brown.donald@epamail.epa.gov
(513) 569-7630
FAX (513) 569-7185
Don Brown
Environmental Engineer
Water Quality Management Branch
National Risk Management Research Laboratory
Office of Research & Development
Mail Address: USEPA (MS-663), Cincinnati. OH 45268
Courier Address: 26 W. Martin L King Dr., 45219
-------
-------
INVENTORY OK CONSTRUCTED
WETLANDS IN THE UNITED STATES
by
Donald S. Brown (USEPA, Cincinnati, OH)
and
Sherwood C. Reed (EEC, Norwich, VT)
Published in f
Water Science & Technology
Vol. 29, No. 4, pp. 309-318, 1994
/
The attached copy is the same as the published paper.
The quality of the graphics in this copy are a little
better than in the journal.
-------
-------
Pergamon //V^ //? n, // ~ Wat.ScL Tech. Vol.29,No.4,pp.309-318,1994.
Copyright © 1994IAWQ
; Printed in Great Britain. All rights reserved.
I 0273-1223/94 $7-00 + 0-00
INVENTORY OF CONSTRUCTED
WETLANDS IN THE UNITED STATES
Donald S. Brown* and Sherwood C. Reed**
* Risk Reduction Engineering Laboratory, U.S. Environmental Protection Aeency,
Cincinnati, OH 45268 USA
** Environmental Engineering Consultants, RR 1, Box 572, Norwich, VT 05055 USA
ABSTRACT
During 1990 and 1991 the U.S. Environmental Protection Agency (EPA) sponsored an effort to identify
existing and planned constructed wetlands in the U.S., and to collect readily available information from
operating systems. In addition to inquiries by telephone and mail, the effort included site visits to over 20
operating subsurface flow constructed wetlands. The inventory documented the presence of over 150
constructed wetland systems for wastewater treatment, including both free water surface (FWS) and
subsurface flow (SF) systems. The majority of the systems identified were SF systems for treating municipal
wastewater. FWS systems were separated into three groups based on the design level of effluent water
quality. SF systems were separated into three groups based on the basic design approach. The inventory
indicated that neither between nor within these groups was there consensus regarding basic hydraulic and
engineering design criteria, system configuration, or any other aspect, such as type of vegetation, size and
type of media, or pretreatmenL Information on location, type of system, design approach, hydraulic and
organic loading rates, costs, and other aspects is presented. Information gathered and "lessons learned" from
the site visits are presented. Insufficient oxygen for nitrification appears to be a problem for both FWS and
SF systems. Insufficient hydraulic design appears to be a problem for SF systems.
KEYWORDS
Constructed wetlands, wastewater treatment, free water surface, subsurface flow.
INTRODUCTION
The U.S. Environmental Protection Agency's (EPA) Risk Reduction Engineering Laboratory (RREL) in
Cincinnati, Ohio currently has several projects related to the use of constructed wetlands for wastewater
treatment Most of the constructed wetlands being built in the U.S. are either of the free water surface type,
or are subsurface flow systems using large gravel or rock media. However, these systems are being built
without the benefit of clear design criteria. Because the lack of basic performance data appears to be the
major obstacle to better design, RREL is undertaking several projects to collect more data. These projects
are focused on subsurface flow wetlands for municipal wastewater treatment. The first project undertaken
was an inventory of constructed wetlands in the United States. Other projects include: monitoring of
operating full-scale systems; pilot scale research work on subsurface flow wetland cells; and evaluating
onsite (individual home) systems.
The inventory is the subject of this paper. The inventory was conducted in two phases. The objective of the
first phase was to locate and collect general information from all constructed wetlands. The only systems
which were specifically excluded from the search were those serving individual homes or mine drainage
309
-------
310
D. S. BROWN and S. C. REED
sites. Data were obtained primarily by mail and telephone during the period January through March 1990.
The first phase was completed in April 1990. The objective of the second phase was to collect existing
design, water quality, operation, and performance data by making site visits to twenty of the subsurface flow
systems. The second phase was completed in October 1991. The inventory had to rely on information that
was readily available, and no new information was generated. Therefore, the information gathered is
incomplete, and the figures that follow may not contain information for all of the wetlands. (For example,
Fig. 1 includes only 90 entries because the starting dates of the other systems were not available.)
. Number of Systems
<'80 <'86 '86 '87 '88 '89 '90 '91 >'91
Stairting Date
Fig. 1. Starling date (actual or projected) constructed wetlands in the United States.
For this inventory, constructed wetlands were defined as wetlands which were specifically designed and
built for wastewater treatment and were located at a site where natural wetlands did not exist at the time of
construction. Constructed wetlands were categorized into two types based on the flow of water in the
system: free water surface (FWS) and subsurface flow (SF). FWS wetlands contain emergent aquatic
vegetation in a shallow bed or channel. The water is exposed to the atmosphere as it flows through the
wetland. SF wetlands contain a permeable medium which supports the root system of the same types of
emergent vegetation. The water level in the bed is maintained below the top of the media so that all flow is
designed to be subsurface. The term subsurface flow is a general term which includes all such constructed
wetlands, including those systems that have also been called "rock-reed filters", "gravel bed wetlands",
"microbial rock-reed filters", and "vegetative submerged bed systems".
INVENTORY RESULTS
"Demographics" of Constructed Wetlands
A total of 143 communities in the U.S. were identified as using, building, designing, or planning constructed
wetlands. A total of 154 wetlands were actually identified because a few communities had more than one
wetland system, or used some combination of both FWS and SF units. Of this total, over 60 operating
systems were found. The use of constructed wetlands for wastewater treatment is a relatively new practice in
the U.S. when compared with the European countries. Constructed wetlands have really been used only
since 1986 (Fig. 1), although a few FWS systems are older. There are almost equal numbers of operating
FWS and SF systems, but SF systems are projected to outnumber FWS systems in the future. The majority
of systems are located in the Mississippi River basin, but they are found or planned throughout the U.S. (Fig.
2).
The majority of systems (70% of FWS and 90% of SF) treat less than 3800 m3/day (Fig. 3). Smaller systems
tend to be SF, which range in size from 5 to 11,400 m3/d (mean = 1150 m3/d, median = 210 m3/d). Larger
systems tend to be FWS, which range from 200 to 76,000 m3/d (mean = 7350 m3/d, median = 1770 m3/d).
-------
312
D. S. BROWN and S. C. REED
Region designs generally used in EPA's Region VI (South-
IV: Central U.S.), derived from work done by Wolverton (1983);
characterized by large rock for media (80 ram or greater
diameter), large length to width ratios (10 or greater), and
higher hydraulic loading rates (greater than 150 L/m2/d)
Othenindependent designs.
L/m2/d
120
100
80
60
40
20
AWT &
Retention
...iiillll
JODjJSS.
& NH3
BOD
"""""&
TSS
HI
Fig. 4. Examples of FWS hydraulic surface area loading rates of constructed wetlands in the United States.
kg BOD/ha/d
60
50
40
30
20
10
BOD & TSS
BOD, TSS, & NH3
AWT & Retention
II--.il •••
il
Fig. 5. Examples of FWS organic surface area loading rates of constructed wetlands in the United States.
-------
Inventory of constructed wetlands
311
Number of Systems
O<5 Hi! 5-10
> 15
Rg. 2. Distribution of construiaed wetlands systems in the United States.
Number of Systems (%)
1
10 102 103 104 105
Design Flow (ms/day)
Fig. 3. Size based on design flow of constructed wetlands in the United States.
Design Characteristics
The primary reason for conducting the inventory was to identify the design criteria used at existing systems.
One of the principal conclusions from the inventory, however, was that there is no consensus regarding the
design of constructed wetlands in the U.S. Design approaches appeared to range from "trial and error" to
semi-empirical to semi-rational using design models based on very limited data. Figures 4 through 8 show
the design information that was available for several operating systems. Each bar in these Figures represents
one wetland for which design criteria were available. (For example, as shown in Figure 4 there were 25 FWS
wetlands for which hydraulic loading criteria were available.)
The data in Figures 4 through 8 is organized into groups for clearer presentation. For ease of viewing the
bars are placed in order from smallest to largest for each group shown. The FWS systems are divided into
three groups according to effluent goals: BOD and TSS removal; BOD, TSS, and NH3 removal; and
advanced wastewater treatment (AWT) or total retention (i.e. no discharge). The SF systems are divided into
three groups according to the basis of design as follows:
TVA:
*ST 29:4-V
guidance provided by the Tennessee Valley Authority (Watson,
et al, 1990), derived from European experience; characterized
by the use of small rock or gravel for media (less than 40 mm
diameter), variable length to width ratios, and lower
hydraulic loading rates (less than 200 L/m2/d)
-------
314 D. S. BROWN and S. CREED
shown there is a wide range of design criteria used. For example, Region VI SF systems are typically loaded
at two to three times the rate of other SF systems or FWS systems, while TV A SF systems are loaded at rates
similar to FWS systems designed for BOD, TSS, and NH3 removal.
There is also no consensus on system configuration, or other details such as depth of water or media, type or
size of media, slope of bed, length to width ratio, level of pretreatment, and type of vegetation. In FWS
systems the water depth ranges from 5 to 90 cm; 30 to 45 cm is typical. In SF systems media depth ranges
from 30 to 76 cm, and sand, peat, gravel and rock have all been used. Gravel and rock size ranges from 6 to
130 mm; 13 to 80 mm is typical. Bottom slopes range from 0 to 1%.
There is disagreement in particular over the desired length to width ratio (L:W) for subsurface flow
wetlands Region VI designs have generally had a high L:W, based on the argument that a high L:W is
needed to ensure plug flow and high levels of performance. TVA and "Other" designs usually consider the
hydraulic loading rate on the vertical cross sectional area of the bed, and this tends to produce wetlands with
a lower L:W. However, as shown in Figure 9 the trend for subsurface flow systems is toward lower L:W
ratios. Length to width ratios for FWS wetlands have remained relatively constant.
.. L:W Ratio
20 " :
15-
• -87 '88 '89 '90 '91
Fig. 9. Trend in L:W for constructed SF wetlands in the United States.
The level of pretreatment varies widely. Facultative lagoons are the most common form of pretreatment and
are used at 41% of the FWS and 44% of the SF systems. Septic tanks are used at 24% of the SF systems.
Aerated lagoons, secondary treatment, and advanced treatment have also been used.
A number of plant species have been used, especially in wanner parts of the country, and about one third of
the systems use a mixture of plant species. The other two-thirds of the systems use bulrush, cattail, or reeds
alone or in combination. One third of the FWS systems use only cattails, and 40% of the SF systems use
only bulrush. Canna lilies, arrowhead, duckweed, reed canary grass, and torpedo grass are some of the other
species that have been used. In addition to the lack of consensus on the type of plants to be used, there was a
lack of consensus on how these plants should be managed. In Region VI states, weeding and annual harvest
are routinely practiced on many SF systems. This is due in part to the use of soft tissue flowering plants
(Canna Lilies, Water Iris, etc), which can be damaged by a slight frost, and decompose very rapidly resulting
in detrimental impact on effluent quality. However, annual harvesting has also been observed on systems
containing Typha, Scirpus and Phragmites. Annual harvesting was not observed at any of the other systems
in the U.S.
Construction Costs
Construction costs were available from only 18 SF systems and 19 FWS systems. SF wetlands cost more per
acre than FWS wetlands due to the cost of the media. However, SF systems are usually smaller than FWS
systems, based on the assumption that the media provides increased surface area for attached bacterial
growth. Therefore, on a unit flow basis (Fig. 10) the costs for both types of svstems are in the same range.
-------
Inventory of constructed wetlands
500-
L/m2/d
313
TVA I Other I
..mill mill
Rg. 6. Examples of SF hydraulic surface area loading rates of constructed wetlands in the United States.
160;
kg BOD/ha/d
Region VI
Fig. 7. Examples of SF organic surface area loading rates of constructed wetlands in the United States.
m3/m2/d
120
Fig. 8. Examples of SF hydraulic cross sectional loading rates of constructed wetlands in the United States.
Figures 4 through 8 show the expected differences between FWS and SF systems. However, these figures
also show that for both types of systems there is no consensus regarding design. Even within the groups
-------
316
D. S. BROWN and S. C. REED
Greanleaves
Philips H.S.
Monterey-
Denham Springs-
CarvlIIe-
Degussa-
60 40 20 0 20 40 6O
Influent (mg/L) Effluent (mg/L)
Fig. 12. TSS of constructed wetlands in the United States from site visits.
Greenleaves-
Phlllps H.S.
Monterey -
Denham Sprlngs-
Carvllle-
Degussa-
10 864202468 10
Influent (mg/L) Effluent (mg/L)
Fig. 13. NH3 of constructed wetlands in the United States from site visits.
Cold Weather Performance
The question of how well constructed wetlands will perform in cold climates is of particular interest in the
U.S. Miller (1988) reported on the performance of a FWS system in Canada, and found the wetland could
survive the winter and continued to function well. The site visit to Monterey, Virginia provided a small
insight on cold weather performance for a subsurface flow wetland. Figures 14 through 16 show the influent
and effluent BOD, TSS, and NH3 data over a one year period. Wastewater temperatures during this period
ranged from 6 to 19 °C. Ambient air temperatures at this location typically drop below -25 °C in the coldest
part of the winter. Figures 14 through 16 show that BOD and TSS removal continued to be quite good
during the winter months, while NH3 removal remained inefficient.
-------
Inventory of constructed wetlands
315
$/m3/d
$600r
$400L-
i
$200 H
AWT NH3 Sec
Free Water Surface
l
llllllllll
Region VI TVA Other
Subsurface Flow
Fig. 10. Capital cost per unit of flow of constructed wetlands in the United States.
DATA FROM SITE VISITS
Monitoring requirements in the U.S. for small treatment facilities are minimal. Most small facilities are
required to monitor only their wastewater dischsirges, and not the raw wastewater entering the treatment
facility or any intermediate locations. Many small facilities are required to collect only one grab sample of
their effluent per month. As a result, these facilities usually do not have data for the influent to their
constructed wetlands, nor do they have data more frequent than the required one per month. A few locations
had collected data more frequently during start-up. A few locations collected both influent and effluent data
for a short study period, typically one year or less. Therefore, data collected during the 20 site visits was
limited.
General Performance
Figures 11 through 13 show the performance of the constructed wetlands for which both influent and
effluent BOD and TSS, or NH3 data were available. The data shown are averages of monthly data over a
period of one year or less. These figures are meant only to show the range of performance; the performance
of SF constructed wetlands is discussed in more detail in another paper at this conference (Reed and Brown,
1992). As reported in literature for other subsurface flow wetlands (Cooper, 1990; Cooper and Findlater,
1990), these wetlands appear to be able to meet the 15 to 20 mg/L BOD and TSS effluent limits that are
typically imposed on them. However, they do not appear to be able to meet typical NH3 effluent limits of 2
to 6 mg/L.
Greenleaves-
Phlllps H.S.
Monterey-|
Denham Sprlngs-
Carvllle-
Degussa-
40 30 20 10 0 10 20 30 40
Influent (mg/L) Effluent (mg/L)
Fig. 11. BOD of constructed wetlands in the United States from site visits.
-------
Inventory of constructed wetlands
50
BOD (mg/L)
Effluent
Jun
Aug Oct
1989
Dec
Feb Apr
1990
50
Fig. 14. BOD performance at Monterey, Virginia.
TSS (mg/L)
Influent
Effluent
Jun
Aug Oct
1989
Dec
Feb Apr
1990
16
Fig. 15. TSS performance at Monterey, Virginia.
NH3(mg/L)
Jun
Aug Oct
1989
Dec Feb Apr
1990
Fig. 16. NH3 performance at Monterey, Virginia.
-------
3Ig D. S. BROWN and S. C. REED
CONCLUSIONS ;
Both free water surface and subsurface flow wetlands have been used throughout the U.S. Based on limited
data from an inventory and selected site visits, both types of wetlands appear to be reliable and cost-effective
methods for wastewater treatment in terms of BOD and TSS removal. However, these systems have not
always been effective for NH3 removal, possibly due to oxygen limitations. Further work is necessary to
optimize the design approach for this wastewater treatment option.
DISCLAIMER
The work described in this paper has been funded by the United States Environmental Protection Agency
through the Agency's Risk Reduction Engineering Laboratory, Cincinnati, Ohio. However, this article has
not been subject to the Agency's review and therefore does not necessarily reflect the views of the Agency,
and no official endorsement should be inferred.
REFERENCES
Cooper, PP., (ed.) (1990). European Design and Operation; Guidelines For Reed Bed Treatment Systems. Water Research
Cooper, P.K^Ind^erJEl.C., (eds.) (1990). Constructed Wetlands in WalerPollution Control. (Adv. Wat. Pollut. Control no 11)
Pergamon Press, New York, NY.
Miller, Gordon (1988). Use of Artificial Cattail Marshes to Trisat Sewagein Northern Ontario, Canada. In: Constructed Wetlands
for WastewaterTreatment. Lewis Publishers, Ann Arbor, MI.
Reed, S.C., Brown, D.S. (1992). Performance of Gravel Bed Wetlands in the United States. Presented at 3rd International
Specialist Conference on Wetland Systems in Water Pollution Control. IAWPRC.
Watson J T, Cboate, KD. and Steiner, G.R. (1990). Petformiince of Constructed Wetland TreatmentSystems at Benton, Hardin,
and Pembroke, Kentucky During the EarlyVegetation Establishment Phase. In: Constructed Wetlands in Water Pollution
Control. (Adv. Wat. Pollut. Control No 11) Pergamon Press, New York, NY.
Wolverton, B.C., McDonald, R.C. and Ruffer, W.R. (1983). Microorganisms and Higher Plants for Wastewater Treatment. J.
Environ. Qual., 12:236-242.
-------
INVENTORY OF CONSTRUCTED WETLANDS IN THE UNITED STATES1
Donald S. Brown* and Sherwood C. Reed**
*Risk Reduction Engineering Laboratory
U.S. Environmental Protection Agency
Cincinnati, OH 45268 USA
**Environmental Engineering Consultants
RR 1, Box 572, Norwich, VT 05055 USA
ABSTRACT
During 1990 and 1991 the U.S. Environmental Protection Agency
(EPA) sponsored an effort to identify existing and planned
constructed wetlands in the U.S., and to collect readily
available information from operating systems. In addition to
inquiries by telephone and mail, the effort included site visits
to over 20 operating subsurface flow constructed wetlands. The
inventory documented the presence of over 150 constructed wetland
systems for wastewater treatment, including both free water
surface (FWS) and subsurface flow (SF) -systems. The majority of
the systems identified were SF systems for treating municipal
wastewater. FWS systems were separated into three groups based
on the design level of effluent water quality. SF systems were
separated into three groups based on the basic design approach.
The inventory indicated that neither between or within these
groups was there consensus regarding basic hydraulic and
engineering design criteria, system configuration, or any other
aspect, such as, type of vegetation, size and type of media, or
pretreatment. Information on location, type of systems, design
approach, hydraulic and organic loading rates, costs, and other
aspects is presented. Information gathered and "lessons learned"
from the site visits are presented. Insufficient oxygen for
nitrification appears to be a problem for both FWS and SF
systems. Insufficient hydraulic design appears to be a problem
for SF systems.
KEYWORDS
Constructed wetlands, wastewater treatment, free water surface,
subsurface flow.
Published in Water Science & Technology, Vol. 29, No. 4,
pp. 309-318, 1994. EPA Document 600/J-94/352.
-------
-------
INTRODUCTION
The U.S. Environmental Protection Agency's (EPA) Risk Reduction
Engineering Laboratory (RREL) in Cincinnati, Ohio currently has
several projects related to the use of constructed wetlands for
wastewater treatment. Most of the constructed wetlands being
built in the U.S. are either of the free water surface type, or
are subsurface flow systems using large gravel or rock media.
However, these systems are being built without the benefit of
clear design criteria. Because the lack of basic performance
data appears to be the major obstacle to better design, RREL is
undertaking several projects to collect more data. These
projects are focused on subsurface flow wetlands for municipal
wastewater treatment. The first project undertaken was an
inventory of constructed wetlands in the United States. Other
projects include: monitoring of operating full-scale systems;
pilot scale research work on subsurface flow wetland cells; and
evaluating onsite (individual home) systems.
The inventory is the subject of this paper. The inventory was
conducted in two phases. The objective of the first phase was to
locate and collect general information from all constructed
wetlands. The only systems which were specifically excluded from
the search were those serving individual homes or mine drainage
sites. Data were obtained primarily by mail and telephone during
the period January through March 1990. The first phase was
completed in April 1990. The objective of the second phase was
to collect existing design, water quality, operation, and
performance data by making site visits to twenty of the
subsurface flow systems. The second phase was completed in
October 1991. The inventory had to rely on information that was
readily available, and no new information was generated.
Therefore, the information gathered is incomplete, and the
figures that follow may not contain information for all of the
wetlands. (For example, Fig. 1 includes only 90 entries because
the starting date of the other systems was not available.)
For this inventory, constructed wetlands were defined as wetlands
which were specifically designed and built for wastewater
treatment and were located at a site where natural wetlands did
not exist at the time of construction. Constructed wetlands were
categorized into two types based on the flow of water in the
system: free water surface (FWS) and subsurface flow (SF). FWS
-------
-------
wetlands contain emergent aquatic vegetation in a shallow bed or
channel. The water is exposed to the atmosphere as it flows
through the wetland. SF wetlands contain a permeable medium
which supports the root system of the same types of emergent
vegetation. The water level in the bed is maintained below the
top of the media so that all flow is designed to be subsurface.
The term subsurface flow is a general term which includes all
such constructed wetlands, including those systems that have also
been called "rock-reed filters", "gravel bed wetlands",
"microbial rock-reed filters", and "vegetative submerged bed
systems".
INVENTORY RESULTS
"Demographics" of Constructed Wetlands
A total of 143 communities in the U.S. were identified as using,
building, designing, or planning constructed wetlands. A total
of 154 wetlands were actually identified because a few
communities had more than one wetland system, or used some
combination of both FWS and SF units. Of this total, over 60
operating systems were found. The use of constructed wetlands
for wastewater treatment is a relatively new practice in the U.S.
when compared with the European countries. Constructed wetlands
have really been used only since 1986 (Fig. 1), although a few
FWS systems are older. There are an almost equal number of
operating FWS and SF systems, but SF systems are projected to
outnumber FWS systems in the future. The majority of systems are
located in the Mississippi River basin, but they are found or
planned throughout the U.S. (Fig. 2).
The majority of systems (70% of FWS and 90% of SF) treat less
than 3800 m3/day (Fig. 3) . Smaller, systems tend to be SF, which
range in size from 5 to 11,400 m3/d (mean = 1150 m3/d, median =
210 m3/d). Larger systems tend to be FWS, which range from 200
to 76,000 m3/d (mean = 7350 m3/d, median = 1770 m3/d).
-------
-------
.Number of Systems
<'80 <'86 '86 '87 '88 '89 '90 '91 >'91
Starting Date
Fig. 1. Starting date
(actual or projected)
Number of Systems
< 5 5- 10
> 15
Fig. 2. Distribution
of systems
Number of Systems (%)
1 10 102 103 104 105
Design Flow (m3/day)
Fig. 3. Size based on
design flow
Design Characteristics
The primary reason for conducting the inventory was to identify
the design criteria used at existing systems. One of the
principal conclusions from the inventory, however, was that there
is no consensus regarding the design of constructed wetlands in
the U.S. Design approaches appeared to range from "trial and
error" to semi-empirical to semi-rational using design models
based on very limited data. Figures 4 through 8 show the design
information that was available for several operating systems.
Each bar in these Figures represents one wetland for which design
criteria were available. (For example, as shown in Figure 4
there were 25 FWS wetlands for which hydraulic loading criteria
were available.)
-------
-------
The data in Figures 4 through 8 is organized into groups for
clearer presentation. For ease of viewing the bars are placed in
order from smallest to largest for each group shown. The FWS
systems are divided into three groups according to effluent
goals: BOD and TSS removal; BOD, TSS, and NH3 removal; and
advanced wastewater treatment (AWT) or total retention (i.e. no
discharge). The SF systems are divided into three groups
according to the basis of design as follows:
TVA:
guidance provided by the Tennessee Valley Authority
(Watson, etal, 1990), derived from European
experience; characterized by the use of small rock or
gravel for media (less than 40 mm diameter), variable
length to width ratios, and lower hydraulic loading
rates (less than 200 L/m2/d).
designs generally used in EPA's Region VI (South-
Central U.S.), derived from work done by Wolverton
(1983); characterized by large rock for media (80 mm
or greater diameter), large length to width ratios
(10 or greater), and higher hydraulic loading rates
(greater than 150 L/m2/d).
Other: independent designs.
Region
VI:
120
100
80
60
40
20
L/m2/d
kg BOD/ha/d
AWT &
Retention
BOD, TSS
& NH3
il.
Fig. 4. FWS hydraulic surface
area loading rates
Fig. 5. FWS organic surface
area loading rate
Figures 4 through 8 show the expected differences between FWS and
SF systems. However, these figures also show that for both types
of systems there is no consensus regarding design. Even within
the groups shown there is a wide range of design criteria used.
-------
-------
500
400
300
200
100
L/m2/d
kg BOD/ha/d
Region V!
TVA
Other
iidllLjil
Fig. 6. SF hydraulic surface
area loading rates
Fig. 7. SF organic surface
area loading rates
120
100
80
60
40
20
0
m3/m2/d
Region Vi
TVA
-juJUL
Other
.Mil
Fig. 8. SF hydraulic cross
sectional loading rate
For example, Region VI SF systems are typically loaded at two to
three times the rate of other SF systems or FWS systems, while
TVA SF systems are loaded at rates similar to FWS systems
designed for BOD, TSS, and NH3 removal.
There is also no consensus on system configuration, or other
details such as depth of water or media, type or size of media,
slope of bed, length to width ratio, level of pretreatment, and
type of vegetation. In FWS systems the water depth ranges from 5
to 90 cm; 30 to 45 cm is typical. In' SF systems media depth
ranges from 30 to 76 cm, and sand, peat, gravel and rock have all
been used. Gravel and rock size ranges from 6 to 130 mm; 13 to
80 mm is typical. Bottom slopes range from 0 to 1%.
There is disagreement in particular over the desired length to
width ratio (L:W) for subsurface flow wetlands. Region VI
designs have generally had a high L:W, based on the argument that
-------
-------
a high L:W is needed to ensure plug flow and high levels of
performance. TVA and "Other" designs usually consider the
hydraulic loading rate on the vertical cross sectional area of
the bed, and this tends to produce wetlands with a lower L:W.
However, as shown in Figure 9 the trend for subsurface flow
systems is toward lower L:W ratios. Length to width ratios for
FWS wetlands have remained relatively constant.
20
15
10
L:W Ratio
i f_i i i...r.J_j
'87 '88
•89
'90
•91
Fig. 9. Trend in L:W for
SF wetlands
The level of pretreatment varies widely. Facultative, lagoons are
the most common form of pretreatment and are used at 41% of the
FWS and 44% of the SF systems. , Septic tanks are used at 24% of
the.SF systems. Aerated lagoons, secondary treatment, and
advanced treatment have also been used.
A number of plant species have been used, especially in warmer
parts of the country, and about one third of the systems use a
mixture of plant species. The other two thirds of the systems
use bulrush, cattail, or reeds alone or in combination. One
third of the FWS systems use only cattails, and 40% of the SF
systems use only bulrush. Canna lilies, arrowhead, duckweed,
reed canary grass, and torpedo grass are some of the other
species that have been used. In addition to the lack of
consensus on the type of plants to be used, there was a lack of
consensus on how these plants should be managed. In Region VI
states, weeding and annual harvest are routinely practiced on
many SF systems. This is due in part to the use of soft tissue
flowering plants (Canna Lilies, Water Iris, etc), which can be
damaged by a slight frost, and decompose very rapidly resulting
in detrimental impact on effluent quality. However, annual
harvesting has also been observed on systems containing Typha,
-------
-------
Scirpus and Phragmites. .Annual harvesting was not observed at
any of the other systems in the U.S.
Construction Costs
Construction costs were available from only 18 SF systems and 19
FWS systems. SF wetlands cost more per acre than FWS wetlands due
to the cost of the media. However, SF systems are usually
smaller than FWS systems, based on the assumption that the media
provides increased surface area .for attached bacterial growth.
Therefore, on a unit flow basis (Fig. 10) the costs for both
types of systems are in the same range.
. $/m3/d
$800
$600
$400
$200
$0'
AWT NH3 Sec
Free Water Surface
JIIMI
Region VI TVA Other
Subsurface Flow
DATA FROM SITE VISITS
Monitoring requirements in the U.S. for small treatment
facilities are minimal. Most small facilities are required to
monitor only their wastewater discharges, and not the raw ''""
wastewater entering the treatment facility or any intermediate
locations. Many small facilities are required to collect only
one grab sample of their effluent per month. As a result, these
facilities usually do not have data for the influent to their
constructed wetlands, nor do they have data more frequent than
the required one per month. A few locations had collected data
more frequently during start-up. A few locations collected both
influent and effluent data for a short study period, typically
one year or less. Therefore, data collected during the 20 site
visits was limited.
General Performance
Figures 11 through 13 show the performance of the constructed
wetlands for which both influent and effluent BOD and TSS, or NH3
8
-------
-------
data were available. The data shown are averages of monthly data
over a period of one year or less. 'These figures are meant only
to show the range of performance; the performance of SF
constructed wetlands is discussed in more detail in another paper
at this conference (Reed and Brown, 1992). As reported in
literature for other subsurface flow wetlands (Cooper,1990;
Cooper and Findlater, 1990), these wetlands appear to be able to
meet the 15 to 20 mg/L BOD and TSS effluent limits that are
typically imposed on them. However, they do not appear to be
able to meet typical NH3 effluent limits of 2 to 6 mg/L.
Greenleaves
Philips H.S.
Monterey
Denham Springs
Carvllle
Degussa
40 30 20 10 0 10 20 30 40
Influent (mg/L) Effluent (mg/L)
Fig. 11. BOD from site visits
Greenleaves
Philips H.S.
Monterey
Denham Springs
Carvllle
Degussa
60 40 20 0 20 40 60
Influent (mg/L) Effluent (mg/L)
Fig. 12. TSS from site visits
Greenleaves
Philips H.S.
Monterey
Denham Springs
Carvllle
Degussa
10 864202468 10 '
Influent (mg/L) Effluent (mg/L)
Fig. 13. NH3 from site visits
Cold Weather Performance
The question of how well constructed wetlands will perform in
cold climates is of particular interest in the U.S. Miller
(1988) reported on the performance of a FWS system in Canada, and
found the wetland could survive the winter and continued to
function well. The site visit to Monterey, Virginia provided a
small insight on cold weather performance for a subsurface flow
wetland. Figures 14 through 16 show the influent and effluent
-------
-------
BOD, TSS, and NH3 data over a one year period. Wastewater
temperatures during this period ranged from 6 to 19 °C. Ambient
air temperatures at this location typically drop below -25 °C in
the coldest part of the winter. Figures 14 through 16 show that
BOD and TSS removal continued to be quite good during the winter
months, while NH3 removal remained inefficient.
BOD (mg/L)
Oct
1989
Dec
Fefa Apr
1990
Fig. 14. BOD performance at
Monterey, Virginia
50
40
30
20
TSS (mg/L)
Influent
Jun
Aug Oct
1989
Dec
Feb Apr
1990
Fig. 15. TSS performance at
Monterey, Virginia
CONCLUSIONS
NH3 (mg/L)
16
14
12
10
8
4
21-
Jun
Effluent
Aug Oct
1989
Dec
Feb Apr
1990
Fig. 16. NH3 performance at
Monterey, Virginia
Both free water surface and subsurface flow wetlands have been
used throughout the U.S. Based on limited data from an inventory
and selected site visits, both types of wetlands appear to be
reliable and cost effective method for wastewater treatment in
terms of BOD and TSS removal. However, these systems have not
always been effective for NH3 removal, possibly due to oxygen
10
-------
-------
limitations. Further work is necessary to optimize the design
approach.for this wastewater treatment option.
DISCLAIMER
The work described in this paper has been funded by the United
States Environmental Protection Agency through the Agency's Risk
Reduction Engineering Laboratory, Cincinnati, Ohio. However,
this article has not been subject to the Agency's review and,
therefore does not necessarily reflect the views of the Agency,
and no official endorsement should be inferred.
REFERENCES
Cooper, P.P., ed. (1990). European Design and Operations
Guidelines For Reed Bed Treatment Systems; Water Research Centre,
Swindon, UK.
Cooper, P.F., Findlater, B.C., ed. (1990). Constructed Wetlands
in Water Pollution Control; Pergamon Press, New York, NY.
Miller, Gordon (1988). Use of Artificial Cattail Marshes to Treat
Sewage in Northern Ontario, Canada; In: Constructed Wetlands for
Wastewater Treatment, Lewis Publishers, Ann Arbor, MI.
Reed, S.C., Brown, D.S. (1992). Performance of Gravel Bed
Wetlands in the United States; In Proceedings: 3rd International
Specialist Conference on Wetland Systems in Water Pollution
Control, IAWPRC (In press).
Watson, J.T., et al (1990). Performance of Constructed Wetland
Treatment Systems at Benton, Hardin, and Pembroke, Kentucky
During the Early Vegetation Establishment Phase; In Proceedings:
Constructed Wetlands in Water Pollution Control, Pergamon Press,
New York, NY.
Wolverton, B.C., et al (1983). Microorganisms and Higher Plants
for Wastewater Treatment; J. Environ. Qual., 12 (2) : 236-242 .
11
-------
------- |