Content uploaded by Roland Clift
Author content
All content in this area was uploaded by Roland Clift on Oct 16, 2019
Content may be subject to copyright.
WATER USE IN LCA
Assessing freshwater use impacts in LCA:
Part I—inventory modelling and characterisation factors
for the main impact pathways
Llorenç Milà i Canals &Jonathan Chenoweth &
Ashok Chapagain &Stuart Orr &Assumpció Antón &
Roland Clift
Received: 31 January 2008 /Accepted: 18 August 2008 / Published online: 9 September 2008
#Springer-Verlag 2008
Abstract
Background, aim and scope Freshwater is a basic resource
for humans; however, its link to human health is seldom
related to lack of physical access to sufficient freshwater,
but rather to poor distribution and access to safe water
supplies. On the other hand, freshwater availability for
aquatic ecosystems is often reduced due to competition
with human uses, potentially leading to impacts on
ecosystem quality. This paper summarises how this specific
resource use can be dealt with in life cycle analysis (LCA).
Main features The main quantifiable impact pathways
linking freshwater use to the available supply are identified,
leading to definition of the flows requiring quantification in
the life cycle inventory (LCI).
Results The LCI needs to distinguish between and quantify
evaporative and non-evaporative uses of ‘blue’and ‘green’
water, along with land use changes leading to changes in
the availability of freshwater. Suitable indicators are
suggested for the two main impact pathways [namely
freshwater ecosystem impact (FEI) and freshwater deple-
tion (FD)], and operational characterisation factors are
provided for a range of countries and situations. For FEI,
indicators relating current freshwater use to the available
freshwater resources (with and without specific consider-
ation of water ecosystem requirements) are suggested. For
FD, the parameters required for evaluation of the common-
ly used abiotic depletion potentials are explored.
Discussion An important value judgement when dealing
with water use impacts is the omission or consideration of
non-evaporative uses of water as impacting ecosystems. We
suggest considering only evaporative uses as a default
procedure, although more precautionary approaches (e.g. an
‘Egalitarian’approach) may also include non-evaporative
uses. Variation in seasonal river flows is not captured in the
approach suggested for FEI, even though abstractions
during droughts may have dramatic consequences for
ecosystems; this has been considered beyond the scope of
LCA.
Conclusions The approach suggested here improves the
representation of impacts associated with freshwater use in
LCA. The information required by the approach is
generally available to LCA practitioners
Int J Life Cycle Assess (2009) 14:28–42
DOI 10.1007/s11367-008-0030-z
Preamble In this series of two papers, the methodological aspects
related to the assessment of freshwater resources use in LCA are
discussed (Part I) and the operational method and characterisation
factors suggested are illustrated for a case study of broccoli produced
in the UK and Spain (Part II).
Electronic supplementary material The online version of this article
(doi:10.1007/s11367-008-0030-z) contains supplementary material,
which is available to authorized users.
L. Milà i Canals (*)
Unilever –Safety & Environmental Assurance Centre,
Colworth Park, Sharnbrook,
Bedfordshire MK44 1LQ, UK
e-mail: Llorenc.Mila-i-Canals@unilever.com
L. Milà i Canals :J. Chenoweth :R. Clift
Centre for Environmental Strategy, University of Surrey,
Guildford,
GU2 7XH Surrey, UK
A. Chapagain :S. Orr
WWF-UK, Panda House,
Weyside Park, Godalming,
GU7 1XR Surrey, UK
A. Antón
IRTA, ctra. Cabrils,
km 2,
08348 Cabrils (Barcelona), Spain
Recommendations and perspectives The widespread use
of the approach suggested here will require some
development (and consensus) by LCI database devel-
opers. Linking the suggested midpoint indicators for FEI
to a damage approach will require further analysis of the
relationship between FEI indicators and ecosystem
health.
Keywords Ecosystem .Evaporative use .FD .FEI .
Freshwater ecosystem impact .LCA .LCI .LCIA .
Virtual water .Water footprint .Water resource
1 Background, aim and scope
Water is a precious and increasingly scarce resource. It is
critical for ecosystem functions (as both habitat and
resource) and equally essential for humans. Water abstract-
ed for human purposes can have significant impacts on
water systems. More than 100,000 species (almost 6% of
all described species) live in freshwater and countless
others depend on freshwater for survival (Dudgeon et al.
2005). Freshwater species and habitats are more imperilled
globally than their terrestrial or marine counterparts (WWF
2006). In the most extreme cases, water scarcity has
resulted in complete ecosystem collapse (Micklin 1988).
Similarly, some major rivers have periodically completely
dried up, including the Rio Grande/Bravo in Mexico and
the Great Ruaha River in Tanzania (WWF 2007).
A lack of adequate access to safe water supplies is dire
from a human health point of view. Globally, nearly two
million people die from diarrhoeal disease every year, with
88% of cases attributed to unsafe water and inadequate
sanitation or hygiene (WHO 2004). However, the water
needed to feed humanity requires significantly higher
amounts than daily drinking and cleaning water. In general,
it takes about 3,400 l per person per day to support a global
average consumption pattern. There is a wide variation in
this amount: from more than 6,000 l per person per day in
many western countries (e.g. USA, Canada, Greece, Italy,
Spain, etc.) to much smaller amounts in developing
countries [e.g. India (2,700 l) and China (1,900 l) per
person per day] (Chapagain and Hoekstra 2004). This is
due mainly to differences in diets, and whilst eradicating
malnutrition by 2025 will require a doubling of water used
for agriculture (Rockstrom et al. 2006), rising affluence in
many emerging countries will raise their averages toward
the western norm.
Measurement of water use therefore provides important
information for attributing responsibility, assessing impact
and developing solutions to water use by companies,
communities and individuals. Different environmental
system analysis tools such as life cycle analysis (LCA)
and virtual water (VW) are able to measure amounts of
water used in the production of various products. However,
both methods currently lack a proper assessment of the
relative scarcity and opportunity cost of water at the point
of production.
Water use in LCA Water use impacts have been underrep-
resented since the start of LCA methodology in the late
1960s, probably due to LCA being developed for industrial
systems (usually less dependent on water resources than
agricultural ones) in water-abundant countries. Basically,
LCA studies report the total amount of water used by the
production system, from cradle (raw material acquisition) to
grave (waste management). In general, such studies do not
even distinguish the source from which water is obtained
nor the way or condition in which water leaves the product
system.
Early examples explicitly addressing water use in LCA
included water-intensive products such as nappies (Johnsson
1994; Sauer et al. 1994). Other references quantified the
total amount of water used in industrial (e.g. Milà i Canals
et al. 2002; Muñoz et al. 2004;2006) and agricultural
systems (e.g. Antón et al. 2005; Milà i Canals et al. 2006;
Coltro et al. 2006; Muñoz et al. 2008). Lundie et al. (2004)
assess water supply as a function and quantify the water
flow associated to different scenarios for water provision in
Sydney (Australia), but do not address the impacts related
to water use. Recently, some authors have suggested ways
to progress beyond merely quantifying the volume of water
used per functional unit. Owens (2002) clarifies the
definitions of different water inputs to (sources) and outputs
from (dispositions) the system, key to a proper inventory of
freshwater use, and suggests the qualitative aspects that
should be factored into water assessment in LCA. Brent
(2004) stresses the need for spatial differentiation, but
suggests simply adding a kilogram of water used with no
further characterisation along the life cycle stages. Other
authors have explored ways to include water use in
agricultural systems through evapotranspiration in the
inventory (e.g. Antón et al. 2005) and/or to incorporate
water use in the impact assessment (life cycle impact
assessment, LCIA) phase (e.g. Heuvelmans et al. 2005).
Bauer and Zapp (2004) offer background information to
relate potential effects of mining on water resources, but do
not provide a means to use this information in the
characterisation stage. In summary, the incorporation of
such improvements in LCA practice has been extremely
limited to date. An obvious reason is that the main
LCIA methods do not provide characterisation factors
for water as an abiotic resource (EDIP 1997: Hauschild
and Wenzel 1998; EI99: Goedkoop and Spriensma 1999;
CML 2001: Guinée et al. 2002). Guinée et al. (2002,part3,
p. 184) conclude that there is no satisfactory method to
Int J Life Cycle Assess (2009) 14:28–42 29
address the habitat aspect of freshwater (related to what
they call ‘desiccation’) in LCA. Even though impacts
resulting from freshwater use are certainly an environ-
mentalissueinmanyproductionsystems,andISO
14044 requires that LCIA reflects ‘a comprehensive list
of environmental issues related to the product system
studied’, no satisfactory method yet exists to include
them in LCA.
1.1 The concept of VW
The concept of VW has evolved since the early 1990s and
refers to the amount of water required to produce a certain
product. VW was introduced by Allan (1998,2001) who
investigated VW imports through the trade of water-
intensive crops as a partial solution to problems of local
water scarcity in the Middle East. Allan suggests that such
trade relieved the need for importing countries to use their
own, often scarce, water resources to produce the same
product. Water is termed as ‘virtual’because the amount of
water physically contained in the final product is negligible
compared to the amount that went into its production. VW
studies have taken on more precise and practical applica-
tions since Hoekstra and Hung (2002), Chapagain and
Hoekstra (2003,2004), Champagain and Orr (2008) began
to quantify and calculate VW flows and related water
footprints.
Originally, Hoekstra and Hung (2002) estimated the
‘blue’water footprint by excluding the ‘green’water use
(use of effective rainfall to produce crops) from domestic
production (see below for definitions). Subsequently,
Chapagain and Hoekstra (2004) included the ‘green’water
footprint related to the consumption of domestic produc-
tion, and Chapagain et al. (2006) included the ‘grey’
component of VW, accounting for the water volumes
needed to dilute waste flows to agreed water quality
standards.
Clearly, there is a need for a more systematic assessment
to characterise the sustainability of freshwater use by
production systems in LCA. Champagain and Orr (2008)
argue that such an indicator could also be useful for VW. In
this paper, we first introduce some terms drawn from LCA
and VW literature in order to avoid confusion by establish-
ing a common terminology (see Section 2). The main
impact pathways resulting from freshwater use are de-
scribed in Section 3; this leads us to the definition of the
relevant water flows that need quantification and assess-
ment in LCA. Section 4offers guidance to quantify such
flows, and in Section 5, we review some of the existing
indicators currently used to characterise the sustainability of
water uses. Characterisation factors for the most promising
indicators are provided in the Electronic supplementary
material in a format ready to use in LCA studies. Finally,
discussion, conclusions and recommendations are provided
in Sections 6,7and 8, respectively.
1
2 Definitions
One possible reason why water has not yet been properly
assessed in LCA is the plethora of forms and routes in
which water enters and exits production systems. This
section suggests some terminology referring to how water
enters and leaves the system.
2.1 Water as an input to the production system
From a resource point of view, water may be a flow (e.g.
rivers, rain), a fund (groundwater) and even a deposit or
stock (fossil water; Bauer and Zapp 2004). Water may enter
the production system from surface bodies such as rivers
and lakes (flow), rainfall (flow), groundwater bodies (fund
or stock) or the sea (e.g. used for cooling or as input to
desalination plants). Owens (2002) further details whether
water is used in-stream (e.g. in a power dam) or is
withdrawn off-stream.
Water occurs in the form of green water (stored as soil
moisture and available for evaporation through crops and
terrestrial vegetation) and blue water (surface or ground-
water). Blue water is the volume of water in ground
(aquifer) and surface water bodies available for abstraction.
The distinction between blue water and green water is
important, as green water is only available for use by plants
at the precise location where it occurs, whereas blue water
is available generally for use in a wide range of human
managed systems, including but not limited to use by
plants. Another possible way for rainwater to enter a system
is through rainwater harvesting. This represents a special
case of storing this form of blue water directly in human
infrastructures and therefore avoiding abstraction from a
natural body although, as for other forms of land use, it
diverts water from replenishing a natural body.
2.2 Water as an output from the production system
As crucial as where water comes from is how water is returned
to nature. Two main paths must be distinguished here:
&Non-evaporative water use (‘water use’according to
Owens 2002: water is returned to the original basin and
may be used by other users after leaving the system).
1
A practical application to illustrate the suggested methodology is
offered in the second part of this paper: Milà i Canals L, Chapagain
AK, Orr S, Chenoweth J (in preparation) Assessing freshwater use
impacts in LCA. Part II: case study for broccoli production in the UK
and Spain.
30 Int J Life Cycle Assess (2009) 14:28–42
&Evaporative water use (‘water consumption’according
to Owens 2002: water is dissipated and not immediately
available after use).
Obviously, non-evaporative water may return to a
different system, which requires a proper definition of the
temporal and spatial system boundaries: When and where
do we consider that water leaves a system and/or becomes
available to other users? The main difference between water
and many other mineral resources is that even dissipated
(evaporated), water will eventually become useful in a
relatively short period of time through rainfall (and then
become renewable surface or groundwater), although this
water cycle generally occurs over vast geographic areas.
Some references (e.g. Mohamed et al. 2005) suggest that
only 10–20% of evaporated water falls as precipitation on
land, the rest being ‘lost’as rainfall on oceans, and thus not
immediately available to other users. This reinforces the
importance of the distinction between evaporative and non-
evaporative uses.
Water transfers between river basins (or, less frequently,
between countries) are a special case of water input/output.
Owens (2002) suggests considering water transfers as an
evaporative use (consumption). However, we see no reason
why this should be the case unless the transferred water is
actually evaporated. We suggest treating transfers (includ-
ing embodied water) from a resource availability perspec-
tive, i.e. the transfer increases the resource in the receiving
basin and decreases it in the providing one. Nevertheless,
global databases of water resources do not subtract water
exports from a country’s available water reserves because
water is actually available before being exported. In effect,
we are not considering any particular environmental impact
related to the resource aspect of transferred water. Indeed,
the transfer itself is associated with environmental impacts
due to energy use, infrastructure construction, etc., but local
effects, e.g. on freshwater ecosystems, are deemed too
spatially specific for LCA.
3 Main impact pathways from effects of freshwater
use on availability
The use of freshwater resources may lead to undesired
impacts such as reduced availability of water for other
users, locally lowering the level of water courses and lakes
with effects on aquatic ecosystems, and ultimately impacts
on human health due to insufficient water availability and
poor water quality. As mentioned previously, in this paper,
we focus on impacts related to the availability of sufficient
freshwater and only suggest research needs for those
quality aspects not yet sufficiently covered in LCA. Two
main aspects of water need to be addressed here: water as a
resource for humans and water as a habitat (resource for
ecosystems; environmental water requirements as defined
by Smakhtin et al. 2004). Related to these two aspects, the
following four main impact pathways may be distinguished
and merit attention in LCA; they are illustrated in Fig. 1:
1. Direct water use leading to changes in freshwater
availability for humans leading to changes in human
health;
2. Direct water use leading to changes in freshwater
availability for ecosystems leading to effects on
ecosystem quality (freshwater ecosystem impact, FEI);
3. Direct groundwater use causing reduced long-term
(fund and stock) freshwater availability (freshwater
depletion, FD);
4. Land use changes leading to changes in the water cycle
(infiltration and runoff) leading to changes in freshwa-
ter availability for ecosystems leading to effects on
ecosystem quality (FEI).
3.1 Effects of freshwater use on human health
Despite the critical nature of water as a resource for
humans, the relationship between natural water resources
availability and human health is not straightforward.
When naturally available renewable water resources per
capita on a national level are compared with basic
indicators of economic development, human health or
well-being, no statistically significant correlations result
(Chenoweth 2008a,b). Besides, the dramatic amount of
human deaths linked to water is mainly caused by poor
water quality and/or sanitation and not by the physical
amount available. We suggest that these aspects should be
properly included in LCA methodology where this is not
currently the case (e.g. water pollution with faecal
bacteria).
This supports the assumption that the link between
human health and water quantity per se is not a
significant issue. As such, we suggest omitting this
aspect from LCA. However, it may be necessary to
include it in specific cases when information about
human deaths caused by lack of adequate water resour-
ces is available.
3.2 Change of water quantity affecting ecosystem health
(FEI)
Whilst critical use of water for human health is generally
guaranteed (see above), this is not the case for ecosystems.
Indeed, ecosystems may be damaged due to excessive
water abstraction for human purposes, through for example,
changes in the groundwater table (with effects to wetlands)
Int J Life Cycle Assess (2009) 14:28–42 31
or changes in the environmental flows of rivers. We suggest
calling this impact ‘freshwater ecosystem impact’(FEI).
In the case of flow and fund water resources, only
evaporative use leads to reduced (temporal and/or spatial)
availability for other users (humans or ecosystems). It is
assumed that the amount of non-evaporative water used
which is subsequently returned to the water source does not
lead to relevant environmental impacts from a resource
perspective and should be disregarded in the LCIA phase.
This is a big assumption, and water extracted from a river
to provide drinking water for a city might actually not
return to the same river directly, thus reducing its flow
(potentially below the defined environmental flow). How-
ever, such a localised perspective is beyond the conven-
tional scope of LCA (see below), and we consider that
water returns to ‘the environment’after a non-evaporative
use rather than to a specific ecosystem. This is a default
(baseline) suggestion, and the option is left open for
practitioners to characterise non-evaporative uses with a
characterisation factor >0 for FEI if deemed appropriate
(see Section 5.2.2).
Because fossil water (stock water resources) would
not occur in an ecosystem if it was not abstracted by
humans, its use does not affect ecosystem health.
Besides, for ecosystems to feel a positive effect from
abstracted fossil water, there would have to be a
continuous abstraction and return to ecosystems, which
is a highly unlikely scenario. We thus recommend not
considering any positive effect from fossil water use on
FEI.
Obviously, changes in the quality of the returned water
may create important impacts and these need to be
considered, but perhaps in other impact categories such as
eutrophication or toxicity. Some flows are not normally
included in LCIA but may be important for some water
uses; an example is heat in the case of cooling water. These
flows should be further studied in the relevant impact
categories, but are not considered in this paper.
From a quantitative point of view, though, it needs to be
stressed that only potential impacts due to the amount of
dissipated resource may be assessed in LCA. Site-specific
local effects, e.g. on flow in a watercourse (see above) or on
wetlands, are usually not accountable in LCA; rather, they
should be addressed during the assessment of specific
projects in the framework of Environmental Impact
Assessment.
Fig. 1 Main impact pathways related to freshwater use. All the
pathways are discussed in this paper, but only those depicted with
solid arrows are considered for LCA. The concepts in circles denote
common denominations in the water footprint field. The numbers refer
to the impact pathways defined in Sections 3.1–3.4
32 Int J Life Cycle Assess (2009) 14:28–42
3.3 Depletion of freshwater resources (FD)
As stated above, water can be a flow, fund or stock resource.
In general, a flow resource such as river water cannot be
depleted but there can only be competition over its use,
whereas depletion may be an issue for fund and stock
resources (Guinée et al. 2002, p. 154). Competition amongst
human users is outside the scope of LCA, whereas competi-
tion for water flows with ecosystems is addressed in the
freshwater ecosystem impact. Conversely, using groundwater
may reduce its availability for future generations, when
aquifers are over-abstracted or fossil water is used, and so
needs to be included as an impact on natural resources.
A special case related to the availability of freshwater
resources is desalination of seawater. Seawater is so
abundant that its use will not cause resource concern in
the foreseeable future. In fact, desalination of seawater may
be considered as a way to increase freshwater resources and
treated as a beneficial effect on freshwater depletion.
Desalinisation brings with it a raft of other environmental
issues (e.g. carbon emissions, energy use, brine discharge,
disturbance to marine ecosystems around water intakes and
brine returns), but these are not related to water as a
resource. In general, a positive feedback with FEI is not
expected because desalinated water is usually used close to
the sea, and any non-evaporative use returned to the
environment will not benefit freshwater ecosystems. This
may however be contemplated in specific cases if the
practitioner has evidence of benefit to freshwater systems.
A different case is rainwater harvesting: Even though
some might consider it a way of increasing freshwater
resources, it should not be considered in the same way as
suggested for desalinated seawater because rainwater would
be available as green or blue water if it was not harvested.
The benefits of a system using harvested rainwater will
appear as a smaller use of green or blue water.
3.4 Changes in the water cycle caused by land use related
to production system (FEI)
Accounting for water stored as soil moisture (green water)
is essential for VW in order to show the total water use of a
crop, to calculate the amount of blue water abstracted and
to show where that water came from in the hydrological
cycle. However, LCA does not account for issues not
affected by the production system. For example, the amount
of solar energy used to grow crops is not accounted because
solar radiation is independent of the crop or production
system, i.e. the land will receive the same insolation
regardless of the type of crop established.
2
In the same
way, a similar amount of soil moisture (stored rain water)
will be used by different crops and/or natural ecosystems
regardless of the production system; therefore, the use of
rainwater does not change the environmental effects that
would occur if the studied system was not established.
However, knowledge of soil moisture (green water) used by
plants is needed to assess the amount of blue water required
to grow those plants. Thus, green water use should be
included to estimate the blue water impacts through
evaporation of irrigation water. The life cycle inventory
(LCI) results thus calculated will be compatible with VW
quantifications, but the amount of green water will
subsequently be disregarded in the LCIA phase because
green water use leads to no environmental impacts.
An issue to consider however is that production systems
may significantly change the amount of rainwater available
to other users through changes in the fractions of rainwater
that follow each one of the three basic paths: infiltration (I),
evapotranspiration (ET) and runoff (R). In general, ET is
the amount ‘lost’from the system (evaporative use),
whereas Iand Rreturn to the system. Owens (2002)
suggests that for in-stream water consumption, “evaporative
losses from reservoirs and canals in excess of unrestricted
river losses”are to be accounted (i.e. the unrestricted river
is considered as the reference system). Likewise, it is here
suggested that changes in evaporation are most significant
in ‘aquatic land uses’such as reservoirs and canals (see
Section 4.2.1). Bauer and Zapp (2004) suggest using the
difference between monthly precipitation and evapotrans-
piration as an estimate of surface and subsurface runoff (i.e.
I+R) and provide values as a world map. However, such an
approach would suggest, rather counter-intuitively, that
sealed land (with largely reduced ET) has a positive effect
on the water cycle due to the increased R. Actually, and
particularly with heavy rainfalls, such rapid returns in the
form of R may have little effect in replenishing aquifers
(see Section 4.3), can lead to increased flooding and be of
no use to ecosystems. Heuvelmans et al. (2005) suggest
using changes in (I−ET) alongside other indicators for land
use impacts on water (namely change in surface runoff and
precipitation surplus). However, the feasibility of their
approach in all life cycle stages is doubtful due to the large
amount of data and modelling required for each land use
along a product’s life cycle.
Therefore, we suggest assessing I+R(as done by Bauer
and Zapp) for ‘non-sealed’land uses and only the potential
changes in Ifor all other uses. The argument is that land
acts as a buffer of the water cycle through I, and this is the
‘ecological quality’of land which should be protected
(Milà i Canals et al. 2007). Note that in specific situations,
increases in Imay lead to adverse effects (e.g. by raising
saline groundwater table), but such site-specific situations are
beyond the scope of LCA and therefore not considered here.
2
Differences in albedo between different crops may be relevant for
global modeling, but are outside the scope of LCA.
Int J Life Cycle Assess (2009) 14:28–42 33
Changes in Iare assessed against the reference system in the
land use impact assessment (Milà i Canals et al. 2007).
This change in groundwater recharge may be linked to
the freshwater ecosystem impact (2.2). Land use may
also be linked to freshwater depletion (2.3) through
groundwater recharge if land use affects the recharge of
overexploited aquifers; we suggest not including this
pathway as a default because it is site-specific. In both
cases, the land use effect may be beneficial (if water
infiltration is increased) or damaging (if infiltration is
reduced).
3.5 Water flows
3
to be quantified in LCI
In summary, the following water flows need to be
accounted in LCI:
From a freshwater ecosystem impact point of view:
&Surface and groundwater evaporative uses: in-stream
evaporation in reservoirs and power dams and off-
stream evaporation of abstracted water through, e.g.
irrigation; in cooling towers, etc. In VW terms:
evaporative blue water.
&Any type of land use occupation and transformation
processes.
From a freshwater depletion point of view:
&Water stocks (groundwater—fossil water) and over
abstracted water funds (groundwater—aquifers): both
evaporative and non-evaporative uses need to be
quantified. This is consistent with existing methods
for abiotic resource depletion such as CML2001, but
in ‘consequence-oriented LCIA methods’(‘type 3’
methodologies according to Lindeijer et al. 2002),
only the amount of resource dissipated (evaporated)
may merit being accounted for (Stewart and Weidema
2005). Our suggestion at this point is to record both
evaporative and non-evaporative use flows separately
so they can then be treated appropriately in LCIA.
Many LCI databases contain flows and datasets for
provision of ‘tap water’; this paper is only concerned with
the elementary flows (coming from/going to nature). As
with energy resources, it will be a matter for public LCI
databases to define the water resource flows related to
‘national water grids’; the impacts of such grids will in turn
depend on the mix of water sources, much as the impacts of
power grids depend on the mix of energy sources and
technologies.
4 LCI modelling
4.1 Calculation of water evaporated from irrigation
The water withdrawal and subsequent evaporation from a
crop field can be calculated following the methodology
used by Champagain and Orr (2008). In brief, the various
steps from their study are as follows:
First, the virtual water content of the primary crop is
calculated as the ratio of water used for crop production to
the yield per unit area. The volume of water used for
production is made up of two components, evaporative and
non-evaporative water. Following Champagain and Orr
(2008), non-evaporative water use is a function of irrigation
losses (implicitly including all losses in storage, convey-
ance systems and field application) and volume of water
rendered unsuitable for use further downstream as a result
of polluted return flows (grey water). They suggest that
environmental effects of grey water are more suitably
addressed in other impact categories.
The evaporative demand is met from soil moisture that is
within the root zone depth of the plant and immediately
available for plant uptake. Depending upon the source of
water maintaining the soil moisture, Chapagain (2006)
distinguishes between green water use (WU
g
) originating
from rainfall on crop land and blue water use (WU
b
)
originating from irrigation water supply.
The components WU
g
and WU
b
depend on the specific
crop evaporation requirement and soil moisture availability
in the field. The crop evaporation requirement (ET
c
[t]) is
calculated using the classical Penman–Monteith equation to
estimate reference crop evaporation following the method-
ology recommended by FAO (Allen et al. 1998). This can
easily be done with, e.g. the publicly available CROPWAT
model (FAO 1992) or alternative models. The USDA Soil
Conservation Service approach has been used to estimate
the effective rainfall (p
eff
[t]), as it is one of the most widely
used methods in estimating effective rainfall in agricultural
water management (Cuenca 1989; Jensen et al. 1990).
Climate data for representative climate stations may be
obtained from the study inventory or alternatively from
CLIMWAT (FAO 1993). Green water use, u
g
[t], is equal to
the minimum of effective rainfall and the crop evaporation
requirement at that time step, and total green water use
(WU
g
) in crop production is calculated by summing up
green water use for each time step over the entire cropping
period, l(day). Green water use is independent of irrigation
water supply and solely depends on the effective rainfall
and crop evaporation requirements, whereas blue water use
depends on crop evaporation requirement, green water
availability and irrigation water supply. The fraction of
ET
c
[t] not met by u
g
[t] is the irrigation requirement (I
r
[t]).
The blue water use (u
b
[t]) is the minimum of irrigation
3
Flow refers here to the element listed in a LCI, i.e. ‘elementary
flow’, and not to the type of resource as in ‘flow, fund or stock’.
34 Int J Life Cycle Assess (2009) 14:28–42
requirement, I
r
[t], and the effective irrigation water supply,
I
eff
[t]. The effective irrigation supply is the part of the
irrigation water supply that is stored as soil moisture and
available for crop evaporation. Blue water use is zero if the
entire crop evaporation requirement is met by the effective
rainfall. Total blue water use (WU
b
) in crop production is
calculated by summing up blue water use for each time step
over the entire cropping period, l(day).
There are inevitable irrigation losses from the local
system, as it is hard to match the blue water demand with
irrigation water supply in time and space. Irrigation losses
(I
loss
[t]) are calculated by subtracting blue water use, u
b
[t],
from irrigation water supply, I
s
[t], if known. Total irrigation
losses from the field over the crop period are calculated by
summing the losses in each time step over the entire crop
growth period, l(day).
4.2 Calculation of water evaporated from other processes
LCA studies include water use in a variety of processes
apart from irrigation (discussed above), many of which will
be essentially non-evaporative. In some, though, part of the
water used is evaporated, and this needs to be estimated in
the LCI in order to provide relevant information for the
LCIA. The following paragraphs suggest ways in which
losses can be estimated for the main industrial processes
causing water evaporation as a proportion of the total water
input to the process.
4.2.1 Evaporation from reservoirs and canals (aquatic land
uses)
Basic data to calculate evaporation from a reservoir per
cubic metre of water used are its area, volume abstracted
per year and potential evaporation in the region (in
millimetre or litres per square metre, common meteorolog-
ical parameter). As a specific example, a 10,000-m
2
reservoir for irrigation providing 900,000 m
3
/year of water
for irrigation in a region with a mean annual evaporation
potential of 1,400 mm has an evaporation loss of 1,400×
10,000/900,000=15.5 l/m
3
of water delivered. This calcu-
lation may be modified to account for devices used to
reduce the evaporation rate, months when the reservoir is
kept empty, etc. when such information is available.
4.2.2 Cooling water
Industrial plants, particularly thermal plants generating
electrical power from fossil or nuclear fuel, are major water
users. Water is used for two essential functions: as the
‘working fluid’driving steam turbines and as coolant in
condensers and other heat exchangers. The water in a
turbine cycle is highly purified and used on a closed-loop
basis: water makeup is therefore small and dependent on
details of plant design and operation. Coolant water is
sometimes abstracted from a water body or river, used once
and then returned at a higher temperature; net coolant water
use is then small. More commonly, the coolant water is
used in a circuit in which it is heated in the condensers and
heat exchangers, to be cooled and reused. Where the plant
is operated in the combined heat and power (CHP) mode,
exporting heat by distributing hot water, the hot water is
circulated to the heat user and returned to the plant; net
water use is then again small. Where the plant exports
steam, the net use depends on whether the steam is vented,
or condensed and reused. Where there is no heat output, the
cooling water is used in a circuit with the heat dissipated by
evaporation in cooling towers. There is then a net water
loss, corresponding to water evaporated.
To a first approximation, the cooling load in a generating
plant using evaporative cooling corresponds to the latent
heat (strictly, enthalpy change) of evaporation of the water
lost. For example, a plant operating with 35% efficiency (i.e.
electrical output energy is 35% of the thermal energy
released from the fuel so that 65% must be dissipated as
‘waste heat’) requires a cooling load of 3,600 × 65/35 kJ per
kilowatt-hour of electrical output, i.e. about 6,700 kJ/kWh.
The latent heat of evaporation of water is about 2,400 kJ/kg
so that the theoretical evaporative loss is 6,700/2,400, i.e.
about 2.8 kg of real water loss per kilowatt-hour. Actual
figures may be somewhat larger or smaller, depending on
details of plant design and operation (including whether any
of the waste heat can be used—CHP is more efficient than
electricity-only generation on water use as well as energy
use). The average figure for nuclear thermal power produc-
tion of 9.01 kg water/kWh(e) given by the Ecoinvent
database (Dones et al. 2004) represents water abstracted;
net loss by evaporation must be established as primary data
for any specific plant, but should be around one third of this
figure even for energetically wasteful evaporative cooling.
4.2.3 Textile drying
The washing stage is one of the most important sources of
water usage in textile products. Most of the water used in
washing is returned as wastewater, and part remains in the
clothes and is partly evaporated through drying. On
average, water content in wet (after centrifugation) clothes
is 0.7 kg/kg clothes (70%), and this goes down to residual
moisture of approximately 0.03 kg/kg clothes (3%) after
conventional drying (Group for Efficient Appliances 1995).
These values thus yield an evaporative use of approximate-
ly 0.67 kg water evaporated per kilogram of dried clothes;
considering a value of about 1.8 m
3
evaporated blue water
per kilogram of seed cotton in the cropping stage
(Chapagain et al. 2006), washing clothes about 100 times
Int J Life Cycle Assess (2009) 14:28–42 35
in their lifetime represents about 3–5% of the evaporative
water use in growing cotton.
4.3 Estimation of land use effects on rainwater infiltration
Table 1suggests some values for the percent of water ‘lost’
due to different land occupation processes. The land uses
listed (derived from the Ecoinvent classification) have been
treated in two different ways:
&In systems generally allowing infiltration (non-sealed
land), both infiltration and runoff (I+R) have been
considered as useful paths for ecosystems; they are
marked as ‘N’land uses. Lost water is ET. Most values
for such uses are derived from a large review of
worldwide water catchments studies (Zhang et al.
1999).
&In systems that are heavily transformed and generally
sealed, only Ihas been considered as useful for
ecosystems, and water ‘lost’is ET+ R. They are marked
as ‘S’land uses in Table 1. All these values are
estimated based on the fact that fully sealed land has
negligible infiltration and assuming an infiltration rate
of 1% on land that is not fully sealed (e.g. construction
site).
This table has to be seen as a gross approximation,
which may be useful as a default set of values but needs
refinement. The assumptions made have been clear in some
cases (e.g. sealed soil reduces infiltration to 0% by
definition), whilst in some others, the justification is less
clear. The table thus offers plausible, rather than accurate,
values.
Table 1provides values for low and high precipitation
areas (less than and >600 mm year
−1
); in general, though,
values for only one gross precipitation group will be needed
for LCA databases. As an example, the last column in
Table 1suggests values of rainfall ‘lost’per m
2
year of
different land occupations in Europe, assuming average
precipitation of 734 mm (Gleick 1993) with forest as the
reference (potential) land use. As a specific example, the
estimated reference rainwater lost from forest is 67%,
whereas in arable land, this is 73%; therefore, the extra loss
due to using arable land is 6% of rainwater, or 44 l m
−2
year
−1
for average precipitation of 734 mm.
5 Characterisation factors for LCIA
5.1 Indicators for FEI
Several indicators are used to compare the sustainability of
water supplies in different countries. Falkenmark (1986)
proposed an indicator based upon water resources (WR) per
capita (WRPC = WR/population) with defined threshold
values for water stress—less than 1,667 (usually rounded to
1,700) m
3
per capita, water scarcity—less than 1,000 m
3
per capita and absolute water scarcity—less than 500 m
3
per capita year. Whilst this indicator has since been adopted
as the standard indicator of water scarcity, it fails to
consider the ability of nations to adapt to reduced per
capita water availability through means such as VW and
does not consider differences in water use patterns between
countries or multiple in-stream uses (Raskin et al. 1997).
Besides, WRPC is intended to apply to human direct
(domestic) use (drinking+sanitary), but this is seldom the
problem: Most water is used in agriculture and then by
industry. This indicator has been used in LCA by Antón et al.
(2005).
Feitelson and Chenoweth (2002) suggest an index of
water sustainability based around the affordability of water
supplies (index of structural water poverty). Its major
shortcoming is that there is no simple or transparent way
to determine real water supply costs in a country, and thus,
data for the index are lacking. Besides, this is an index for
social impacts related to water, rather than environmental
impacts as sought by LCA.
A more useful index for determining the environmental
sustainability of water supply and water stress is the water
use per resource (WUPR = WU/WR) indicator put forward
by Raskin et al. (1997). This index compares the percentage
of available water resources being withdrawn from natural
water bodies. This index does not address water quality
issues, but it highlights the water remaining for in-stream
usage or further development and/or ecosystems, a factor
disregarded by the standard WRPC indicator. Therefore,
WUPR is a good indicator for potential impact on aquatic
ecosystems. A high WUPR indicates serious water stress as
most available water is being used. Additionally, because of
climate variability, the higher the water exploitation ratio,
the greater the chances of water shortages during dry years.
The WUPR ratio thus indicates the ‘marginal impacts of
water usage’: The impacts of providing one extra unit of
water increase as the proportion of resources already used
increases. This is true both from an ecosystems perspective
and from a resource provision perspective (e.g. desalination
becomes an option only after most of the easily available
resources are used). Data for this indicator are readily
available for most of the world at a national level. River
basin level data may be more relevant from a FEI
perspective, and LCA practitioners are encouraged to find
and use an appropriate level of precision for the foreground
system. The WUPR indicator may be used directly as a
characterisation factor: Multiplying by the percent of water
used gives bigger weight to water used in countries/regions
where a bigger proportion is used. Alternatively, more
36 Int J Life Cycle Assess (2009) 14:28–42
Table 1 Effects of land occupation on usable proportion of precipitation, considering Ecoinvent’s land occupation flows
Ecoinvent land occupation flows Land use Type Rainfall <600 mm/year Rainfall >600 mm/year ‘Lost precipitation’mm/m
2
year
assuming 734 mm/year rainfall
(Gleick 1993)andforestas
reference land use
Percentage of ‘lost’
precipitation
Sample size Percentage of ‘lost’
precipitation
Sample size
Occupation, arable, non-
rrigated
a
N9310731944
Occupation, construction site S 99 Est. 99 Est. 235
Occupation, dump site S 99 Est. 99 Est. 235
Occupation, dump site, benthos –
b
–––
Occupation, forest, intensive
a
N 83 4 67 36 0
Occupation, forest, intensive, normal
a
N 83 4 67 36 0
Occupation, industrial area
c
S 99 Est. 98 Est. 228
Occupation, industrial area, benthos –
b
–––
Occupation, industrial area, built up S 100 Est. 100 Est. 242
Occupation, industrial area, vegetation
d
N 94 Est. 73 Est. 44
Occupation, mineral extraction site S 100 Est. 100 Est. 242
Occupation, pasture and meadow, extensive
a
N9415733544
Occupation, pasture and meadow, intensive
a
N9415733544
Occupation, permanent crop, fruit, intensive
a
N951967510
Occupation, shrub land, sclerophyllous
a
N 92 3 64 7 −22
Occupation, traffic area, rail embankment N 95 Est. 95 Est. 206
Occupation, traffic area, rail network N 95 Est. 95 Est. 206
Occupation, traffic area, road embankment N 95 Est. 95 Est. 206
Occupation, traffic area, road network S 100 Est. 100 Est. 242
Occupation, urban, discontinuously built
c
S 99 Est. 98 Est. 228
Occupation, water bodies, artificial –
b
–––
Occupation, water courses, artificial –
b
–––
a
Data derived from Zhang et al. (1999)
b
‘Aquatic land uses’have been considered in a different way (see text)
c
Assumed to have 5% vegetated area
d
Considered as pasture
Int J Life Cycle Assess (2009) 14:28–42 37
sophisticated characterisation factors (CF) may be con-
structed to describe step changes using ‘threshold values’
and/or non-linear relationships (see Section 5.1.2).
An alternative indicator for environmental water stress is
explored by Smakhtin et al. (2004), who make a first
attempt at estimating the environmental water requirements
(EWR) for all world river basins. They then combine EWR
with the water resources available and their use (i.e. WUPR
defined per river basin) by subtracting EWR from the
available resources to derive a water stress indicator (WSI =
WU/(WR−EWR)). A more accurate indication of the water
resources available for further human use after ‘reserving’
the necessary resource for ecosystems (EWR) can thus be
obtained. However, the use of this indicator is at an early
stage and lack of data might hamper its use in LCA.
5.1.1 Developing factors for the suggested indicators
Table A-1in Electronic supplementary material provides
values for the WRPC and WUPR indicators for most
countries. It has been compiled using data from the FAO
Aquastat database (FAO 2004) and the UNDP human
development indicators (United Nations Development
Programme 2006). The basic parameters required to
construct the indicators (population, water resources and
use) should be found for smaller geographical areas when
available and appropriate for CF calculation.
Table A-2in Electronic supplementary material provides
values for WSI for the world’s main river basins (Smakhtin
et al. 2004).
5.1.2 Possible further sophistication: thresholds and damage
approach
As noted above, the WRPC, WUPR or WSI ratios may be
used directly as a characterisation factor for the amount of
water evaporated. However, it may be argued that the
potential for impacts is unlikely to follow a linear
relationship with such ratios; for example, a double-
exponential relationship may be more appropriate, with
low effects for low WUPR and a maximum effect before
the 100% value to reflect the likelihood that most
ecosystems will be damaged even before all water is used
by humans.
The construction of such relationships requires extensive
research on the effects on aquatic ecosystems of increasing
appropriation of freshwater for human uses. Alcamo et al.
(2000) note that there is no objective basis for setting stress
thresholds for the water exploitation index (WUPR);
however, they suggest that a withdrawal to availability
(WUPR) ratio exceeding 0.8 indicates very high stress; a
ratio between 0.4 and 0.8 indicates high stress, 0.2 to 0.4
medium stress, 0.1 to 0.2 low stress and below 0.1 no
stress. The European Environment Agency has essentially
adopted these thresholds but combines the medium, high
and very high stress categories into a single ‘stressed’
category (European Environment Agency 2003). Smakhtin
et al. (2004) discuss the definition and meaning of such
thresholds and suggest that 20–50% of the available river
water volume should be left for ecosystems (stressing that
this value is arbitrary and ecosystem-/river basin-dependent).
Further development of the WUPR or WSI effect on
ecosystem health (i.e. the ‘dose–response curve’) is required
to link freshwater use to damage-level indicators such as
PAF (potentially affected fraction of species). Smakhtin et al.
(2004) review on-going initiatives to enlarge the knowledge
base on aquatic species diversity, which would be necessary
to derive a relationship between WUPR or WSI and PAF.
However, they suggest that total water abstraction, rather
than only the evaporative uses, should be considered as
impacting ecosystems, arguing that it is normally unclear
how much of the withdrawn water actually returns to the
abstracted system (and in what condition). As explained
in Section 3.5, we propose considering evaporative use
flows only as a baseline approach; nevertheless, alterna-
tive world views (e.g. an ‘egalitarian view’according to
the cultural theory adopted in the EcoIndicator 99, see
Goedkoop and Spriensma 1999) might prefer taking a
more conservative attitude, including also non-evaporative
uses for FEI.
Table 2provides a description of FEI as a new impact
category following the format suggested by Guinée et al.
(2002), with WSI as the recommended characterisation
model.
5.2 Characterisation factors for FD
Given that water is an abiotic resource and that it may, in
some circumstances, be at least temporally and spatially
depleted, the abiotic depletion potential (ADP; Guinée and
Heijungs 1995) suggested as a baseline method for abiotic
resources depletion in the CML 2001 guide (Guinée et al.
2002) seems the most appropriate approach.
5.2.1 Developing ADP factors for freshwater
Adapting the ADP formula (Guinée et al. 2002, p. 544)
with the possibility of regeneration of water funds (in a
similar way as with biotic resources, see p. 546), we get:
ADP
i¼ERiRRi
Ri
ðÞ
2RSb
ðÞ
2
DRSb
ð1Þ
where ADP
i
is the abiotic depletion potential of resource i
(e.g. groundwater from aquifer x); ER
i
is the extraction rate
of resource i;RR
i
is the regeneration rate of resource i;R
i
is
38 Int J Life Cycle Assess (2009) 14:28–42
the ultimate reserve of resource i;DR
Sb
is the deaccumu-
lation rate of the reference resource for ADP (Sb,
Antimony); R
Sb
is the ultimate reserve of the reference
resource for ADP (Sb, Antimony). Note that underexploited
groundwater bodies (i.e. with RR>ER) would yield a
negative characterisation factor: Such cases would not lead
to depletion of freshwater resources and therefore should be
neglected in the assessment.
The problem with this approach is that groundwater
reserves are seldom quantified in terms of their relative
abundance compared to potential use, with the exception
of small aquifers (Prof R. Llamas, November 2007,
personal communication). When assessed on a country
level, values tend to be very uncertain, and in any case,
RR is often bigger than ER. For example, Hernández-
Mora et al. (2007) provide a very detailed assessment of
groundwater resources in Spain, but give only a rough
estimate of the country’s reserves at 150,000–
300,000 mm
3
and highlight that some specific aquifers
are known to be overexploited. If there is knowledge that
the relevant aquifer is being over-abstracted, or that fossil
water is being used, then the LCA practitioner should find
the necessary values to develop ADP factors for the specific
water bodies in question. For instance, Custodio (2002)
provides a detailed examination of many cases of reported
overexploited aquifers around the world. As an illustra-
tion, he cites the case of California where an overexploi-
tation (ER−RR) of 2.5×10
9
m
3
year
−1
and estimated total
reserves of 1,600× 10
9
m
3
are reported (of which only
140×10
9
m
3
are assumed usable). In Almeria (Spain), the
same author reports a depletion rate (ER−RR) of 50 ×
10
6
m
3
year
−1
and total reserves of 1,100×10
6
m
3
(of
which 750×10
6
m
3
are usable; approximately 15 years to
depletion). Applying the ADP formula above, and using as
a reference Antimony’sR
2
/DR of 5.69×10
−24
, yields the
following ADP for groundwater (gw):
ADPgw;California ¼1:72 105kg Sb eq=kg:
ADPgw;Almeria ¼7:26 109kg Sb eqkg:
These ADP factors are much higher than ADP for most
other abiotic resources listed in Guinée et al. (2002). This
indicates that water may be comparatively more vulnerable
to depletion locally than other resources. Thus, when water
from overexploited aquifers is involved, it may dominate
the ADP category. However, in most cases, groundwater
will be found to be renewable (i.e. RR>ER) and neglected
from the abiotic resource depletion impact category.
5.2.2 Possible further sophistication: damage approach
‘Depleted’(dissipated) water actually returns to being
naturally available in a short period of time. However, if
one acknowledges that water may be temporally and locally
depleted, consequence-oriented approaches such as the
surplus energy to obtain the resource when the natural
source has been exhausted may be used. In the case of
water, desalination may be considered the ultimate backup
technology (Stewart and Weidema 2005). In the 1970s,
early reverse osmosis plants used as much as 20 kWh of
electricity per cubic metre, but this was reduced to as little
as 3.5 kWh/m
3
by the end of the 1990s (Fritzmann et al.
2007). The theoretical amount of energy required for
desalination of seawater, regardless of the technology used,
has been estimated to be less than 1 kWh/m
3
(Avlonitis et al.
2003).
Desalination technologies are continuing to evolve, and
it is hard to predict how far efficiency might improve in the
Table 2 Description of the necessary components of FEI according to ISO 14044 (point 4.4.2.2.)
Component Description
LCI results assigned to impact
category
Described in Section 3.5
Characterisation model 3 options suggested. Recommended model: WSI
WSI ¼WU
WREWR
where WSI is the water stress indicator (Smakhtin et al. 2004); WU (water use) is the amount of water
abstracted for human uses; WR are the renewable water resources; and EWR are the Environmental
Water Requirements (Smakhtin et al. 2004)
Characterisation factors WSI for main world river basins are provided in Table A-2in Electronic supplementary material
Category indicator m
3
of ‘ecosystem-equivalent’water, referring to the volume of water likely to be affecting freshwater
ecosystems. It does not seem appropriate to utilise a reference ecosystem such as ‘m
3
Amazon-water-
equivalents’. From a damage approach, the PAF of species may be used once a relationship is worked
out between WSI and PAF
Category endpoint Freshwater ecosystems
Environmental relevance There is a good linkage between the category indicator results and the endpoint
Int J Life Cycle Assess (2009) 14:28–42 39
future. The ammonia–carbon dioxide forward osmosis
process, which has been demonstrated in the lab and for
which a pilot-scale plant is currently under construction
(Elimelech 2007), has an estimated energy requirement of
0.84 kWh of energy per cubic metre of water desalinated
(McGinnis and Elimelech 2007). Much of this energy may
be sourced as low temperature heat (as low as 40°), with
only 0.25 kWh of energy required as electrical power.
From the above discussion, the lower and upper limits
for the energy requirements (Stewart and Weidema 2005)
for water ultimate backup technology (desalination) are
0.84 and 3.5 kWh/m
3
.
6 Discussion
It is crucial for LCI to distinguish between evaporative and
non-evaporative uses of freshwater. The paper provides
guidance to quantify both types of uses for the main
processes commonly assessed with LCA.
Some value judgements have been identified through the
paper: One of the main issues is whether non-evaporative use
of water should be addressed at all in impact assessment
(LCIA). Although we suggest considering only evaporative
losses as affecting freshwater ecosystems, a more precau-
tionary (egalitarian) approach might call for inclusion of
non-evaporative use as well. Indeed, considering that non-
evaporative uses of water represent no impact on FEI may
yield an underestimation of local effects (e.g. when water
abstracted from a river does not return to the same river):
Locally, the effects on river flows might be devastating.
A new midpoint impact category, FEI, has been suggested,
which could be linked to ecosystem health at a damage level.
Three different indicators have been assessed for FEI, of
which WSI appears as the most useful at the present state of
understanding. However, information on water use for
different river basins is not usually available in LCA (except
for the foreground system). Average values for larger regions
(e.g. Europe) may be necessary, although possibly not so
representative from an ecosystem point of view.
It needs to be highlighted that variation in seasonal flows is
not considered in the FEI indicators suggested (which are
based on annual resource estimates): During drought periods,
any abstraction may have dramatic consequences for ecosys-
tems, but this is not yet implemented in the approach. On the
other hand, time dependency is implicitly included in the VW
equations for evaporative use of water in agriculture.
The uncertainty and inherent variability in most param-
eters for the calculation of water characterisation factors are
big, and the stakes are particularly high when inter-country
comparisons leading to sensitive commercial decisions are
to be made. For the USA, for example, estimates of internal
renewable water resources range from 1,890 to 3,760 km
3
depending upon the source (FAO 2003). The Aquastat
database (FAO 2004) lists the USA as having 2,818.4 km
3
.
This closely links to the spatial dependency issue, as
pointed out by Owens (2002). This paper suggests a simple
yet plausible set of LCI models and LCIA characterisation
factors that may be used as a default. It is left to the
practitioner’s responsibility to provide more detailed data
for the foreground level if relevant, following the frame-
work defined in this paper to derive new characterisation
factors for smaller reference areas such as river basins.
The approach suggested in this paper has important
requirements for LCI databases with respect to the number
of elementary flows that need to be recorded. We argue that
this requirement is justified by the relevance of impacts on
freshwater ecosystems, whilst we recognise that some
decisions will need to be made on where the spatial
differentiation is justified. Besides, it needs to be decided
whether water flows are recorded on a national level or a
river basin level; information currently exists for both, and
the latter has more environmental meaning. In any case,
‘average’characterisation factors will be required for
regions such as Europe, North America or Asia because
site-generic LCI databases will not have more precise
information for specific manufacturing sites. Recent meth-
odological developments based on the WUPR suggest
grouping water flows in six classes according to scarcity
(Frischknecht 2008); this would significantly increase the
possibilities of applying such an approach, with a small
compromise on resolution of relative FEI. In the case of
FD, the impact is so localised that it will probably affect
only known cases of aquifer over-abstraction in the
foreground system; modelling such cases should not cause
a problem for LCI databases.
7 Conclusions
Current work to calculate VW associated with delivery of
goods and services, particularly for biotic production
systems, can help to compile LCI information. In
particular, calculations of blue and green water use are
valuable for both VW calculations and LCI. However,
the requirements for LCA differ from VW estimation, e.
g., green water, essential in VW calculations to show the
total water use of a crop, receives a characterisation
factor of zero in LCIA.
This paper addresses primarily quantitative aspects
related to freshwater use, but highlights some qualitative
issues that are not yet properly assessed in LCIA, namely
impacts on aquatic ecosystems due to changes in water
temperature (e.g. from cooling water) and impacts on
human health from microbiological pollution (usually in
less developed countries).
40 Int J Life Cycle Assess (2009) 14:28–42
The information required for systematic assessment of
freshwater resources in LCA and VW is generally available,
although it will require some degree of consensus on the level
of detail and extensive effort from LCA database developers.
This paper has proposed indicators that will improve the
representation in LCA of impacts arising from use of
freshwater resources and will be useful to assess the
comparative merits/threats posed by water-intensive prod-
ucts such as food or feedstock for bioenergy sourced from
different regions. Such characterisation may also be useful
for VW studies aiming to highlight the potential consequen-
ces of trade on source countries, in terms of impacts on
freshwater ecosystems and long-term freshwater availability.
One new midpoint-level impact category has been
suggested, freshwater ecosystem impact, which addresses
the potential effects on aquatic ecosystems caused by
changes in freshwater availability. CF have been proposed
for the midpoint level approach as well as some hints to
derive related factors for a damage level approach.
In addition, the use of groundwater resources should be
considered alongside other resources in categories for
abiotic resource depletion, following current methods to
derive CF from use and replenishment rates. Preliminary
indications are that groundwater depletion can be much
more significant than depletion of other abiotic resources
when aquifers are over abstracted.
8 Perspectives/needs for further research
Future development of the WSI to measure FEI will require
measurements of WSI at the scale of individual river basins,
along with data on all principal groundwater bodies at least
at the scale of main geographical regions and preferably at a
more localised scale. To provide threshold values for this
impact category, better understanding is needed of the
relationship between WSI and other indicators such as the
PAF of freshwater species.
Acknowledgements Dr Vladimir Smakhtin and Prof Ramón Llamas
have provided very useful input to this paper. Dr. Milà i Canals has been
funded by the RELU project RES-224-25-0044 (http://www.bangor.ac.
uk/relu), and also acknowledges support from GIRO CT (http://www.
giroct.net) during the elaboration of this paper. The authors appreciate
the useful comments provided by two anonymous reviewers.
References
Alcamo J, Henrich T, Rösch T (2000) World Water in 2025—global
modeling and scenario analysis for the World Commission on
Water for the 21st Century. Centre for Environmental System
Research, University of Kassel, Kassel. Available at http://www.
usf.uni-kassel.de/usf/archiv/dokumente/kwws/kwws.2.pdf
Allan JA (1998) Virtual water: a strategic resource global solutions to
regional deficits. Groundwater 36(4):545–546
Allan JA (2001) The Middle East water question: hydropolitics and
the global economy. I.B. Tauris, London
Allen RG, Pereira LS, Raes D, Smith M (1998) Crop evapotranspiration—
guidelines for computing crop water requirements. FAO, Rome
Antón MA, Montero JI, Lorenzo P, Muñoz P, Castells F (2005) Use of
water in LCA: case study involving protected horticulture in the
Mediterranean. In: Castells F, Rieradevall J (eds) LCM2005.
Innovation by life cycle management. Barcelona, Spain, pp 394–
398
Avlonitis SA, Kouroumbas K, Vlachakis N (2003) Energy consump-
tion and membrane replacement cost for seawater RO desalina-
tion plants. Desalination 157:151–158
Bauer C, Zapp P (2004) Generic characterisation factors for land use
and water consumption. In: Dubreuil A (ed) Life cycle
assessment of metals—issues and research directions. SETAC,
Pensacola, USA, pp 147–152
Brent AC (2004) A life cycle impact assessment procedure with resource
groups as areas of protection. Int J Life Cycle Assess 9(3):172–179
Chapagain AK, Hoekstra AY (2003) Virtual water flows between
nations in relation to trade in livestock and livestock products.
Value of Water Research Report Series No. 13, UNESCO-IHE
Chapagain AK, Hoekstra AY (2004) Water footprints of nations.
Value of Water Research Report Series No. 16. UNESCO-IHE,
Delft, the Netherlands
Chapagain AK (2006) Globalisation of water: opportunities and
threats of virtual water trade. Balkema, The Netherlands
Chapagain AK, Orr S (2008) An improved water footprint methodology
to link global consumption to local water resources: a case study of
Spanish tomato consumption. J Environ Manage (in press)
Chapagain AK, Hoekstra AY, Savenije HHG, Gautam R (2006) The
water footprint of cotton consumption: an assessment of the impact
of worldwide consumption of cotton products on the water resources
in the cotton producing countries. Ecol Econ 60(1):186
Chenoweth J (2008a) Minimum water requirement for social and
economic development. Desalination 229(1–3):245–256
Chenoweth J (2008b) Re-assessing the standard indicator of water
scarcity. Water Int 33(1):5–18
Coltro L, Mourad AL, Oliveira PAPLV, Baddini JPOA, Kletecke RM
(2006) Environmental profile of Brazilian green coffee. Int J Life
Cycle Assess 11(1):16–21
Cuenca RH (1989) Irrigation system design: an engineering approach.
Prentice Hall, Englewood Cliffs, New Jersey
Custodio E (2002) Aquifer overexploitation: what does it mean.
Hydrogeol J 10:254–277
Dones R, Faist M, Frischknecht R, Heck T, Jungbluth N (2004) Life
cycle inventories of energy systems: results for current systems in
Switzerland and other UCTE countries. Final report Ecoinvent
2000 No. 5, Paul Scherrer Institut Villigen, Swiss Centre for Life
Cycle Inventories, Duebendorf (Switzerland, for Ecoinvent
members only)
Dudgeon D, Arthington AH, Gessner MO, Kawabata Z, Knowler DJ,
Leveque C, Naiman RJ, Prieur-Richard A, Soto D, Stiassny MLJ,
Sullivan CA (2005) Freshwater biodiversity: importance, threats,
status and conservation challenges. Biol Rev 81:163–182
Elimelech M (2007) Yale constructs forward osmosis desalination
pilot plant. Membr Technol 1:7–8
European Environment Agency (2003) Indicator fact sheet: (WQ01c)
Water exploitation index. European Environment Agency,
Copenhagen. http://themes.eea.europa.eu/Specific_media/water/
indicators/WQ01c%2C2003.1001/WEI_101003v2.pdf
Falkenmark M (1986) Fresh water—time for a modified approach.
Ambio 15:192–200
Feitelson E, Chenoweth J (2002) Water poverty: towards a meaningful
indicator. Water Policy 4(3):263–281
Int J Life Cycle Assess (2009) 14:28–42 41
FAO (1992) CROPWAT: a computer program for irrigation planning
and management. Irrigation and Drainage, Paper NO. 46, FAO,
Rome, http://www.fao.org
FAO (1993) CLIMWAT for CROPWAT: a climatic database for
irrigation planning and management. Irrigation and Drainage,
Paper No. 49, FAO, Rome, http://www.fao.org
FAO (2003) Review of world water resources by country. Food and
Agriculture Organization, Rome
FAO (2004) Aquastat: FAO’s information system on water and
agriculture. Land and water development division, food and
agriculture organization, http://www.fao.org/ag/agl/aglw/aquastat/
dbase/index.stm
Frischknecht R (2008) Regionalised assessment of fresh water
abstraction within the ecological scarcity method 2006. 35th
Discussion Forum: Assessment of Water Use within LCA. 5th
June 2008. ETH Zurich, Zurich, Switzerland, http://www.lcainfo.
ch/df/default.htm
Fritzmann C, Löwenberg J, Wintgens T, Melin T (2007) State-of-the-
art of reverse osmosis desalination. Desalination 216:1–76
Gleick PH (1993) Water in crisis: a guide to the world’s fresh water
resources. Oxford University Press, New York, USA
Goedkoop M, Spriensma R (1999) The Eco-Indicator 99. A damage
oriented method for life cycle impact assessment. Methodology
Report, Third edition June 2001. http://www.pre.nl/eco-indicator99/
ei99-reports.htm
Group for efficient appliances (1995) Washing machines, driers and
dishwashers. Danish Energy Agency, http://www.ens.dk/
graphics/publikationer/Energibesparelser_UK/gea2_col.pdf
Guinée JB, Heijungs R (1995) A proposal for the definition of
resource equivalency factors for use in product life cycle
assessment. Environ Toxicol Chem 14(5):917–925
Guinée JB, Gorrée M, Heijungs R, Huppes G, Kleijn R, de Koning A,
van Oers L, Wegener Sleeswijk A, Suh S, Udo de Haes HA, de
Bruijn JA, van Duin R, Huijbregts MAJ (2002) Handbook on life
cycle assessment. Operational guide to the ISO standards. Series:
eco-efficiency in industry and science. Kluwer, Dordrecht, The
Netherlands
Hauschild MZ, Wenzel H (1998) Environmental assessment of
products. Vol. 2: scientific background. Chapman and Hall (ed),
University Press, Cambridge, UK
Hernández-Mora N, Martínez Cortina L, Llamas MR, Custodio E
(2007) Groundwater issues in southern EU member states. Spain
Country Report. Draft of a near future article presented on April
19, 2007 in the second meeting of the EASAC working group in
the headquarters of the ARECES Foundation in Madrid
Heuvelmans G, Muys B, Feyen J (2005) Extending the life cycle
methodology to cover impacts of land use systems on the water
balance. Int J Life Cycle Assess 10:113–119
Hoekstra AY, Hung PQ (2002) Virtual water trade: a quantification of
virtual water flows between nations in relation to international
crop trade. Value of Water Research Report Series No. 11.
UNESCO-IHE, Delft, The Netherlands
Jensen ME, Burman RD, Allen RG (1990) Evapotranspiration and
irrigation water requirements. ASCE-manuals and reports on
engineering practice 70. American Society of Civil Engineers,
New York
Johnson BN (1994) Inventory of land management inputs for
producing absorbent fiber for diapers: a comparison of cotton
and softwood land management. For Prod J 44:39–45
Lindeijer E, Müller-Wenk R, Steen B (eds) (2002) Impact assessment
of resources and land use. In: Udo de Haes HA, Finnveden G,
Goedkoop M, Hauschild M, Hertwich EG, Hofstetter P, Jolliet O,
Klöpffer W, Krewitt W, Lindeijer EW, Müller-Wenk R, Olsen SI,
Pennington DW, Potting J, Steen B (eds) Life cycle impact
assessment: striving towards best practice. SETAC, Pensacola,
USA, pp 11–64
Lundie S, Peters GM, Beavis PC (2004) Life cycle assessment for
sustainable metropolitan water systems planning. Environ Sci
Technol 38(13):3465–3473
McGinnis R, Elimelech M (2007) Energy requirements of ammonia–
carbon dioxide forward osmosis desalination. Desalination
207:370–382
Micklin PP (1988) Desiccation of the Aral Sea: a water management
disaster in the Soviet Union. Science 241:1170–1176
Milà i Canals L, Domènech X, Rieradevall J, Fullana P, Puig R (2002)
Use of life cycle assessment for the establishment of the
ecological criteria for the Catalan eco-label of leather. Int J Life
Cycle Assess 7(1):39–46
Milà i Canals L, Burnip GM, Cowell SJ (2006) Evaluation of the
environmental impacts of apple production using life cycle
assessment (LCA): case study in New Zealand. Agric Ecosyst
Environ 114:226–238
Milà i Canals L, Bauer C, Depestele J, Dubreuil A, Freiermuth
Knuchel R, Gaillard G, Michelsen O, Müller-Wenk R, Rydgren B
(2007) Key elements in a framework for land use impact
assessment in LCA. Int J Life Cycle Assess 12(1):5–15
Mohamed YA, van den Hurk BJJM, Savenije HHG, Bastiaanssen
WGM (2005) Hydroclimatology of the Nile: results from a
regional climate model. Hydrol Earth Syst Sci 9(3):263–278
Muñoz I, Rieradevall J, Domènech X, Milà i Canals L (2004) LCA
application to integrated waste management planning in Gipuzkoa
(Spain). Int J Life Cycle Assess 9(4):272–280
Muñoz I, Rieradevall J, Domènech X, Gazulla C (2006) Using
LCA to assess eco-design in the automotive sector: case study
of a polyolefinic door panel. Int J Life Cycle Assess 11(5):323–
334
Muñoz I, Milà i Canals L, Clift R (2008) Consider a spherical man—a
simple model to include human excretion in life cycle assessment
of food products. J Ind Ecol (in press)
Owens JW (2002) Water resources in life-cycle impact assessment.
Considerations in choosing category indicators. J Ind Ecol 5:37–54
Raskin P, Gliek P, Kirshen P, Pontius G, Strzepek K (1997) Water
futures: assessment of long-range patterns and problems. Swed-
ish Environment Institute/United Nations, Stockholm
Rockstrom J, Lannerstad M, Falkenmark M (2006) Assessing the water
challenge of a new green revolution in developing countries. http://
www.pnas.org/cgi/content/abstract/0605739104v1
Sauer BJ, Hilderbrandt CC, Franklin WE, Hunt RG (1994) Resource
and environmental profile analysis of children’s diaper systems.
Environ Toxicol Chem 13:1003–1009
Smakhtin V, Revenga C, Döll P (2004) Taking into account environ-
mental water requirements in global-scale water resources assess-
ments. Comprehensive Assessment Report 2, http://www.iwmi.
cgiar.org/assessment/FILES/pdf/publications/ResearchReports/
CARR2.pdf
Stewart M, Weidema B (2005) A consistent framework for assessing
the impacts from resource use. A focus on resource functionality.
Int J Life Cycle Assess 10(4):240–247
United Nations Development Programme (2006) Human development
report 2006. Oxford University Press, New York, http://hdr.undp.
org/hdr2006/report.cfm
WHO (World Health Organization) (2004) Water, sanitation and
hygiene links to health: facts and figures. Geneva, Switzer-
land, http://www.who.int/water_sanitation_health/factsfigures
2005.pdf
WWF (2006) Living planet report 2006. WWF International, Switzerland
WWF (2007) 10 rivers most at risk report. http://www.panda.org/
about_wwf/what_we_do/freshwater/problems/river_decline/
10_rivers_risk/index.cfm
Zhang L, Dawes WR, Walker GR (1999) Predicting the effect of vegetation
changes on catchment average water balance. Technical Report 99/
12. Cooperative Research Centre for Catchment Hydrology
42 Int J Life Cycle Assess (2009) 14:28–42
