Article pubs.acs.org/est
Water Footprint of European Cars: Potential Impacts of Water Consumption along Automobile Life Cycles Markus Berger,*,† Jens Warsen,‡ Stephan Krinke,‡ Vanessa Bach,† and Matthias Finkbeiner† †
Technische Universität Berlin, Department of Environmental Technology, Chair of Sustainable Engineering, Office Z1, Strasse des 17. Juni 135, 10623 Berlin, Germany ‡ Volkswagen AG, Group Research, Environmental Affairs Product, P.O. Box 011/1774, 38436 Wolfsburg, Germany S Supporting Information *
ABSTRACT: Due to global increase of freshwater scarcity, knowledge about water consumption in product life cycles is important. This study analyzes water consumption and the resulting impacts of Volkswagen’s car models Polo, Golf, and Passat and represents the first application of impactoriented water footprint methods on complex industrial products. Freshwater consumption throughout the cars’ life cycles is allocated to material groups and assigned to countries according to import mix shares or location of production sites. Based on these regionalized water inventories, consequences for human health, ecosystems, and resources are determined by using recently developed impact assessment methods. Water consumption along the life cycles of the three cars ranges from 52 to 83 m3/car, of which more than 95% is consumed in the production phase, mainly resulting from producing iron, steel, precious metals, and polymers. Results show that water consumption takes place in 43 countries worldwide and that only 10% is consumed directly at Volkswagen’s production sites. Although impacts on health tend to be dominated by water consumption in South Africa and Mozambique, resulting from the production of precious metals and aluminum, consequences for ecosystems and resources are mainly caused by water consumption of material production in Europe.
■
INTRODUCTION “Yet Another Footprint to Worry About: Water” was a headline in The Wall Street Journal1 with regard to the foundation of the Water Footprint Network 2 in 2008, which published surprisingly high figures of 70 L virtual water consumption3 per apple or 2700 L per cotton T-shirt. Starting from such volumetric tools, which simply aggregate consumptions of ground and surface water (blue water), soil moisture (green water4), and volumes of polluted freshwater (gray water), substantial methodological developments were undertaken recently. Modern impact-oriented water footprinting methods, which were reviewed in a previous work,5 characterize water consumption based on parameters such as local scarcity or sensitivity of population and ecosystems and model complex impact pathways. However, these methods were hardly tested or applied in complex industrial product systems. So far, most water footprint studies published focus on agricultural products such as food,6,7 natural fibers,8,9 or bioenergy10 and biofuels.11 Volkswagen has been analyzing the environmental effects of its cars and components by means of life cycle assessment (LCA)12,13 for many years.14 However, due to lack of data and appropriate impact assessment models, the consumption of freshwater has not yet been considered. © 2012 American Chemical Society
Therefore, the aim of this study is to analyze freshwater consumption along the life cycles of three Volkswagen car models on both inventory and impact assessment levels. First, regionalized water inventories are determined, showing country-specific water consumption figures, for the Polo 1.2 turbocharged direct injection (TDI), Golf 1.6 TDI, and Passat 2.0 TDI (model year 2010). Based on these inventories, seven impact assessment methods, which represent different levels of sophistication and model different impact pathways, are applied. Further objectives comprise the discussion and comparison of impact assessment results, the identification of significant life cycle stages and processes, and the analysis of sensitivity of results to altered regionalization scenarios. Finally, the potential damages resulting from water consumption are compared to damages caused by other environmental interferences, such as resource use and emissions, in order to estimate the relevance of water for the automotive industry. Received: Revised: Accepted: Published: 4091
November 9, 2011 January 28, 2012 March 5, 2012 March 5, 2012 dx.doi.org/10.1021/es2040043 | Environ. Sci. Technol. 2012, 46, 4091−4099
Environmental Science & Technology
■
Article
METHODOLOGY
The production phase comprised the mining of raw materials such as iron ore, the production of materials such as steel, the processing of components like trunk lids, and final assembly of the car. In the use phase, crude oil production and the refinery of the diesel required to run the cars for 150 000 km in the New European Driving Cycle (NEDC)16 were included. Accordingly, diesel production of 5700 L (Polo), 6750 L (Golf), and 8550 L (Passat) was considered. Washing of cars was not included in this analysis as it highly depends on individual use, personal attitude, and the technique applied, for which no reliable water consumption (evaporation) figures are available. Moreover, the washing of cars was not included in Volkswagen’s existing LCA studies upon which this study is based. Because we intended to compare impacts resulting from water consumption to damages caused by other environmental interferences, consistent system boundaries had to be ensured. Finally, it was expected that the contribution of car washing to water evaporation is rather small, but this assumption should be validated in future studies. The end-of-life phase was modeled in accordance with Volkswagen’s SiCon recycling process.17 It comprises the draining of fluids, the removal of batteries, catalysts, and spare parts, the shredding of the remaining body, the treatment and recycling of shredding residues, and the disposal of wastes. In contrast to conventional recycling approaches, this process allows for the recycling of nonmetallic shredder residues as well and enables recycling rates of 95% by weight. Environmental credits gained from the use of secondary materials in subsequent product systems were not considered. The process of modeling a life cycle inventory (LCI) of a whole car is very complex due to the fact that it involves registering thousands of components, together with any related upstream supply chains and production processes. Therefore, Volkswagen developed the slimLCI interface system,18 which enables a consistent data collection and automated modeling of the LCI in the GaBi LCA software.19 A description of the slimLCI procedure along with representative extracts of the LCI model created in the LCA software can be found in the Supporting Information. The LCI data used in this study were originally determined for the environmental commendations of the Polo, Golf, and Passat,20−22 which are reviewed by independent experts according to ISO 14040/44 (2006).12,13 To apply impact assessment methods evaluating the consequences of the water consumption determined from the LCI models, the basic volume is not enough. Regionalized water inventories, which state the location where water consumption occurs, are needed to consider regional water scarcity conditions, the vulnerability of ecosystems, or socioeconomic parameters affecting the sensitivity to water scarcity induced health damages.5 Such geographically explicit water inventories were determined in a top-down approach. First, the car’s total water consumption was divided into the shares consumed by the life cycle stages production, use, and end-oflife. For further specification, the water consumed in the production phase is assigned to manufacturing steps and to 15 material groups (specified in the German material classification in motor vehicle construction standard23). Finally, the water consumption caused by the manufacturing steps and material groups was allocated to specific countries based on production mixes, location of suppliers, production sites, etc. For instance, water consumption in the painting, final assembly, and recycling of the Polo, Golf, and Passat was assigned to Volkswagen’s production sites and recycling operators in Germany and Spain. Water consumed by the
Water Inventory. Figure 1 shows the system boundaries in which water consumption was analyzed along the life cycles of the three cars, comprising the production, use, and end-of-life phases. It should be noted that water consumption denotes only the fraction of total water use that is not returned to the same river basin from which it was withdrawn due to evaporation, product integration, or discharge into other watersheds and seawater.15 Transports between process steps, the generation of electricity, and the production of auxiliary materials were within the scope of this study but not shown explicity in Figure 1.
Figure 1. Life cycle of the cars along which water consumption was analyzed. 4092
dx.doi.org/10.1021/es2040043 | Environ. Sci. Technol. 2012, 46, 4091−4099
Environmental Science & Technology
Article
material group aluminum was allocated to countries contributing to the European aluminum import mix proportionally to their import shares.24 For polymers, water consumption was further divided into water consumed in oil extraction, refinery, polymerization, and component fabrications based on generic data available in the GaBi database.19 These water consumption shares were assigned to countries based on the European import mixes of crude oil,19 the location of refineries and polymerization plants,25 and on the location of plastic component productions conducted at suppliers and Volkswagen production sites. Impact Assessment. Several impact assessment methods are available to evaluate consequences for human health, ecosystems, and resources resulting from water consumption. In addition to spatially explicit water inventories, some methods require information on the type of watercourse,26 water qualities,27 or time of use28 as different watercourses fulfill different ecological functions, different qualities enable different uses, and scarcity can vary throughout the year. However, such high inventory requirements are hard to satisfy as the desired information is not available in today’s LCI databases.19,25 For that reason we were restricted to apply methods for which regionalized water inventories are sufficient. To allow for a comprehensive impact assessment the following methods, representing different levels of sophistication and assessing different consequences, have been selected: The ecological scarcity method,29 which assesses water consumption based on physical water scarcity, measured in eco-points/m3. The impact assessment method of Motoshita et al. (2011),30 which evaluates damage to human health caused by infectious diseases resulting from polluted water uptake as a consequence of domestic water scarcity, expressed in disability adjusted life years (DALY)/m3. The method of Pfister et al. (2009),9 comprising 5 characterization models: Freshwater deprivation, which assesses freshwater consumption based on physical water scarcity (dimensionless); Damage to human health, which addresses health impacts resulting from malnutrition as a consequence of agricultural water shortage, quantified in DALY/m3; Damage to ecosystem quality, which evaluates ecological consequences resulting from decreased biodiversity due to water shortage, measured in potentially disappeared fraction of species (PDF) m2 year/m3; Damage to resources, which assesses depletion of freshwater resources as a consequence of water uses exceeding renewability rates, expressed in MJsurplus energy/m3; and Overall damage, which aggregates impacts determined in the three previous characterization models to a single-score result, quantified in points/m3. It should be noted that the ecological scarcity method uses in general an average, policy-based approach while Motoshita et al.30 and Pfister et al.9 follow a marginal impact approach.31 The characterization factors which were used to calculate the impacts resulting from water consumption in different countries are provided in the Supporting Information of the methods of Motoshita et al.30 and Pfister et al.9 Since Motoshita et al.30 determined damage factors only for domestic water consumption, these factors were multiplied by a countryspecific ratio of domestic to total water use32 in order to allow for an assessment of general water consumption. Therefore, the method is termed Motoshita et al. (2011)* in the following. For the ecological scarcity method only few characterization factors for OECD countries are provided. Factors for nonOECD countries were calculated using hydrological data from AQUASTAT.32
In general, country-specific characterization factors were used in this study. As they reflect hydraulic conditions in countries with inhomogeneous water scarcity, like Spain or the United States, more realistically, it would be preferable to use watershed-specific factors throughout this study. However, this data is not available and cannot be generated in our topdown approach described above. Moreover, depending on the car, between 45 and 60% of water consumption occurs in countries with similar water scarcity throughout the territory such as Germany, Sweden, or Russia. Yet, in order to specify the impacts of water consumed at Volkswagens production sites in Pamplona, Spain (Polo), Wolfsburg, Germany (Golf), and Emden, Germany (Passat) watershed-specific characterization factors were applied. In the ecological scarcity method, characterization factors for the catchment areas Ebro (Pamplona, Spain), Weser (Wolfsburg, Germany), and Ems (Emden, Germany) were determined using data from the WaterGAP 2 model.33 For the five characterization models provided by Pfister and colleagues,9 watershed-specific factors were derived from a Google Earth34 layer provided by the authors.35 Only for the method of Motoshita et al. (2011)* no site-specific characterization factors, which are influenced by parameters such as house connection rate to water supply and sanitation, were determined as they are very low and similar throughout Europe. Uncertainties and Sensitivity Analysis. Uncertainties result from the LCI modeling, LCI databases, assumptions to establish the regionalized water inventories, and impact assessment models. Inevitable uncertainties resulting from the LCI modeling of complex industrial products are discussed in Koffler et al. (2008)18 but cannot be quantified in the scope of this study. Although efforts have been made to provide highquality water consumption data, limitations occur due to the lack of data in current LCI databases. For instance, the ecoinvent database25 reports only freshwater withdrawals but no wastewater discharges in its data sets and, thus, allows only for the determination of water use but not consumption. Data sets from the GaBi database19 used in this study contain consistent water in- and output figures;36 however, often only for the foreground system and for energy production but not for processes in the background system such as mining. As water consumption in electricity production provides a major share of total water consumption in industrial processes,37 the data sets can be used for water footprint calculations but may underrepresent the real water consumption. These uncertainties, which result from partly lacking water consumption figures in the background system of LCI data sets, can hardly be quantified without detailed insight into aggregated data sets available in the GaBi database. Also methodological uncertainties of the impact assessment methods cannot be assessed statistically, but they are addressed by applying seven different characterization models and comparing the results obtained. There are few uncertainties from the determination of regionalized water inventories for the manufacturing steps in the foreground system. Yet, several assumptions were necessary to geographically differentiate water consumed in the background system: It was assumed that materials are purchased according to average import mix sharesspecific information was not included. By assigning the water consumption of material groups to countries based on import mix shares, it is presumed that the water intensity for producing a certain material is equal in all countries. The assumption that material 4093
dx.doi.org/10.1021/es2040043 | Environ. Sci. Technol. 2012, 46, 4091−4099
Environmental Science & Technology
Article
production is accomplished exclusively in countries contributing to the import mix neglects the fact that minor volumes of water consumed during the production of auxiliary products might have been consumed in countries other than those included in the mix. Because regionalized water inventories are a prerequisite for all impact assessment methods, the uncertainties mentioned above cannot be avoided, however, they can be quantified by means of sensitivity analysis. As several different possibilities concerning the geographical differentiation are feasible, we decided to set up a minimum and a maximum scarcity scenario. In the minimum scarcity (min-s) scenario the individual water consumption of the 15 material groups was assigned to the countries in the corresponding import and production mixes which show the lowest physical water scarcity. In contrast, in the maximum scarcity (max-s) scenario the material group specific water consumption was fully allocated to the water scarcest country in the respective import and production mixes. Physical water scarcity was measured by means of the withdrawal-to-availability (WTA) ratio, which relates annual freshwater use to the renewable water supply in a country. Water consumption in the foreground system such as the manufacturing and recycling of the Polo, Golf, and Passat remained assigned to Germany and Spain in both scenarios. As it might be too optimistic or too pessimistic to assume that all materials were derived exclusively from the countries of the lowest or the highest water scarcity, the scenarios should be regarded as boundaries between which realistic options are possible.
Figure 2. Relative contributions of life cycle stages to total results in the impact categories eutrophication (EP), ozone layer depletion (ODP), photochemical ozone creation (POCP), global warming (GWP), acidification (AP), and in the water consumption inventory (WC) for the Golf 1.6 TDI.
share of special metals (gold, silver, and platinum group metals (PGM)) is higher than in conventional impact categories. However, these figures tend to overestimate the actual share of special metals, as their supply has been modeled with 100% of primary material. Yet, Volkswagen has been running an effective catalysts recycling program for years, which helps to recover and recycle PGM in a closed loop system. Nevertheless, the fact that less than 1 kg of precious metals is responsible for more than 20% of the overall water consumption throughout a Golf’s life cycle highlights the large material specific water consumption of these materials. After assigning the water consumption of materials and production steps to countries based on import mixes, location of production sites, etc., regionalized water inventories were established for the three cars. As shown in Figure 4, water consumption takes place in 43 countries worldwide. Less than 10% is consumed directly at the production sites in Pamplona, Wolfsburg, and Emden resulting mainly from painting and evaporation of cooling water. Hence, more than 90% of the water consumption along the cars’ life cycles is caused by the material and energy production in the background system. Detailed geographically explicit water inventories for the Polo, Golf, and Passat are available in the Supporting Information, showing country-specific water consumption figures for production (separated by the assembly and production of 15 material groups), use, and end-of-life. Impact Assessment. Based on the regionalized water inventories, the impact assessment models of the ecological scarcity method, Motoshita et al.30 and Pfister et al.9 were applied in order to evaluate consequences resulting from water consumption in different countries. Figure 5 shows the results obtained by means of the water inventory and impact assessment methods normalized to the Polo for the default, min-s, and max-s scenarios. Because absolute results for the Polo differ among the scenarios, Figure 5 only allows for comparing the cars within one impact category and scenario. Comparisons of results obtained per car and impact category in different scenarios are shown separately in Figure 6. The results of the ecological scarcity method and the impact category freshwater deprivation depend on two factors: the volume of water consumed and the physical water scarcity at the place of consumption. Whereas the ecological scarcity method uses the WTA ratio as a weighting factor directly, freshwater deprivation uses a water stress index (WSI) as a
■
RESULTS AND DISCUSSION Water Inventory. The water consumption along the life cycles of the three cars amounts to 51.7 m3 (Polo 1.2 TDI), 62.4 m3 (Golf 1.6 TDI), and 82.9 m3 (Passat 2.0 TDI). It should be noted that our figures are lower than previously reported data of 400 m3 virtual water consumption per car.38 This result was calculated based on economic input−output tables determining the water consumption per US$ of industrial product.2 Hence, it only represents a rather rough estimate of an average industrial product and is based on economic, not physical data. This fundamental difference in the modeling approach does not allow for a detailed discussion of differences. Determining the water consumption of the main life cycle stages, revealed that about 95% of the water is consumed in the production phase of all three cars (Figure 2). This is in contrast with most other environmental interferences evaluated in Volkswagen’s LCA studies, like eutrophication or global warming,39 which are usually dominated by the car’s use phase.21 Hence, it can be seen that different processes than fossil fuel consumption are relevant from a water perspective. Yet, it should be remembered that these results were obtained assuming the use of fossil diesel, whichaccording to our datahas a low water consumption of 0.005 L/MJ compared to biodiesel consuming 217−335 L/MJ on global average depending on the crop used.10 A significance analysis was accomplished to identify the contributions of individual materials and manufacturing steps to the impact assessment and water consumption results of the production phase for the Golf. As shown in Figure 3, steel and iron materials, as well as polymers, contribute equally strong to most impact categories and water consumption (70−80%). In contrast, the contribution of light metals (aluminum and magnesium alloys) to total water consumption is lower and the 4094
dx.doi.org/10.1021/es2040043 | Environ. Sci. Technol. 2012, 46, 4091−4099
Environmental Science & Technology
Article
Figure 3. Relative contributions of material groups to production impacts in the impact categories eutrophication (EP), ozone layer depletion (ODP), photochemical ozone creation (POCP), global warming (GWP), acidification (AP), and in the water consumption inventory (WC) for the Golf 1.6 TDI.
Comparison Among Cars. By showing the results of the water inventory and impact assessment methods normalized to the Polo, Figure 5 allows for a comparison of the three cars. In the default scenario it can be seen that the increased water consumption of the Golf and Passat are to a similar extent reflected by the ecological scarcity method and the model of Motoshita et al. (2011)*, showing that these methods lead to similar conclusions as the inventory in this scenario. Yet, in the categories developed by Pfister et al.,9 the impacts of the Polo and Golf are regarded as rather similar despite different water consumption. This can be explained by two facts. First, similar water consumption is weighted higher at the Polo’s production site in Spain than at the Golf’s production site in Germany. This compensates the advantages of the lower water consumptions in the material production resulting from the reduced weight of the Polo in comparison to the Golf. Second, some impact categories, especially the one developed by Pfister and colleagues9 measuring damages to human health, are dominated by the water consumption of the PGM production in South Africa. As the PGM contents of the Polo and Golf are comparable, results of these impact categories are similar, too. Since the Passat contains more PGM than the Polo and Golf, the same reasoning can explain the higher impacts in the human health categories. In contrast, the water consumption in South Africa hardly affects damage to resources since WTA is below 1 in most watersheds, which, according to Pfister et al.,9 means that no depletion of water resources occurs. For that reason the Passat scores only slightly worse in this impact category due to the larger water consumption of the larger material production. Sensitivity Analysis. Beause the regionalization of water inventories contains several assumptions and impact assessment results strongly depend on local aspects, a sensitivity analysis was accomplished. The water consumption of the material groups was assigned to the countries with the lowest and the highest physical water scarcity available in the material-specific import or production mixes. Figure 6 displays the differences between the max/min-s scenarios and the default scenario in relation to the default scenario. Hence, a difference of +100% means that impacts have doubled compared to the default
characterization factor which is based on WTA, but additionally considers seasonal variation of water availability.9 Despite different proportions, both methods are dominated by the water consumption in similar countries mainly Germany (due to high volumes), as well as Spain, Belgium, and South Africa (due to high scarcity). Figure S3 in the Supporting Information shows the water consumption in areas of different water stress throughout the lifecycles of the three cars. Whereas the method of Pfister and colleagues9 assessing health damages from malnutrition considers physical water scarcity and socio-economic aspects, the method of Motoshita et al. (2011)* measuring health damages from infectious diseases considers only socio-economic aspects. As physical water scarcity is high and the level of development is rather low, human health impacts measured according to Pfister et al.9 are dominated by the water consumed in South Africa resulting from the PGM production. Damages determined in the method of Motoshita et al. (2011)* are mainly caused by relatively low amounts of water (78−191 L) consumed in the aluminum production in Mozambique. In contrast, due to high sanitation standards and a high degree of development, the water consumption in countries like Spain or Australia does not cause damages to human health, despite high physical water scarcity in these countries. Ecosystem damage denotes the loss of biodiversity and is influenced by water scarcity and the local sensitivity of vascular plants.9 Again, the water consumption in South Africa dominates the impact assessment result with 56% (Golf) to 67% (Passat). Damages caused by the depletion of resources only occur in countries where water withdrawal exceeds the renewability rate (WTA > 1). As this is not the case in Central Europe, where most of the water is consumed, large shares of water consumption do not contribute to resource damage. This impact category is dominated by the water consumption in Spain and Ukraine which contribute 55% (Passat) to 67% (Polo) to the overall result depending on the car. In addition to this comparison of the results obtained by different impact assessment methods, a critical evaluation of the underlying characterization models can be found in Berger and Finkbeiner.40 4095
dx.doi.org/10.1021/es2040043 | Environ. Sci. Technol. 2012, 46, 4091−4099
Environmental Science & Technology
Article
models with the exception of Motoshita et al. (2011)*, where the min-s scenario leads to higher impacts than the max-s scenario. This can be explained by the fact that this method is only sensitive to socio-economic parameters and does not account for physical water scarcity which was used as an indicator to define min-s and max-s. Comparison to Other Environmental Interferences. An advantage of the ecological scarcity method is that it enables a comparison of water consumption related impacts to impacts resulting from other environmental interventions like the consumption of fossil and mineral resources or emissions to air, water, and soil. In a similar way, the method of Pfister et al.9 allows for a comparison to other environmental damages determined by means of the eco-indicator 99 method.41 Also impacts on human health determined according to Motoshita et al. (2011)* can be compared to health damages calculated by eco-indicator 99. In that way the contribution of water consumption to total impacts caused along automotive life cycles and, thus, the relevance of water consumption to the automotive industry can be estimated. As shown in Figure 7 for the production of the Golf, water consumption affects the total results to an extent of 0−7% depending on the impact category. While similar results were identified for the Polo slightly higher percentages ranging from 0 to 13% were obtained for the Passat (Figure S4 in the Supporting Information). For all cars it was shown that water consumption mainly affects damages to ecosystems rather than damages to human health and resources. This can be explained by the fact that most of the water consumption occurs in Europe, where the level of development avoids water induced health damages. Moreover, water use in Europe usually does not exceed the renewability rate, which prevents the depletion of freshwater resources. The rather low contribution of water consumption to the overall damage can be explained by the application of normalization and subsequent weighting of the three damage categories (human health 40%, ecosystems 40%, resources 20%) in the eco-indicator 99 methodology. It should be noted that these figures only reflect the production phase of cars. When considering the whole life cycle, the share of water related damages decreases even more as other environmental interferences increase significantly while water consumption and related impacts stay rather constant. Furthermore, a comparison of damages to human health and ecosystem quality resulting from water consumption and freshwater pollution in the production of the Golf was accomplished. Impacts caused by water pollution result from the emission of carcinogenic substances in the category human health and from the release of acidifying, eutrophying, and ecotoxic substances in the category ecosystem quality.41 As shown in Figure S5 in the Supporting Information, damages from water consumption are smaller than damages resulting from emissions into freshwater in both damage categories. Recommendations. Even though water consumption tends to be of minor relevance for European passenger cars run with petrol-based fuels, it can cause significant impacts in agricultural products such as food,6,7 natural fibers,8,9 or biofuels.10,11 However, conclusions drawn from water footprint studies that rely on current LCI databases have to be handled with care. First, data sets either only contain water use figures (ecoinvent) or tend to underestimate water consumption due to the partly ignorance of water consumed in background processes (GaBi). Second, as water flows are not geographically differentiated, uncertainties resulting from the top-down
Figure 4. Global water consumption throughout the life cycles of (a) the Polo 1.2 TDI, (b) the Golf 1.6 TDI, and (c) the Passat 2.0 TDI.
scenario. A result of 0% indicates no changes and a result of −100% means that impacts are zero in this scenario. More than 70% of water consumption derives from the production of steel and iron materials as well as polymers, which occurs mainly in Central and Northern Europe (except the iron ore and crude oil production). Hence, local shifting of large water consumption between the scenarios takes place within Europe. Only the shift of water consumption deriving from the PGM production between South Africa and Russia in the max-s and min-s scenarios causes significant changes in the regionalized inventories outside Europe. As water scarcity, vulnerability of ecosystems, and socio-economic parameters are rather similar throughout Central and Northern Europe, variations in the impact assessment results were mainly caused by the shifting of water consumption between South Africa and Russia. As it can be seen in Figure 6, variations from the default scenario range from −100 to +150% in all impact assessment 4096
dx.doi.org/10.1021/es2040043 | Environ. Sci. Technol. 2012, 46, 4091−4099
Environmental Science & Technology
Article
Figure 5. Relative comparison of results on the inventory and impact assessment levels for the Polo, Golf, and Passat normalized to the results obtained for the Polo in the default scenario (bars), the min-s scenario (circles), and the max-s scenario (diamonds).
regionalization of water inventories are added. Such uncertainties could be avoided if spatially differentiated water flows were available in the LCA databases, as it is already common practice for fossil energy carriers to consider different calorific values.19 The general procedure to establish regionalized water inventories presented in this work can be used as an approximation, enabling the application of impact assessment models in complex product systems. We refrained from publishing fixed allocation keys, as the assignment of water consumption to specific countries took Volkswagen’s specific supply situation into account. Hence, it should be decided for each case study if for a certain material the global production mix, a national import mix, or the location of a specific supplier should be used. In terms of method development we currently observe a trend toward sophisticated end-point impact assessment methods.5 Even though these methodological developments are appreciated from a scientific point of view, such methods often require even more inventory data. In addition to the spatial differentiation of water flows, the type of watercourse used,26 quality data,27 or even temporal information28 need to be known. However, such information is hardly available and costly to collect, especially if complex background systems are involved. Some authors developed water categories42 and provide default characterization factors27 in order to reduce inventory demands. Yet, often not even these requirements can be fulfilled by today’s inventory databases. Therefore, efforts should be put into the development of both more detailed inventory data sets and robust and applicable impact assessment methods, in order to promote the important assessment of water consumption and its consequences in LCA and other disciplines.
Figure 6. Relative differences in impact assessment results between max/min-s scenarios and the default scenario in relation to results obtained in the default scenario.
■
Figure 7. Relative contribution of water consumption in the production of a Golf 1.6 TDI to the total impacts according to the ecological scarcity method and the impact assessment models of ecoindicator 99 (hierarchist approach) and Motoshita et al. (2011)* and Pfister et al. (2009).9
ASSOCIATED CONTENT
S Supporting Information *
Additional results, figures, and country specific water consumption of all life cycle stages. This material is available free of charge via the Internet at http://pubs.acs.org. 4097
dx.doi.org/10.1021/es2040043 | Environ. Sci. Technol. 2012, 46, 4091−4099
Environmental Science & Technology
■
Article
(16) European Union. Directive 98/69/EC of the European Parliament and of the Council of 13 October 1998 relating to measures to be taken against air pollution by emissions from motor vehicles and amending Council Directive 70/220/EEC. Europian Parliament: Brussels, Belgium, 1998; http://eur-lex.europa.eu/LexUriServ/LexUriServ. do?uri=CONSLEG:1998L0069:19981228:EN:PDF. (17) Volkswagen AG. Life Cycle Assessment of End-of-Life Vehicle Treatment; Group Research, Environment Affairs Product: Wolfsburg, Germany, 2005; http://www.volkswagen.com/vwcms/master_public/ virtualmaster/en2/unternehmen/environmental_commendations. html. (18) Koffler, C.; Krinke, S.; Schebek, L.; Buchgeister, J. Volkswagen slimLCI - a procedure for stream-lined inventory modelling within Life Cycle Assessment (LCA) of vehicles. Int. J. Veh. Des. 2008, 46 (2), 172−188 (special issue). (19) PE International Website. http://www.gabi-software.com. (20) Volkswagen AG. The Polo - Environmental Commendation; Group Research - Environment Affairs Product: Wolfsburg, Germany, 2010; http://www.volkswagen.com/vwcms/master_public/ virtualmaster/en2/unternehmen/environmental_commendations. html. (21) Volkswagen AG. The Golf - Environmental Commendation; Group Research - Environment Affairs Product: Wolfsburg, Germany, 2010; http://www.volkswagen.com/vwcms/master_public/ virtualmaster/en2/unternehmen/environmental_commendations. html. (22) Volkswagen AG. The Passat - Environmental Commendation; Group Research - Environment Affairs Product: Wolfsburg, Germany, 2010; http://www.volkswagen.com/vwcms/master_public/ virtualmaster/en2/unternehmen/environmental_commendations. html. (23) German Association of the Automotive Industry. Material Classification in Motor Vehicle Construction; Structure and Nomenclature (VDA 231-106); Verband der Automobilindustrie e.V. (VDA) (German Association of the Automotive Industry): Frankfurt, Germany, 1997. (24) Leroy, C. Provision of LCI data in the European aluminium industry - Methods and examples. Int. J. Life Cycle Assess. 2009, 14 (Suppl 1), 10−44. (25) Ecoinvent Centre Website; http://www.ecoinvent.org/. (26) Mila i Canals, L.; Chenoweth, J.; Chapagain, A.; Orr, S.; Anton, A.; Clift, R. Assessing freshwater use in LCA: Part I - inventory modelling and characterisation factors for the main impact pathways. Int. J. Life Cycle Assess. 2008, 14 (1), 28−42. (27) Boulay, A.-M.; Bulle, C.; Bayart, J.-B.; Deschenes, L.; Margni, M. Regional Characterization of Freshwater Use in LCA: Modelling Direct Impacts on Human Health. Environ. Sci. Technol. 2011, 45 (20), 8948−8957. (28) Hoekstra, A. Y.; Mekonnen, M. M. Global Water Scarcity: The Monthly Blue Water Footprint Compared to Blue Water Availability for the World’s Major River Basins; Value of Water Research Report Series 53; UNESCO-IHE: Delft, The Netherlands, 2011. (29) Frischknecht, R.; Steiner, R.; Jungbluth, N. The Ecological Scarcity Method - Eco-Factors 2006 - A Method for Impact Assessment in LCA; Environmental studies no. 0906; Federal Office for the Environment: Bern, Swizerland, 2009. (30) Motoshita, M.; Itsubo, N.; Inaba, A. Development of impact factors on damage to health by infectious diseases caused by domestic water scarcity. Int. J. Life Cycle Assess. 2011, 16 (1), 65−73. (31) Pennington, D. W.; Potting, J.; Finnveden, G.; Lindeijer, E.; Jolliet, O.; Rydberg, T.; Rebitzer, G. Life cycle assessment Part 2: Current impact assessment practice. Environ. Int. 2004, 30 (5), 721− 739. (32) FAO Website; http://www.fao.org/nr/water/aquastat/main/ index.stm. (33) Alcamo, J.; Doll, P.; Heinrichs, T.; Kaspar, F.; Lehner, B.; Rosch, T.; Siebert, S. Development and testing of the WaterGAP 2 global model of water use and availability. Hydrol. Sci. J. 2003, 48 (3), 317− 337.
AUTHOR INFORMATION
Corresponding Author
*E-mail:
[email protected]; phone: +49.(0)30.31425084; fax: +49.(0)30.314-21720. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS The research grant for this study was provided by the Volkswagen AG, Group Research. We thank the Center for Environmental Systems Research (CESR) at the University of Kassel for providing watershed specific WTA ratios for the basins Ebro, Weser, and Ems. Moreover, the provision of updated characterization factors by Dr. Masaharu Motoshita from the National Institute of Advanced Industrial Science and Technology (AIST) and fruitful discussions with Dr. Stephan Pfister from the Institute of Environmental Engineering at the Swiss Federal Institute of Technology Zurich are highly appreciated.
■
REFERENCES
(1) Alter, A. Yet Another 'Footprint; to Worry About: Water. The Wall Street Journal 17 February 2009. (2) Water Footprint Network Website; http://www.waterfootprint. org. (3) Allan, J. A. Virtual water: A strategic resource, global solutions to regional deficits. Ground Water 1998, 36 (4), 545−546. (4) Falkenmark, M. The massive water scarcity now threatening Africa: Why isn’t it being addressed? Ambio 1989, 18 (2), 112−118. (5) Berger, M.; Finkbeiner, M. Water footprinting - How to address water use in life cycle assessment? Sustainability 2010, 2 (4), 919−944. (6) Milà i Canals, L.; Chapagain, A.; Orr, S.; Chenoweth, J.; Anton, A.; Clift, R. Assessing freshwater use impacts in LCA, part 2: Case study of broccoli production in the UK and Spain. Int. J. Life Cycle Assess. 2010, 15 (6), 598−607. (7) Chapagain, A. K.; Hoekstra, A. Y. The blue, green and grey water footprint of rice from production and consumption perspectives. Ecol. Econ. 2011, 70 (4), 749−758. (8) Chapagain, A. K.; Hoekstra, A. Y.; Savenije, H. H. G.; Gautam, R. 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. 2006, 60 (1), 186−203. (9) Pfister, S.; Koehler, A.; Hellweg, S. Assessing the environmental impacts of freshwater consumption in LCA. Environ. Sci. Technol. 2009, 43 (11), 4098−4104. (10) Gerbens-Leenes, W.; Hoekstra, A. Y.; Van der Meer, T. H. The water footprint of bioenergy. Proc. Natl. Acad. Sci. U.S.A. 2009, 106 (25), 10219−10223. (11) Gerbens-Leenes, W.; Hoekstra, A. Y. The water footprint of biofuel-based transport. Energy Environ. Sci. 2011, 4 (8), 2658−2668. (12) ISO 14040. Environmental management - Life cycle assessment Principles and framework (ISO 14040:2006); International Organisation for Standardisation: Geneva, Switzerland, 2006. (13) ISO 14044. Environmental management - Life cycle assessment Requirements and guidelines (ISO 14044:2006); International Organisation for Standardisation: Geneva, Switzerland, 2006. (14) Volkswagen, A. G. The Life Cycle of a Car - Environmental Commendations Document Progress; Group Research, Environment Affairs Product: Wolfsburg, Germany, 2010; http://www.volkswagen. com/vwcms/master_public/virtualmaster/en2/unternehmen/ environmental_commendations.html. (15) Bayart, J. B.; Bulle, C.; Koehler, A.; Margni, M.; Pfister, S.; Vince, F.; Deschenes, L. A framework for assessing off-stream freshwater use in LCA. Int. J. Life Cycle Assess. 2010, 15 (5), 439−453. 4098
dx.doi.org/10.1021/es2040043 | Environ. Sci. Technol. 2012, 46, 4091−4099
Environmental Science & Technology
Article
(34) Google Inc. Website; http://earth.google.com/. (35) Swiss Federal Institute of Technology Zurich Website; http:// www.ifu.ethz.ch/staff/stpfiste. (36) PE International, Water footprint workshop - 27 October 2010. Stuttgart, Germany, 2010. (37) Pfister, S.; Saner, D.; Koehler, A. The environmental relevance of freshwater consumption in global power production. Int. J. Life Cycle Assess. 2011, 16 (6), 580−591. (38) Die Zeit Website; http://www.zeit.de/online/2009/25/ infografik-wasser. (39) Guinee, J. B., Ed. Life Cycle Assessment - An Operational Guide to the ISO Standards; Kulwer Academic Publishers: Dordrecht, The Netherlands, 2002. (40) Berger, M.; Finkbeiner, M. Methodological challenges in volumetric and impact oriented water footprints. J. Ind. Ecol. 2012, DOI: , DOI: 10.1111/j.1530-9290.2012.00495.x. (41) Goedkoop, M.; Spriensma, R. The Eco-indicator 99 - A damage oriented method for Life Cycle Impact Assessment, Methodology Report; Product Ecology Consultants (PRe): Amersfoort, The Netherlands, 2001. (42) Boulay, A.-M.; Bouchard, C.; Bulle, C.; Deschenes, L.; Margni, M. Categorizing water for LCA inventory. Int. J. Life Cycle Assess. 2011, 16 (7), 639−651.
4099
dx.doi.org/10.1021/es2040043 | Environ. Sci. Technol. 2012, 46, 4091−4099