~ Pergamon Vol 40. No.3. pp. 291-300.

1999
Wal . SCI. Tech.
<OI9991AWQ
Pubhshed by ElsevierScience Ud
Printed in Great Britain. All rightsreserved
0273-1223199 520.00 + 0.00
PII: S0273-1223(99)00448-5

HYDRAULICS EFFICIENCY OF
CONSTRUCTED WETLANDS AND PONDS
1. Persson*, N. L. G. Somes** and T. H. F. Wong**
• DepartmentofHydraulics. Chalmers University. 41296 Gtneborg, Sweden
•• DepartmentofCivilEngineering. Monash University. 900 DandenongRoad.
CaufieldEast. Vic 3145. Australia

ABSTRACT

Constructed ponds and wetlands are widely used in urban design to serve a number of functions including
storm water management. The design of constructed wetlands for stormwater management involves a number
of multi-disciplinary inputs. Fundamental to their sustainable operation are the proper control of the
hydrologic regime of the wetland and optimal flow hydrodynamics within the wetland . Many of the problems
encountered in constructed wetlands can be minimised or avoided by good engineering design principles.
Poor wetland hydrodynamics are often identified as a major contributor to wetland management problems.
Ponds and wetlands with a high hydraulic efficiency are expected to promote full utilisation of the available
detention storage and near plug flow conditions. The shape and layout of urban ponds and wetlands are often
varied to suit the landscape and to satisfy aesthetic requirements as an urban water feature. These can be
achieved while maintaining an effect ive stormwater treatment outcome if steps are taken to ensure that the
hydrodynamic behaviour of the system is not severely compromised. A cons istent measure is required to
allow the effects of design features to be evaluated against this criterion. This paper introduces a new measure
for hydraulic efficiency that combines existing measures of flow uniform ity and effective volume . Case
studies are presented on the use of this measure to assess the effects of different pond and wetland shapes .
locations of inlet and outlet. botanical layouts and basin morphology on the flow hydrodynamics.
<0 1999 IAWQ Published by Elsevier Science Ltd. All nghts reserved

KEYWORDS

Detention period. flow hydrodynamics, ponds. stormwater, constructed wetlands

INTRODUCTION

Constructed wetlands and water pollution control ponds are becoming widely used for the treatment of
storm water and combined sewer overflows (CSO). These systems are surface flow detention systems which
are often collectively referred to as constructed wetlands. However, it is useful to provide some distinction
between ponds and wetlands as they do have significantly hydrologic and hydraulic characteristics and
promote different water quality treatment processes. Ponds are generally small artificial bodies of open
water with a small range of water level fluctuation . Emergent aquatic macrophytes are normally restricted to
the margins because of water depth, although submerged plants may occur in the open water. Constructed
wetlands are shallow detention systems, which regularly fill and drain and are typically extensively vegetated
with emergent aquatic macrophytes.

The design of constructed wetlands and ponds requires multi-disciplinary input involving biological and
ecological sciences, aquatic chemistry. engineering hydrology and flow hydraulics . In addition to stormwater
quality improvements, constructed wetlands and ponds are often used in urban design as water features to
291
292 J. PERSSON et al.

serve as part of the urban landform and to provide recreational amenities. Many such systems have been
constructed, often with insufficient consideration to their requirement for proper hydrological and hydraulic
des ign. Insufficient provision of storage volume and unsatisfactory hydrologic and hydrodynamic control are
the main causal factors for poor performance of constructed wetlands and ponds as water pollution control
facilities (Reed, 1995). These deficiencies have also resulted in many of these urban wetlands and ponds
becoming a long-term liability to the community.

In relation to best practice in wetland design, Wong et al. (1998) suggest that three principal design
components need to be addressed, ie.

I . hydrologic effectiveness ;
2. hydraulic efficiency; and
3. facilitation and optimisation of water quality treatment processes.

Although each of the above processes are in some way inter-related, a systematic design procedure would
address these three principal components in the above general order . The first design objective would be to
facilitate an optimal rate of capture and detention of stormwater runoff by the wetland, ie. to optimise the
hydrologic effectiveness of the wetland in accordance to the rainfall characteristics of the catchment (Wong
and Somes, 1995). The second design objective is to ensure that stormwater inflow into the wetland is well
distributed throughout the wetland by avoiding physical elements which may lead to the occurrence of short-
circu its and poor utilisation of the available detention storage . A high system hydraulic efficiency can be
achieved by proper defin ition of the shape and depth of the wetland or pond, and the loc ations and types of
inflow and outflow structures. Special flow divers ion features using vegetation. basin bathymetry design and
other hydraulic measures may have to be introduced to facilitate even distribution of flow throughout the
wetland or pond . Optimal hydrologic effectiveness and hydraulic efficiency provide the most appropriate
conditions for promoting the necessary biological and chemical processes of stormwater treatment.

Constructed ponds and wetlands arc often used in urban design as landscape amenities and landscape
architects are prominent designers of these systems. The shape and layout of urban ponds and wetlands need
to suit the landscape and satisfy aesthetic requirements as a urban water feature. These can be achieved
while maintaining an effective stormwater treatment outcome if steps are taken to ensure that the
hydrodynamic behav iour of the system is not severely compromised. A consistent and reliable measure of
hydraulic efficiency is necessary to enable wetland and pond des igns to be evaluated for their expected
hydrodynamic performance. Desired variations in their geometric properties, bathymetry, vegetation layout,
and locations and type of inlet and outlet structures to satisfy landscape design objectives need to be assessed
and optimised to achieve the best possible hydrodynamic conditions in these systems. This paper describes a
method for determining the hydraulic efficiency of a constructed wetland or pond . The method could be
used as a tool to compare existing ponds and wetlands, or as a design tool to compare the efficiencies of
different des ign alternatives prior to construction . The measure of hydraulic efficiency (A) is tested against a
wide range of hypothetical pond shapes and inlet/outlet locations and flow dynamics computed for a research
wetland in the field. Depth integrated, modeling techniques were used to determine tracer responses for a
number of combinations of pond and wetland design features and their hydraulic efficiencies evaluated.

BACKGROUND

The hydraulic efficiency of ponds and wetlands needs to reflect two basic features in the hydrodynamic
performance of a stormwater detention system. The first is the ability to distribute the inflow evenly across
the detention system and the second is the amount of mixing or re-eirculation, ie. deviations from plug flow.
Figure I shows a number of tracer responses in a detent ion system and is used to illustrate these two basic
features . The figure shows three tracer concentration-time distributions in response to a spike tracer input
and steady flow conditions. All three cases have similar mean detention periods but will have significantly
different hydraulic efficiencies owing to their differences in the range of detention period experience by
individual parcels of tracer entering the system .
 Hydraulic efficiency of constructed wetlands and ponds 293

Kadlec and Knight (1996) describe a distribution function of hydraulic residence time, referred to as the
Retention Time Distribution (RTD) function to reflect the degree in which the hydraulic residence time
varies. Under plug flow conditions, the concentration-time distribution is simply a spike with a very small
standard deviation about the mean residence time as shown in Figure I. This suggests that all individual
parcels of tracer entering the wetland experience a similar period of detention. For continuously stirred flow
condition, the concentration-time distribution takes the form of an exponential function where the effect of
flow dilution in steady flow conditions progressively reduces the tracer concentration at the outflow.

....-
III ----VIQ

•
1•
g
~

Tbw..

Figure I TIlustration of Tracer Concentration-Time Distribution

Plug or continuously stirred flow conditions never occur in natural systems and the concentration-time
distribution of natural wetland systems lies somewhere in between the distributions of plug flow and fully
mixed flow conditions. According to Kadlec and Knight (1996), flow hydrodynamics within a wetland
system may be modelled as a combination of plug flow (ie. a time delay before tracer outflow is observed)
and a number of continuously stirred tanks reactors (CSTRs). A single CSTR will result in a pollutant
hydraulic residence time distribution represented by an exponential function while plug-flow condition is the
result of the number of CSTRs in series approaches infinity. The concentration-time distribution takes the
form of a positively skewed distribution function with the tail of the distribution extending as flow condition
for the entire detention system approaches fully mixed condition. The extent to which flow conditions depart
from an idealised plug flow condition is reflected in the spread of the distribution function. Generally, an
outflow concentration distribution with a large standard deviation suggest the presence of short-circuit flow
paths and flow re-circulating zones. In some cases. the combined effect of short-circuit flow paths and re-
circulating zones can result in the outflow concentration-time distribution exhibiting multiple peaks, or in
other cases in a flat extended peak.

Tracer response curves for several surface flow wetlarrds were examined by Kadlec and Knight (1996) and
found them all to be reasonably represented by 3 CSTRs in series in spite of their different shapes. Wind
induced mixing was thought to be an important factor in reducing the number of CSTRs in series even in
systems with large length to width ratios. However. closer examination of the observed tracer response
curves clearly showed some of the more regularly shaped wetlands to have promoted better tracer residence
time distributions and that it is possible to delineate systems with higher hydraulic efficiencies in the group
examined by Kadlec and Knight (1996).

Figure 1 also shows that the mean detention periods of the three concentration-time distributions were less
than the nominal detention period, computed as the ratio of the volume over the discharge (ie. V/Q). This is
attributed to not all the available storage volume has been utilised in the detention of the tracer. ie. the
effective volume is less than the nominal storage volume. A large difference between the observed mean
detention period and the nominal detention period again suggests the presence of zones of stagnation (ie.
ineffective detention zones) in the system.
294 J. PERSSON et al.

It is evident that both near-plug flow and effective volume utilisation conditions are necessary to promote
good hydraulic efficiency. Detention systems featuring near-plug flow conditions alone may not reflect good
wetland design if the presence of a dominant short-circuit flow path results in the majority of the pollutant
being rapidly transported to the outlet, albeit in a near-plug flow manner. While all parcels of pollutants
experiences similar detention periods, these detention periods are significantly less that what would have
been the case had all the storage volume been utilised in the detention of the pollutants. Similarly, systems
that yield a mean detention period close to the nominal detention period (VIQ) but with a flat concentration-
time distribution are also exhibiting poor hydraulic efficiency. In such cases, the individual parcel of
pollutants experiences highly varied detention periods about the nominal detention period.

MEASURES OF FLOW HYDRODYNAMICS

Current measures of hydraulic efficiency are directed at one of the two features discussed above with none
combining the two hydraulic performance criteria. Some common measures of the shape of the pollutant
hydraulic residence time distribution involve the computation of the mean detention period and the variance
(cr) or standard deviation of the hydraulic residence time distribution, ego

(I)

where E(t) is the pollutant hydraulic residence time distribution;

tmean is the mean detention time but is often assumed to be the same as the nominal
detention period (to =V/Q)

Others have used variations of the above, including the use of a dispersion number (D) as defined in the
Wehner-Wilhelm equation to characterise the degree of non-ideal flow conditions within a detention system.
A dispersion number of zero suggests plug flow conditions and the number approaches infinity for fully
mixed conditions. There are a number of ways, in which the dispersion number can be computed, many of
which utilise some measure of the variance of the hydraulic residence time distribution. For example,
Levenspiel (1972) provided the following expression relating the dispersion number to the coefficient of
variation (alt mean ) of the pollutant hydraulic residence time distribution, ie.

(2)

According to Fogler (1992), the number of CSTRs in series is simply the inverse of the square of the
coefficient of variation of the pollutant hydraulic residence time distribution, ie.

N=lt: I (3)

The above measures require a quantification of the range of detention times within a wetland to allow
calculation of the mean and standard deviation of the tracer response. The retention time distribution (RTD)
is quantified by either field tracer study or numerical simulation of the passage of a conservative tracer
through the wetland or pond. Many field tracer responses are characterised by positively skewed
distributions with long tails and, as a result, the calculation of means and standard deviations of the RID can
vary significantly depending on the selected end point of the field measurement. In field studies, the end
point is either defined by the detection limit of the tracer or a pre-specified minimum tracer recovery
criterion, with unaccounted tracer assumed to be lost in the detention system. Numerical models can
preserve the mass balance of the tracer input with the results often characterised by extremely long tails. It is
 Hydraulic efficiency of constructed wetlands and ponds 295

therefore necessary to arbitrarily select an endpoint . the position of which can affect the statistics of the
distribution particularly the variance significantly .

To overcome these shortcomings, a number of methods have been developed to define the mean and
standard deviation, ie.

l mean = Iso (4)

and
(J =.!.(
2
184 - 1 16 )
Iso
(5)

th
where t84, tse and tis are the 84th , 50'h and 16 percentiles of the hydraulic residence time distribution.
Equation 5 assumes the tracer response is normally distributed, which is never the case in wetland and pond
systems. Kadlec and Knight (1996) suggested an alternative equation to calculate the number of CSTRs in
series. This was based on defining the difference in the time of the peak outflow concentration and the mean
detention time, ie.

N=_I_n_ (6)
In - I p

th
Ta and Brignal (1998) utilised the ratio of the 16 percentile detention time over the 50'h percentile detention
time as a measure of the extent of short-circuiting (S) in a pond system, ie.

s=~ (7)
t so

None of the above measures explicitly account for any loss in effective detention volume and merely
examine the shape of the pollutant hydraulic residence time distribution to determine the measure of flow
hydrodynamic conditions in the wetland or pond. Thackston et al. (1987), on the other hand. utilised the
ratio of the mean detention period over the nominal detention period to measure the effective volume ratio
utilisat ion of detention systems. ie.

t mean V.ffecrive
e=--=--- (8)
In VIOlaI

HYDRAULIC EFFICIENCY (A.)

A review of measures adopted by others to define flow hydrodynamics in ponds and wetlands found them to
inadequately described the two hydrodynamic criteria in the previous sections and illustrated in Figure I. It
is proposed that an alternative measure. based on the product of the effective volume ratio and a term
involving the equivalent number of tanks in series. be formulated . The later factor describes the shape of the
RTD, with N equal to 1 for continuously stirred flow and 00 for plug flow. This new measure of
hydrodynamic conditions in ponds and wetlands is referred to as the Hydraulic Efficiency (A.) and is
expressed as follows :

(9)

In the above expression. the first term defines the effective volume ratio of the detention system and the
second term involves the term N. which is simply the number of CSTRs in series. Both terms have a range
296 J. PERSSON et al.

of 0 to I providing equal weighting for effective volume and pollutant hydraulic residence time distr ibution
The resulting expression for hydraulic efficiency is simply the ratio of the time of the peak outflow
concentration to the nominal detention period (V/Q) . The measure can thus be readily derived from observer
outflow pollutographs of wetland and pond systems and does not have the problems associated with defining
the mean detention period (t mean ) .

Two- and three-dimensional models can often provide valuable insights into the hydrodynamic behaviour 01
open water systems. Two-dimensional depth-averaged models are commonly applied to simulate flow
patterns in shallow water bodies such as floodplains in flood invest igations. Their appropriate use ir
modelling flow hydrodynamics in wetland and pond systems, which have sign ificantly different flow
characteristics from floodplain flow cond itions during flood events, was investigated by Somes et al. (1996 :
and Bishop (1999) . That study found two-dimensional depth-averaged models to be capable of reproducing
observed flow pattern. defined by measured velocity vectors. in a research wetland provided careful attentior
is placed on modelling the diffusion process. Adamsson et at. (1999) , compared the results obtained from I
two-dimensional depth integrated model (MIKE 21). a three-dimensional model (Fluent) and a physica
model and found both numerical models were able to reproduce the hydrodynamics of the physical model.

SIMULATIONS OF POND SHAPES AND INLET/OUTLET CONFIGURATIONS

INFLUENCES

In a recent study. Persson (1999) investigated the influence of pond shape. inlet/outlet locations and
inlet/outlet type on the hydrodynamics of these systems. In that study, 13 h~thetical ponds were:
investigated as shown in Figure 2. Each of these systems had approximately 2700 m in volume and a depth
of 1.5 m. A two-dimensional depth integrated hydraulic model, MIKE-2I, was used to simulate the progress
through the system of a spike of conservative trace injected at the inlet. The result ing RIDs were then used
to compute various hydrodynamic measures such as the effect ive volume ratio and the amount of mixing.
Table I lists the results of the simulation including the corresponding values of A..

The cases considered may be categorised into the following three groups; (i) good hydraulic efficiency with
A. > 0.75; (ii) satisfactory hydraulic efficiency with 0.5< A. 5 0.75; and (iii) poor hydraulic efficiency where A
5..0.5. Table 2 lists the cases in these categories.

Table I Hydrodynamic Measures of Hypothetical Cases

Case S e I-lIN tp
A=-
tn

A 0.29 0.74 0.41 0.30

B 0.24 0.79 0.33 0.26
C 0.10 0.46 0.23 0.11
D 0.16 0.34 0.52 0.18
E 0.68 0.89 0.85 0.76
G 0.72 1.0 0.76 0.76
H 0.10 0.44 0.25 0.11
I 0.30 1.0 0.41 0.41
J 0.87 1.0 0.90 0.90
K 0.34 0.78 0.46 0.36
0 0.25 0.73 0.35 0.26
p 0.57 0.96 0.64 0.61
Q 0.50 0.93 0.64 0.59
 Hydraulic efficiency of constructed wetlands and ponds 297

At l G
[t1J 0

+ 0

f
Bf+ H

-EJ p -Eo t
cO 1+ + Q-B +
D

t r J+

KflJ
3+

t Figure 2
+ The thirteen pond shapes and configurations investigated.

Table 2 Ranking of hypothetical ponds according to A.

Category Cases
Poor Hydraulic Efficiency A, B, C, D, I, H, K & 0
Satisfactory Hydraulic Efficiency P&Q
Good Hydraulic Efficiency E, G, & J

Elongated pond shapes (Case J) or baffled systems (Case G) clearly provided very high hydraulic efficiency
although care needs to be applied in designing elongated shapes to ensure that the increased flow velocity
associated with the narrower cross section would not lead to resuspension and remobilisation of settled
material. Case E, which spread inflow across the wetland, also provided a high hydraulic efficiency.

The simulations showed that designs involving a length to wide ratio of 4: I or less, and with point inflow
and outflow, (Cases A, B, D, H & I) will not promote good hydraulic efficiency unless steps are taken to
evenly distribute the inflow across the width of the detention storage. The introduction of a small island in
front of the inlet (Case P) was found to more than double the hydraulic efficiency (compared to Case B) of
the system. The use of a submerged berm at the inlet, or a flow distribution inlet structure, was found to
have a similar effect as for the case of a small island near the inlet with the distributed inlet providing a clear
advantage. While no simulations were carried out for the case of a distributed outlet, it was suggested that a
single outlet point would not affect the hydraulic efficiency significantly.

L-shaped ponds with an effective length to width ratio of 3:1 was found to yield unsatisfactory hydraulic
efficiency with the result being only marginal better than the case of a rectangular-shaped pond with a 2: I
length to width ratio.

As expected, poor placement of the outlet (Case C) and length to width ratio (Case H) resulted in the lowest
hydraulic efficiency. Similarly, the location of an island at the side of the system (Case 0) does not improve
flow distribution in the system.
298 J. PERSSON et al.

SIMULAnONS OF WETLAND MORPHOLOGY AND VEGETATION INFLUENCES

Somes et al. (1998) undertook an investigation to examine available options to improve the hydraulic
efficiency of a research wetland in South Gippsland, Australia. In that study, a 2-dimensional hydraulic
model, MIKE-21, was used to simulate the effects of modifications to the wetland morphology and
vegetation layout on the flow hydrodynamics. The existing wetland was constructed on an old creek line and
this is reflected in its bathymetry, with a channel connecting the inlet and outlet. The bathymetry and
original planting of the wetland have limited the distribution of emergent macrophytes to the fringes of the
wetland. The model was calibrated against field measurement of flow velocities as reported by Somes et al.
(1997). In examining possible modification options, 5 cases were simulated in addition to the existing base
case. These were (i) fully vegetating the wetland; (ii) a series of aquatic benches (banded bathymetry) with
fringing vegetation; (iii) a meandering low flow channel (labyrinth bathymetry) and full vegetation; (iv)
aquatic benches (banded bathymetry and full vegetation); and (v) uniform depth and full vegetation.

Figure 4 plots the tracer responses for the simulations where the horizontal axis represents the time since the
injection of the tracer and the vertical axis plots the outlet tracer concentration. The vertical line at 3:40
represents the theoretical exit time of a tracer slug if ideal flow conditions (VIQ) occurred in the wetland.
The base case (Figure 4) indicates a significant proportion of the tracer exits well before the nominal
detention time. Table 3 lists the six cases and their corresponding values of the two terms used to define the
Hydraulic Efficiency (A.). The results clearly show that systems with fringing vegetation and natural
bathymetry have low effective volume (0.4). The natural creek bed and fringing vegetation was found to
have resulted in a preferential flow path along the central region of the wetland. How condition at this
central region exhibited relatively low flow mixing (or a high degree of near-plug flow conditions) with an
estimated number of CSTRs in series of 5.7. However, the combined effect of effective volume and flow
mixing resulted in the A. value of 0.32, being the lowest of the cases considered.

The option of using a meandering low flow channel (labyrinth bathymetry) and full vegetation improved the
effective volume but was found to have promoted two preferential flow paths as reflected in the presence of
two peak concentrations in the outflow pollutograph. This has led to a lower number of CSTRs in series but
the improvements to the effective volume was found to outweigh the increased mixing with a resulting A
value of 0.52.

16000
-ldelllFlow

V._
-+-BueCue
....... Op"on 1 • Na",,,1Bolhymel>y ..d FunVtjela1lOll -oc-- Op"on 2 • B..dedBolhymay ol F _ V.....1lOll
14000 ....... Opeen 3 • Labymlh BoIhym.try -.d FullV.,.ta1lOll __ OpROO 4 • B..&odBolhymay ol Full
..... Op_ 5 • Trap_,dol Bolhymelry ol FullV.aeta"oo

i 12000

10000
'1
~ 8000
U
b
~
6000
]
:3 4000

2000

0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40

Figure 4 Tracer response from MIKE-21 Modelling - Monash University Research Wetland
 Hydraulic efficiency of constructed wetlands and ponds 299

The simulat ions found modification to the vegetation layout within the wetland to be most effective in
improving flow hydrodynamic conditions . The option of full vegetation without associated morpholog ical
modifications was found to have doubled the base case hydraulic efficiency from 0.32 to 0.64. However, as
discussed by Somes et al. (1996), maintaining a sustainable botanical structure is not a simple matter and
requires particular attention to be given to the hydrologic regime control of the system and the matching of
vegetation type to the wetness gradient.

Further improvements to the hydraulic efficiency for a fully vegetated system can be achieved with
modification to the bathymetry to promote uniform flow conditions across the full cross section of the
wetland by either using a series of submerged aquatic benches or to shape the basin bathymetry as
trapezoidal cross sections to provide uniform flow depth throughout the system. As noted in Figure 4, the
shapes of the outlet pollutographs for the banded bathymetry and uniform depth with full vegetation cases
are similar and thus is expected to have similar hydraulic efficiencies. Both options gave high hydraulic
efficiency values of approximately 0.75 .

The derived values of the number of CSTRs in series (N) for the banded bathymetry and uniform depth fully
vegetated cases examined were 23 and 12 respectively and highlight the importance in providing an upper
and lower bound to the terms defining hydraulic efficiency. While the values of N are significantly different,
the influence of N on the shape of the RTD is exponential and at large N values, the improvements are often
marginal in spite of a large increase in N.

Table 3 Hydrodynamic Measures of Monash University Research Wetland

~
Case N

Base case 0.39 5.7 0.32

Full Vege tation 0.68 15 0.64
Banded Bathymetry &Fr inging Vegetation 0.48 5.0 0.38
Labyrinth Bathymetry & Full Vegetation 0.77 3.1 0.52
Banded Bathymetry & Full Vegetation 0.83 23 0.76
Uniform Depth & Full Vegetation 0.80 12 0 .74

CONCLUSION

A simple measure for defining the flow hydrodynamic characteristics of ponds and wetlands has been
developed to allow the effect of variations in the design of these systems to be evaluated. The need to
provide such measures stems from the recognition that ponds and wetlands are becoming widely adopted in
urban design to provide functions beyond stormwater treatment. Their shape and layout will be varied
depending on the landscape and the design features to meet aesthetic requirements. Insufficient
considerations of the effects of these changes on flow hydrodynamics have been a major cause of poor
performance of these systems. The Hydraulic Efficiency measure (A) was found to provide a good balance in
assessing the hydrodynamic performance of detention systems against the uniformity of flow and the
effective utilisation of the available detention storage volume.

REFERENCES

Adamsson, A., 1. Persson and S. Lyngfelt, (1999), Numerical Simulation and Large-Scale Physical
Modelling of Flow in a Detention Basin. paper submitted to the 8th International Conference on
Urban Storm Drainage, 30 August - 3 September 1999, Sydney, Australia.
Bishop, W. (1999), Modelling Hydrodynamics in Wetlands, MengSc thesis, Department of Civil
Engineering, Monash University, 101p.
300 1. PERSSON et al.

Fogler. H.S.,(1992), Elements of Chemical Reaction Engineering, 2nd Edition, Englewood Cliffs, NJ
Prentice-Hall.
Kadlec, Rand R Knight, (1996), Treatment wetlands, CRC Press, Lewis Publishers, 892p.
Persson, J, (1999), The Hydraulic Performance of Ponds of Various Layouts, paper submitted to Int. Jnl. 01
Urban Water, UK.
Levenspiel, 0.,1972, Chamical Reaction Engineering. 2nd Edition, New York: John Wiley & Sons.
Reed, S.C., Crites, R.W and Middlebrooks EJ, 1995, Natural Systems for Waste Management anG
Treatment, 2nd Edition, Mc Graw-HiII, New York.
Somes, N L G, Breen, P E and Wong, T H F, (1996), Integrated Hydrologic and Botanical Design 0
Stormwater Control Wetlands, Proceedings of the 5th International Conference on Wetland System!
for Water Pollution Control, Vienna, Austria, Vol I, pp 1II/4.1-1II/4.8, September 1996.
Somes, N L G, Bishop, W A and Wong, T H F, (1997), Numerical Simulation of Wetland Hydrodynamics.
Modsim97, Hobart, pp 385-390.
Somes, N L G, Persson, J, and Wong, T H F, (1998) "Influence of Wetland Design Parameters on the
Hydrodynamics of Stormwater Wetlands" Hydrastorm, Adelaide, 27 -30 September, 1998 pp 123·
128.
Ta, C T and Brignal, W J, (1998), Application of computational fluid dynamics technique to storage
reservoir studies, Water and Science Technology, Vol. 37, No.2, pp 219-226.
Thackston, E L, Shields, F D (Jr) and Schroeder, P R, (1987), Residence time Distributions of Shallow
Basins, J. of Environmental Engineering, Vol. 113, No.6, pp. 1319-1332.
Wong, T H F and Somes, N L G, (1995), A Stochastic Approach to Designing Wetlands for Stormwatei
Pollution Control, Water and Science Technology, Vol. 32, No. I, pp. 145-151.
Wong, T H F, Breen, P F, Somes, N L G and Lloyd, S D, (1998), Managing Urban Stormwater using
Constructed Wetlands, Industry Report 98n, Cooperative Research Centre for Catchmenl
Hydrology, November 1998, 4Op.