Content uploaded by Tom Burkitt
Author content
All content in this area was uploaded by Tom Burkitt on May 18, 2020
Content may be subject to copyright.
Assessing the environmental impact of two options for small-
scale wastewater treatment: comparing a reedbed and an
aerated biological filter using a life cycle approach
Andrew Dixon
a,
*, Matthew Simon
a
, Tom Burkitt
b,1
a
Life Cycle Design Group, School of Engineering, Sheffield Hallam University, Sheffield S1 1WB, UK
b
Oceans-ESU Limited, 4 Cumberland House, Greenside Lane, Bradford BD8 9TF, UK
Received 2 July 2002; received in revised form 28 January 2003; accepted 30 January 2003
Abstract
Different options for wastewater treatment have different performance characteristics and also different direct
impacts on the environment. These impacts occur over the whole life cycle of the treatment system. Natural wastewater
systems such as the reedbed may offer reduced whole life cycle environmental impacts compared to systems with more
technical and material sophistication. This paper describes a study of the life cycle impacts of two options for small-
scale wastewater treatment which are a horizontal flow reedbed system and a package bio-filtration plant. The study is
limited to impacts during the construction and operation phases. Energy use, CO
2
emission and solid emissions were
chosen as the environmental aspects. It was found that overall the reedbed and conventional systems were quite similar
in terms of life cycle energy use. Transport occurring during construction and operational maintenance was a key
contributor to energy use and CO
2
emissions. The environmental impact of the reedbed reduced if the soil excavated on
site was suitable for re-use in the infill. The effect on the results of the different assumptions made during the study was
checked using sensitivity analysis. This analysis also helped us to identify recommendations for reducing the life cycle
impact of the reedbed system.
#2003 Published by Elsevier B.V.
Keywords: Reedbed; Wastewater; Treatment; Life cycle assessment
1. Introduction
Wastewater treatment systems have been de-
signed to minimise the environmental impacts of
discharging untreated wastewater. Different op-
tions for wastewater treatment have different
performance characteristics and also different
direct impacts on the environment. Some systems
have high energy usage, some use materials that
have a high embodied energy (e.g. plastics) others
occupy a lot of land. If minimisation of environ-
mental impacts is one of the main functions of
wastewater treatment systems then they should be
designed so that their total impact on the environ-
* Corresponding author. Tel.: /44-114-225-3091; fax: /44-
114-225-3433.
1
Tel.: /44-1274-483-301; fax: /44-1274-483-003.
Ecological Engineering 20 (2003) 297 /308
www.elsevier.com/locate/ecoleng
0925-8574/03/$ - see front matter #2003 Published by Elsevier B.V.
doi:10.1016/S0925-8574(03)00007-7
ment is reduced; the whole life cycle of the system
must be considered. The life cycle approach to
environmental assessment and design is not stan-
dard practise employed in the UK water industry.
Natural wastewater systems such as the reedbed
may offer reduced whole life cycle environmental
impacts compared to systems with more technical
and material sophistication. Treatment of waste-
water using reedbeds is not as common as
conventional options such as aerated filter treat-
ment, although both can offer similar performance
in terms of discharge water quality.
This paper describes a study of certain environ-
mental impacts caused by the construction and
operation of an aerated biological filter treatment
system and a reed bed system using a life cycle
approach. The study has been undertaken with the
cooperation and assistance of an environmental
engineering firm and with further information
provided by their regular suppliers.
1.1. Reedbed systems for wastewater treatment
Reedbeds have been used for wastewater treat-
ment for some years in the UK, although not
extensively. Reedbed performance is sensitiveto
local conditions, wastewater characteristics and
design. There are many texts that describe reedbed
performance and design issues including Cooper
and Green (1995) and Worrall et al. (1998). The
value of reedbeds in terms of bio-diversity, public
amenity, habitat, and aesthetics is discussed in
other papers (Worrall et al., 1997;Shutes, 2001).
An extensive annotated bibliography of con-
structed wetlands research has been compiled by
Moerman and Muirhead (1994).
1.2. Life cycle assessment and wastewater systems
Life cycle assessment (LCA) is a method for
investigating the environmental impacts of a
product or system over its whole life cycle, it
includes extraction of raw materials, processing,
manufacture, use and end of life within its scope,
refer to the ISO14040 series for detailed informa-
tion. LCA has been used to explore the sustain-
ability of wastewater systems (Emmerson et al.,
1995;Dennison et al., 1998;Tillman et al., 1998;
Ashley et al., 1999;Mels et al., 1999;Brix, 1999;
Balkema et al., 2001 and others). Arguments in
these papers tend to support the recommendation
of Emmerson et al. (1995) that LCA should be
more widely used in the water industry as a
decision aid in environmental policy and in
environmental improvement activities. Brix
(1999) highlights the quantitative nature of LCA
and its deficiency in appreciating qualitativevalues
such as amenity, habitat and aesthetic considera-
tions, such issues are discussed in Worrall et al.
(1997) and Shutes (2001).
Brix (1999) conducted a preliminary review of
life cycle impacts of so called green alternatives to
conventional wastewater treatment and found he
was unable to judge the ‘greenness’ of the different
options based solely on the energy requirements
and nutrient recycling. Brix considered several
systems including the Stensund aquaculture sys-
tem, the Living machines extended aeration plant,
batch reactor, and oxidation ditches. Our study
focuses on just two options for small-scale waste-
water treatment and explores life cycle impacts of
materials, energy usage and transport. Gaterell
and Lester (2000) suggest that a majority of
environmental burdens associated with a range
of wastewater treatment scenarios are due to a
limited number of key system inputs and outputs,
such as energy use in operation.
Key points extracted from the papers on LCA
and water systems included:
.Results of LCA’s are sometimes unexpected or
difficult to anticipate.
.LCA’s are data intensive; they require a lot of
high quality data.
.The boundary conditions of an LCA must be
defined with care since they may havea
significant effect on the LCA results. Key issues
are spatial boundaries, time scale over which
life cycle comparison is made, scale at which
comparison is made and the level of detail that
the study goes into.
.LCA’s are necessarily specific and limited. They
relate to physical issues and do not account
fully for other spheres of value such as bio-
diversity, public amenity, habitat, and aes-
thetics.
A. Dixon et al. / Ecological Engineering 20 (2003) 297 /308298
2. Methods
2.1. Goal and scope of life cycle study
The study is based upon a hypothetical waste-
water treatment application put forward by an
environmental engineering company. The com-
pany has provided the study with information on
system design based upon their experience in
similar projects. The aim has been to complete a
comparative life cycle study of a reedbed system
and an equivalent conventional wastewater treat-
ment system based upon a simplified LCA
method. The study includes a life cycle inventory
assessment and a sensitivity analysis. It does not
include any evaluation of environmental impacts
such as assessing the relative importance of global
warming potential compared to land use. The
purpose of the study is to provide insights into
the environmental impact of reedbed system de-
sign, construction and operation for sewage treat-
ment and provide a comparison with an existing
and more conventional equivalent wastewater
treatment technology. The study has not followed
the full ISO 14041 but has been informed by that
standard. Economic factors are not considered in
this study.
2.2. Functions of the systems
The function of each system described in this
paper is to treat sewage effluent to an acceptable
discharge standard so that it may be discharged to
the natural environment e.g. a watercourse. Typi-
cally discharge consent to surface waters set by the
Environment Agency are for 10 mg/l BOD, 25 mg/
l suspended solids, and 5mg/l ammonia. The
functional unit has been defined as a number of
‘population equivalents’, a measure of daily waste-
water production. One population equivalent
(p.e.) is the same as a dry weather flow (DWF)
of 0.2 m
3
/day. Population equivalents havebeen
used in other LCA studies of wastewater treatment
plant (e.g. Roeleveld et al., 1997). The two systems
are compared at three different scales. In terms of
time scale, Lundin and Morrison (2002) recom-
mends that the sustainability of wastewater sys-
tems be considered over a period of 50 /100 years.
However, Emerson reports that 15 years is the
span considered as a useful life of a plant by
Anglian Water, although Emerson suggests this
might not extend to other UK utilities. The period
of comparison in this study has been set at 10 year
based upon the operational life of the oldest
reedbed designed by the company. There is
evidence to suggest that reedbeds can operate
effectively for significantly longer than this if
they are well designed and managed for example
the reedbed at Othfresen, Germany by Kikuth was
established in 1974. Longer term operation of reed
beds is dependent on sediment adsorption, litter
and peat accumulation and solids accretion rates.
Most uncertainty surrounds the longevity of the
Phosphorous treatment aspect of the reedbed.
Evidence from natural wetlands (Bavor et al.,
1995) suggest that P removal can continue for
several decades. Other immobilisable material
such as heavy metals may pose additional pro-
blems.
2.3. Contender product systems
The product systems represent two options that
the engineer would typically consider as conten-
ders for a particular treatment scenario. The
product systems to be studied are:
.Reed bed systems incorporating a reed bed and
septic tank (RBS),
.An aerated filter treatment unit (AF).
The design of the reedbed unit and choice of AF
are based upon meeting the same water quality
criteria and for the same functional unit (i.e.
population equivalents). The AF unit is only one
of many possible options for small-scale waste-
water treatment units however it is an option that
the engineer has selected in previous projects of a
similar type. Other treatment units or methods
would have different life cycle inventories and
environmental impacts.
2.4. Product system boundaries
In this study the focus is on materials, transport
and operation, the system boundaries are shown in
A. Dixon et al. / Ecological Engineering 20 (2003) 297 /308 299
Fig. 1. They include manufacture, installation and
operation of the systems but not end-of-life. The
physical boundaries of each treatment system are
defined by the inlet pipe as it enters each system
and the outlet pipe as it discharges to the environ-
ment (see Fig. 1). The ultimate disposal of sludge
from the treatment units is not considered
although the transport required to removeit
from the site to a place of disposal is included.
Only the direct impacts from using machines, such
as trucks, heat guns, pumps etc. is taken into
account. The full life cycle impact of those
machines is excluded. End-of-life is not considered
for any of the systems.
2.5. Environmental aspects
This study focuses on a limited range of
environmental aspects, these are:
.Energy use,
.CO
2
emissions,
.Solid emissions,
.Land use (as used directly by the product
systems).
Energy use and solid emissions (sometimes
termed as solid waste, depending on end-of-life
of material) are a standard measure in LCA
methodology. CO
2
emissions are included given
their relevance to the reed beds potential to act as a
carbon sink, they also enrich the embodied energy
data. Land use is unusual in an LCA but the direct
footprint is considered here given the obvious
difference in product system footprint between
the reedbed and conventional systems. Full LCA
studies involve a detailed assessment of many
environmental aspects that are not included within
this short study such as other emissions to air,
resource depletion and emissions to water. Em-
merson et al. (1995) and Lundin et al. (2000) have
carried out detailed studies of larger scale waste-
water treatment plant.
2.6. Life cycle data
Design data for the reedbed system has been
provided by an environmental engineering com-
pany. The technical sales departments of the
engineers regular suppliers were contacted to
provide data for the prefabricated treatment units.
Data concerning the embodied environmental
aspects of materials, transport use and other
processes was taken from SimaPro software (Pre
consultants, 2002) and inserted into a spreadsheet
model. Estimates based on experience, published
texts and discussions with the client are made for
items of data that could not be found elsewhere.
Such assumptions are described in the text and
their influence on the results is tested in the
sensitivity analysis. The following sections describe
the materials, construction and operation of the
two system types.
Fig. 1. Boundaries of life cycle study.
A. Dixon et al. / Ecological Engineering 20 (2003) 297 /308300
2.6.1. Outline of materials used in reedbed system
design
The reedbed system is a horizontal flow unit
incorporating a septic and reedbed connected by
uPVC pipes and an electric pump. The septic tank
unit is made of hot pressed GRP (glass reinforced
polyester) moulded sections, bolted together with
some corrosion resistant steel reinforcement. It is
laid on a poured concrete slab and the hole is
backfilled with concrete. The reedbed incorporates
a 1 mm LDPE liner (e.g. Monarflex) protected by
a layer of sand and a geotextile to prevent the
movement. The infill or growing medium is a
mixture of soil, sand and compost and is laid in the
excavation to a depth of between 500 /1000 mm
according to local conditions. Soil from the
excavation is re-used in the infill providing quality
is suitable, this is not always the case and is
dependent on local site conditions. A gravel-lined
trench is sited at the inlet extending for the whole
width of the bed. The outlet chamber is made up of
prefabricated concrete drainage pipe sections to a
height of around 1.5 m, founded on a poured
concrete slab. This chamber contains a flexipipe
water level control system operated by a galva-
nised steel chain. A precast concrete lid covers the
chamber, with a steel manhole penetrating the
cover for access. The reeds are grown offsite at the
nursery in the North East of England. They are
delivered by van and planted by hand. The
reedbed system is commissioned by priming the
bed with ‘fresh’ water. The commissioning process
takes 2/4 weeks depending on bed size. The
engineering company prefer to use locally avail-
able recycled aggregates and compost if possible.
Nevertheless, the worst case of virgin materials
throughout is assessed herein.
Life cycle data for most of these materials was
available from SimaPro, however it was necessary
to estimate data for the geotextile and the materi-
als associated with growing the reeds in the
nursery.
2.6.2. Outline of materials used in conventional
system designs
The conventional treatment unit used for the
smallest scale application is a 2 tank, conical
shaped sewage treatment system incorporating a
central aeration chamber and a surrounding
clarifier chamber. The unit is made out of GRP
(glass reinforced polyester) modules, each hand
moulded (spray lamination) and then bolted
together and contains galvanised steel reinforce-
ment. The manufacturers specify a 2.1-m diameter,
2.74-m unit that has an empty weight of 210 kg.
The unit contains polyethylene baffles and some
uPVC pipe-work. It is laid on a poured concrete
slab and the hole is backfilled with concrete.
The two larger units for 60 and 200 p.e. are
made out of GRP sections moulded by hand using
a spray laminating technique. The sections are
bolted together and reinforced with corrosion
resistant steel. The units include a primary settling
tank and an aeration tank, filled with polypropy-
lene rings, which act as a growing medium. The
aeration and pumping requirements of the system
are met by an externally sited electrically powered
‘blower’ or compressor. The tank is laid on a
poured concrete slab in an excavation which is
then backfilled with concrete.
The technical sales team for the larger treatment
unit manufacturers were only able to provide
rough estimates of the mass of their units. These
estimates were checked by against estimates using
the spatial dimensions provided by the manufac-
turer (e.g. diameter depth, thickness of material).
This was an area of uncertainty which is explored
in the sensitivity analysis.
2.6.3. Transport of materials
The engineering company typically sources its
materials from a distributor/supplier local to the
construction site. Ideally materials present on the
site will be used for construction, and materials
delivered to site are preferably recycled from other
applications. However some items are only avail-
able from a specialist distributor, who would have
to deliver over a longer distance. An estimate of
distance for delivery and return was used in each
case. In reality these distances vary considerably
depending on site location in relation to appro-
priate distributors. The significance of this uncer-
tainty was explored in the sensitivity analysis.
All materials were assumed to be delivered by 16
T truck with the exception of the reeds that were
assumed to be delivered by van ( B/3.5 T). The
A. Dixon et al. / Ecological Engineering 20 (2003) 297 /308 301
data for transport related emissions refer to a
typical 50% loading factor */meaning that a de-
livery vehicle is full on its outward journey and
empty on its return journey. It was assumed that
the prefabricated treatment units were all delivered
over a national distance (rather than local) even
though the units would in reality be routed
through a local distributor.
For the purposes of this study, it has been
assumed that 100% of the materials required for
construction of the reed bed are required off site.
Under many build conditions, a design utilising
onsite materials and more substantial excavation
into the ground would be adopted. Consequently,
the reed bed design compared here against the AF
unit sets a baseline as a probable ‘worst case’
scenario in terms of off-site material requirements.
The associated change in life cycle impacts asso-
ciated with re-using on site materials is explored.
2.6.4. Construction of treatment systems
The environmental impacts of the construction
stage include the energy and fuel use and emissions
from powered machines used in the construction
process. It has been assumed that excavated soil
will be removed from site is a solid emission and
sent to waste (e.g. landfill), this is the worst case
scenario for the reedbed system as mentioned in
the previous paragraph. The machines used on site
were JCB (or similar), heat gun (to weld the LDPE
liner), pump (to pump water in to commission
system).
There are a number of assumptions involved in
assessing the impacts of using the JCB, the
significance of which is explored in the sensitivity
analysis. It was assumed that the JCB would have
similar fuel consumption characteristics as a 16-T
diesel truck. Estimates of machine operation were
given in units of hours e.g. excavation of reedbed
would take approximately 1 day. Environmental
aspect data for operation of the truck was in units
of kg CO
2
per tonne.km (t.km). A conversion
factor was used to relate hours of JCB usage to
t.km (the default value was set at 1 h /10 t.km).
The high uncertainty of estimating a value for this
factor was explored in the sensitivity assessment.
Estimates with a higher degree of certainty were
made for the power rating of the heat gun and
commissioning pump.
2.6.5. System operation and maintenance
The reedbed system requires modest mainte-
nance once it is up and running. The main task is
emptying the septic tank, the frequency depends
upon on loading rates and is typically three times a
year. Other tasks include examining plant health,
removing weeds, unblocking pipes and taking
water samples for analysis. Non-specialist labour
is suitable for the majority of these tasks. It is
assumed that sludge is taken by tanker to a local
sewage treatment works for disposal. The engi-
neers offer a service of site visits during the first
few years of operation of the RBS, to advise on
operational aspects and maintenance require-
ments. The pump, a 0.3 kW for the small scale,
0.75 kW for the medium and 1.5 kW for the large
operates on demand and is estimated at the
equivalent of 2 h continuous running per day. It
is assumed that reeds are not harvested and the
bed is not disturbed throughout the 10-year
operation cycle. The need to carefully manage
the reedbeds is backed up by evidence from case
studies in the UK and overseas (Shutes, 2001).
Maintenance of the conventional systems ex-
tends to removing the sludge from the tanks and
maintaining mechanical parts of the system such
as the compressor. The smaller scale system
produces a reduced amount of sludge which only
needs to be cleared out once every 2 years, this is
due to the extended aeration function. The two
larger systems need to be de-sludged at 60 /90-day
intervals depending on loading rates. The com-
pressors operate 24 h a day, the energy rating of
the smallest system is 0.14 kW, the medium one is
0.55 kW, and the largest one is 0.917 kW.
2.6.6. Reeds as a carbon sink
The growing reeds act as a carbon sink, locking
away atmospheric carbon in their structure. The
rate at which they do this is estimated rate as 3.3
kg/m
2
/year with an accuracy of 9/15% (Vanyarkho
and Arkebauer, 1995). The sensitivity coefficient
for this variable is /1.3. This means a 10%
increase in CO
2
uptake leads to a 13% decrease
of CO
2
emitted during the life cycle.
A. Dixon et al. / Ecological Engineering 20 (2003) 297 /308302
2.6.7. Influence of scale on design and operation
The conventional and reedbed systems have
been compared at different scales (12, 60 and 200
p.e..) The reedbed has the same basic design for
each scale the only difference is a proportionate
increase in surface area. The septic tank remains
the same capacity for each size of system. The
surface area increase is related to the increase in
influent (roughly five times the population equiva-
lent). The depth of the bed remains the same at
each scale. In the spreadsheet the increase in scale
is modelled by applying a scale factor to the small
scale unit. The factor for the liners, mass of
materials in the reedbed itself and the number of
reeds is calculated by dividing the population
equivalent of the reedbed under consideration by
the original 12 p.e. bed (60 p.e. 60/12 /5, and 200
p.e. 200/12/16.67). The pump was estimated as
being the same as recommended by the manufac-
turer of the prefabricated AF for the 60 and 200
p.e. treatment plants. The outlet chamber (and
associated materials) was increased only slightly
for the larger scales, by a factor of 1.5. The design
of the largest scale (200 p.e.) reed bed system uses
two reedbeds operating in parallel, each with an
area of 500 m
2
to give a total area of 1000m
2
.
There are only minor design differences between
the AF systems of different scales. Manufacturers
published technical specifications provided pump
power consumption and enough information to
estimate excavation volume for each scale. The
treatment system for the smallest scale AF system
is slightly different in design to the two larger scale
systems as it does not include the BioRings filter
media. However, its basic operation is the same
and the materials used in its construction are the
same.
2.6.8. UK energy mix
The UK energy mix was based on data from the
UK Department of Trade and Industry (2002) and
comprises 30% coal, 40% natural gas, 23% nuclear
and the remaining 7% is imports, renewables and
oil. Emissions from UK energy generation were
rated at 232 g of CO
2
per kW h of energy used
based on SimaPro data. It was assumed that this
mix was constant for the period of comparison.
3. Results and discussion
A summary of the results is given in Tables 1 /3.
All results relate to 10 years of system operation.
Transport refers to transport of materials. Trans-
port to and from site for the purposes of main-
tenance or sludge removal are included under
operation. The results should be considered in
the context of the various assumptions made in
carrying out the study.
3.1. CO
2
emissions
The most significant contribution to CO
2
emis-
sions of the reedbed systems is from the vehicles
used in the maintenance visits during the operation
phase, accounting for over 20% of the overall life
cycle CO
2
from the RBS at 12 p.e. The CO
2
embodied in the reedbed system materials, parti-
cularly the GRP septic tank unit and concrete
foundations accounts for 17% of total CO
2
at 12
p.e. At the smaller scale pump operation accounts
for 13% of emissions of the RBS compared to 71%
for the AF system. The action of the reedbed as a
carbon sink reduces its overall CO
2
emission to a
value below that of the equivalent scale conven-
tional system. At the 12 p.e. scale the RBS has a
significantly lower CO
2
emission than the AF and
at the two larger scales the RBS actually absorbs
more CO
2
than it emits, whilst the emissions of the
AF increase roughly in proportion to the scale.
The results for CO
2
for the conventional plant
are comparable to those of (Lundin et al., 2000)
who calculated an emission of 5.2 kg per p.e./year
compared to our result for 200 p.e. system of 12.2
kg/p.e. year. (2.0 kg for small, 7.6 kg for medium).
In both cases it is the CO
2
emissions due energy
consumed in operation phase that account for the
majority of the total emissions.
3.2. Energy utilisation
The life cycle energy utilisation of the two
systems follow similar patterns to the CO
2
emis-
sions. The energy embodied in the reedbed system
materials and used in the transport of those
materials (incl. soil waste) exceeds that of the
conventional systems at all scales (Fig. 2). How-
A. Dixon et al. / Ecological Engineering 20 (2003) 297 /308 303
ever, the conventional systems have significantly
higher energy utilisation in their operation, due to
the aeration and pumping equipment that they
employ (Fig. 3). The results concur with those of
Brix (1999) that energy requirements of con-
structed wetlands are typically low, based on a
light pump duty only. Gaterell and Lester (2000)
found that consumption of energy during the
operational phase of the life cycle was the
most significant environmental impact of the
systems that they investigated typical 1000 p.e.
systems).
3.3. Solid emissions
The main sources of solid emission are from the
soil excavated during construction and sludge
production during operation. Excavated material
is particular significant for the reed bed system.
The operation of the 12 p.e. conventional system
produces a reduced amount of sludge due to the
extended aeration technique that it employs. The
two larger AF units (60 and 200 p.e.) produce
significantly more sludge which contributes the
larger part of the total solid emission of these
system. End-of-life of the systems is not consid-
ered, Gaterell and Lester (2000) suggest that the
contribution of demolition phase to life cycle
environmental impacts is minimal.
Dennison et al. (1999) report that disposal of
sludge from wastewater treatment is potentially a
significant life cycle environmental impact of the
system, citing land use and the transport asso-
ciated with sludge disposal as key contributors to
the impact. A study of the most sustainable sludge
disposal option by Suh and Rousseaux (2002)
indicated that a favoured option was a combina-
tion of anaerobic digestion and application to
agricultural land. Major environmental impacts
arise from heavy metals in the sludge together with
emissions from vehicles used to transport the
sludge. These environmental impacts could be
reduced by dewatering the sludge from the septic
tank and treating onsite on the reedbed system as
is practised in some reedbed systems in continental
Europe (Pauly, 1992).
Table 1
Summary of results for life cycle assessment (12 p.e.)
System Process Materials Transport Operation Total
CO
2
(kg) RBS 77 1287 793 /225 1931
CO
2
(kg) AF 23 716 136 3162 4037
Energy (MJ) RBS 137 16,771 9433 17,280 43,621
Energy (MJ) AF 29 4807 1723 46,001 52,560
Solid emissions (kg) RBS 0 77,372 0 90,000 167,372
Solid emissions (kg) AF 0 21,753 0 20,000 41,753
RBS, reedbed system; AF is the conventional package treatment plant.
Table 2
Summary of results for life cycle assessment (60 p.e.)
System (type/p.e.) Process Materials Transport Operation Total
CO
2
(kg) RBS 273 2925 3413 /7383 /773
CO
2
(kg) AF 23 2040 351 12,771 15,184
Energy (MJ) RBS 373 60,569 40,186 29,106 130,233
Energy (MJ) AF 29 13,002 4431 177,372 194,835
Solid emissions (kg) RBS 0 320,239 0 90,000 410,239
Solid emissions (kg) AF 0 53,724 0 300,000 353,724
RBS, reedbed system; AF is the conventional package treatment plant.
A. Dixon et al. / Ecological Engineering 20 (2003) 297 /308304
3.4. Land use
The reed bed has a significantly larger footprint
at all scales. The conventional plant only occupy-
ing 6% of the land area at the smallest scale, 3% at
the 60 p.e. plant and only 2.5% of the largest plant
Table 4. This could be a major issue in areas of
reduced land availability. However, as mentioned
earlier it is possible that the reedbed could perform
multiple functions in its land use, e.g. habitat for
wildlife, use of contaminated land sites.
3.5. Options for reductions of life cycle impacts
The study reveals aspects of system design,
construction and operation that could be targeted
to reduce their environmental impact. These
aspects and directions for their improvement
follow:
.Increase efficiency of pumping and aeration
systems,
.Re-use onsite soil during construction of the
reedbed system,
.Use efficient vehicles to transport maintenance
personnel,
.Develop a management plan that can utilise
local labour,
.Consider application of sludge for local agri-
cultural purposes or dewater and treat on
reedbed,
.Use recycled materials in construction (e.g.
aggregates).
Table 3
Summary of results for life cycle assessment (200 p.e.)
System (type/p.e.) Process Materials Transport Operation Total
CO
2
(kg) RBS 831 7586 11,068 /29,279 /9794
CO
2
(kg) AF 23 3977 812 20,907 25,718
Energy (MJ) RBS 1174 187,451 130,034 48,816 367,475
Energy (MJ) AF 29 27,265 10,251 294,156 331,702
Solid emissions (kg) RBS 0 1,030,521 0 90,000 1,120,521
Solid emissions (kg) AF 0 131,593 0 450,000 581,593
RBS, reedbed system; AF is the conventional package treatment plant.
Fig. 2. Life cycle energy utilisation for Reedbed system at 12
p.e.
Fig. 3. Life cycle energy utilisation for Aerated filter system at
12 p.e. Table 4
Land use by product systems during operation phase of life
cycle
12 p.e. 60 p.e. 200 p.e.
RB (m
2
) 60 300 1002
Conventional (m
2
) 3.5 10.2 24.9
A. Dixon et al. / Ecological Engineering 20 (2003) 297 /308 305
4. Sensitivity of model output for CO
2
emissions to
changes in value of key variables
The approach to investigating model sensitivity
to changes in value of key variables has been
carried out only for the smallest scale system (12
p.e) and only for CO
2
emissions. It is anticipated
that the results of the sensitivity assessment carried
out here will be a good indicator of the significance
of sensitivity issues for other indicators and scales.
The sensitivity of a small number of variables is
investigated. These variables have been chosen in
part because they directly contribute to the most
significant life cycle impacts (e.g. transport) or
their values are of uncertain accuracy and wide
variability.
4.1. Calculating sensitivity coefficients
The following equation (Eq. (1)) is a means of
working out a dimensionless sensitivity coefficient
that indicates the sensitivity of a particular model
output to changes in the variable being considered.
Sensitivity Coefficient
(Outputhigh Outputlow)=Outputdefault
(Inputhigh Inputlow)=Inputdefault
(1)
where Input is the value of the input variable (e.g.
mass of concrete) and Output is the value of the
output indicator (e.g. energy utilisation).
.The CO
2
emissions of the reedbed system are
significantly more sensitive to the variables for
transport distance than the conventional treat-
ment system. For example a 10% increase in
local distance leads to a 6% increase in the CO
2
output for the reedbed system, but only a 0.6%
increase in CO
2
output for the conventional
system.
.The sensitivity coefficient for the reedbed pump
power rating is 0.33, which indicates a signifi-
cant sensitivity (there would be a similar result
for change in UK energy mix or change in daily
pump duty).
.Replacing PVC pipes with clay type pipes in the
reedbed system results in a 3% reduction in
total CO
2
.
.The growing reeds act as a carbon sink, locking
away atmospheric carbon in their structure. The
rate at which they do this is estimated rate as 3.3
kg/m
2
/year with an accuracy of 9/15%. The
sensitivity coefficient for the rate at which the
reeds take up CO
2
is /1.3. This means a 10%
increase in CO
2
uptake leads to a 13% decrease
of CO
2
emitted during the life cycle for the
smaller scale reedbed system.
.Total emissions were insensitive to changes in
variables for CO
2
emission relating to Geotex-
tiles and reed growing materials which no data
for emissions were available.
.The model output for CO
2
emissions is some-
what sensitive to the value embodied energy of
soil (sensitivity coefficient /0.11).
.The model output for CO
2
emissions is not
sensitive to the conversion factor used to
convert tonne.km of truck use to hours of
JCB use (sensitivity coefficient /0.04).
.The model output for CO
2
emissions is sensitive
to the assumed volume of sludge removed from
the septic tank (sensitivity coefficient /0.21).
.The model output for CO
2
emissions is sensitive
to the operational lifetime of the reedbed system
(sensitivity coefficient //0.44). A 10% exten-
sion of operation lifetime leads to a decrease in
CO
2
emissions by 4.4%, (due to the carbon
uptake of the reedbed.
4.2. Re-use of excavated soil
The study revealed that a major life cycle impact
of the reedbed was due to transport of materials to
and from site. The impacts can be reduced
considerably by re-using the excavated soil in the
reedbed infill, only possible if the soil is of suitable
quality. If 50% of the reedbed infill was provided
by the soil excavated on site then the total CO
2
emission of the RBS would be decreased by 13%,
and the energy utilisation would be decreased by a
more modest 6%. If all of the reedbed infill could
be provided by onsite soil then CO
2
emissions are
reduced by a total of 25% and the energy utilisa-
tion by 11%. The reduction in waste soil would be
around 20 T for the 50% re-use and over 40T for
the 100% re-use.
A. Dixon et al. / Ecological Engineering 20 (2003) 297 /308306
4.3. Summary of sensitivity assessment
The model output was found to be particularly
sensitive to the following variables.
.Uptake of CO
2
by reeds.
.Local and national transport distance.
.Re-use of soil.
.Power rating of reedbed pump.
.Embodied CO
2
of imported soil for infill.
.Volume of sludge removed from septic tank.
.Operational lifetime.
These sensitivity results refer only to the reedbed
system designed to treat 12 p.e. and should be
considered along with the results presented in
Section 3. A fuller sensitivity assessment might
result in discovering other variable to which model
output is particularly sensitive. A special note
should be made of the sensitivity of the model
output to changes in the operational lifetime.
5. Conclusions
Overall the reed bed system and conventional
system were quite similar in terms of embodied
energy, although the conventional system bene-
fited the most from economies of scale. The reed
bed had significantly lower overall CO
2
emissions
than the conventional system, due to carbon
uptake by the reedbed. The reedbed produced
more solid emissions than the conventional sys-
tems overall. Excavated soil and sludge from the
septic tank were the main contributors to solid
emissions for reedbed and conventional systems.
The excavated soil of the reed bed becoming
proportionally more significant at larger scales,
whilst the sludge produced by the conventional
units became the most significant factor in solid
emissions for those systems. The reedbed had a
significantly larger footprint than the conventional
system.
The life cycle assessment revealed that signifi-
cant proportions of embodied energy and CO
2
emissions were due to the transport of materials
and for travelling to and from site in order to carry
out maintenance. Also, the pumping requirements
of a system were significant. In the case of the
conventional system pumping was the dominant
life cycle component, exceeding construction, ma-
terials and transport in its contribution to the
inventory. Other key items were system compo-
nents that were made out of plastic e.g. septic tank,
Diamond unit, and uPVC pipes. Concrete also
contributed a significant proportion of embodied
energy and CO
2
. It was discovered that model
output was particularly sensitive to changes in the
CO
2
uptake rate of the reedbed, which in any case
was a significant mitigating factor in the life cycle
CO
2
emissions, offsetting the high emissions from
materials and transport. The model output was
sensitive to the assumption concerning the embo-
died energy of imported soil for the infill. It is
important to note that the net CO
2
emissions of
the smallest reedbed reduced when the operational
lifetime was increased from 10 years.
Acknowledgements
The authors would like to thank Jeremy Horsell
of Lifecycling for carrying out a peer review of the
work and also to thank the reviewers for their
comments. The work was funded by a European
SMART award through the Environmental Busi-
ness Network, Sheffield.
References
Ashley, R.M., Souter, N., Butler, D., Davies, J., Dunkerley, J.,
Hendry, S., 1999. Assessment of the sustainability of
alternatives for the disposal of domestic sanitary waste.
Water Sci. Technol. 39 (5), 251 /258.
Balkema, A.J., Preisig, H.A., Otterpohl, R., Lambert, A.J.,
Weijers, S.R., 2001. Developing a model based decision
support tool for the identification of sustainable treatment
options for domestic wastewater. Water Sci. Technol. 43 (7),
265 /270.
Bavor, H.J., Roser, D.J., Adcock, P.W., 1995. Challenges for
the development of advanced constructed wetlands technol-
ogy. Water Sci. Technol. 32 (3), 13 /20.
Brix, H., 1999. How green are aquaculture, constructed wet-
lands and conventional wastewater treatment systems?
Water Sci. Technol. 40 (3), 45 /50.
Cooper, P., Green, B., 1995. Reedbed treatment systems for
sewage treatment in the UK*/the first 10 years experience.
Water Sci. Technol. 32 (3), 317 /327.
A. Dixon et al. / Ecological Engineering 20 (2003) 297 /308 307
Dennison, F.J., Azapagic, A., Clift, R., Colbourne, J.S., 1998.
Assessing management options for wastewater treatment
works in the context of life cycle assessment. Water Sci.
Technol. 38 (11), 23 /30.
Dennison, F.J., Azapagic, A., Clift, R., Colbourne, J.S., 1999.
Life cycle assessment: comparing strategic options for the
mains infrastructure */Part I. Water Sci. Technol. 39 (10 /
11), 315/319.
Emmerson, R.H.C., Morse, G.K., Lester, J.N., Edge, D.R.,
1995. Life-cycle analysis of small-scale sewage treatment
processes. J. Chart. Inst. Water Environ. Manag. 9 (3),
317 /325.
Gaterell, M.R., Lester, J.N., 2000. Establishing the true costs of
environmental protection and enhancement in the aquatic
environment. Sci. Tot. Environ. 249, 25/37.
Lundin, M., Morrison, G.M., 2002. An LCA based procedure
for development of environmental sustainability indicators
for wastewater systems. Urban Water 4 (2), 145 /152.
Lundin, M., Bengtsson, M., Molander, S., 2000. Life cycle
assessment of wastewater systems: Influence of system
boundaries and scale on calculated environmental loads.
Environ. Sci. Technol. 34, 180 /186.
Mels, A.R., Arjen, F., van Nieuwenhuijzen, J., van der Graaf,
H.J.M., Klapwijk, B., de Koning, J., Rulkens, W.H., 1999.
Sustainability criteria as a tool in the development of new
sewage treatment methods. Water Sci. Technol. 39 (5), 243 /
250.
Moerman, D., Muirhead, D., 1994. Constructed wetlands for
wastewater treatment: annotated bibliography, funded by
the Canada/Nova Scotia Agreement on the Agricultural
Component of the Green Plan Sustainable Agri-Food
Waste Management Program Project GP-022. http://
www.gov.ns.ca/nsaf/rs/envman/randd/wetlands/annobib.txt
(accessed 20/5/02).
Roeleveld, P.J., Klapwijk, A., Eggels, P.G., Rulkens, W.H., van
Starkenburg, W., 1997. Sustainability of municipal waste-
water treatment. Water Sci. Technol. 35 (10), 221 /228.
Pauly, U., 1992. The equipping, development, performance and
behaviour of a root zone system with integrated sludge
treatment. PhD thesis, University of Kassel, Germany.
Shutes, R.B.E., 2001. Artificial wetlands and water quality
improvement. Environ. Int. 26, 441 /447.
Suh, Y.J., Rousseaux, P., 2002. An LCA of alternative waste-
water sludge treatment scenarios. Resources Conserv.
Recycling 35 (3), 191 /200.
Tillman, A.-M., Lundstro
¨m, H., Svingby, M., 1998. Life cycle
assessment of municipal waste water systems. Int. J. LCA 3
(3), 145 /157.
Vanyarkho, O., Arkebauer, T.J., 1995. Responses of Phrag-
mites australis and Scirpus acutus leaf and canopy gas
exchange to environmental conditions. Ecological Society of
America 80th Annual Meeting, Snowbird, UT.
Worrall, P., Peberdy, K., Millett, M.C., 1997. Constructed
wetlands and nature conservation. Water Sci. Technol. 35
(5), 205 /213.
Worrall, P., Peberdy, K., McGinn, H., 1998. Construction and
preliminary performance of reedbed treatment systems at
Castle Espie Wildfowl and Wetlands Trust Centre, North-
ern Ireland. J. Chart. Inst. Water Environ. Manag. 12, 86/
91.
A. Dixon et al. / Ecological Engineering 20 (2003) 297 /308308
