treating paper mill wastewaters (Knight et al., 1994).
No acute toxicity to either organism was found. Inuent sam-
ples were chronically toxic to the water ea in all cases, but
reduced by the wetlands in 10 of 11 comparisons. No chronic
toxicity to fathead minnows was observed in any of the wet-
land outow samples.
A number of petroleum facilities have reported the reduc-
tion of toxicity by treatment wetlands (API, 1999). These
include reneries at Richmond, California (Duda, 1992), and
St. Charles, Louisiana (Hawkins et al., 1995), as well as oil
transfer terminals and oil sand processing facilities (Bishay
et al., 1995).
The Tres Rios constructed wetlands at Phoenix, Arizona,
have been extensively studied for toxicity reduction (Wetland
Solutions, Inc., 2003). Acute and chronic toxicity in both
incoming pretreated water and the wetland outows was nearly
nonexistent or difcult to detect. It was concluded that use of
full-scale constructed wetlands for nal efuent polishing of
the 91st Avenue WWTP nal efuent reduced the frequency
and magnitude of WET. The wetlands typically reduced low
levels of toxicity when they occurred in the WWTP efuent,
and did not add to toxicity. However, high levels of toxicity
passed through the wetlands with minor alteration, probably
due to the very short hydraulic residence times in this system.
In summary, there exist studies indicating that treatment
wetlands generally reduce levels of acute and chronic toxic-
ity if present in the inow. However, the limited amount of
wetland data, variation in inuent toxicity, and the variability
among test protocols preclude forecast modeling of wetland

effects on WET. It is not currently possible to calculate the
required size of a treatment wetland for a specied level of
toxicity reduction.
16.5 CLIMATE AND THE WATER BUDGET
Performance-based design rests upon the water budget for
the system. The amount of water to be treated is only one
part of the hydrologic setting for the project, with precipita-
tion, evapotranspiration, and seepage being the other prin-
cipal components. These variables are in turn components
of the site climate. Another critical component of climate is
temperature, which is a driving force for the rates of some
wetland processes.
CLIMATE
Climate cannot be controlled when selecting a specic site
for a constructed or natural wetland project. However, cli-
mate is important during project planning and alternative
selection because it affects the type and size of wetland that
will be used. Latitude is the most critical determinant of
climate because it determines seasonal temperature ranges.
Other climatic factors that are important during project plan-
ning include rainfall, insolation, wind direction and veloci-
ties, and evapotranspiration. In the United States, climatic
information can be obtained from the National Aeronautic
and Atmospheric Administration (NOAA) at the following
Web address:
National Climatic Data Center
U.S. Department of Commerce
/>html
For the entire world, a summary overview of annual patterns
of temperature and precipitation may be obtained (free of

charge) at:
Climates of the World
National Oceanic and Atmospheric Administration
National Climatic Data Center
Asheville, North Carolina

climatesoftheworld.pdf
Weather and climate data for specic sites across the planet
can also be obtained through Web sites such as Weather
Underground ( />Wetland hydrology is affected not only by the volume
of applied wastewater but also by the net balance between
rainfall and evapotranspiration. Wetland designers in areas
with high net precipitation must take this incremental added
water volume into account when deciding on wetland vol-
ume, ow control structures, and freeboard of berms. In
© 2009 by Taylor & Francis Group, LLC
Design Basis 611
areas of high evapotranspiration, wetland design calculations
need to consider water loss to the atmosphere. In some cases,
the decision may be made to allow the wetland to leak, and
seepage can become an important feature of the water budget
and have a strong effect on treatment calculations. Chapter 2
presents examples of water balances for FWS wetland treat-
ment systems.
RAINFALL
Figure 2.2 provides a map of the annual average precipitation
for the United States. More detailed information concerning
the average monthly net precipitation values for the project
location should be used to estimate the typical monthly water
balance for a continuous ow wetland under extreme seasonal

conditions. Historical values of monthly mean precipitation
may be found, for 1–10 divisions per state, in the following
free Internet publication:
Climatology of the United States No. 85
Division Normals and Standard Deviations of Tem-
perature, Precipitation, and Heating and Cooling
Degree Days
National Oceanic and Atmospheric Administration
National Climatic Data Center
Asheville, North Carolina

climatenormals/climatenormals.pl
In cold climates, snowfall occurs in the winter months,
and accumulates on the wetland during the early winter.
The spring melting period causes this stored water to enter
the wetland, often as a fairly brief pulse lasting a few days
only. Snowfall and snow depth records are available from
the National Climatic Data Center (NCDC) for the United
States.
As discussed in Chapter 14, rainfall is the primary driv-
ing force for stormwater wetlands, and yet more detailed
data is therefore desirable. Site-specic information on the
intensity and duration of rains, and the interevent spacing, is
needed to perform design calculations.
EVAPOTRANSPIRATION
The mechanisms by which a wetland loses water to the atmo-
sphere have been detailed in Chapter 2. Methods of estima-
tion and data sources have also been detailed in that chapter.
As an example of the kind of information that wetland
designers have available, Figure 16.7 shows the various cli-

mate zones of the state of California. Table 16.10 lists the
monthly ET for each zone. Note that arid regions have an
annual ET loss of 1,819 mm/yr, whereas wet, coastal regions
have only 838 mm/yr. The peak ET rates are in midsummer,
and range up to 9 mm/d for the warm arid zones. For states
that do not have irrigation information services, the designer
can obtain pan evaporation data for locations in the United
States from the NCDC.
The transpiration component of ET moves downward into
the root zone of a FWS wetland, and thus represents a with-
drawal of water and pollutants from the surface water body.
In that regard, this transpiration ux resembles inltration,
and can be treated as such in FWS design. In SSF systems,
water passes through the rhizosphere, and consequently both
evaporation and transpiration losses are withdrawn from the
owing water.
As discussed in Chapter 2, evaporation affects both the
ow rate (detention time) and the concentrations of pollut-
ants (evaporative concentration). Therefore, it is inaccurate to
merely average the inows and outows to account for effects,
because that misses the evaporative concentration effect. The
ow is slowed, and pollutants become more concentrated.
The reverse is true for rainfall: velocities are increased, and
pollutants are diluted. The mass balances used in this book
for design correctly account for both velocity and dilution
effects.
Extreme examples of systems in which ET is a domi-
nant factor include lightly loaded wetlands in arid climates.
The Incline Village, Nevada, system evaporates nearly all the
incoming water, and ET was the key factor in project design

(Kadlec et al., 1990). The Saginaw, Michigan, project oper-
ates only in the summer, during which there is an excess
of ET over rainfall even for this moderately wet climate
(Kadlec, 2003c). In another example, on the order of half the
water entering the Roblin, Manitoba, FWS wetland system
is, by design, lost to net evapotranspiration (PFRA and Town
of Roblin, 2003).
SEEPAGE
Although many treatment wetlands are contained, either with
clay or synthetic liners, there are a number of situations in
which containment is not needed or not feasible. Any water
that leaks vertically downward, or horizontally outward
under and through containment berms, represents a water
loss that slows ow and contributes to pollutant loss from the
wetland water body. Inltrated water moves into and with the
regional groundwater, and may reemerge at nearby locations
as groundwater discharge. During passage through the soils
under and around the wetland, these seepage waters may
receive some additional treatment, by microbial degradation
or adsorption, for example. However, not all potential con-
taminants are subject to soil and aquifer treatment.
The design procedures described herein pertain to the
wetland water body and the planned outows that occur as
discharges from structures. This is not to say that inltra-
tion ows can be ignored; instead, the projects with signif-
icant seepage are to be viewed as dual-discharge projects.
The seepage should meet groundwater requirements, and the
structure outows should meet the appropriate goals, which
may be for surface water recipients or inltration.
Examples of systems in which ET is a nontrivial fac-

tor include the stormwater treatment areas of South Florida
(Nungesser and Chimney, 2006), which lose and/or gain up
to about 20% of the ow from groundwater. The New River
© 2009 by Taylor & Francis Group, LLC
612 Treatment Wetlands
wetlands in Southern California inltrated 47% of the inow
at the Brawley site, and 39% at the Imperial site, over a four-
year period of operation (TTI and WMS, 2006).
The wetlands at the Sacramento, California, site (Nolte
and Associates, 1998) inltrated 12.4 ± 4.0% of the inow
for four cells in 1998. The inltration rates for the site soils
were determined prior to the project to be less than 0.15 cm/h
for two loams, and 0.15–0.50 cm/h for two clays. For the four
wetland cells, in their fth year of operation, the rates were
0.028 ± 0.008 cm/h, considerably less than those of the par-
ent soils.
Three of the four wetlands at the Tres Rios, Arizona,
site had minimal inltration, but the fourth was built on very
coarse river gravel bed materials (Wass, Gerke and Associ-
ates, 2001), and the majority of the water leaked vertically
downward. Of interest in this project are the trends in the
inltration rates over the rst few years of the project history
(Figure 16.8). Wetlands H1 and H2 were built on former agri-
cultural soils, and possessed modest inltration losses of 1.5
and 1.9 cm/d, respectively. Wetland C2 was built on coarse
river gravel with a topping of agricultural soil, and the agri-
cultural soils sealed the wetland to about the same leakage
rate as H1 and H2: 1.7 cm/d. Wetland C1 was built on coarse
river gravel, and leaked at an average rate of 19.7 cm/d. There
was a decrease in the inltration in the cobble wetland C1,

from 25.3 to 14.0 cm/d based on the trend line. However, the
other three cells maintained a relatively constant inltration
rate. Based on these limited results, only minimal self-sealing
should be anticipated in treatment wetlands. If inltration is
not an allowable operating condition, the use of clay or syn-
thetic liners should be considered.
TEMPERATURE
As shown in the design equations provided in Chapter 9,
water temperature is important when sizing wetland treat-
ment systems for removal of all forms of nitrogen. Figure 16.9
provides a map of mean annual temperatures in the United
States. However, more detail is desirable, and monthly mean
temperature is a good estimator of monthly mean water tem-
perature (see Chapter 4), which allows the interpretation of
thermal effects on some microbial components of treatment.
1
1
1
2
2
2
8
4
2
1
4
6
9
16
16

16
16
6
6
12
17
17
18
18
18
15
14
14
14
13
5
10
15
12
12
11
14
14
14
14
1
1
1
1
1

3
13
10
7
3
3
3
FIGURE 16.7 Reference evapotranspiration for the state of California. See Table 16.10 for the annual pattern for each zone. (Adapted from
California Irrigation Management Information System; />© 2009 by Taylor & Francis Group, LLC
Design Basis 613
Historical values of monthly mean air temperatures may be
found, for 1–10 divisions per state, in Climatology of the
United States No. 85 (previously discussed in the subsection
titled “Rainfall” in this chapter).
For planning purposes, the long-term average tempera-
ture during the coldest month of the year has been found to
be a good estimator of the critical low water temperature that
will be experienced in a wetland treatment system. For areas
TABLE 16.10
Monthly Reference Evapotranspiration by Zone for the State of California
Zone
Jan
(mm/mo)
Feb
(mm/mo)
Mar
(mm/mo)
Apr
(mm/mo)
May

(mm/mo)
Jun
(mm/mo)
Jul
(mm/mo)
Aug
(mm/mo)
Sep
(mm/mo)
Oct
(mm/mo)
Nov
(mm/mo)
Dec
(mm/mo)
Total
(mm/yr)
1 24 36 63 84 102 114 118 102 84 63 30 16 836
2 31 43 79 99 118 130 126 118 99 71 46 31 991
3 47 57 94 122 134 145 142 134 107 87 61 47 1,176
4 47 57 87 114 134 145 150 142 114 87 61 47 1,184
5 24 43 71 107 142 160 165 150 114 79 38 24 1,115
6 47 57 87 122 142 160 165 157 122 94 61 47 1,262
7 16 36 63 99 134 160 189 165 122 71 30 16 1,101
8 31 43 87 122 157 175 189 165 130 87 46 24 1,255
9 55 71 102 130 150 168 189 173 145 102 69 47 1,401
10 24 43 79 114 150 183 205 181 130 79 38 24 1,248
11 39 57 79 114 150 183 205 189 145 94 53 39 1,347
12 31 50 87 130 173 198 205 181 137 94 46 24 1,356
13 31 50 79 122 165 198 228 197 145 94 46 24 1,379

14 39 57 94 130 173 198 220 197 145 102 53 39 1,449
15 31 57 94 145 189 206 220 197 145 102 53 31 1,472
16 39 64 102 221 197 221 236 213 160 110 61 39 1,664
17 47 71 118 152 205 229 252 220 168 110 69 47 1,688
18 63 85 134 175 220 244 244 220 175 126 76 55 1,819
Note: See Figure
16.7 for the locations of the zones.
Source: California Irrigation Management Information System; />Cell Soil
Infiltration Rate
(cm/d)
Percent of
Inflow
C1 Cobble 19.7 74
C2 Silty loam over cobble 1.7 12
H1 Silty loam 1.5 12
H2 Silty loam 1.9 14
FIGURE 16.8 Inltration rates at the Tres Rios FWS wetlands.
y = –0.47x + 30.9
R
2
= 0.2007
y = –0.023x + 2.43
R
2
= 0.034
0.1
1
10
100
1,000

12 18 24 30 36
Months
Infiltration (cm/d)
C1
C2
H1
H2
Linear (C1)
Linear (H2)
© 2009 by Taylor & Francis Group, LLC
614 Treatment Wetlands
in which the average monthly temperature is less than zero,
it can be assumed that the minimum wetland operational
temperature will be slightly above zero under an ice cover
(see Chapter 4 for more information).
16.6 SELECTION OF WETLAND TYPE
The type of wetland selected varies with a number of fac-
tors, and the best choice may require alternatives evaluation.
There are some over-arching concepts that can assist in the
selection.
“NATURAL” VERSUS “ENGINEERED” SYSTEMS
Treatment wetlands can be designed to be “au natural” or
complex. Examples of designs along a full spectrum from
almost zero fossil-fuel input to highly engineered and man-
aged systems can be found around the world. In this book
we term these design and operation strategies as Type A and
Type B wetlands. Type A systems are at the most natural,
low-energy end of the scale. They have the minimum number
of structures, and may often be gravity fed. Vegetation may
be allowed to proceed by natural recruitment (Figure 16.10).

Type B wetlands are higher energy and more highly engineered
systems. Pumps are likely to be present, for recycle as well as
other water transfers, and piping can be complex. Special media
50°
50
50
50
40
40
40
50
50
50
60
60
55
55
60
60
180°
70°
60°
160°W
10
15
30
35
40
40
40

45
30
25
20
15
10
N
N
140°
70
70
65
65
45
45
55
55
55
50
50
50
50
55
60
55
50
50
50
45
45

40
40
55
60
50
45
45
45
45
50
50
40
40
45
45
65
70
70
75
75
65
70
70
70
75.6
77.2
74.9
74.6
53.4
75.9

69.7
75.0
22°
20°
160°
105° 95° 85°
Contour Interval: 5°F
Based on Normal Period 1961–1990
75° 65°115°125° W
155°
55.5
74.0
72.4
75.6
N
W
45
40
60
60
70
65
45
45
45
45
45
40 40 40
40°
30°

20°
25
20
FIGURE 16.9 Mean annual air temperatures in the United States (°C  (°F – 32)/1.8). (From National Climatic Data Center (2007)
at />may be involved, and the system may be mechanically aerated
in some cases. Vegetation is likely to be highly specic, such
as a monoculture of planted Phragmites (Figure 16.11). There
are good reasons to slide on this scale of design and manage-
ment complexity, depending on project goals.
Type A wetlands are the preferred solution when:
There is sufcient treatment potential with a Type
A approach.
Relatively inexpensive land area is abundant and
land is not a limiting factor.
There is insignicant potential for unauthorized
human interaction with the wastewater in the
wetland.
Habitat and aesthetic benets outweigh the poten-
tial for nuisance conditions related to mosquitoes
or other wildlife.
Type A wetlands are typically constructed FWS marshes,
relatively low-tech HSSF wetlands, and natural treatment
wetlands. They are always less costly to operate than Type
B treatment wetlands, and may be less costly to construct,
depending on the local situation.
The Type B treatment wetland design and management
strategy is preferred when one or more of the factors listed
previously are limiting the design. Specic examples of Type
B treatment wetlands include:
•

•
•
•
© 2009 by Taylor & Francis Group, LLC
Design Basis 615
HSSF wetlands augmented with saturable media
intended for phosphorus or metals removal
Addition of aerators in constructed HSSF wetlands
for improvement of nitrication
Addition of VF components to HSSF constructed
wetlands to improve nitrication (potentially with
ow recirculation)
FWS or HSSF wetlands and aquatic plant treat-
ment systems with plant harvesting
High solids treatment wetlands with frequently
cleaned forebays, or frequent refurbishment and
restarting, as for sludge reed beds
Type B wetlands are more expensive to operate, and may be
more expensive to construct, compared to Type A wetlands,
•
•
•
•
•
and consequently there needs to be a clearly demonstrable
need for the added complexity and energy consumption.
FWS OR HSSF?
The two most prevalent types of treatment wetland, espe-
cially during the early history of the technology, are FWS
and HSSF wetlands. The mind-set in North America was

often to emulate the natural habitat wetlands (duck marshes)
for ecological reasons. In Europe, there prevailed, and still
does, the concept that HSSF ow systems are more efcient,
and will require a smaller footprint for an equivalent level of
treatment performance.
An analysis of performance data suggests that HSSF
wetlands are not more efcient than FWS systems. Earlier
FIGURE 16.10 Type A constructed wetlands may be built by berming off a plot of ground and allowing natural regrowth to vegetate the
system. This wetland at Onaway, Michigan, is used for ultra-polishing of a lagoon efuent that has had post-sand ltration and phosphorus
removal by chemical precipitation.
FIGURE 16.11 This small-scale vertical ow wetland system, near Aarhus, Denmark, consists of a below-ground sedimentation tank, two
vertical ow Phragmites beds, three phosphorus removal beds, four pumps in pump wells, and complex interconnected piping. It qualies
as a Type B constructed wetland system. (Photo courtesy H. Brix.)
© 2009 by Taylor & Francis Group, LLC
616 Treatment Wetlands
analyses, such as that in Kadlec and Knight (1996), suffer
from two difculties: the North American experience of 15
years ago was very limited in numbers of systems and the
length of their operations, and the calibration ranges of FWS
and HSSF systems were quite different. To compare types,
FWS calibrations had to be extrapolated to higher load-
ings, and HSSF, to lower loadings. As is now known, such
extrapolations are not warranted, and can be greatly in error
(Kadlec, 2000).
At the time of this writing, there are much more exten-
sive and numerous data sets available, so that comparisons
need not rely on extrapolations. The foundation for accom-
plishing this comparison has been set down in Part I for the
various common pollutants. There are two obvious avenues
of comparison: loading charts and k-values, which were both

used in Kadlec and Knight (1996). However, both can now
be updated based on much more information. In the case of
loading charts, the (desirable) overlap in loading ranges has
increased. In the case of k-values, the underlying hydraulics
can now be better accounted.
Superpositions of the loading data for HSSF, FWS, and
VF
wetlands are shown in Figure 16.12, 16.13, and 16.14 for
BOD, ammonia, and fecal coliforms, respectively. Other pol-
lutants behave similarly. It is clear from these scatter plots
that the respective data clouds overlap virtually entirely for
HSSF and FWS wetlands. There is no difference when they
are compared on this areal basis. At the level of this broad
perspective, this means that a square meter of FWS will
do the same job as a square meter of HSSF. However, VF
wetland data are set apart on these graphs, with VF systems
typically providing lower outlet concentrations at equivalent
loadings. This has, of course, been well documented in the
literature for reduction of ammonia.
The loading chart method does not separately account
for changes in inlet concentrations and hydraulic loading
(or detention time); but rst-order areal rate coefcients do.
Therefore, it is useful to examine the various calibrations that
ha
ve been set forth in Part I. Table 16.11 provides the cen-
tral percentiles of the distributions of k-values for the various
common contaminant targets. For BOD, FWS rate coef-
cients are higher, except for the tertiary application zone.
FWS is also better for NH
4

-N, TKN, and TN, but slightly
worse for NO
x
-N. The horizontal subsurface systems appear
slightly better for fecal coliform reduction. Overall, the mes-
sage is that there is not a large difference between k-values
for FWS and HSSF wetlands, and the advantage may be to
either type, depending on which contaminant is under con-
sideration. From the standpoint of treatment performance per
unit area, as embodied in the k-values, there is little discrim-
ination. This is supportive of the same conclusion reached
from the loading chart approach.
VF wetlands appear to be slightly better for removal of
BOD and ammonia, but because the nitrate reduction is not
much different, the reduction of total nitrogen is not superior.
This is also supportive of the same conclusion reached from
the loading chart approach.
FIGURE 16.12 Overlay of FWS, HSSF, and VF BOD loadings and efuent concentrations.



    





 



!
© 2009 by Taylor & Francis Group, LLC
Design Basis 617
As discussed in Chapter 23, it is not just economics alone
that favor the choice of HSSF wetlands. Factors other than
reduction performance are also important in the selection
process. The principal reasons for selecting the HSSF option
over the FWS option are concerns about:
Human health via contact with untreated waste-water
Mosquito control
Odor control
Minimizing wildlife interactions within the wetland
•
•
•
•
In the United States, human health regulations do not allow
the potential for contact between humans and raw wastewa-
ter. The perceived risk of contact with pathogens is too great.
This in turn means that no treatment system can allow such
contact. The raw sewage facilities at mechanical plants and
lagoons are strongly fenced, with required setbacks from
human activities. If wetlands are to be used for raw waste-
water (or primary efuent), and are unprotected from human
access, then the HSSF option is necessary to ensure that there
will be no human contact with the water being treated.
1.E – 01
1.E + 00
1.E + 01
1.E + 02

1.E + 03
1.E + 04
1.E + 05
1.E + 06
1.E + 07
1.E – 02 1.E – 01 1.E + 00 1.E + 01 1.E + 02 1.E + 03 1.E + 04 1.E + 05 1.E + 06 1.E + 07
1.E + 08
Inlet Fecal Coliform (#/100 mL)
Outlet Fecal Coliform (#/100 mL)
FWS
HSSF
VF
FIGURE 16.14 Overlay of FWS, HSSF, and VF inuent–efuent fecal coliform concentrations.
FIGURE 16.13 Overlay of FWS, HSSF, and VF ammonia loadings and efuent concentrations.
HSSF
FWS
VF
NH
4
-N Concentration Out (mg/L)
0.01
0.1
1
10
100
1,000
NH
4
-N Load In (g/m
2

· d)
0.0001 0.001 0.01 0.1 1 10 100 1,000
© 2009 by Taylor & Francis Group, LLC