See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/227006774

Design Methodology of Free Water Surface Constructed Wetlands

Article in Water Resources Management · January 2004

DOI: 10.1007/s11269-004-6480-6

CITATIONS READS

36 4,868

2 authors, including:

Vassilios A. Tsihrintzis
National Technical University of Athens
252 PUBLICATIONS 8,332 CITATIONS

SEE PROFILE

All content following this page was uploaded by Vassilios A. Tsihrintzis on 20 May 2014.

The user has requested enhancement of the downloaded file.

 Water Resources Management 18: 541–565, 2004.


C 2004 Kluwer Academic Publishers. Printed in the Netherlands.
541

Design Methodology of Free Water Surface

Constructed Wetlands

MARIA A. ECONOMOPOULOU and VASSILIOS A. TSIHRINTZIS∗

Laboratory of Ecological Engineering and Technology, Department of Environmental Engineering,
Democritus University of Thrace, Xanthi 67100, Greece
(∗author for correspondence, e-mail: [email protected]; [email protected])

(Received: 7 January 2003; in final form: 1 April 2004)

Abstract. Simple criteria, guidelines and models are established for free water surface (FWS) con-
structed wetland selection and preliminary sizing. The analysis employs models for FWS constructed
wetland design, considering simultaneously the removal requirements and the hydraulics of the
system. On the basis of these models, a step-by-step methodology is developed outlining the de-
sign procedure for new and performance evaluation for existing FWS constructed wetland systems.
This methodology is combined with simple equations predicting the maximum wetland capacity
in summer, so as to assist designers in sizing installations in tourist areas with increased summer
populations. Furthermore, this methodology is further simplified, based on sensitivity analysis of
the unit area requirements for wastewaters of various strengths, and various design conditions and
performance criteria. In addition, comparison of the unit area requirements of FWS constructed wet-
land systems, subsurface flow (SF) constructed wetland systems and stabilization pond systems for
wastewaters of various strengths and design conditions, provides designers with general guidelines
concerning the preliminary selection between alternative natural treatment systems in areas where
the use of natural systems is favored because of their low-cost, simple operation and high removal
performance.

Key words: design methodology, free water surface constructed wetlands, seasonal performance,
sensitivity analysis, stabilization ponds, subsurface flow constructed wetlands, unit area requirement,
wastewater treatment

1. Introduction
Natural treatment systems are characterized by low-maintenance, simple and re-
liable operation and high removal efficiencies. These systems are highly favored
in small to medium communities, where the resources and the skilled personnel
required for the operation of conventional systems are often limited (USEPA, 1988;
Sartoris et al., 2000). Furthermore, they are considered a favorable treatment al-
ternative for the production of effluents that can be reused for irrigation, limiting
the fresh water consumption and the possibility for pollution of receiving waters
(WHO, 1989). Therefore, emphasis is placed on the development of practical de-
sign and analysis procedures that promote the proper evaluation and use of natural
systems (e.g., Economopoulou and Tsihrintzis, 2003).
542 M. A. ECONOMOPOULOU AND V. A. TSIHRINTZIS

Constructed wetlands have been extensively used in the treatment of waste-

waters. A number of constructed wetland projects have been reported to treat var-
ious types of wastes (e.g., Hammer, 1989; Schueler, 1992; Moshiri, 1993; Ol-
son, 1993; Etnier and Guterstam, 1997; Mulamoottil et al., 1999; Dialynas et al.,
2002; Mander and Jenssen, 2002, 2003; among others). Several natural wetland
restoration projects have also been reported (e.g., Mitsch and Gosselink, 1986;
Hammer, 1992; Minnesota Board of Water and Soil Resources, 1993; Tsihrintzis
et al., 1995a, 1995b, 1996, 1998, among others), and various design procedures
have been presented (e.g., USEPA, 1988; Tchobanoglous and Burton, 1991; Reed
et al., 1995; Kadlec and Knight, 1996; Tsihrintzis and Madiedo, 2000; Tsihrintzis,
2001; Economopoulou and Tsihrintzis, 2003; among others).
Economopoulou and Tsihrintzis (2003) have presented a step-by-step method-
ology for designing subsurface flow (SF) constructed wetlands for wastewater treat-
ment. This paper presents a convenient procedure that can be used in the design and
sizing of new free water surface (FWS) constructed wetland systems, on the basis
of performance criteria, without going through iterations. This procedure can also
be used to assess the removal efficiency of existing systems against local design
parameters.

2. Pollutant Removal Theory

The present study employs the constructed wetland design theory and typical kinetic
rate constants for municipal wastewater in computing pollutant removal efficiencies.
The following assumptions have been made:
• The water temperature can be assumed approximately equal to the mean ambi-
ent temperature. This is a reasonable assumption for relatively warm climates
(Kadlec and Knight, 1996).
• The removal rates for BOD and nitrogen in FWS constructed wetland systems
are typically based on first-order kinetics and on the assumptions of plug flow,
and are based on the models proposed by the USEPA (1988) and Reed et al.
(1995), which have been used in the design of most constructed wetland systems
in the U.S. and Europe (Chen et al., 1999; Economopoulou and Tsihrintzis,
2003).
BOD and nitrogen removal rates in FWS constructed wetlands are estimated
by the following general Equation (1) (Reed et al., 1995), whereas, coliform and
phosphorus removals by general Equation (2) (Kadlec and Knight, 1996):

Ce
= e−K T t (1)
Ci
Ce −
K1
=e hl
(2)
Ci
DESIGN METHODOLOGY OF FREE WATER SURFACE CONSTRUCTED WETLANDS 543

Table I. Pollutant removal equations and rate constants for FWS constructed wetlands

Equation Rate constant

Pollutant used Rate constant units

BOD (1)∗ K T = 0.678(1.06)T -20 [d–1 ]

Fecal coliforms (2)∗∗ K 1 = 0.3 [m d–1 ]
Nitrogen
Nitrification (1)∗ K T = 0.0389T 0 < T < 1◦ C [d–1 ]
K T = 0.1367(1.15) T −10 1 < T < 10◦ C [d–1 ]
K T = 0.2187(1.048) T -20 T > 10 ◦ C [d–1 ]
Denitrification (1)∗ K T = 0.023T 0 < T < 1◦ C [d–1 ]
K T = 1.15(T – 20) T > 1 ◦C [d–1 ]
Phosphorus (2)∗∗ K 1 = 0.0273 [m d–1 ]
∗ by Reed et al. (1995), ∗∗ by Kadlec and Knight (1996).

In these two general equations: Ce is the pollutant effluent concentration[mg

L−1 of BOD, nitrogen or phosphorus, or number of fecal coliforms/100mL]; Ci is
the pollutant influent concentration [mg L−1 of BOD, nitrogen or phosphorus, or
number of fecal coliforms/100mL]; K T is a reaction rate parameter [d−1 ] depen-
dent on the water temperature T [◦ C], and the pollutant of interest (Table I); K 1
is a reaction rate constant [m d−1 ] dependent on the pollutant of interest (Table
I); h l is the hydraulic loading rate[m d−1 ]; and t is the hydraulic residence time
(HRT) in the system [d]. The last two parameters are defined by the following
equations:

Q
hl = (3)
A
V Ayφ yφ
t = = = (4)
Q Q hl

where Q is the design flow rate [m3 d–1 ], assumed constant; A is the mean surface
area of the system [m2 ]; V is the system volume [m3 ]; y is the flow depth [m];
φ is the fractional porosity, which expresses the space available for water to flow
through the vegetation and litter in the FWS constructed wetland system (Reed
et al., 1995).
The estimation of BOD removal in FWS constructed wetlands is simpler com-
pared to SF constructed wetlands (Economopoulou and Tsihrintzis, 2003), since the
water temperature T can be assumed equal to the ambient temperature Tα (Kadlec
and Knight, 1996), and thus t can be computed directly from Equation (1), setting
T = Ta and using the appropriate kinetic rate constant from Table I. Of course, this
assumption is valid only for relatively warm climates with no ice cover.
544 M. A. ECONOMOPOULOU AND V. A. TSIHRINTZIS

Reed et al. (1995), based on the operational experience of several constructed

wetland systems in the United States, proposed that the organic loading in FWS
constructed wetland systems should not exceed a 10 g m−2 d−1 limit value, which
can be expressed by the following relation (Economopoulou and Tsihrintzis, 2003):

Q
Ci ≤ 10 (5)
A

Since the area A [m2 ] is not known prior to designing a new FWS constructed
wetland system, the estimation of the organic loading requires a trial-and-error
procedure.
The removal of coliforms in FWS constructed wetlands is due to physical sep-
aration of the particles and die-off (Kadlec and Knight, 1996; Reed et al., 1995),
and is estimated from Equation (2) and the appropriate rate constant value listed
in Table I. The removal of bacteria, viruses and helminth eggs has also been found
to be satisfactory in arid climates (Khatiwada and Polprasert, 1999; Mandi et al.,
1998), so effluent wastewaters could be used even for irrigation providing minimum
risk to workers and consumers.
The removal of nitrogen based on nitrification and denitrification is highly de-
pendent on water temperature (USEPA, 1988; Reed et al., 1995; Kadlec and Knight,
1996). This is also in accordance with the works of Bachand and Horne (2000a,
2000b) and Reilly et al. (2000), who studied the denitrification of wastewater in
FWS constructed wetlands, and concluded that nitrate removal is highly dependent
on water temperature and organic carbon availability.
It is assumed that the total Kjeldahl nitrogen entering the system is converted to
ammonia, which is partly converted to nitrate by nitrification and subsequently is re-
moved by denitrification (Reed et al., 1995; Economopoulou and Tsihrintzis, 2003).
The sum of the unconverted ammonia after nitrification (Ce,amm ) and unconverted
nitrate after denitrification (Ce,nitr ) represent the total nitrogen (TN) remaining in
the system. Nitrification and denitrification are both described by Equation (1),
with the appropriate rate constant value from Table I. A minimum design HRT of
about 6 to 8 days is advisable, to guarantee oxygen availability for nitrification and
maximize ammonia removal efficiency (Reed et al., 1995; Reed and Brown, 1993).
The removal of phosphorus is estimated from Equation (2) and the appropriate
rate constant listed in Table I and is believed to range between 30 to 50% in the
long term.

3. Hydraulic Design Theory

Kadlec (1990) and Kadlec and Knight (1996) proposed the following general equa-
tion for the hydraulic design of FWS constructed wetland systems:

Q = a W yb Sc (6)
DESIGN METHODOLOGY OF FREE WATER SURFACE CONSTRUCTED WETLANDS 545

where: Q is the flow rate [m3 d−1 ]; W is the wetland width [m]; a, b and c are
coefficients assuming the following values: a = 107 d−1 m−1 for dense vegetation,
a = 5× 107 d−1 m−1 for sparse vegetation, b = 3.0 and c = 1.0; y is the depth of
flow [m], which usually ranges from 0.1 to 0.6 m (Reed et al., 1995); and S is the
water surface slope [m/m], estimated by the following equation:

γy
S= (7)
L

where γ is the fraction of the depth serving as head differential (Reed et al., 1995);
and L is the wetland length [m].
Equation (6), with the values of the coefficients a, b and c mentioned above,
seems to describe accurate flow through wetland vegetation (Kadlec, 1990). Al-
ternatively, Manning’s equation can also be used, if a, b and c in Equation (6)
are set equal to 1/n, 5/3 and 1/2, respectively, with n as the Manning’s roughness
coefficient. A comprehensive approach on wetland flow resistance, presented by
Tsihrintzis and Madiedo (2000) and Tsihrintzis (2001), has shown that the Man-
ning’s coefficient values for wetland vegetation are significantly higher than those
for turbulent open channel flows controlled by skin friction. Indeed, flow resistance
in wetland vegetation is highly dependent on the type, height and density of vege-
tation, the diameter and flexibility of the vegetation’s stem, the depth of flow, the
depth of litter layer, etc. (Kadlec, 1990; Kadlec and Knight, 1996; Tsihrintzis and
Madiedo, 2000; Tsihrintzis, 2001).
If Manning equation is to be used, then the Manning’s roughness coefficient n
can be estimated with one of the following three methods:
1. Using the data, procedure and design graphs provided by Tsihrintzis and
Madiedo (2000) and Tsihrintzis (2001), where n is presented as a function
of parameter VR (product of mean velocity and hydraulic radius), vegetation
density and other factors.
2. Using a logarithmic diagram developed by Kadlec and Knight (1996), based on
information from existing wetlands, which allows for the preliminary estimation
of Manning’s roughness coefficient n as a function of flow depth. To simplify
the use of this diagram, Economopoulou and Tsihrintzis (2000) have developed
two regression lines of the following general equation:
n = β1 y β2 (8)
where: β 1 = 0.1564 for sparse vegetation and 1.09 for dense vegetation; and
β 2 = −1.356 for sparse vegetation and −1.436 for dense vegetation.
3. Using the general Equation (8), but with the following values for coefficients β 1
and β 2 , proposed by Reed et al. (1995): β 2 = −0.5; β 1 = 0.4 m1/6 ·s for sparse
and low-standing vegetation with flow depth y > 0.4 m; β 1 = 1.6 m1/6 ·s for
moderately dense vegetation with flow depth y in the range 0.3 to 0.4 m; and
546 M. A. ECONOMOPOULOU AND V. A. TSIHRINTZIS

β 1 = 6.4 m1/6 ·s for very dense vegetation with litter layer and with y < 0.3 m.
In most FWS constructed wetlands, β 1 ranges from 1 m1/6 ·s (sparse vegetation)
to 4 m1/6 ·s (dense vegetation).
It is noticed that in most FWS constructed wetland systems there is water sur-
face control at the outlet. Thus, the depth may not be normal close to the outlet.
However, it will tend to normal depth further upstream, particularly if the wetland
is designed with a large length to width ratio, something generally recommended.
In any case, the downstream control depth can be set close to normal depth. As
suggested by Kadlec and Knight (1996), the aspect ratio L:W should be greater
than 2:1 to ensure plug flow conditions. However, very high ratios may result
in overflow problems due to resistance increase as a result of the gradual accu-
mulation of vegetation litter. Commonly used aspect ratios are between 2:1 and
5:1.

4. Simplified FWS Constructed Wetland Design

As seen in the previous section, the design equation based on total nitrogen, the
estimation of FWS system’s flow depth and the estimation of the organic loading
cannot be solved directly, but require a trial-and-error procedure. Therefore, to
avoid this, the theory and equations presented in the previous sections are revisited
here. As a result, appropriate nomographs and equations have been prepared and
are presented in this section, which simplify the procedure and provide a conve-
nient method for sizing new or for evaluating the performance of existing FWS
constructed wetlands.

4.1. NITROGEN

As mentioned in Section 2, effluent total nitrogen is the sum of residual ammonia

from nitrification and remaining nitrate from denitrification. Thus, Equation (1)
for nitrification and Equation (1) for denitrification were mathematically combined
into the following equation for total nitrogen removal efficiency (Economopoulou
and Tsihrintzis, 2003)
Ce
= e−K TN t + e−K TD t − e−K TN t e−K TD t (9)
Ci
where: Ci is the total Kjeldahl nitrogen inflowing concentration (assumed all con-
verted to ammonia) [mg L−1 ], and Ce is the total nitrogen outflowing concentra-
tion [mg L–1 ]; K TN and K TD are the reaction rate parameters for nitrification and
denitrification [d−1 ], respectively, taken from Table I as a function of the design
temperature T . In a new constructed wetland sizing problem, the required t [d] can
be obtained from Equation (9), as a function of the total nitrogen removal ratio
and T [◦ C]. However, Equation (9) cannot be solved directly for t, but only by a
DESIGN METHODOLOGY OF FREE WATER SURFACE CONSTRUCTED WETLANDS 547

Figure 1. Nitrogen removal efficiency in FWS constructed wetlands.

trial-and-error procedure. To simplify the computation, the nomograph of Figure 1

provides a graphical solution to Equation (9), allowing for direct estimation of t for
nitrogen removal as a function of T and Ce /Ci .

4.2. HYDRAULIC DESIGN

On the basis of Equations (6), (7) and (4) the following general equation is derived
(Economopoulou and Tsihrintzis, 2000):
2c
2 2b+3c−1
L
Qϕ 1− c+1
a c+1 y c+1 = W
2c 2c (10)
1− c+1
t γ c+1

If the Manning’s equation is used, then the coefficients a, b and c are equal
to 1/n, 5/3 and 1/2, respectively. Their introduction into Equation (10), with n es-
timated by the general Equation (8) and the values of the constants β 1 and β 2
equal to those proposed by Economopoulou and Tsihrintzis (2000) for sparse
and dense vegetation, respectively, yields the following equation for the flow
548 M. A. ECONOMOPOULOU AND V. A. TSIHRINTZIS

depth y:

 L 3 0.0764
0.0006 Q3ϕ
y= W
(for sparse vegetation) (11)
t γ2
864004

 3 0.0745
1.4110 WL Q 3 ϕ
y= (for dense vegetation) (12)
864004 t γ2

Similarly, the introduction of the coefficients a = 1/n, b = 5/3 and c = 1/2 in

Equation (10), with n estimated by the general Equation (8) and the values of the
constants β1 and β2 equal to those proposed by Reed et al. (1995), results in:

 L 3 3/29
β14 Q3ϕ
y= W
(13)
864004 t γ2

Finally, the introduction of the values b = 3.0, c = 1.0 with a = 5× 107

d m−1 (sparse vegetation) or a = 107 d−1 m−1 (dense vegetation) (Kadlec, 1990;
−1

Kadlec and Knight, 1996) into Equation (10), yields the following equation for y:

1/4
1 WL Q
y= (14)
a γ

4.3. ESTIMATION OF THE ORGANIC LOADING

When designing a new FWS constructed wetland system, the area A of the system
is not known in advance, thus it would be preferable to use a more convenient
equation to check the organic loading in the wetland, so as to avoid time-consuming
iterations. If BOD removal is required, then Equations (1) and (5) are solved for
A and the two equations are set equal and solved for K T , yielding the following
formula:
 
10 ln CCei
KT ≤ − (15)
Ci yφ
Equation (15), which is equivalent to Equation (5), depends on flow depth y
rather than the system’s area A. As it will be seen later in this paper, the organic
loading limit plays an important role, as it usually controls the design of the wetland
system for higher temperatures.

5. Method Description and Validation

This section presents a procedure for rapid sizing of a new FWS constructed
wetland, based on specified BOD, pathogen, nitrogen and phosphorus removal
DESIGN METHODOLOGY OF FREE WATER SURFACE CONSTRUCTED WETLANDS 549

efficiency, as well as for determining the removal performance of existing con-

structed wetland systems, using the presented theory and the nomograph of Figure
1. In hydraulic design, Equation (14) is proposed because of its simplicity (it does
not require a trial-and-error procedure), and based on the results of the sensitivity
analysis presented in the following Section 7.1.

5.1. DESIGN OF NEW FWS CONSTRUCTED WETLANDS

In sizing a new FWS constructed wetland, the following parameters are typically
known: the design values of Q, Ce /Ci for one or more pollutants, T , L:W (preferably
between 2:1 to 5:1) and the vegetation porosity φ (typically in the range of 0.65 to
0.75) (Reed et al., 1995). To size the system with the above variables specified, the
following procedure is used:
1. The value of y is computed from Equation (14), which should be in the range
0.1 m to 0.6 m. If y > 0.6 m, two parallel systems are considered or the ratio
L:W is reduced (Reed et al., 1995). For the computed value of y, it is checked
if Equation (15) holds. It has to be noted that Equations (11)–(13) can also
be used, but a trial-and-error procedure has to be followed, since t enters in
these equations.
2. If BOD removal is required, then the hydraulic residence time tBOD [d] is esti-
mated on the basis of Equation (1) and the appropriate reaction rate constant of
Table I.
3. If coliform removal is required, then the hydraulic residence time tCOLI [d] is
estimated on the basis of Equations (2) and (4) and the appropriate reaction rate
from Table I.
4. If nitrogen removal is required, then the hydraulic residence time tTN [d] is
estimated on the basis of Figure 1, or Equation (9), and the appropriate reaction
rate from Table I.
5. If phosphorus removal is required, then the hydraulic residence time tPHOS [d]
is estimated on the basis of Equations (2) and (4) and the appropriate reaction
rate from Table I.
6. From the above steps up to four hydraulic residence time values are esti-
mated, depending on the removal requirements of the pollutants. The de-
sign value of the hydraulic residence time is the maximum of these four
values.
7. Steps 1–5 are repeated for winter and summer conditions (i.e., lowest winter and
summer season temperatures) resulting in two hydraulic residence time values.
On the basis of these two values and the corresponding populations, the design
conditions are determined.
8. From the estimated surface area (Steps 1–7) and the chosen aspect ratio L:W ,
the values of W and L are easily computed.
550 M. A. ECONOMOPOULOU AND V. A. TSIHRINTZIS

5.2. PERFORMANCE EVALUATION OF EXISTING FWS CONSTRUCTED WETLANDS

In the removal performance analysis of an existing constructed wetland the design

values of Q, T , L, W and φ are known. To calculate the system performance with
the above variables specified, the following procedure is used:
1. The value of y is computed from Equation (14) or Equations (11)–(13), and it
is checked whether y ≤ 0.6 m. Equation (15) is checked to see if the organic
loading is smaller than 10 g m−2 d−1 , and if not, the influent flowrate Q should
be reduced accordingly.
2. The values of t and h l are computed from Equations (3) and (4) and, to ensure
satisfactory NH3 removal efficiency, it is checked whether t > 6 to 8 d.
3. The removal efficiencies for BOD, coliforms, nitrogen, and phosphorus are
obtained from Equations (1), (2) as appropriate, or the nomograph of Figure 1
with the appropriate rate constants found in Table I.

5.3. METHODOLOGY VALIDATION WITH EXISTING DATA

Available data were collected concerning the area and the operating conditions of
existing FWS constructed wetland systems in the U.S. from an existing database
(Kadlec and Knight, 1996). It has to be emphasized that available data in the liter-
ature on such systems are limited, and existing studies usually contain incomplete
information which makes validation studies difficult. The data used is summarized
in Table II along with the estimated areas from the methodology presented in this
paper (Section 5.1). The selected constructed wetland systems that were consid-
ered for the validation process had a mean air temperature higher than 5 ◦ C. For the
design methodology the water temperature was considered equal to the mean air
temperature. Information on the influent and effluent concentrations of at least one
pollutant was available. Since other design parameter information was not avail-
able, the following assumptions were made in all cases in estimating the system’s
surface area: φ = 0.65, γ = 0.1 and use of Equation (14) with a = 107 d−1 m−1
(dense vegetation). Computations were made for two aspect ratios, which cover the
common range, namely L:W = 2:1 and 5:1.
A way to assess the validation is through the use of scattergrams where predicted
quantities are plotted against observed ones. In the scattergram, a regression straight
line of the following form is fitted through the data: Ap = r A, where Ap is the
predicted wetland surface area [ha] and A is the real (as constructed) wetland
surface area [ha]. The slope r of this regression line is compared with the 1:1 slope
line (perfect match). The value of the slope r is a measure of the over (r > 1.0)
or underestimation (r < 1.0) of the predicted values compared to the real data.
In addition, the square of the correlation coefficient R 2 of the regression line is
computed. The lower the value of R 2 falls below 1.0, the worse the data correlation
is (i.e., the greatest the scatter). Therefore, best calibration requires that values for
both slope r and correlation coefficient R 2 be as close to 1.0 as possible.
Table II. Summary of data from existing constructed FWS wetlands (Kadlec and Knight, 1996) and listing of the predicted area based on these
data

Wetland surface area

Concentrations Predicted∗ Ap [ha]
System name T [◦ C] Q [m3 d–1 ] Influent[mg L–1 ] Effluent [mg L–1 ] Real A [ha] L:W = 2:1 L:W = 5:1
Bear Bay, SC, USA 6.0 876.63 Ci,BOD = 13.50 Ce,BOD = 1.90 28.3 18.2 14.5
Ci,TN = 17.58 Ce,TN = 2.35
Ci,PHOS = 3.88 Ce,PHOS = 0.40
Boggy Gut, SC, USA 6.0 5826.78 Ci,BOD = 6.30 Ce,BOD = 3.00 20.2 48.0 38.2
Ci,TN = 11.22 Ce,TN = 3.53
Ci,PHOS = 4.26 Ce,PHOS = 3.35
Fort Deposit, AL, USA 9.0 673.73 Ci,BOD = 32.90 Ce,BOD = 6.90 6.0 2.4 1.9
Gustine, CA, USA 9.0 128.71 Ci,BOD = 173.30 Ce,BOD = 36.70 0.4 0.7 0.5
Johnson City, TX, USA 10.0 114.00 Ci,BOD = 20.0 Ce,BOD = 8.00 0.1 0.3 0.3
Ci,PHOS = 7.00 Ce,PHOS = 5.50
Mt View CA, USA 5.0 2820.98 Ci,BOD = 33.40 Ce,BOD = 23.10 4.3 2.1 1.6
W. Jackson County, USA 5.0 6268.00 Ci,BOD = 25.93 Ce,BOD = 7.40 22.7 12.8 10.2
∗ Note:
In area estimate Ap , it was assumed: ϕ = 0.65, γ = 0.1 and use of Equation (14) with a = 107 d–1 m–1 (dense vegetation). Two
DESIGN METHODOLOGY OF FREE WATER SURFACE CONSTRUCTED WETLANDS

computations were made with aspect ratios L:W = 2:1 and 5:1.
551
552 M. A. ECONOMOPOULOU AND V. A. TSIHRINTZIS

Figure 2. Linear correlation between real (as constructed) and predicted wetland area values
for existing FWS constructed wetland systems.

Results for the selected wetland sites are presented in Figure 2. The slope r of
the regression line, based on the methodology presented in this paper, has the value
of 1.01 for L:W = 2:1 and 0.80 for L:W = 5:1. Since this range of L:W values
is the common one, the predictions of the estimated values are considered good.
The R 2 value is the same for both values of the aspect ratio (0.45); it is not very
high, showing scattering of the data around the line. However, the agreement is
considered overall satisfactory, taking into account that, as mentioned earlier, the
exact values of the design parameters T, φ, L:W , γ and a for each system were not
available and assumptions were made.

6. Season-Dependent System Capacities

The increased capacity of a constructed wetland system during the hot summer
months makes possible wastewater treatment from larger summer populations,
while still providing the required removal efficiency in the winter. This can be
a significant advantage over conventional systems in areas where the summer
DESIGN METHODOLOGY OF FREE WATER SURFACE CONSTRUCTED WETLANDS 553

temperatures are high and tourism can increase the permanent (winter) popula-
tion. FWS constructed wetland systems can accept higher loads during summer,
when the design is controlled by BOD or nitrogen removal, since in both cases the
removal efficiency depends on temperature.
In relation to BOD, if Equation (15) is not fulfilled under winter or summer
conditions, then PS /PW =1, where PW and PS are the winter and summer popula-
tions, respectively. If Equation (15) is fulfilled under winter but not under summer
conditions, then Equation (5) can be expressed as:

Ci Q S
AS = (16)
10
Similarly, Equation (1) for BOD removal under winter conditions can be ex-
pressed as:

ln(Ce /Ci ) Q W
AW = − (17)
K TW yW ϕ

where: TW is the winter design water temperature [◦ C]; yW is the winter flow depth
[m]; and K TW is the reaction rate parameter [d–1 ] for BOD removal (Table I) under
winter conditions. Since AS = AW , Equations (16) and (17) are set equal and the
resulting equation is solved for PS /PW :

PS 10 ln(Ce /Ci )
=− (18)
PW Ci K TW yW φ

If relation (15) is fulfilled under summer but not under winter conditions, then
similarly to the procedure used for the estimation of Equation (18) the ratio PS /PW
can be estimated from the following equation:

PS Ci K TS yS φ
=− (19)
PW 10 ln (Ce /Ci )

where: TS is the summer design water temperature [◦ C]; yS is the summer flow
depth [m]; and K TS is the reaction rate parameter [d−1 ] for BOD removal (Table I)
under summer conditions.
If Equation (15) is fulfilled under winter and summer conditions, then, similarly
to the procedure used for the estimation of Equation (18), the ratio PS /PW can be
estimated from the following equation:

PS K TS yS
= (20)
PW K TW yW

If the ratio of the actual summer to winter population exceeds the computed
ratio PS /PW for the given summer and winter design water temperatures, on the
554 M. A. ECONOMOPOULOU AND V. A. TSIHRINTZIS

basis of Equations (18) or (19) or (20), then summer conditions control the design.
Otherwise, winter conditions control the design and the constructed wetland can
serve a summer population up to PS .
The ratio of summer to winter population based on nitrogen removal for FWS
constructed wetland systems is not a simple expression. One way, however, to
compute this ratio is the following. On the basis of Equation (4), and assuming that
A, y and φ do not change, one gets:

PS tTN,W
= (21)
PW tTN,S

where tTN,W and tTN,S are winter and summer hydraulic residence times required
for nitrogen removal. Values for tTN,W and tTN,S can be computed from Figure 1,
as a function of winter and summer temperatures TW and TS , and the ratio PS /PW
can be computed from Equation (21).
For FWS systems, y is a function of Q (Equation (14)), and therefore, population.
In other words, the summer flow will be deeper if PS > PW . Since the summer
depth yS is larger than the winter depth yW , the summer value γ S of parameter γ
is lower than the winter value γ W of γ , because for the same bed slope (assuming
normal flow) it is:

γS yS = γW yW (22)

7. Sensitivity Analysis
7.1. FLOW DEPTH

The sensitivity of Equations (11)–(14) used for flow depth estimation is analyzed
as a function of the population served (Figure 3) and the ratio L:W (Figure 4).
Figure 3a presents the flow depth y [m] as a function of the population P [capita],
assuming L:W = 2:1, the unit daily BOD production β = 50 g capita–1 d–1 , the
unit flow rate q = 0.15 m3 capita–1 d–1 and t = 5 d. Figure 3b presents the flow
depth y [m] as a function of the population P [capita], assuming L:W = 5:1, the
unit daily BOD production β = 50 g capita–1 d–1 , the unit flow rate q = 0.15 m3
capita–1 d–1 and t = 5 d. The two graphs cover the common range of aspect ratio
L:W (between 2:1 and 5:1).
Graphs 4a, 4b and 4c in Figure 4 present the flow depth y [m] as a function of the
ratio L:W , assuming three populations served, namely P = 1000, 5000 and 10000
capita, respectively. For all three, β = 50 g capita–1 d–1 , q = 0.15 m3 /capita−1 d−1
and t = 5 d. As mentioned, in Equation (13), for sparse vegetation β 1 = 1 m1/6 · s
and for dense vegetation β 1 = 4 m1/6 ·s, and in Equation (14) for sparse vegetation
a = 5× 107 d–1 m–1 and for dense vegetation a = 107 d–1 m–1 .
DESIGN METHODOLOGY OF FREE WATER SURFACE CONSTRUCTED WETLANDS 555

Figure 3. Flow depth of the FWS constructed wetland system, as function of population P
[capita] for two aspect ratios L:W : (a) L:W = 2:1; and (b) L:W = 5:1.

From Figure 3 one can conclude the following:

1. The flow depth increases as population increases and as L:W increases.
2. For sparse vegetation [Equations (11), (13) with β 1 = 1m 1/6 ·s and (14) with a
= 5 × 107 d−1 m−1 ] and for dense vegetation [Equations (12), (13) with β 1 = 4
m1/6 ·s and (14) with a = 107 d−1 ·m−1 ], it is evident that the flow depth y
increases by approximately 50 to 70% when the vegetation changes from sparse
to dense. This conclusion is independent of the aspect ratio L:W , and it is valid
for a wide range of hydraulic residence time and population values. Therefore, in
designing the FWS constructed wetland system, the use of equations for sparse
vegetation would, based on this conclusion and Equation (4), yield a larger (more
conservative) area requirement for a given hydraulic residence time instead of
556 M. A. ECONOMOPOULOU AND V. A. TSIHRINTZIS

Figure 4. Flow depth of the FWS constructed wetland system, as function of aspect ratio L:W
for three population P values: (a) P = 1000 capita; (b) P = 5000 capita; and (c) P = 10000
capita.
DESIGN METHODOLOGY OF FREE WATER SURFACE CONSTRUCTED WETLANDS 557

those for dense vegetation. Or, in reverse, a certain system’s performance would
increase as the vegetation becomes denser.
3. For the commonly used range of L:W and for sparse vegetation the highest flow
depths are estimated on the basis of Equation (14) (a = 5 × 107 d−1 m−1 ); the
lowest flow depths are estimated on the basis of Equation (11) for population
more than about 2000 capita (L:W = 5:1) to 4000 capita (L:W = 2:1), otherwise
by Equation (13) (β 1 = 1 m1/6 ·s). For dense vegetation and L:W = 2:1, the
highest depth is predicted on the basis of Equation (12); for L:W = 5:1, the
highest depth is predicted on the basis of Equation (12) up to a population of
about 4000 capita and Equation (13) (β 1 = 4 m1/6 ·s) for higher populations. The
lower flow depth is estimated on the basis of Equation (14) (a = 107 d−1 m−1 )
for a population greater than about 3500 capita (L:W = 2:1) or 1250 capita
(L:W = 5:1), otherwise by Equation (13) (β 1 = 4 m1/6 ·s).
From Figure 4 one can conclude the following:
1. For sparse vegetation, the highest flow depth is estimated on the basis of Equation
(14) (a = 5× 107 d−1 m−1 ) for the entire population range; for dense vegetation,
the highest flow depth is estimated on the basis of Equation (12) for P = 1000
capita, and on the basis of Equation (13) for large populations and aspect ratios.
2. The flow depth increases as the ratio L:W increases. Therefore, increased values
of L:W result in increased values of y, which, based on Equation (4), would
yield a smaller area requirement (more economic design) for a given hydraulic
residence time. Nevertheless, as mentioned, it is not recommended to use L:W >
5:1, and increased depths can be achieved with water surface control at the outlet
of the system.
3. A system serving a larger population would be more economic, in terms of
unit area (m2 /capita) requirements, because of the increased depth (see also the
following Section 7.2).
From this sensitivity analysis of the alternative equations used for the estimation
of the flow depth it is evident that differences in flow depth result in differences in
the required area, depending on the selected equation. Thus, one can conclude that
the research on flow resistance in free water surface constructed wetland systems is
not yet complete and probably the more conservative design should be employed.
In this study, Equation (14) was selected for use in the following sections, as this
equation estimates, for dense vegetation, relatively reduced values of flow depth
and consequently relatively increased values of the required area, which leads to a
conservative design. Another advantage of this equation is that it can be used for
the estimation of the wetland’s flow depth without time-consuming iterations.

7.2. WETLAND UNIT AREA

For a FWS constructed wetland system four typical performance criteria can be
recognized, which correspond to different disposal options:
558 M. A. ECONOMOPOULOU AND V. A. TSIHRINTZIS

Figure 5. Unit area requirements of FWS constructed wetlands, depending on performance

criteria I, II, III and IV, assuming β = 50 g capita–1 d–1 , and treating: (a) weak; (b) typical;
and (c) strong domestic wastewater.
(Continued on next page )
DESIGN METHODOLOGY OF FREE WATER SURFACE CONSTRUCTED WETLANDS 559

Figure 5. (Continued)

• Performance criterion I corresponds to a BOD effluent concentration of 30 mg

L−1 (e.g., for effluents discharged into rivers);
• Performance criterion II corresponds to a fecal coliform effluent concentration
of 1000/100 mL (e.g., for effluents used for irrigation or discharged near the
coastline);
• Performance criterion III corresponds to a total nitrogen effluent concentrations
of 5 mg L−1 (e.g., for effluents discharged into lakes); and
• Performance criterion IV corresponds to a phosphorus effluent concentration of
3 mg L−1 (e.g., for effluents discharged into lakes).
The graphs 5a, 5b and 5c in Figure 5, yield the unit area requirements [m2
capita–1 ] of FWS constructed wetland systems designed to meet the above defined
performance criteria I, II, III and IV, as a function of the water temperature T , based
on treatment of weak, typical and strong municipal wastewater, respectively. Weak
wastewaters have the following characteristics: influent BOD, total nitrogen and
phosphorus concentrations of 120 mg L−1 , 12 mg L−1 and 5 mg L−1 , respectively.
For typical wastewaters, the influent BOD, total nitrogen and phosphorus concen-
trations are 330 mg L−1 , 50 mg L−1 and 11 mg L−1 , respectively. Finally, for strong
wastewaters, the influent BOD, total nitrogen and phosphorus concentrations are
550 mg L−1 , 90 mg L−1 and 15 mg L−1 , respectively. In all cases the total coliform
560 M. A. ECONOMOPOULOU AND V. A. TSIHRINTZIS

number is 108 /100mL, the daily BOD contribution is β = 50 g capita−1 d−1 and the
vegetation porosity φ = 0.65. For the computation of the unit area requirements,
and for all performance criteria, the maximum organic loading limit was set equal
to 10 g m−2 d−1 .
Figure 5 can help designers in analyzing the design as a function of the
wastewater type treated, the design temperature and the performance criteria I,
II, III and IV. From this figure, one can draw the following conclusions:
1. For BOD and nitrogen removal (performance criteria I and III), the required unit
area increases as population increases.
2. For BOD removal (performance criterion I), the required unit area tends to be-
come independent of temperature for high-design temperatures and large pop-
ulations served, as the design of the wetland is controlled by the maximum
organic limit of 10 g m−2 d−1 (Equation 15).
3. For BOD and coliform removal (performance criteria I and II) of strong wastew-
aters, the design of the FWS constructed wetland system is controlled by per-
formance criterion I, as performance criterion II is fulfilled for any T.
4. For coliform and phosphorus removal (performance criteria II and IV) of typical
to strong wastewaters, the design of the FWS constructed wetland system is
controlled by performance criterion IV, as performance criterion II is fulfilled
for any T .
5. For coliform and phosphorus removal (performance criteria II and IV) of weak
wastewaters, the design of the FWS constructed wetland system is controlled
by performance criterion II, as performance criterion IV is fulfilled for any T.
6. For BOD, coliform, phosphorus and nitrogen removal (performance criteria I, II,
III and IV) of typical to strong wastewaters, the design of the FWS constructed
wetland system is controlled by performance criterion III, as performance criteria
I, II and IV are fulfilled for T < 25 ◦ C.
Figure 5 can help designers in formulating an overview of the system’s behavior
relative to the wastewater type treated, the design temperature and the performance
criteria I, II, III and IV. From Figure 5 one can make the following simplifications
in the design procedure presented above:
1. For BOD and coliform removal (performance criteria I and II) of strong
wastewaters, Steps 3, 6 and 7 can be omitted since the design of the FWS
is based on BOD removal. When the area requirements are constant, indepen-
dent of temperature, then the conditions controlling the design of the wetland are
determined on the basis of maximum population between summer and winter.
If the area requirements for BOD removal is function of the design temperature,
then the conditions controlling the design of the wetlands are determined on the
basis of Equations (18) or (19) or (20) or (21).
2. For coliform and phosphorus removal (performance criteria II and IV) of typical
to strong wastewaters, steps 3, 6 and 7 can be omitted, since the design of the FWS
DESIGN METHODOLOGY OF FREE WATER SURFACE CONSTRUCTED WETLANDS 561

Figure 6. Unit area requirements (performance criteria I and II) of FWS and SF constructed
wetland systems, and stabilization pond systems, assuming β = 50 g capita–1 d–1 and BOD
influent concentrations of 120 mg L–1 , 330 mg L–1 , and 550 mg L–1 .

constructed wetland system is based on phosphorus removal. The conditions

controlling the design of the wetland are determined on the basis of maximum
population between winter and summer.
3. For coliform and phosphorus removal (performance criteria II and IV) of weak
wastewaters, steps 5, 6 and 7 can be omitted, since the design of the FWS
constructed wetland system is based on coliform removal. The conditions con-
trolling the design of the wetland are determined on the basis of maximum
population between winter and summer.
4. For BOD, coliform, phosphorus and nitrogen removal (performance criteria I,
II, III and IV) of typical to strong wastewaters and for T < 25 ◦ C, steps 2,
3, 5 and 6 can be omitted, since the design of the FWS constructed wetland
system is based on nitrogen removal. The conditions controlling the design of
the system are determined on the basis of maximum hydraulic residence time
value for nitrogen removal, and population served between winter and summer
(Equation 21).
562 M. A. ECONOMOPOULOU AND V. A. TSIHRINTZIS

8. Comparison of FWS Constructed Wetlands with Other Aquatic Natural

Systems
To guide planners in a preliminary selection between SF and FWS constructed
wetland systems and stabilization pond systems that provide secondary treatment
(BOD concentrations and number of fecal coliform effluent less than 30 mg L−1 and
1000/100 mL, respectively, i.e., performance criteria I and II) the unit area require-
ments are compared for various influent BOD concentrations. Figure 6 presents
the unit area requirements for these three aquatic natural systems for BOD influ-
ent concentrations Ci ranging from 120 mg L−1 to 550 mg L−1 , coliform influent
concentration equal to 108 /100 mL and daily BOD contribution equal to β = 50 g
capita–1 d–1 .
The design of the FWS system is based on the procedure presented above as-
suming P = 1000 capita and φ = 0.65. The design of the SF system is based on
the design procedure presented by Economopoulou and Tsihrintzis (2003), assum-
ing y = 0.6 m and φ = 0.38. The design of the stabilization pond system, which
comprises one anaerobic, one facultative and several maturation ponds, is based on
the design procedure presented by Economopoulou and Tsihrintzis (2004), which
estimates the optimum area of the system designed to meet the desired performance
criteria. For the estimation of the optimum stabilization pond area the anaerobic,
facultative and maturation pond depths are assumed equal to 3 m, 1.5 m and 1.5 m,
respectively.
From Figure 6 it is evident that constructed wetland systems are favored over
stabilization ponds for high BOD influent concentrations, small differences between
summer and winter populations and low design temperatures. Thus, in developing
countries where Ci is high, and in regions with temperate or cold climate and
nearly constant population throughout the year, constructed wetland systems should
always be considered as an alternative control system.

9. Application Example
The following illustrative design example is based on the steps described in Section
5.1: A coastal city in the Mediterranean has a winter population PW of 2000 capita.
The unit wastewater flow q is 0.15 m3 capita−1 d−1 (therefore, Q is 300 m3 d−1 ),
the design winter air temperature Ta,W is 6 ◦ C. The BOD concentration of the
untreated influent is 330 mg L−1 (about 50 g capita−1 d−1 ) and after preliminary
treatment the BOD concentration is reduced to 220 mg L−1 . The fecal coliform
number is 108 /100 mL. Total nitrogen and phosphorus influent concentrations are
30 and 7 mg L−1 , respectively. The performance criteria only consider BOD and
coliforms. The required BOD effluent concentration is 30 mg L−1 , whereas the
effluent number of fecal coliforms should not exceed 1000/100 mL. The design
summer air temperature Ta,S is 18 ◦ C and the summer population rises to 2500
capita. Under these design conditions size the FWS constructed wetland system
that provides the required removal efficiency throughout the year.
DESIGN METHODOLOGY OF FREE WATER SURFACE CONSTRUCTED WETLANDS 563

The solution is according to the following steps:

1. From Equation (14), y = 0.173 m given Q = 300 m3 d−1 and assuming
L:W = 3:1, γ = 0.1 and a = 107 m−1 d−1 . Relation (15) is fulfilled under
winter conditions, assuming φ = 0.65.
2. Since PS > PW then from Equations (14) and (22) γ S = 0.09 and yS = 0.183
m. Relation (15) is fulfilled under summer conditions, assuming φ = 0.65. The
actual ratio PS /PW is smaller than the ratio estimated from Equation (20), thus
winter conditions control the design of the wetland for BOD removal. Given
Ce /Ci = 0.14 and Ta = 6 ◦ C, Equation (1) yields tBOD = 6.64 d, which from
Equation (4), results in ABOD = 17, 704 m2 .
3. Since PS > PW , then summer conditions control the design of the wetland for
coliform removal. Hence, on the basis of Equation (2) and the appropriate values
of the reaction rate constant listed in Table I, one estimates (h l )−1 = 38.46 d
m−1 , and Equation (4), under summer conditions, yields ACOLI = 14, 391 m2 .
4. The value of ABOD , for BOD removal in winter, is higher than the value of ACOLI ,
for coliform removal in summer, therefore, BOD removal controls the design
under winter conditions.
5. For L:W = 3:1, the length L = 230.5 m whereas the width W = 76.8 m.
6. The effluent concentrations for nitrogen and phosphorus estimated from
Figure 1 and Equation (1) under winter conditions are: 22.6 mg L−1 and
1.3 mg L−1 , respectively. Similarly, the effluent concentrations for nitrogen
and phosphorus under summer conditions are: 9.5 mg L−1 and 1.9 mg L−1 ,
respectively.

10. Summary and Conclusions

A simplified step-by-step approach was presented allowing for the design of new
and performance evaluation of existing FWS constructed wetland systems. The area
requirements are based on the removal of BOD, coliforms, nitrogen and phosphorus.
Alternative equations are presented for hydraulic computations. The methodology
was validated through the comparison of estimated surface areas with areas of
existing systems.
An overview of the FWS system’s behavior over a wide range of conditions was
offered by sensitivity analysis of the unit area requirements for different types of
wastewaters and performance criteria. The results of this analysis simplify consider-
ably the design methodology. In addition, for tourist areas with increased summer
populations, simple equations were derived for the estimation of the maximum
seasonal population that can be served considering the system’s performance and
capacity dependency on the seasonal climatic conditions. In addition, the sensitivity
of the unit area requirements of SF and FWS constructed wetlands, and stabilization
ponds was presented. This analysis can facilitate enforcement agencies in the pre-
liminary screening of existing or perspective constructed wetland systems based on
564 M. A. ECONOMOPOULOU AND V. A. TSIHRINTZIS

their performance, or help environmental planners in properly evaluating alternative

wastewater control and disposal options.

Acknowledgments
The first author gratefully acknowledges the financial support from the National
Scholarship Foundation (I.K.Y.) of Greece.

References
Bachand, P. A. M. and Horne, A. J., 2000a, ‘Denitrification in constructed free-water surface wetlands:
I. Very high nitrate removal rates in a macrocosm study’, Ecol. Eng. 14, 9–15.
Bachand, P. A. M. and Horne, A. J., 2000b, ‘Denitrification in constructed free-water surface wetlands:
II. Effects of vegetation and temperature’, Ecol. Eng. 14, 17–32.
Chen, S., Wang, G. T., and Xue, S. K., 1999, ‘Modeling BOD removal in constructed wetlands with
mixing cell method’, J. Environ. Eng. 125(1), 64–71.
Dialynas, G. E., Kefalakis, N., Dialynas, E. G., and Angelakis, A. N., 2002, ‘Performance of the first
free water surface constructed wetland in Crete, Greece’, in Proceedings of the IWA Regional Sym-
posium on Water Recycling in Mediterranean Region, 26–29 September, Iraklio, Crete, Greece,
pp. 371–378.
Economopoulou, M. A. and Tsihrintzis, V. A., 2000, ‘Hydraulic design of constructed wetlands’, in
Proceedings of the 8th National Conference of the Hellenic Hydrotechnical Society, 19–22 April,
Athens, Greece, pp. 235–242 (in Greek).
Economopoulou, M. A. and Tsihrintzis, V. A., 2003, ‘Design methodology and area sensitivity analysis
of horizontal subsurface flow constructed wetlands’, Water Resour. Manage. 17(2), 147–174.
Economopoulou, M. A. and Tsihrintzis, V. A., 2004, ‘A Rapid method for design optimization and
performance evaluation of stabilization ponds’, (submitted).
Etnier, C. and Guterstam, B. (ed), 1997, Ecological Engineering for Wastewater Treatment, 2nd edn.,
Lewis Publishers, Boca Raton, Florida, U.S.A.
Hammer, D. A. (ed), 1989, Constructed Wetlands for Wastewater Treatment, Municipal, Industrial,
Agricultural, Lewis Publishers, Boca Raton, Florida, U.S.A.
Hammer, D. A., 1992, Creating Freshwater Wetlands, Lewis Publishers, Boca Raton, Florida,
U.S.A.
Kadlec, R. H., 1990, ‘Overland flow in wetlands: Vegetation resistance’, J. Hydraulic Eng., ASCE
116(5), 691–706.
Kadlec, H. R. and Knight, R. L., 1996, Treatment Wetlands, Lewis Publishers, Boca Raton,
U.S.A.
Khatiwada, N. R. and Polprasert, C., 1999, ‘Kinetics of fecal coliform removal in constructed wet-
lands’, Water Sci. Technol. 40(3), 109–116.
Mander, U. and Jenssen, P., 2002, Natural Wetlands for Wastewater Treatment in Cold Climates, WIT
Press, Southampton, U.K.
Mander, U. and Jenssen, P., 2003, Constructed Wetlands for Wastewater Treatment in Cold Climates,
WIT Press, Southampton, U.K.
Mandi, L., Bouhoum, K., and Ouazzani, N., 1998, ‘Application of constructed wetlands for domestic
wastewater treatment in an arid climate’, Water Sci. Technol. 38(1), 379–387.
Minnesota Board of Water and Soil Resources, 1993, Administration Manual for the Minnesota
Wetland Conservation Act, Wetland Conservation Act of 1991’, The Minnesota Association of
Soil and Water Conservation Districts, St. Paul, MN, USA.
 DESIGN METHODOLOGY OF FREE WATER SURFACE CONSTRUCTED WETLANDS 565

G. A. Moshiri (ed), 1993, Constructed Wetlands for Water Quality Improvement, Lewis Publishers,
Boca Raton, Florida, U.S.A.
Mitsch, W. J. and Gosselink, J. G., 1986, Wetlands, Van Nostrand Reinhold, New York, NY, U.S.A.
Mulamoottil, G., McBean, E. A., and Rovers, F., 1999, Constructed Wetlands for the Treatment of
Landfill Leachates, Lewis Publishers, Boca Raton, Florida, U.S.A.
Olson, R. K., 1993, Created and Natural Wetlands for Controling Nonpoint Source Pollution, C.K.
Smoley-CRC Press, Boca Raton, Florida, U.S.A.
Reed, S. C. and Brown, D., 1993, ‘Constructed wetland design – The second generation’, Natural
System Digest, Water Environment Federation Digest Series.
Reed, S. C., Crites, R. W., and Middlebrooks, E. J., 1995, Natural Systems for Waste Management
and Treatment, 2nd edn. McGraw-Hill, Inc, New York, U.S.A.
Reilly, J. F., Horne, A. J., and Miller, C. D., 2000, ‘Nitrate removal for a drinking water supply
with large free-surface constructed wetlands prior to groundwater recharge’, Ecol. Eng., 14,
33–47.
Sartoris, J. J., Thullen, J. S., Barber, L. B., and Salas, D. E., 2000, ‘Investigation of nitrogen trans-
formations in a Southern California constructed wastewater treatment wetland’, Ecol. Eng. 14,
49–65.
Schueler, T. R., 1992, Design of Stormwater Wetland Systems: Guidelines for Creating Diverse and
Effective Stormwater Wetland Systems in the Mid-Atlantic Region, Department of Environmental
Programs, Metropolitan Washington Council of Governments.
Tchobanoglous, G. and Burton, F. L., 1991, Wastewater Engineering: Treatment, Disposal and Reuse,
Metcalf & Eddy, Inc., McGraw-Hill, Inc., New York, NY, U.S.A.
Tsihrintzis, V. A., 2001, ‘Discussion on variation of roughness coefficients for unsubmerged and
submerged vegetation’, J. Hydraulic Eng., ASCE 127(3), 241–244.
Tsihrintzis, V. A., John, D. L., and Tremblay, P. J., 1998, ‘Hydrodynamic modeling of wetlands for
flood detention’, Water Resour. Manage. EWRA 12(4), 251–269.
Tsihrintzis, V. A. and Madiedo, E. E., 2000, ‘Hydraulic resistance determination in marsh wetlands’,
Water Resour. Manage. 14(4), 285–309.
Tsihrintzis, V. A., Vasarhelyi, G. M., and Lipa, J., 1995a, ‘Hydrodynamic and constituent transport
modeling of coastal wetlands’, J. Mar. Environ. Eng., Gordon and Beach Publishing Co., 1(4),
295–314.
Tsihrintzis, V. A., Vasarhelyi, G. M., and Lipa, J., 1995b, ‘Multiobjective approaches in freshwater
wetland restoration and design’, Water Int., IWRA 20(2), 98–105.
Tsihrintzis, V. A., Vasarhelyi, G. M., and Lipa, J., 1996, ‘Ballona Wetland: A multi-objective salt
marsh restoration plan’, Water, Mar. Energy, ICE, London, U.K. 118(2), 131–144.
USEPA (U.S. Environmental Protection Agency), 1988, Design Manual: Constructed Wetlands and
Aquatic Plant Systems for Municipal Wastewater Treatment. EPA/625/1-88/022. Cincinnati, OH,
Environmental Protection Agency.
WHO (World Health Organization), 1989, Health Guidelines for the Use of Wastewater in Agriculture
and Aquaculture, Technical Report Series 778, World Health Organization, Geneva (1989).

View publication stats