Water Accounting and Vulnerability Evaluation - American Chemical

Mar 24, 2014 - ABSTRACT: Aiming to enhance the analysis of water consumption and resulting consequences along the supply chain of products, the water ...
0 downloads 11 Views 3MB Size
Article pubs.acs.org/est

Water Accounting and Vulnerability Evaluation (WAVE): Considering Atmospheric Evaporation Recycling and the Risk of Freshwater Depletion in Water Footprinting Markus Berger,*,† Ruud van der Ent,‡ Stephanie Eisner,§ Vanessa Bach,† and Matthias Finkbeiner† †

Technische Universität Berlin, Chair of Sustainable Engineering, Office Z1, Strasse des 17. Juni 135, 10623 Berlin, Germany Delft University of Technology, Department of Water Management, Postbus 5, 2600 AA Delft, The Netherlands § University of Kassel, Center for Environmental Systems Research, Wilhelmshöher Allee 47, 34109 Kassel, Germany ‡

S Supporting Information *

ABSTRACT: Aiming to enhance the analysis of water consumption and resulting consequences along the supply chain of products, the water accounting and vulnerability evaluation (WAVE) model is introduced. On the accounting level, atmospheric evaporation recycling within drainage basins is considered for the first time, which can reduce water consumption volumes by up to 32%. Rather than predicting impacts, WAVE analyzes the vulnerability of basins to freshwater depletion. Based on local blue water scarcity, the water depletion index (WDI) denotes the risk that water consumption can lead to depletion of freshwater resources. Water scarcity is determined by relating annual water consumption to availability in more than 11 000 basins. Additionally, WDI accounts for the presence of lakes and aquifers which have been neglected in water scarcity assessments so far. By setting WDI to the highest value in (semi)arid basins, absolute freshwater shortage is taken into account in addition to relative scarcity. This avoids mathematical artifacts of previous indicators which turn zero in deserts if consumption is zero. As illustrated in a case study of biofuels, WAVE can help to interpret volumetric water footprint figures and, thus, promotes a sustainable use of global freshwater resources.



After a comprehensive review13 and test of volumetric and impact oriented WF methods in an industrial case study,14 methodological challenges have been identified in both approaches.15 According to the current definition, water consumption denotes the fraction of total water use (withdrawal) not returning to the originating drainage basin due to product integration, discharge into seawater and other basins, and evapo(transpi)ration.16 However, this definition neglects the fact that significant shares of evaporated water can return via precipitation within short time and length scales17 and, thus, should be regarded as water use but not consumption. To analyze the severity of a volume of water consumed, many impact assessment methods3,18,19 use scarcity-based impact factors which rely on the withdrawal-to-availability (WTA) ratio.20 However, WTA contains several shortcomings. First, withdrawals can include large shares of cooling water that are returned to the basin immediately and, therefore, lead to an overestimation of water scarcity in industrial regions. Second, ground and surface water stocks are neglected as “availability” includes runoff only. Yet, aquifers and lakes can buffer temporal water scarcity and, thus, are important when impacts of water consumption are to be assessed. Third, the WTA ratio turns

INTRODUCTION

During the past century water use was growing twice as fast as the world’s population, which has resulted in 1.2 billion people living in water scarce regions today.1 The analysis of water use and associated impacts throughout the supply chains of products, that is, the water footprint (WF),2 can serve as a relevant tool to mitigate water stress. Volumetric approaches, such as “virtual water”, consider the consumption of ground and surface water (blue water), the evapo(transpi)ration of rainwater (green water), and the pollution of freshwater (gray water).3 By revealing surprisingly high volumes, like 140 L per cup of coffee4 or 2700 L per cotton T-shirt,5 consumers have been made aware of the amounts of water consumed or polluted during the production of daily goods. Even the WF of nations and global virtual water imports and exports have been analyzed based on WF estimates of products, consumption patterns, and trade statistics.6,7 Despite the relevance of global freshwater appropriation figures, there is an ongoing debate whether the WF should be a volumetric indicator or an impact based measure like the carbon footprint.8 While some authors highlight the necessity of impact factors, as 1 m3 of water consumption in Canada does not compare to 1 m3 of water consumed in Mexico,9,10 other scientists argue that a product’s volumetric footprint is more important since freshwater is a global resource virtually traded via products.11,12 © 2014 American Chemical Society

Received: Revised: Accepted: Published: 4521

September 13, 2013 February 21, 2014 March 24, 2014 March 24, 2014 dx.doi.org/10.1021/es404994t | Environ. Sci. Technol. 2014, 48, 4521−4528

Environmental Science & Technology

Article

Figure 1. Water inventory flows along the life cycle of a product considered in WAVE.

zero when the numerator (withdrawal) is zero15 as it assesses relative scarcity only, but neglects absolute shortage of water. Consequently, very dry regions, like the Sahel zone, can be regarded uncritical if withdrawal is close to zero due to the absence of population or industry. Tackling the above-mentioned challenges, the water accounting and vulnerability evaluation (WAVE) model is introduced. It should be noted that there is no connection to a different model describing the connection between “substances, water and agrochemicals in the soil, crop and vadose environment”21 which is also termed WAVE. On the inventory level, a new water accounting approach considers atmospheric evaporation recycling effects and, therefore, allows for a determination of more realistic water consumption figures. In order to translate volumes into potential impacts, the vulnerability of a basin to freshwater resource depletion is evaluated by means of a new blue water scarcity indicator. Hence, WAVE will help to interpret volumetric virtual water studies and can be used as an inventory and characterization model in life cycle assessment (LCA)22 and water footprinting.23

are determined by multiplying the volumes of evapo(transpi)ration (Ei) and synthetically created vapor (Vi) with the basin internal evaporation recycling ratio (BIER) and the runoff fraction (α). ER i = Ei ·BIER n·α VR i = Vi ·BIER n·α

According to van der Ent and Savenije, BIER is estimated based on the length of the basin in the direction of the main moisture flux (x) and the average local length scale of the evaporation recycling process (λ). BIER =

∑ (FWi − WWi − ER i − VR i)

−x

(4)

The length scale of the evaporation recycling process (λ) has been calculated based on an atmospheric water accounting model considering evaporation, precipitation, winds, and humidity.17,24 For simplification each basin is assumed to be a square with x representing the side length determined via the square root of the basin’s surface area. BIER has been determined for more than 11 000 basins on a global level (Figure 2) and varies from 0% in the Sahel zone to 38% in the Congo basin. Thus, significant shares of the water consumed in a product system due to evapo(transpi)ration can be returned to the originating drainage basin via precipitation. In the same way, water vapor created in chemical reactions can be returned to the basin of origin to a noticeable extent. Since the share of evaporation recycling increases with distance, large drainage basins show higher BIER values than small basins when λ is constant. It should be noted that only a fraction of the evaporation recycling, which is returned to the originating basin via precipitation, will be available as ground or surface water. Since WAVE focuses on blue water only, the runoff fraction (α) is implemented. Based on data derived from the hydrological model WaterGAP2,25,26 α is determined by relating the longterm average runoff (R), that is, groundwater recharge and surface runoff, to the total precipitation (P) within a drainage basin (eq 5).

(1)

WCeff,n is determined by subtracting total wastewater discharges (WWi), evapo(transpi)ration recycling (ERi) and synthetically created vapor recycling (VRi) from freshwater withdrawals (FWi) occurring within basin n. WCeff, n =

( −λx ) + λ

−x − λ ·exp

WATER ACCOUNTING MODEL A new inventory method for the accounting of water use is introduced. In addition to freshwater withdrawals and wastewater discharges considered in existing inventory schemes,13 the new water accounting model explicitly considers the share of withdrawal which is consumed due to evapo(transpi)ration. Moreover, vapor created synthetically in chemical reactions, e.g. by burning fossil fuels, is regarded in an explicit way (Figure 1). In order to consider the effects of atmospheric evaporation recycling,17 the effective water consumption (WCeff), which represents the sum of effective water consumptions in each basin n (WCeff,n), is introduced.

∑ WCeff,n

(3b) 17



WCeff =

(3a)

(2)

As shown in eqs 3a and 3b, the volumes of evapo(transpi)ration and synthetically created vapor recycled within a basin

α= 4522

R P

(5) dx.doi.org/10.1021/es404994t | Environ. Sci. Technol. 2014, 48, 4521−4528

Environmental Science & Technology

Article

Figure 2. Basin internal evaporation recycling (BIER) ratios denoting the fractions of evaporated water returning to the originating basins via precipitation.

As shown in Figure S1 in the Supporting Information (SI), α is highest (>80%) in basins located in Alaska and the Himalayas and in the Amazon basin. The resulting hydrologically effective basin internal evaporation recycling (BIERhydrol‑eff), which is obtained by multiplying BIER with α, is shown in SI Figure S2. Since α is comparably low in Central Africa, large BIER ratios determined in, for example, the Congo basin (38%) are reduced when considering the hydrologically effective fraction (BIERhydrol‑eff = 11%). Even though BIERhydrol‑eff is below 5% in most of the world’s drainage basins, it reduces blue water consumption significantly in basins in the Himalayas, Alaska, southeast Asia, and the North of South America (10−33%, SI Figure S2).

as characterization factors for impact assessment in water footprinting and LCA. RFD =

∑ (WCeff, n ·WDIn)

(6)

Tackling the shortcomings related to WTA, WDI is based on the consumption-to-availability (CTA) ratio,28 which relates annual water consumption (C) to annual availability (A). The freshwater availability of a drainage basin (A) expresses the annually renewable freshwater volumes within the basin which can be quantified by means of runoff (plus upstream inflows if the basin is divided into subcatchments). Data for C and A is available in WaterGAP2 for more than 11 000 basins on a global level. As shown in eq 7, CTA is modified in two steps.



VULNERABILITY EVALUATION MODEL In order to assess consequences resulting from water consumption, many impact assessment models developed in LCA try to describe impacts on the areas of protection resources, ecosystems, and human health.13 Some authors model concrete cause-effect chains, like water consumption leads to less water available for irrigation, leading to less productive agriculture, leading to health impacts due to malnutrition.19,27 However, such end point models rely on various assumptions. Moreover, regressions used to describe impact pathways are often of low statistical significance. This shows that the relation between water consumption and impactsespecially on human health and ecosystemsis not straightforward and depends on multiple variables. Therefore, this work focuses on freshwater resources only and evaluates the regional vulnerability of drainage basins to blue water depletion. This vulnerability approach distinguishes the WAVE model from conventional characterization models describing consequences on resources, such as ref 19. It is not intended to “predict” impacts on freshwater resources but to denote the risk that water consumption in a certain region will lead to freshwater depletion. This risk of freshwater depletion (RFD) can be determined by multiplying the effective water consumption in each basin with its corresponding water depletion index (WDI). WDI denotes the vulnerability of drainage basins to freshwater depletion based on physical blue water scarcity. It can be used to interpret volumetric water footprints on a qualitative level or

CTA* =

C ·AFGWS A + SWS

(7)

First, annually usable surface water stocks (SWS) are added to A in order to consider lakes, wetlands, and dams in the scarcity index. While storage volumes of dams (Vdam) are available directly,29 volumes of lakes and wetlands are determined by multiplying their surface areas (Alake/wetland) per basin30 with an effective depth (deff = 5 m for lakes and 2 m for wetlands). In order to combine the volumes of dams, lakes, and wetlands (km3) with the flows C and A (km3/a), an annually usable fraction of 1% of the total volumes is used in the determination of SWS (eq 8). This means that ground and surface water stocks can be used for at least 100 years, even if no renewability occurs. SWS =

∑i (Vdam, i + (Alake/wetland, i ·deff, i)) 100 years

(8)

The effective depths of lakes and wetlands are derived from WaterGAP225,26 which represents the only data source for this kind of information on a global level. Just as the time horizon of 100 years, they can be regarded as conservative estimates which are exceeded in many basins. Such methodological choices are in line with the vulnerability approach applied in this work, which aims at assessing the risk that freshwater depletion can occur rather than predicting impacts. In order to evaluate the 4523

dx.doi.org/10.1021/es404994t | Environ. Sci. Technol. 2014, 48, 4521−4528

Environmental Science & Technology

Article

influence of deff and the time horizon on the final result, a comprehensive sensitivity analysis is accomplished in the SI. In contrast to SWS, volumes of groundwater stocks (GWS) are not available on a global level. Therefore, an adjustment factor (AFGWS) is introduced, which reduces the scarcity ratio based on the availability of groundwater. Using data provided by WHYMAP,31 AFGWS is defined based on geological structure and annual recharge as shown in Table 1. The scarcity

Since CTA* expresses a ratio of consumption to availability, the resulting WDI takes into account relative freshwater scarcity only. In order to consider absolute freshwater shortage as well, WDI is set to 1 per se in semiarid and arid basins34 shown in SI Figure S4. This setting is relevant as freshwater resources are highly vulnerable to depletion in (semi)arid regions regardless of the relative scarcity. As shown in Figure 3, WDI is at the highest level in many drainage basins located in Central Asia, India, Saudi Arabia, Australia, Northern and Southern Africa, Mexico, the southwest of the U.S., and the Andes. In contrast, little or no freshwater resource depletion is caused by water consumption in basins located in Russia, Canada, Northern Europe, or around the equator.

Table 1. Adjustment Factors for Ground Water Stocks (AFGWS) Reducing Water Scarcity Based on Geological Structure and Annual Recharge geological structure major groundwater basin complex hydrogeological structure

annual recharge (mm)

scarcity reduction

AFGWS

>300 100−300 >300 100−300

10.0% 7.5% 5.0% 2.5% 0.0%

0.900 0.925 0.950 0.975 1.000

others



CASE STUDY The methodology developed above is tested by means of an existing WF study of bioethanol produced from sugar cane in five producing countries.35 Table S1 in the SI shows the blue water consumption (evapotranspiration of blue irrigation water) required to produce 1 GJ of bioethanol from sugar cane in Columbia, Mexico, Thailand, Australia, and Zambia based on ref 32. By means of country specific factors for BIERhydrol‑eff and WDI, the effective water consumption (WCeff) and the risk of freshwater depletion (RFD) are determined (SI Table S1). In order to compare the results obtained by means of the WAVE model to those obtained by other impact assessment methods, potential impacts resulting from WCeff are additionally evaluated by means of the models developed by Pfister et al. (2009)19 and Frischknecht et al. (2009).18 Figure 4 shows the water consumption figures on a relative scale and presents the reductions in WCeff compared to WC resulting from the consideration of BIERhydrol‑eff. Moreover, potential impacts determined by means of WAVE as well as by the models of Pfister et al. (2009) and Frischknecht et al. (2009) are shown normalized to the highest result in each category (for absolute results see SI Table S1). It should be noted that a consideration on the country level has its limitations as plants may be grown in particular basin whose hydrogeological situation may differ from the country average. For example sugar cane from Australia is mainly produced in the coastal areas36 showing less severe water stress than the country average. However, as ref 35 provides data on

reduction rates have been derived from discussions with the developers of the WHYMAP.32 In line with the vulnerability approach, moderate reduction rates are selected. Their influence on the final result has been analyzed in a sensitivity analysis presented in the SI. Fossil groundwater stocks are excluded from this analysis as they cannot be quantified on a global level and it is not sure that they can be accessed in every part of the world. The factor WDI aims at assessing a basin’s vulnerability to freshwater depletion based on CTA* as shown in eq 9. It can be understood as an equivalent volume of depleted water resulting from a volume of water consumption. Similar to existing scarcity indexes,19,27 the logistic function plotted in SI Figure S3 leads to nonlinear transformation of physical water scarcity into vulnerability to freshwater depletion. This is important as in the upper and lower ranges of CTA* doubled scarcity does not necessarily lead to doubled vulnerability to depletion. WDI turns 1 above a CTA* of 0.25, which is regarded as the threshold of extreme water stress.33 1 WDI = −40·CTA * 1 1+e −1 (9) 0.01

(

)

Figure 3. Factors WDI expressing the vulnerability of basins to freshwater resource depletion [m3depleted/m3consumed]. 4524

dx.doi.org/10.1021/es404994t | Environ. Sci. Technol. 2014, 48, 4521−4528

Environmental Science & Technology

Article

Figure 4. Relative presentation of blue water consumption required to produce 1 GJ of bioethanol from sugar cane, reduction of water consumption due to consideration of BIERhydrol‑eff, and potential impacts determined by means of WAVE and by the impact assessment methods of Pfister et al. (2009) and Frischknecht et al. (2009).

The three assessment methods also lead to different conclusions regarding bioethanol produced from sugar cane in Australia. While the required irrigation water consumption of 23 m3/GJ causes the highest risk of freshwater depletion in WAVE, it is considered less relevant than production in Mexico and Thailand in the other impact assessment models. In the method of Frischknecht et al. (2009), suggested by the European Union for the product environmental footprint,38 water consumption in Australia (23 eco-points/m3) is even regarded far less relevant than in countries abounding in water like Germany (910 eco-points/m3). These unexpected results can be explained by the fact that the methods of Frischknecht et al. (2009) and Pfister et al. (2009) consider relative freshwater scarcity only. Yet, even though only a comparably small fraction of Australia’s water availability is used, the country suffers from absolute freshwater shortage. This highlights the relevance of considering absolute scarcity by means of aridity in the WAVE model.

the country and state level and since we provide BIER and WDI on the level of countries and river basins, the country level is the lowest common denominator to be used in this study. Moreover, the testing of the WAVE model and the comparison to other impact assessment methods is regarded more important than the absolute result in this case. On the inventory level, the consideration of the hydrologically effective evaporation recycling by means of BIERhydrol‑eff leads to a reduction of water consumption between 0% in Australia and 10% in Columbia. If the total basin internal evaporation recycling (BIER) is taken into account, the recycled fractions of evapotranspirated irrigation water will increase significantly (up to 24% in Zambia). The assessment of potential impacts resulting from irrigation water consumption in the countries considered leads to different conclusions than a volumetric analysis. All three assessment methods come to the result that the largest water consumption in Zambia (39m3/GJ) actually causes the lowest impacts, as no water scarcity is detected in this country on an annual basis. However, this also shows the limitations of an annual assessment method in a seasonal product system. As discussed in detail in the following chapter, water consumption and water scarcity can vary throughout the yearespecially in countries with dry and wet seasons.37 The relatively high water consumption of bioethanol produced in Thailand (18 m3/GJ) is evaluated differently in the assessment models. While the risk of freshwater depletion (RFD) is considered low in the WAVE model, significant impacts are expected in the methods of Pfister et al. (2009) and Frischknecht et al. (2009). The reason for this can be found in the underlying methodologies. While a consumption-toavailability ratio is considered in WAVE, the two other impact assessment models are based on a withdrawal-to-availability ratio according to which Thailand is much more water scarce.



DISCUSSION Scope. The WAVE model developed in this work focuses on blue water consumption which occurs mainly due to evapo(transpi)ration or product integration of ground and surface water. While the vulnerability evaluation model is restricted to assess blue water consumption only, BIER and BIERhyrdol‑eff can also be used to assess the basin internal recycling of plant evapotranspiration (green water consumption). As the WAVE model does not consider water quality degradation in its default version, two possibilities of including water quality aspects are presented. When the method described here is applied in an LCA study, freshwater pollution is assessed by means of impact categories like eutrophication, acidification, or human and eco-toxicity. However, similar to 4525

dx.doi.org/10.1021/es404994t | Environ. Sci. Technol. 2014, 48, 4521−4528

Environmental Science & Technology

Article

simplification that basins are of quadratic shape leads to an under or overestimation of BIER and BIER100 depending on the actual shape and the prevailing wind directions (SI Figure S6). The consideration of basin internal evaporation recycling leads to an interesting effect when considering agricultural product systems: Even the evapotranspiration of green water by agricultural plants can cause blue water benefits. The reason is that parts of this green water evapotranspiration will be recycled within the drainage basin (BIER), of which parts will be hydrologically effective (α). Moreover, the accounting model developed explicitly considers the emission of water vapor created in chemical reactions and its partial return due to atmospheric moisture recycling effects (Figure 1). Consequently, the combustion of fossil fuels may lead to negative effective water consumption if the synthetically created vapor recycling is higher than the difference between freshwater withdrawals, wastewater discharges, and evaporation recycling (eq 2). So far, WAVE has considered the evapo(transpi)ration recycling within drainage basins only leading to a global average BIER of 1%. However, the average continental evapotranspiration returning as continental precipitation amounts to 57%.24 Thus, the examination of basin internal evaporation recycling (BIER) effects should be extended to a fate of evaporation analysis15 which considers the fractions of evapo(transpi)ration returning to other basins as well. Vulnerability Evaluation Model. In this work the vulnerability of a drainage basin to freshwater depletion is evaluated. Based on physical water scarcity, the water depletion index (WDI) denotes the risk that water consumption leads to freshwater depletion. In contrast to most other water scarcity indicators used as impact factors in water footprinting, WDI is based on a consumption instead of withdrawal-to-availability ratio. Even though withdrawal implicitly accounts for quality degradation as well, a consumption based indicator expresses water shortage more realistically as large shares of cooling water, which are returned with low quality degradation due to temperature increase, are excluded. Moreover, for the first time ground and surface water stocks are included in a water scarcity indicator. As shown in SI Figure S7, the consideration of aquifers, lakes and wetlands leads to a scarcity reduction of up to 10% in many basins around the globe; especially in Canada, Central Africa, Central Europe, South America, and Russia. Even higher reductions of more than 80% are achieved in small basins in Alaska and the Himalayas. Thus, the consideration of ground and surface water stocks leads to a further scarcity reduction in regions which are under low water stress anyway. Even though the logistic function diminishes this reduction effect in the final WDI result (SI Figure S3), the difference between water scarce regions, like Saudi Arabia, and regions abounding in water, such as Canada, is increased. This higher precision in water scarcity assessment is especially relevant when comparing water consumption in different regions. By setting WDI to the highest value (1.00) in arid and semiarid basins, WAVE considers absolute freshwater shortage in addition to relative scarcity. This helps to avoid the mathematical artifact that dry regions are regarded uncritical if consumption is close to zero. SI Figure S8 presents the influence of this setting on the final WDI result. Without this consideration of absolute freshwater shortage, most of the arid

the gray water footprint, this approach does not consider the water quality of freshwater inputs. Therefore, a possibility of determining the quality corrected effective water consumption (WCq,eff,n) which can be used as the basis for calculating the quality corrected risk of freshwater depletion (QRFD), is presented in the SI. In the WAVE model, evaporation recycling and freshwater scarcity are determined based on annual averages. As mentioned in the Case Study, this is a limitation as climatic conditions influencing evaporation recycling as well as the hydrological situation in a basin might vary throughout a year. Especially the combination of these effects can be relevant in semiarid drainage basins, as BIER may be high in the rainy season when water is abundant but low in the dry season when water scarcity is of concern. For this reason, some authors introduce monthly scarcity factors39 which are especially relevant for agricultural products grown during particular seasons. However, such approaches require temporally explicit inventory data, which is difficult to obtain − especially if complex background systems are involved. As an alternative, consumption weighted annual averages of monthly scarcity factors can be used.40 Yet, this does not overcome the key methodological challenge of a monthly scarcity assessment: the consideration of intermonthly storage capacities which can buffer water scarce periods throughout the year.40 Moreover, the temporal resolution of water scarcity assessments also determines the required spatial resolution. Large basins can have flow times of several months from spring to mouth, which makes a monthly assessment difficult. Therefore, this work refrains from providing monthly or weighted annual BIER and WDI factors but acknowledges them as an important subject to future research. Water Accounting Model. The accounting approach presented in this work considers the basin internal recycling of the share of water withdrawal consumed due to evapo(transpi)ration. Even though this complies with the definition of water consumption,16 it may appear arguable whether the reduction of actual water consumption is reasonable in large basins, like the Danube, where evaporation recycling can occur after hundreds of kilometers. In our opinion such an approach is justified for three reasons. First, a river basin delineation of WaterGAP2 is used which divides the world’s 34 largest drainage basins into subcatchments. This avoids extremely long evaporation recycling distances that would otherwise occur in, for example, the Congo basin. Second, in several countries, like the U.S. or Australia, water is transported from withdrawal to use through pipelines over long distances. The fraction of withdrawal returned to the river also reduces the water consumption volume in such cases. Hence, atmospheric water transport should be treated in the same way as anthropogenic transport. Third, in many basins the downwind transport of evaporated water is in fact an upstream transport from a river perspective as, for example, in the Amazon basin.24 This means that recycled evaporation is actually returning water, which is available for a possible second consumption. Nevertheless, additional BIER100 ratios are determined as a sensitivity check by restricting the evaporation recycling distances (x) to 100 km (eq 4). As shown in SI Figure S5, BIER100 is significantly lower than BIER (Figure 1) with a maximum of 19% determined in a Colombian drainage basin. Apart from Australia, Northern Africa, Saudi Arabia, and large regions in Central Asia BIER100 ranges from 1 to 5% in most of the world’s basins. Yet, it should be noted that the model 4526

dx.doi.org/10.1021/es404994t | Environ. Sci. Technol. 2014, 48, 4521−4528

Environmental Science & Technology

Article

(3) Hoekstra, A. Y.; Chapagain, A. K.; Aldaya, M. M.; Mekonnen, M. M. The Water Footprint Assessment ManualSetting the Global Standard; Earthscan: London, Washington, DC, 2011. (4) Chapagain, A. K.; Hoekstra, A. Y. The water footprint of coffee and tea consumption in the Netherlands. Ecol. Econ. 2007, 1 (64), 109−118. (5) 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. (6) Hoekstra, A. Y.; Mekonnen, M. M. The water footprint of humanity. Proc. Natl. Acad. Sci. U.S.A. 2012, 109 (9), 3232−3237. (7) Suweis, S.; Rinaldo, A.; Maritana, A.; D’Odorico, P. Watercontrolled wealth of nations. Proc. Natl. Acad. Sci. U.S.A. 2012, 110 (11), 4230−4233. (8) Finkbeiner, M. Carbon footprintingOpportunities and threats. Int. J. LCA 2009, 14, 91−94. (9) Pfister, S.; Hellweg, S. The water ‘‘shoesize’’ vs. footprint of bioenergy. Proc. Natl. Acad. Sci. U.S.A. 2009, 106 (35), E93−E94. (10) Ridoutt, B. G.; Huang, J. Environmental relevanceThe key to understanding water footprints. Proc. Natl. Acad. Sci. U.S.A. 2012, 109 (22), E1424. (11) Hoekstra, A. Y.; Gerbens-Leenes, W.; van der Meer, T. H. Reply to Pfister and Hellweg: Water footprint accounting, impact assessment, and life-cycle assessment. Proc. Natl. Acad. Sci. U.S.A. 2009, 106 (40), 114. (12) Hoekstra, A. Y.; Mekonnen, M. M. Reply to Ridoutt and Huang: From water footprint assessment to policy. Proc. Natl. Acad. Sci. U.S.A. 2012, 109 (22), E1425. (13) Berger, M.; Finkbeiner, M. Water footprintingHow to address water use in life cycle assessment? Sustainability 2010, 2 (4), 919−944. (14) Berger, M.; Warsen, J.; Krinke, S.; Bach, V.; Finkbeiner, M. Water footprint of European cars: Potential impacts of water consumption along automobile life cycles. Environ. Sci. Technol. 2012, 46 (7), 4091−4099. (15) Berger, M.; Finkbeiner, M. Methodological challenges in volumetric and impact oriented water footprints. J. Ind. Ecol. 2013, 17 (1), 79−89. (16) 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. (17) van der Ent, R. J.; Savenije, H. H. G. Length and time scales of atmospheric moisture recycling. Atmos. Chem. Phys. 2011, 11, 1853− 1863. (18) Frischknecht, R.; Steiner, R.; Jungbluth, N. The Ecological Scarcity MethodEco-Factors 2006A Method for Impact Assessment in LCA; Federal Office for the Environment: Bern, Swizerland, 2009. (19) Pfister, S.; Koehler, A.; Hellweg, S. Assessing the environmental impacts of freshwater consumption in LCA. Environ. Sci. Technol. 2009, 43 (11), 4098−4104. (20) Alcamo, J.; Flörke, M.; Märker, M. Future long-term changes in global water resources driven by socio-economic and climatic changes. Hydrol. Sci. J. 2007, 52 (2), 247−275. (21) Timmerman, A.; Feyen, J. The WAVE model and its application; Simulation of the substances water and agrochemicals in the soil, crop and vadose environment. Revista Corpoica 4 (1), 36-41. (22) ISO 14044. Environmental ManagementLife Cycle Assessment Requirements and Guidelines (ISO 14044:2006); International Organisation for Standardisation, Ed.; Geneva, Switzerland, 2006. (23) Boulay, A.-M.; Hoekstra, A. Y.; Vionnet, S. Complementarities of Water-Focused Life Cycle Assessment and Water Footprint Assessment. Environ. Sci. Technol. 2013, 47 (21), 11926−11927. (24) van der Ent, R. J.; Savenije, H. H. G.; Bettina, S.; Steele-Dunne, S. C. Origin and fate of atmospheric moisture over continents. Water Resour. Res. 2010, 46. (25) Döll, P.; Kaspar, F.; Lehner, B. A global hydrological model for deriving water availability indicators: Model tuning and validation. J. Hydrol. 2003, 270 (1−2), 105−134.

and semiarid basins (SI Figure S4) would have significantly lower WDI results than those shown in Figure 3. Especially, impacts from water consumption in the Sahel zone or in Australia would be close to zero, as it is the case in existing impact assessment methods like Frischknecht et al. (2009). So far, WAVE has only assessed the vulnerability of drainage basins to freshwater depletion from a blue water resource perspective. In future research also the vulnerability to human health and ecosystem impacts should be analyzed. By means of sensitivity factors the risk that water consumption in water scarce regions can lead to impacts could be analyzed. However, especially when assessing the vulnerability to health impacts, the consumption and availability of green water need to be considered in combination with blue water. This combined approach is needed as both types of water are equally important for food production and there are many countries which suffer from blue water scarcity but have enough green water to grow crops.41 Application. BIER, BIERhydrol-eff, and WDI are determined on the level of drainage basins, as they reflect hydrologic conditions best. Since inventory information is often not available on such a detailed geographic resolution, all factors are provided on the country level as well. In order to promote the applicability of the WAVE model, BIER, BIERhydrol-eff, and WDI are made available free of charge on both drainage basin and country levels in a Google Earth layer and spreadsheet, respectively: http://www.see.tu-berlin.de/wave/parameter/en/. For the determination of country specific factors consumption weighted averages are used. This adds higher weight to those basin fractions within a country which contribute a higher share to the country’s total consumption. Uncertainties related to the creation of country averages are discussed in the SI and quantified in the spreadsheet.



ASSOCIATED CONTENT

S Supporting Information *

Additional explanations, figures, and tables are available in the Supporting Information. This material is available free of charge via the Internet at http://pubs.acs.org.



AUTHOR INFORMATION

Corresponding Author

*Phone: +49.(0)30.314-25084; fax: +49.(0)30.314-21720; email: [email protected]. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS



REFERENCES

We express sincere thanks to Stephan Pfister (Swiss Federal Institute of Technology Zurich), Martina Flörke (Center for Environmental Systems Research, Kassel), as well as Andrea Richts and Wilhelm Struckmeier (Federal Institute for Geosciences and Natural Resources, Hannover) for providing datasets used in this work and many fruitful discussions.

(1) United Nations Water and Food and Agricultural Organization Coping with water scarcitychallenge of the twenty first century. (2) ISO FDIS 14046. Water FootprintPrinciples, Requirements and Guidance; International Organization for Standardization, Ed.; Geneva, Switzerland, 2012. 4527

dx.doi.org/10.1021/es404994t | Environ. Sci. Technol. 2014, 48, 4521−4528

Environmental Science & Technology

Article

(26) Flörke, M.; Kynast, E.; Bärlund, I.; Eisner, S.; Wimmer, F.; Alcamo, J. Domestic and industrial water uses of the past 60 years as a mirror of socio-economic development: A global simulation study. Global Environ. Change 2013, 23 (1), 144−156. (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. 2011b, 45 (20), 8948− 8957. (28) 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. 2011b, 45 (20), 8948− 8957. (29) Lehner, B.; et al. High resolution mapping of the world’s reservoirs and dams for sustainable river flow management. Front. Ecol. Environ. 2011, 9 (9), 494−502. (30) Lehner, B.; Döll, P. Development and validation of a global database of lakes, reservoirs and wetlands. J. Hydrol. 2004, 296 (1−4), 1−22. (31) Richts, A.; Struckmeier, W.; Zaepke, M. WHYMAP and the groundwater resources of the world 1:25,000,000. In Sustaining Groundwater Resources; Jones, J. A. A., Ed.; Springer: Heidelberg, 2011; pp 159−173. (32) Struckmeier, W.; Richts, A. Personal communication (February 19, 2013). Federal Institute for Geosciences and Natural Resources, 2013 (33) Richter, B. D.; Davis, M. M.; Apse, C.; Konrad, C. Short communication: A presumptive standard for environmental flow protection. River Res. Appl. 2011, DOI: 10.1002/rra.1511. (34) United Nations Environment Programme. World Atlas of Desertification, 2 ed.; Arnold: London, 1997; p 192. (35) Mekonnen, M. M.; Hoekstra, A. Y. The green, blue and grey water footprint of crops and derived crop products. Hydrol. Earth Syst. Sci. 2011, 15 (5), 1577−1600. (36) Griggs, P. D. Deforestation and sugar cane growing in Eastern Australia. Environ. Hist. 2007, 13 (3), 255−283. (37) Savenije, H. H. G. Water scarcity indicators; the deception of the numbers. Phys. Chem. Earth, Part B 2000, 25 (3), 199−204. (38) European Union Commission Recommendation of 9 April 2013 on the use of common methods to measure and communicate the life cycle environmental performance of products and organisations. Off. J. Eur. Union 2013, L124, 56. (39) Hoekstra, A. Y.; Mekonnen, M. M.; Chapagain, A. K.; Mathews, R. E.; Richter, B. D. Global monthly water scarcity: Blue water footprints versus blue water availability. PLoS One 2012, 7 (2), e32688. (40) Pfister, S.; Baumann, J., Monthly characterization factors for water consumption and application to temporally explicit cereals inventory. In 8th International Conference on LCA in the Agri-Food Sector, Rennes, France, 2−4 October 2012, 2012. (41) Rockström, J.; Falkenmark, M.; Karlberg, L.; Hoff, H.; Rost, S.; Gerten, D. Future water availability for global food production: The potential of green water for increasing resilience to global change. Water Resour. Res. 2009, 45 (7), W00A12.

4528

dx.doi.org/10.1021/es404994t | Environ. Sci. Technol. 2014, 48, 4521−4528