BIBLIOTHEQUE SYNTHESIS

Environmental impacts evaluations methods of water use

LOISEAU Eléonore E-mail: [email protected]

February 2010

AgroParisTech–ENGREF centre de Montpellier B.P.44494 – 34093 MONTPELLIER Cedex 5 Tél. : (33) 4 67 04 71 00 Fax : (33) 4 67 04 71 01

Equipe ELSA (Environmental Lifecycle and Sustainability Assessment) Montpellier SupAgro 5 Place Viala 34000 Montpellier Tél. : (33) 4 99 61 21 71

ABSTRACT The aim of the study is to analyze how water use is taken into account in environmental impact assessment methodologies. The study will focus on the Life Cycle Assessment methodology, which quantifies all the consumptions and emissions of a good or a service from cradle to grave. Firstly, we will show that LCA considers water as a compartment like air or soil, and thus quantifies the environmental impacts done on this compartment (eutrophication, acidification, aquatic ecosystem toxicology). Although water is already studied in LCA as a compartment, water is also a resource such as petrol or copper. Methodologies have been developed to assess this kind of resources; nonetheless water is a specific resource with a lot of particularities (lifecycle, origins, use, and so on). In order to fill this gap, researchers have proposed new indicators. They have been defined thanks to the concepts of Virtual Water and Water Footprint, and by using water stress indicators. Consequently, after describing these concepts, we will show how they have been applied to LCA and what are the next challenges on LCA. Keywords: Life Cycle Assessment (LCA), Virtual Water (VW), Water Footprint, eutrophication, acidification, human health, quality of ecosystem, resource depletion, water stress indicator

RESUME L’objectif de cette synthèse est d’étudier les méthodes d’évaluation des impacts environnementaux prenant en compte les différents usages de l’eau. Cette étude fera plus particulièrement le point sur la méthodologie de l’ACV, qui quantifie toutes les émissions et les consommations des biens et des services tout au long de leur cycle de vie (c’est à dire « du berceau à la tombe »). Dans un premier temps, il faudra définir le concept de l’ACV et montrer que l’ACV considère l’eau comme un compartiment, au même titre que le sol et l’air, et donc quantifie les impacts subis par l’eau (eutrophisation, acidification, écotoxicologie aquatique). Bien que l’eau soit déjà étudiée dans l’ACV en tant que compartiment, l’eau est également une ressource comme le pétrole ou le cuivre. Néanmoins, l’eau est une ressource à part, ayant des caractéristiques propres (cycle de vie, origines, usages, etc.). Afin de prendre en compte ces aspects, des équipes de recherche ont proposé de nouveaux indicateurs. Ils ont été définis sur la base des approches de l’eau virtuelle et de l’empreinte sur l’eau, et en utilisant les indicateurs de stress hydrique. Par conséquent, après une présentation de ces concepts, il sera montré comment ils ont été intégrés dans la méthodologie de l’ACV et quels sont les futurs axes de recherche à développer. Mots clé: Analyse du Cycle de Vie (ACV), Eau virtuelle, Empreinte sur l’eau, indicateur de stress hydrique, eutrophisation, acidification, santé humaine, qualité des écosystèmes, épuisement des ressources

2

Introduction _______________________________________________________________ 4 1.

Assessment of environmental impacts undergone by the compartment water in LCA _ 5

1.1.

LCA Principles ________________________________________________________ 5

1.2.

Definition of impacts ___________________________________________________ 6

1.2.1.

Quantification of midpoint impacts _______________________________________ 7

1.2.1.1.

Acidification _______________________________________________________ 7

1.2.1.2.

Eutrophication______________________________________________________ 8

1.2.2.

Quantification of damages (endpoint impacts) ______________________________ 8

1.2.2.1.

Human health ______________________________________________________ 9

1.2.2.2.

Quality of ecosystems ________________________________________________ 9

1.2.2.3.

Resources depletion__________________________________________________ 9

2.

The methodological contribution of Virtual Water _____________________________ 9

2.1.

Basis of virtual water ___________________________________________________ 9

2.2.

The three components of virtual water ____________________________________ 10

2.3.

Calculation of the three components of virtual water_________________________ 10

2.3.1.

Green water use (ug(t)) _______________________________________________ 10

2.3.2.

Blue water use (ub(t))_________________________________________________ 10

2.3.3.

Grey water use (or polluted volume of water resource, up(t))__________________ 10

3.

Recognition of water as a resource in LCA __________________________________ 11

3.1.

Proposition of midpoint impacts _________________________________________ 11

3.1.1.

Impacts of water use on human health____________________________________ 12

3.1.2.

Impact of freshwater use on ecosystem (Freshwater Ecosystem Impact, FEI) _____ 12

3.1.2.1. Water stress index developed by Falkenmarks & al (Water Resource Per Capita, WRPC) 12 3.1.2.2.

Index of Water Use Per Resource (WUPR)_______________________________ 12

3.1.2.3.

Water Stress Indicator (WSI) _________________________________________ 12

3.1.3. 3.2.

Freshwater Depletion (FD) ____________________________________________ 13 Adjust of endpoint impacts______________________________________________ 13

3.2.1.

Water Stress Index (WSI) ______________________________________________ 13

3.2.2.

Damages to human health _____________________________________________ 13

3.2.3.

Damages to ecosystems quality _________________________________________ 14

3.2.4.

Damages to resources ________________________________________________ 14

3.3.

Results______________________________________________________________ 14

Conclusion _______________________________________________________________ 16 Bibliography ______________________________________________________________ 17

3

Introduction

Freshwater is a scarce and precious resource on Earth. It is essential for human being and for the stability of ecosystems (as a resource and as a habitat). So, more than 100 000 species (which stand for more than 6% of the total of identified species on earth) live in freshwater and countless other need freshwater to survive. These ecosystems are vulnerable to shortages of water, and that can lead to their extinction. For human being, the lack of available good standard water can also be life threatening. Each year in the world, nearly two billion of people die from diarrhoeal diseases, with 88% of the cases attributed to a water of poor quality and unadapted to sanitary needs (WHO 2004). However, the amounts of water needed for domestic uses are insignificant compare to the amounts used by agriculture to feed humanity. By taking into account the domestic uses and the amounts of water contained in foodstuffs, in the mean, a person consumes 3400 liters a day (there is a great variability between countries, from 2700 liters in India to 6000 liters in USA). According to data, amounts of water used in agriculture should be doubled in order to eradicate malnutrition in 2025 (Mila i Canals & al. 2009). Consequently, environmental impact assessment methods of water use are vital to estimate the scarcity of that resource, to analyze human’s influences on this compartment and to develop solutions in order to improve its allocation and to optimize its consumption. There are a lot of environmental impact assessment methods. Two of many are the Life Cycle Assessment (LCA) approach and the Virtual Water (VW) concept. They seem to form a part of the rising ones to answer to the issue of assessing water use at a global scale. LCA is a multi criterion evaluation method, compare to virtual water, which focuses on water as a resource. LCA is a method in development whose aim is to improve its indicators. For example, water uses impacts (on ecosystems, on human health, …) are not well studied because there are complex. To fill this gap, searchers have built new indicators on the basis of the concept of virtual water. Thereby, they open the way for a better recognition of this resource which can be (rain) or not to be renewable (fossil water). The purpose of this synthesis is to evaluate the situation on the recognition of water in LCA. It will present the indicators, which help to assess the environmental impacts on the compartment “water”. It will also show the recent data on the development of new indicators analyzing water as a resource.

4

1. Assessment of environmental impacts undergone by the compartment water in LCA 1.1. LCA Principles There are many methods of environmental assessment. Depending on the scale, the objectives and the potential users, the method will differ. LCA is an approach focused on a "product" or a "sector". It allows comparisons between patterns of production and consumption, identification of pollution’s transfers and awareness of the environment in production systems. This method addresses to governments, businesses, institutions, retailers, consumers, etc.. According to ISO 14040 (1997), LCA is a “technical evaluation of environmental aspects and potential environmental impacts associated with a product system”. The product system is the lifecycle of the product. The inventory and the potential impacts are considered throughout the lifecycle of the product: the acquisition of raw material, its production, use and destruction. It is also called "cradle to grave". For the implementation of the method, a methodological framework has been defined and formalized in ISO 14,040 to 14,043 (ISO 1997, 1998a, 1998b, 2000). It includes four steps:
Goal and scope definitions Its aim is to identify for whom (target group, government, department of a company, etc.) and why the study is performed (identification of key impacts of a product, improvement of an existing product, choice of one product over another, drawing-up of an environmental policy and so on). This step also defines the function of a product and its Functional Unit (FU), which quantifies the function performed by the product studied. The environmental impacts will be reported to this unit (Boeglin & Veuillet 2005). For example, the functional unit of a packing bag is to pack 9000 liters of cargo (ECOBILAN PriceWaterHouseCoopers). Direct comparison of the environmental impacts of a bag A and a bag B is meaningless and could lead to misinterpretation (eg, bag B is 20% more polluting than bag A, but if A has a 30% smaller capacity than the bag B, it is impossible to conclude that a bag is less polluting than another. Therefore it is necessary to use a functional unit). According to the study’s aim, the UF can be chosen differently, and this will affect the results and the interpretation (Basset-Mens 2005).
The lifecycle inventory (LCI) During this step, the system’s boundaries are defined. The resources consumed and the emissions to the air, to the water and to the soil are quantified for each stage of the product’s lifecycle. Once quantified, these inventory data are aggregated throughout the life cycle and expressed by UF. Each data contains spatial and temporal characteristics that are lost in this aggregation.
The lifecycle impact assessment (LCIA) Once the substances have been listed, they are converted into impacts. This step is dealt with a part later in the synthesis.
Interpretation It is a key step that assesses the robustness of the results, the choices and the assumptions. The initial objectives of the study are taken to evaluate the results and offer conclusions and adapted recommendations (Basset-Mens 2005).

5

Figure 1 presents the four steps of the LCA framework as suggested by the ISO norm. As shown in this diagram, the LCA is an iterative process. The choices and assumptions proposed in the study can be modified by acquisition of new information (CIRAIG 2005)

Figure Erreur! Argument de commutateur inconnu. Methodological framework of LCA : the four steps of a LCA (ISO 14040, 1997)

1.2. Definition of impacts Even if the conceptual framework of LCA is unique, with its four steps (norm ISO), there are differences in the definition of the impacts and in their evaluation. Generally, there are two types of impacts. The first type is the midpoint impact (impact in the middle of the causal chain), which quantifies global effects from released or consumed substances. The second type is the endpoint impact, at the end of the causal chain. Its aim is to quantify the potential damages caused by global effects (figure 2).

S O2 Emissions

Acid rain

Acidified Lakes

Dead fish

Biodiversity loss Endpoint

Midpoint

Damages

Impacts

A c a u s a l i t y c h a i n d u e t oe S mOi s s i o n s

2 Figure Erreur! Argument de commutateur inconnu. Domino effects : sulfur dioxide emissions (Roux 2009)

There is a long list of midpoint impacts according to the methodology employed. The most widespread are greenhouse effects, ozone couch degradation, photochemical oxidation, toxicity, acidification, eutrophication, and so on. The advantage of midpoint impacts is that they are more accurate as they are at the beginning of the causal chain. However, endpoint impacts are sometimes preferred as they are more comprehensible for non-informed public. In the literature, there is a consensus between the different methods of impacts’ evaluation (greenhouse effect, acidification, eutrophication, ozone, …). Yet, other impacts are always in discussion like toxicity, ecotoxicity and resource depletion.

6

1.2.1. Quantification of midpoint impacts Emissions and extractions are weighed in each type of impacts they contribute. In order to weigh, characterization factor are defined. They show the relative importance of emissions (or extractions) of a substance for a given midpoint impact. These factors have to be validated scientifically and well quantified. The released masses are multiplied by these factors and added for each type of impacts so as to get a score of midpoint impact, often expressed in kilograms of a reference substance. For example, all greenhouse gas emissions (methane, CFC, CO2, …) are stated in kg CO2 equivalent (Jolliet & al. 2005). For the compartment “water”, there are many midpoint impacts. The most studied are acidification and eutrophication. 1.2.1.1.Acidification Some compounds released in the air (SO2, NOx) may be oxidized and transformed in acids (sulfuric acids, nitric acid). Rainfalls may wash these acids and cause acid rains. Then, the acid compounds are present in streaming waters and in watercourses. That acidification leads to significant impacts on flora and fauna. The indicator of acidification is built as the indicator of climate change, by taking as reference the contribution of SO2 to acidification. Thus, all emissions of compounds, which are likely to acidify the environment, are stated in kg SO2 equivalent. The characterization factor for the impact acidification is the Acidification Potential (AP). The Acidification Potential of a substance i (APi) is defined as that:

APi =

Number of protons produced b y 1 k g o f s u b s t a n c e i Number of prtons produced by 1 kg of SO2

Thereby, if the value of AP of a substance is superior to 1, the substance is more acidifying than SO2, since it releases more protons. Thanks to the AP’s values of released acidifying substances to the air, it is possible to calculate the impact acidification: Acidification = É i APi . mi mi : m a s s o f s u b s traenl ecaes ie todt h e e n v i r o n m(w ea n t e, rsoil, air)

These calculations are valid for optimal conditions. Thereby, the acidification potential calculates the maximum potential of acidification of substances. However, in reality, the impact acidification is dependant on local conditions. For example, if the mineralization rate of the receiving environment is high, the AP will be lower than the one estimated in optimal conditions. Moreover, the Acidification Potential of a substance may be also reduced if there is a buffer zone in the receiving environment, or if protons are absorbed or wiped out by biomass. These cases are particularly true for NOx and NHx, where the real acidification value can vary from 0% to 100% of Acidification Potential. To overcome these differences due to local conditions, several methods have been suggested:
Neglect the emissions in non-sensitive areas to acidification
Weigh the emissions according to the degree of sensitivity of areas where molecules are discharged
Analyze and compare the scores of two extreme scenarios (minimal and maximal acidification)

7


Take into account the characteristics of the receiving environment and the transport of acid compounds released in the air Amongst these four propositions, the last one seems to be the more relevant. Indeed, it integrates the characteristics of the receiving environment and also the transport modeling of acid compounds. These compounds, depending on weather conditions, can travel hundreds of kilometers from their emission sites (Guinée & al. 2001). 1.2.1.2.Eutrophication It corresponds to an excess of nutrients (nitrogen and phosphorus) in aquatic and terrestrial ecosystems. It can lead to a change in the composition of species living in the environment and to an overproduction of biomass. The characterization factor is the Eutrophication Potential (EP). The Potential Eutrophication of a substance i (EPi) shows the potential effect of a substance on biomass formation, by comparing it to the effect of phosphate PO43-. It is a dimensionless number.

EPi =

Potential contributions to eutrophication of 1 mole o f s u b s t a n c eMolar i / mass of i Potential contributions to eutrophication of 1 mole of PO 43- / M o l a rmass of PO 43-

The impact eutrophication is stated in kg PO43- equivalents. It is calculated like this : Eutrophication = É i EPi . mi mi : m a s s o f s u b s tr ae nl ecaes ie todt h e e n v i r o n m(we an t e, rsoil, air)

This approach has been adopted for two main reasons. First of all, it enables to get a global characterization factor, independent of local conditions. Secondly, it does not take into account the characteristics of the receiving environment (freshwater, seawater, groundwater, air). Nevertheless, these aspects can be criticized, and two propositions have been made to study local conditions. The first one is to divide ecosystems in subcategories (aquatic, terrestrial, default in nitrogen, default in phosphorus). The second one is to study modeling of impacts at a regional scale. That work has been done for NOx and NHx but only for the compartment “air”. Yet, eutrophication has above all significant impacts on aquatic environment. Consequently, two methods exist to quantify terrestrial and aquatic ecosystems but the scores cannot simply be summed. Finally, Guinée & all suggest to keep the method showed above because it quantifies the impact eutrophication for all environmental compartments (Guinée & al. 2001).

1.2.2. Quantification of damages (endpoint impacts) The characterization of damages enables to evaluate the midpoint impacts contribution to one or several categories of final damages (human health, quality of ecosystems and resources depletion). Thus, it is necessary to quantify the damages generated by unit of each reference substance. It is like creating characterization factors of damages by unit of each reference substance defined in the step of impacts characterization. These characterization factors are multiplied by the scores of midpoind impacts so as to get the score for each category of damages (Jolliet & al. 2005). There are several methods to characterize damages. One of the most employed is Eco Indicator 99. It contains three endpoint impacts (also called areas of protection). 8

1.2.2.1.Human health It is expressed in equivalent of life years lost (Disability Adjusted Life Years, DALY). It is a unit developed by the World Health Organization (WHO). It enables to weigh morbidity and mortality. Morbidity links effects on health to the number of years lived with disablement (Years Lived Disabled, YLD). Mortality evaluates the number of years of life lost (Years of Life Lost, YLL). Morbidity and mortality are summed in that way : 1 year of life lost = 4 years live disabled (Goedkoop & Spriensma 2000). 1.2.2.2.Quality of ecosystems It is expressed in percentage of disappeared species in specific areas because of the environmental pressure. Several midpoint impacts can damage ecosystems. Thus, it is important to find a single unit. Ecotoxicity is stated in percentage of species present in the environment, which are affected and under a toxic stress (Potentially Affected Fraction of species, PAF). As it is not an observable damage, a conversion factor is used to translate toxic stress in a more observable damage. It is a gross factor, expressed by the percentage of disappeared species (Potentially Damage Fraction of species, PDF). Acidification and eutrophication are also expressed in PDF, but only for vascular plants. It is the same pattern for the impacts due to land use (Jolliet et al. 2005). 1.2.2.3.Resources depletion This damage is characterized by the surplus of energy needed for future removals. So as to calculate that energy, for each type of resource, the method refers to a back-up technology, able to remove it. For instance, the technology employed for water is desalination plant (Goedkoop & Spriensma 2000).

2. The methodological contribution of Virtual Water Water is already taken into account in the LCA framework. It evaluates midpoint and endpoint impacts undergone by the compartment “water”. However, water as a resource may have impacts on ecosystems, on human health and on its own depletion. These damages have not yet been studied in LCA. To fill this gap, studies showed the interest of the methodology of virtual water. 2.1. Basis of virtual water J. A. Allan has developed that concept since the 90’s so as to count the volume of water traded through global markets. Several methods of accountancy of volumes of traded water exist. One of the most usual definitions is “the quantity of water consumed throughout the production of a good”. For instance, to account the volume of water contained in a foodstuff, a simulation of its needs is used. It is about a net consumption of water. Water is said “virtual” because the real quantity of water contained in a product is more lower than the quantities employed to product it (especially in agriculture) (A.K. Chapagain & Orr 2009). One aim of that concept was to find a tool to manage tensions linked to water use in some countries. Indeed, in a country where freshwater resource is low, the importation of high water consumers’ products is a good mean to loosen pressure (Fernandez 2008). Virtual water can be employed to calculate water’s consumption of a country, the concept is called water footprint. It is equal to the total domestic consumption of a country, minus its

9

exportations of virtual water and plus its importations of virtual water (A.K. Chapagain & Orr 2009). 2.2. The three components of virtual water The concept has been above all developed for agriculture. This sector of activity is very dependant on freshwater resource. A lot of researches enable to quantify and calculate global flows of virtual water thanks to indicators. Beyond evaluating the sustainability of exportsbased areas, the developed indicators have as goal to point the origin of the resource (A.K. Chapagain & Orr 2009). Thereby, virtual water has been divided in three components :
Green water : it comes from rainfalls and it is naturally present in the ground. Its quantity depends on weather conditions and land use.
Blue water : it is available water for withdrawals (ex: irrigation). The mobilization and the management of blue water ask a lot of means (financial and economical). Thus, it can be the source of pressure.
Grey water (or polluted) : It is a theorical amount of water that would be required to dilute polluants emitted during the production process to such an extent that the quality of the ambient water would remain below agreed water quality standards. 2.3. Calculation of the three components of virtual water 2.3.1. Green water use (ug(t)) It is equal to the minimum effective rainfall, peff (t), and the crop evaporation requirement for each time step over the entire length of crop period. The effective rainfall is the part of rainfall, that contributes to satisfy the needs in water of crops or/and the net need of irrigation water (in mm) (FAO 1978). The crop evaporation requirement estimates the evaporation of a crop growing in optimal conditions (A.K. Chapagain & Orr 2009). 2.3.2. Blue water use (ub(t)) It is the minimum irrigation requirement, Ir(t), and the effective irrigation water supply, Ieff(t), for each time step over the entire length of crop period. 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 (A.K. Chapagain & Orr 2009). 2.3.3. Grey water use (or polluted volume of water resource, up(t)) It is equal to the value of the ratio of the polluant mass (in ton) on the allowed concentration of this polluant in the media (in ton/m3) (A.K. Chapagain & Orr 2009). To put in application these concepts, important means have been requested so as to generate databases for a lot of countries. Consequently, the components of virtual water and their databases constitute a reference for the development of indicators on water use in LCA.

10

3. Recognition of water as a resource in LCA The impacts on the compartment “water” have already been developed in LCA methodologies. However, the recognition of water as resource in LCA lacks because of its complexity. Indeed, two water’s characteristics are significant on impacts of water use on environment. There are:
The origin of freshwater (renewable or not renewable water, sensitive environment to lack of freshwater, etc.)
Water use : they are classified in four categories. For each category, it is more or less easy to quantify the impacts (Pfister & al. 2009) : o In stream use : is the use of water in situ (e.g. for a dam for hydroelectric power or navigational transport on a river) o off-stream use : is the use of water that requires removal from the natural body of water or groundwater aquifer (e.g. pumping or diversion for municipal, agricultural, or industrial uses) o Degradative (or non evaporative) use : defines the withdrawal and discharge into the same watershed after quality alteration o Consumptive (or evaporative) use : is the use of freshwater when release does not occur because of evaporation, product integration, or discharge into different watersheds or seas.

W a t e in r LCA from a resource point of view

? Boundaries ? SYSTEM (Water use)

Inputs

Water resources

• flow (r i v e r, slakes, rain)

Outputs

Water Use

• fund (groundwater)

Water available for other systems

• Evaporative • Non-évaporative

• stock (fossil water)

Figure Erreur! Argument de commutateur inconnu. Resources of water and its different use (from Mila i Canals & al. 2009)

The issue of the resource and its use which is asked here is thus quantitative : what is the water use and which part is restored to environment. 3.1. Proposition of midpoint impacts (Mila i Canals & al. 2009) Mila i Canals and his team have searched to take into account the impacts on environment caused by freshwater use. They worked on the entire causal chain. They began to describe midpoint impacts.

11

3.1.1. Impacts of water use on human health There is no significant evidence of correlation between quantities of available water by person ad by year in a country, and its human development index. Moreover, the issues of health due to water are more linked to its quality than to its quantity. From these observations, Mila i Canals & al did not take into account the impacts that water use could have on human health. 3.1.2. Impact of freshwater use on ecosystem (Freshwater Ecosystem Impact, FEI) This a new category of midpoint impacts. It can be linked to the protection area “quality of ecosystem” in the causal chain. Indeed, when men withdraw water in water bodies, it can lead to a lower quantity of water for ecosystems. According to water pressure in withdrawal area, it can damage more or less the existing biodiversity. Thereby, Mila i Canals & all proposed three indicators to describe water pressure at a country or watershed level. 3.1.2.1.Water stress index developed by Falkenmarks & al (Water Resource Per Capita, WRPC) It is based on the estimation of renewable available water quantities by person and by year. This estimation is compared to the individual water need calculated with the references of a developed country with a semi-arid climate. 1700 m3/inhabitant/year is the limit below which the shortages of water are only temporal and local. Below 1000 m3/inhabitant/year, shortages of water affect the economical development of a country, the heath and the well-being of inhabitants. Finally, below 500 m3/inhabitant/year, water availabilities are a main constrain to life (Fernandez 2007). However, this index does not take into account the capacity of a country to adapt and to import water from other countries (contrary to what is proposed in the concept of virtual water). Besides, this index considers only domestic use although the biggest water consumer is agriculture. 3.1.2.2.Index of Water Use Per Resource (WUPR) This index is the ratio between quantities of water use and quantities of available freshwater. This ratio shows the percentage of available water for other uses than the ones for human. A significant value of this index points a severe water stress because almost all available water has been used. Thus, it is a good indicator of potential impacts that could damage aquatic ecosystem. Moreover, thank to the FAO Aquastat database, data for this indicator are available for most countries. 3.1.2.3.Water Stress Indicator (WSI) (Smakhtin & al. 2004) WSI =

WU WR - EWR

W U (W a t e Use) r : a m o u n tof s w i t h d r a w n w afor t e hr u m a nuse WR W ( a t e r R e s o u) r :camounts e of f r e s h w a t e r r e s o u r c e E W R E( n v i r o n m e n t a l W a t e r R e q u)i :r w em a teenrt sr e q u i r e mfor e n et sc o s y s t e m s s u r v i v a l

This indicator seems to be more relevant to evaluate water stress impacts on ecosystems. However, in order to calculate this indicator, data are only available at watershed level. Even though, this scale is well apposite for analysis, data are not compatible with inventory data of LCA whose scale is at countries level. Finally, three indicators have been proposed to describe the Freshwater Ecosystem Impact midpoint impact category. The indicator WUPR and WSI seem to be the most relevant. Nevertheless, it is necessary to carry on researches so as to link the FEI impact to the “quality 12

of ecosystems” endpoint impact, whose unit is PAF (Potentially Affected Fractions of species) (Mila i Canals & al. 2009). 3.1.3. Freshwater Depletion (FD) Given that freshwater is an abiotic resource and that it can be depleted at some time and space levels, Mila i Canals & al consider that the Guinée approach on abiotic resources (Guinée & al. 2001) is the most relevant to assess freshwater depletion. Therefore, the Abiotic Depletion Potential (ADP) has to be adapted to freshwater resource. This approach raises an issue on data availability. Indeed, data on groundwater are rare. Consequently, it is necessary to adapt ADP at each study. If there are scientific evidences that an aquifer is overexploited or that withdrawn water is fossil, data have to be reevaluated. Besides, contrary to other resources, which have global impacts, water can be depleted at a small scale and has high impacts on local ecosystems.

3.2. Adjust of endpoint impacts (Pfister & al. 2009) Contrary to Mila i Canals and its crew, Pfister & al studied endpoint impacts. They focused on three protection areas : human health, quality of ecosystems and resource depletion. Damages are calculated using the Eco Indicator 99 framework. 3.2.1. Water Stress Index (WSI) To develop apposite and assessable indicators, the authors distinguished evaporative use and non-evaporative use. They decided to focus only on evaporative use (or consumptive water use, WUconsumptive) to make easier the assessment. Indeed, in the case of agriculture, virtual water, thanks to its databases, gives the information to evaluate WUconsumptive (only blue water is taken into account). Moreover, quantifications are done at a regionalized scale and not a global scale. To determine endpoint impacts for the three areas of protection, the authors have created a common midpoint impact, called “water loss”. Therefore, they have defined a characterization factor, the water stress index. It points the consumed water fraction by a use, which deprives other freshwater users. This index is calculated from the water stress value (ratio of total annual freshwater withdrawals to hydrological availability). To better take into account local conditions, water stress is corrected. It integrates annual and monthly rainfalls variations and constructions (such as dams). 3.2.2. Damages to human health Lack of water can be the source of poor hygienic conditions, but at a bigger scale, it can lead to irrigation decrease and thus to local malnutrition. Hygienic issues are very dependant ofnlocal conditions whereas irrigation impacts are easier to understand. It is why, only consequences of a low availability of water on agriculture are assessed in LCA. These damages are expressed in DALY. ÆHHmalnutrition = WSIi . WU%agriculture,i . HDFmalnutrition, i . WRmalnutrition-1 . DFmalnutrition . WUconsumptive, i C Fmalnutrition,i V a l u ewhich s a r ec a l c u l a t e d t h at ondkas t a b a s of e sV i r t u a l W a t e r r

13

WRmalnutrition : water requirement to prevent malnutrition (m3/year/person) HDFmalnutrition, i : human development factor. It links human development index to vulnerability to malnutrition DFmalnutrition: damage factor. It expresses damages due to malnutrition (DALY / year / person) WRmalnutrition and DFmalnutrition are independent of the studied country. Yet, HDI is a calculated value for each country. CFmalnutrition,i is the characterization factor of human health damage. It expresses the expected damage when an additional unit of water is consumed (DALY/m3). 3.2.3. Damages to ecosystems quality It is expressed in PDF. This is a complex value to calculate. To get round this difficulty, PDF values are generally assessed as vulnerability of vascular plant species biodiversity (VPSB). The vulnerability of that kind of plants is significantly correlated to net primary production (NPP). It stands for the quantity of carbon, which is caught and stocked by plants while photosynthesis. Thereby, the fraction of NPP, which is limited by freshwater availability, represents the vulnerability of an ecosystem to a lack of water. It is used as an approximation of PDF.

ÆEQ = CF E Q . WU consumptive = NPP wat-lim . WU consumptive / P CFEQ is the characterization factor of damages to ecosystems quality (m2/m3 in one year). P is the mean of annual rainfalls (m/year). Therefore, the ratio WUconsumptive/P denotes the theoretical area-time equivalent which would be needed to recover the amount of consumed water by natural precipitations. 3.2.4. Damages to resources Extraction of fossil water or overexploitation of water bodies (such as Aral Sea) can cause freshwater depletion. Back-up technologies principle is employed to estimate damages to resources. It calculates needed quantities of energy to reestablish freshwater quality. In the case of freshwater, desalination plants are the back-up technology employed.

ĘR = E desalination . Fdepletion . WUconsumptive Edesalination is the energy required for seawater desalination (MJ/m3) and Fdepletion is the fraction of freshwater consumption that contributes to depletion. Fdepletion serves also as characterization factor for the midpoint indicator “freshwater depletion”. It depends on local conditions in a watershed, so it would be different according to countries.

3.3. Results Recent researches on the quantification of environmental impacts due to freshwater use have enabled to develop new methodologies. The two studies (Mila i Canals & al and Pfister & al) on this subject show that the problematic can be addressed under two axes. Indeed, Mila i Canals proposed a method which is based on midpoint impacts assessment whereas Pfister developed an endpoint impacts assessment. Weighing up the pros and cons, 14

endpoint impacts method proposes indicators with high incertitude and the definitions employed do not meet general approval inside scientific community (Cooney 2009). In the other hand, midpoint impacts method needs to make clear the links between effects in causal chains (such as between water stress index and freshwater ecosystem impact) (Mila i Canals & al. 2009). Nevertheless, the two studies take credit for taking into account freshwater use damages and trying to quantify according LCA framework (Cooney 2009).. Both of them have, as starting point, the purpose of water footprint. It makes easier the lifecycle inventory phase. However, this purpose is disputed. It ignores the primary role of water footprint, which brings two pieces of information. First data on water footprints of products, consumers and producers inform the discourse about sustainable, equitable and efficient freshwater use and allocation. Second, water footprint accounts help to estimate local environmental, social and economic impacts (Hoekstra & al. 2009). To better assess water use impacts, the two methods need to improve data quality so as to decrease incertitudes. They have to take into account temporal variations (seasons, dry years, and so on) and regional factors (such as freshwater availability, water infrastructure, rainfall, and consumption patterns at a specific location). It is why, it is relevant to redefine the level of assessment. In the case of freshwater, the watershed level is more appropriate for the assessment, because hydrological processes are connected within watersheds (Cooney 2009). Besides, in order to be exhaustive, it is necessary to study qualitative impacts, which can damage available water amounts. Indeed, if a river is polluted, as well as damage ecosystems quality, it will deprive water for other use (Mila i Canals et al. 2009).

15

Conclusion Thanks to its rigorous methodological framework and its multi criterion assessment, LCA is, nowadays, a well-renowned environmental impacts evaluation method (Basset-Mens 2005). Nonetheless, even if the aim of LCA is to be exhaustive, it does not integrate freshwater as a resource. Indeed, it assesses well impacts on water as an environmental compartment (eutrophication, acidification, …). Yet, the recognition of impacts due to water use defaults. To fill this gap, searchers studies this lack closely. They have proposed two axes of reflection. One focuses on midpoint impacts method and the other on endpoint impacts method. Both of them have advantages and drawbacks, but they have in common the purpose of water footprint as starting point. The links between impacts in causal chain need to be consolidated to improve these two axes of reflection. Consequently, it is necessary to generate reliable databases, to define more accurate analysis levels and to take into account all the effects of water use (water pollution, increase of water temperature, and so on). Besides, LCA is one of many environmental impacts evaluation methods. It has ambitious aims but it can answer to all questions. It assesses all the impacts of a sector or a product at a global level without studying their local impacts. Thus, it is essential to couple it with other tools so as to get all the information before making a decision. Thereby, in a lot of cases an impacts study enables to assess the consequences of a decision at local level (Pfister & al. 2009). However, at a territory level, these local or global impacts studies do not give keys to elaborate sustainable settlement political policies. To meet this requirement, the concept of Industrial Ecology has come to the fore since the 90’s. It takes a leaf out of the cyclic functioning of natural ecosystems, in order to recreate at industrial level, an organization, which is characterized by low flows of energy and material, and by a high rate of recycling. In this methodological framework, LCA constitutes a judicious tool of diagnostic (Adoue 2007). Industrial ecology enables to develop land settlement scenarios and LCA evaluates their environmental performance. These concepts are recent, but perspectives of development seem to be promising.

16

Bibliography Adoue, C., 2007. Mettre en oeuvre l'écologie industrielle Presses polytechniques et universitaires romandes., Lausanne. Basset-Mens, C., 2005. Propositions pour une adaptation de l'Analyse de Cycle de Vie aux systèmes de production agricole. Mise en oeuvre pour l'évaluation environnementale de la production porcine. Thèse en Science de l'Environnement. Agrocampus Rennes, 242 p. Available at: http://www.rennes.inra.fr/umrsas/media/fichiers/these/these_claudine_basset_mens__ 1. Boeglin, N. & Veuillet, D., 2005. Introduction à l'Analyse de Cycle de Vie (ACV). Chapagain, A. & Orr , S., 2009. An improved water footprint methodology linking global consumption to local water ressources : A case of Spanish tomatoes . Journal of Environmental Management, (90), 1219-1228. CIRAIG, 2005. CIRAIG - Centre interuniversitaire de recherche sur le cycle de vie des produits, procédés et services. ACV. Available at: http://www.ciraig.org/fr/acv.html [Accédé Décembre 1, 2009]. Cooney, C., 2009. LCA finally takes water into account. Environmental Science and Technology. ECOBILAN PriceWaterHouseCoopers, www.eco-conception.fr. Bilan environnemental de 4 types de sacs de caisse. Available at: http://www.eco-conception.fr/sacs_de_caisse.html [Accédé Novembre 30, 2009]. FAO, 1978. A. FACTEURS AGRONOMIQUES. Cycle de croissance et période végétative. Available at: http://www.fao.org/docrep/003/S8500F/s8500f0e.htm#TopOfPage [Accédé Décembre 1, 2009]. Fernandez, S., 2007. L'eau virtuelle dans les pays Méditerranéens : un indicateur pour contribuer à l'analyse des questions de gestion et de répartition de l'eau en situation de pénurie?, Zaragossa: Plan Bleu. Fernandez, S., 2008. L'eau virtuelle, quelle réalité? Maghreb-Machrek, (196), 45-62. Goedkoop, M. & Spriensma, R., 2000. The Eco-indicator 99. A damage oriented method for Life Cycle Impact Assessment, Amersfoort: PRé Consultants. Guinée, J.B., Gorrée, M., Heijungs, R., Huppes, G., Kleijn, R., de Koning, A., van Oers, L., Wegener Sleeswijk, A., Suh, S., Udo de Haes, H.A., de Bruijn, H., va, Duin, R., & Huijbregts, M.A.J., 2001. Life Cycle Assessment : An operational guide to ISOstandards, Leiden: University of Leiden.

17

Hoekstra, A.Y., Gerbens-Leenes, W. & van der Meer, T.H., 2009. Reply to Pfister and Hellweg: Water footprint accounting, impact assessment, and lifecycle assessment. Proceedings of the National Academy of Sciences, 106(40), 114. ISO International Standard 14040, 1997. Management environnemental - Analyse du Cycle de Vie - Principes et cadre. International Organisation for Standardisation (ISO), Genève, Suisse. Jolliet, O., Saadé, M. & Crettaz, P., 2005. Analyse du cycle de vie : comprendre et réaliser un écobilan, Lausanne: Les Presses polytechniques et universitaires romandes. Mila i Canals, L., Chenoweth, J., Chapagain, A., Orr, S., Antón, A., & Clift, R., 2009. Assessing freshwater use impacts in LCA: Part I—inventory modelling and characterisation factors for the main impact pathways . International Journal Life Cycle Assessment, (42), 28-42. Pfister, S., Koehler, A. & Hellweg , S., 2009. Assessing the Environmental Impacts of Freshwater Consumption in LCA . Environmental Sciences and Technology, 43(11), 4098-4104. Roux, P., 2009. Evaluation environnementale, gestion de l’eau et analyse du Cycle de Vie. Présentation Réunion UMR G-eau & Pôle ELSA. Cemagref Montpellier. Smakhtin, V., Revenga, C. & Döll, P., 2004. Taking into account environmental water requirements in global-scale water resources assessments, IWMI. Available at: http://www.iwmi.cgiar.org/assessment/FILES/pdf/publications/ResearchReports/CAR R2.pdf. WHO (World Health Organization), 2004.Water sanitation and hygiene links to health : facts and figures. Available at : http://www.who.int/water_sanitation_health/factsfigures2005.pdf

18