!
!
!
!
!
Master’s Thesis, 60 ECTS!
Social-ecological Resilience for Sustainable Development Master’s programme 2013/15, 120 ECTS
!
!
!!
!
!
Aquatic food production and resource management - Freshwater use in Chinese
aquaculture
!
!!
!!
!
Lara!D.!Mateos!
!
!
!
!
!
!
!
!!Stockholm!Resilience!Centre!
!!!!Research!for!Biosphere!Stewardship!and!Innovation!
Aquatic food production and resource management - Freshwater use in Chinese aquaculture
Lara D. Mateos
Supervisor: Dr. Max Troell
Co-Supervisors: Dr. Lisa Deutsch & Patrik J. G. Henriksson
!
ABSTRACT
!
Aquaculture will play an important role for future seafood supply, but its increasing dependency on freshwater resources may pose a challenge for its growth. This thesis explores the freshwater footprint of cultured aquatic animals, using the global aquaculture giant, China, as a case study. Main objectives were to: a) perform a preliminary estimation of the freshwater footprint (m3 tonne-1) at the national/regional scale, using the conceptual framework and methodology of the Water Footprint Network (WFN), b) identify key methodological aspects and variables specifically related to measuring the water footprint in aquaculture, c) analyse and discuss water consumption to inform future sustainable water management strategies, through a deeper understanding of Chinese aquaculture as a social- ecological system. Results show that aquaculture’s freshwater footprint is similar to terrestrial animal production systems, with an average of 14952 m3 tonne-1. Water consumption mainly takes place at the farm through evaporation from freshwater ponds, and dilution of freshwater in brackish water ponds. Indirect water footprint through feed consumption is mainly influenced by the composition of ingredients, and the assimilation efficiency of the different species. The trend is towards intensification of production, and this has the potential to lower water consumption per yield, however, increased consumption of higher quality feed in such systems may work in the opposite direction. Key sustainability aspects that require further attention within the WFN’s methodology include cross-scale interactions between the focal scale with its upper and lower scales, and the interconnectedness of water scarcity issues to other resource uses and associated impacts. An integrated framework is needed to allow the comparison and aggregation of indices across the three pillars of sustainability. Here, the transdisciplinarity of the SES approach can help create sustainability criteria that reflect water consumption impacts in a more integrated way.
ACKNOWLEDGEMENTS
First and foremost, I would like to thank my supervisor Dr. Max Troell, owner of the original idea, who offered me his support and advice, as well as understanding along the way. I am grateful for the positive attitude and all the opportunities given. My co-supervisors Dr. Lisa Deutsch for the useful feedback and guidance, and Patrik Henriksson, for all the time invested in helping me, and providing me with very useful background data. A special thank you to Lan Wang and Patrick Keys for a wake up call and energy surge, as well as the crucial evaporation data behind the calculations. I express my gratitude to Dr. Marc Metian for his valuable input into the calculations related to feed, and Dr. Wenbo Zhang for providing me with much needed data on Chinese production, without whom I would have not been able to understand it. I am grateful to Emma Sundström for her tips on QGIS as well as her calm energy and good predisposition. Of course, I cannot proceed without mentioning my incredible thesis group (Carmen Seco Pérez, Kate Williman, and Roweena Patel), for the unlimited support through thick and thin. Knowing we could count on each other has been a keystone these months. A special thank you to Carmen Seco Pérez, whose everlasting disposition to help and bounce ideas at, has been incredibly helpful. Thank you for not allowing me to forget the reason why we are studying sustainable development. And last but not least, my great friends from the “thesis cave”, for the laughter and tears down in the basement, and the long long hours of great discussions. Thanks to all of you for making this experience a wonderful and most of all, enriching one.
TABLE OF CONTENTS
1. INTRODUCTION ... 1
2. THEORETICAL FRAMEWORK ... 4
2.1 The concept of Sustainable Development ... 4
2.2 The Social-Ecological Systems approach to Sustainable Development ... 4
2.3 Water footprint accounting as a sustainability evaluation method ... 6
3. CASE STUDY – China ... 7
3.1 China – the global aquaculture giant ... 7
3.2 The sustainability of China’s aquaculture reliance on scarce freshwater resources. ... 8
4. AIM & RESEARCH QUESTIONS ... 10
5. METHODS ... 11
5.1 The directwater footprint of Chinese aquaculture. ... 11
5.1.1 Semi-closed water systems ... 12
a. Evaporation from water surfaces ... 13
b. Infiltration losses. ... 14
c. Water drainage and exchange ... 15
5.1.2 Open water systems ... 16
5.2 Indirect water footprint of Chinese aquaculture. ... 17
a. Feed composition ... 17
b. Origin of ingredients ... 17
c. Water footprint allocation method ... 18
d. Economic Feed Conversion ratio (eFCR) ... 19
6. RESULTS ... 20
6.1 Directwater footprint: The water footprint of Chinese fresh and brackish water aquaculture ponds. ... 20
6.1.1 Evaporation water footprint in freshwater ponds. ... 20
6.1.2 Infiltration losses in freshwater ponds. ... 22
6.1.3 Dilution losses in brackish water ponds. ... 22
6.2 Indirect water footprint: The water footprint of Chinese aquaculture feeds ... 23
6.2.1 Quantity and composition of plant-based aquaculture feeds. ... 23
6.2.2 The origin of the ingredients. ... 24
6.2.3 The water footprint of feed ingredients. ... 24
6.2.4 The water footprint of different species groups. ... 27
6.3 Linking the direct and indirect water footprints ... 32
6.3.1 Freshwater pond aquaculture. ... 32
6.3.2 Brackish water pond aquaculture. ... 32
6.3.3 Marine aquaculture. ... 33
7. DISCUSSION ... 34
7.1 Understanding the largest water-consuming processes within an aquaculture production system ... 34
7.1.1 Direct water footprint of freshwater ponds ... 34
7.1.2 Direct water footprint of brackish water ponds. ... 34
7.1.3 Marine aquaculture ... 35
7.2 Short-comings of the current estimates ... 37
7.3 The adequacy of the water footprint methodology to inform about sustainability issues within the aquaculture sector ... 38
7.3.1 The appropriate scale for sustainability ... 39
7.3.2 A single impact approach ... 39
7.3.3 The three water flows ... 40
7.4 The overlap between aquaculture and water scarcity exemplified in the Chinese case. ... 41
8. CONCLUSION ... 44
LITERATURE CITED ... 45
APPENDIX I - Tables and figures ... 54
Methods ... 54
Results ... 55
APPENDIX II – Glossary ... 57
APPENDIX III – Proposed future research frontiers. ... 59
LIST OF FIGURES
!
Figure 1. The nested circles model of sustainability (Image source: Rockström and Klum, 2012). ... 5
!
Figure 2. a) Total aquaculture production in 2012 by weight (%) comparing China with the five continents, b) the top 13 countries in aquaculture production of fish, crustaceans, molluscs, etc. (FishStat, 2012). ... 8
!
Figure 3. Different types and components of the water footprint assessment (Hoekstra et al., 2011). ... 12
!
Figure 4. Schematic representation of the main processes through which water can be
consumed or relocated in an aquaculture pond. ... 13
!
Figure 5. a) The evaporation component of the blue water footprint of freshwater ponds and b) the freshwater pond’s yield of Chinese aquaculture, averaged at the province level. .. 21
!
Figure 6. Average annual evaporation rates in the Chinese provinces for the years 1999 to 2011 (Wang-Erlandsson et al., 2014). ... 22
!
Figure 7. The green, blue and grey water footprint of producing the ingredients using a) mass allocation and b) economic allocation. Wheat, starch (wheat starch) and wheat flour were assume to be the same. The same applies to wheat by-products and wheat bran. ... 26
!
Figure 8. Plant ingredients found in the diets of the different species groups farmed in
Chinese aquaculture for the year 2010. Background data from Tacon et al. (2011). ... 28
!
Figure 9. Freshwater footprint of different species groups cultured in Chinese aquaculture in 2010 using a) mass allocation and b) economic allocation. Their Economic Feed
Conversion ratio (eFCR) from Tacon et al. (2011) is shown in brackets. ... 31
!
Figure 10. Blue water scarcity in the river basins that fall under Chinese territory between 1996 and 2005. <100% = Low (the blue water footprint is < 20% of natural runoff and does not exceed blue water availability); 100-150% = Moderate (the blue water footprint
= 20-30% of natural runoff); 150-200% = Significant (the blue water footprint = 30-40%
of natural runoff); >200 = Severe (The monthly blue water footprint >40% of natural runoff). Background data from Hoekstra et al. (2012). ... 42
!
Figure 11. Virtual water balance per economic region and the net direction of virtual water flows, only showing the largest net virtual water flows (<2Gm3 year-1) (Zhao et al., 2015). ... 43
LIST OF TABLES
!
Table 1. Ingredients used in Chinese aquaculture feeds in 2010. ... 23
!
Table 2. Quantity used and imports of key ingredient groups found in aquaculture feeds in China. In red are the ingredient groups with considerably high imported quantities
(http://faostat3.fao.org/download/T/TP/E). Countries from where ≥ 95% of the ingredient groups were imported from in 2011 are identified
(http://faostat3.fao.org/download/T/TM/E). ... 24
!
Table 3. Final freshwater footprint of freshwater aquaculture in China, disaggregated into the green, blue and grey water footprints, and divided into the most important water
consuming processes. Numbers in brackets refer to results obtained using economic allocation of the water footprint as oppose to mass allocation. ... 33
!
Table 4. Final freshwater footprint of brackish and marine aquaculture in China,
disaggregated into the green, blue and grey water footprints, and divided into the most important water consuming processes. Numbers in brackets refer to results obtained using economic allocation of the water footprint as oppose to mass allocation. ... 33
!
Table 5. Water footprint of producing terrestrial animal meat (Mekonnen and Hoekstra, 2010). ... 37
!
1. INTRODUCTION
Inappropriate resource management has led us to unprecedented deterioration of ecosystems on a global scale e.g. land degradation, fish stocks collapse, etc., putting the long-term wellbeing of human society at risk (Rockström et al., 2009; Steffen et al., 2015). Managing renewable resources sustainably is key for satisfying present and future human needs, and this constitutes a sustainable development goal to be defined by the United Nations this year (United Nations General Assembly, 2014). One way to direct our society towards more sustainable practices is through footprint accounting. This method assesses the resource consumption and waste production associated with human demands on nature (Mathews et al., 2000; Giovannini, 2004; Tyedmers and Pelletier, 2007), and one of the most environmentally demanding human needs is food provision (Steinfeld, 2006). Food consumption (i.e. kcal per capita per day) and animal-derived protein demands are expected to continue increasing (FAO, 2011), with a wealthier global population reaching 9 billion by the middle of this century (Godfray et al., 2010). Therefore, ensuring the physical and economic access to people’s dietary needs and preferences is the major concern of our time (World Food Summit, 1996). Fish constitutes a food source that can offer a high nutritional profile at an affordable price (Roos et al., 2006). In 2011, 2.9 billion people obtained 20% of their animal-derived protein from fish, with an additional 4.3 billion people obtaining 15% of it (FAO, 2014). Fish consumption has grown at an annual rate of 3.6% since 1961, twice that of human population growth (WHO, 2014).
From the total production of aquatic animals in 2012, 58% was sourced from capture fisheries (FAO, 2012). However, world fisheries landings stagnated since the 1990s (WRI, 2015), which contrasts aquaculture’s average growth rate of 8% per year since the 1970s (FAO, 2009), the fastest growing animal food sector in the world. Nowadays, 57% of wild marine fish stocks are fully exploited and 30% are overexploited, indicating a continued decline of fish stocks if fisheries governance does not improve it (WRI, 2015). Aquaculture therefore holds a significant potential to play a key role in global future food security (Troell et al., 2014). Nevertheless, the sector’s fast development faces social and environmental challenges through its substantial reliance on natural resources as inputs and raw materials (Beveridge et al., 2013; Troell et al., 2014).
Aquaculture affects freshwater resources through water consumption and pollution (Boyd, 1985, Verdegem et al., 2006; Henriksson et al., 2014; Wang et al., 2014), thus impacting on freshwater availability. Water availability is not only fundamental to food production (Wada and Bierkens, 2014), which already accounts for 80% of the global human freshwater use (Carr et al., 2012), but its accessibility is also a Human Right (United Nations General Assembly, 2010). Water is also a vital provisioning resource (MA, 2005), necessary to support the growing industrial and urbanization trends (Mung, 2015), but most importantly, the basic functioning of ecosystems (MA, 2005). Water scarcity is increasingly recognized as a threat for sustainable development (Oki and Kanae, 2006; Liu and Savenije, 2008).
Currently around one billion people live in water scarce areas, with the potential to reach 3.5 billion in ten years if water resource management is not improved (WRI, 2015). At the same time, global water management faces important challenges such as climate change-induced shifts in the hydrological cycle (Rockström et al., 2009). Fluctuations in freshwater quantity and quality can generate socioeconomic cascading effects, impacting across scales in society (Gordon et al., 2008). Therefore, reaching adequate water governance constitutes the core of many important international initiatives. For example it is one of the six global challenges the World Resource Institute has set for its new four-year strategy (WRI, 2014-2017), and the World Economic Forum identified the “water crises” as one of its five global risks (WRI, 2015).
Aquaculture’s water consumption occurs in situ in its production units i.e. culture systems, but also indirectly through consumption of feeds containing agriculture crops (Verdegem and Bosma, 2009; Troell et al., 2015). Current research efforts that aim at identifying alternative aquaculture feeds involves replacing fish protein-based to plant protein-based feeds (Rust et al., 2011), which will further increase freshwater use by adding pressure on agriculture crops (Boissy et al., 2011; Troell et al., 2015). To quantify the freshwater footprint of products, processes, consumers, businesses, and/or nations, the Water Footprint Network (WFN) (http://www.waterfootprint.org) has developed a comprehensive assessment methodology (Hoekstra et al., 2011). It provides a vast database of water footprints, free and easily accessible by the general public. Nevertheless, and very surprisingly, fish production and aquaculture practices are not included in their databases, despite their global importance. In fact, aquaculture has been absent from many international debates on resource use or food and nutrition security.
This thesis aims to explore the freshwater footprint of aquatic animals produced through aquaculture in China. The conceptual framework and methodology developed by the WFN will be used as a basis, with additional dimensions added to capture the specificity of aquaculture production systems. The results describe the freshwater consumption as water volume per production unit at farm gate (m3 tonne-1), which will be discussed using the social- ecological systems approach to sustainable development. A special focus will also be placed on water scarcity issues and the links between resource management strategies, and the current and future Chinese aquaculture production.
2. THEORETICAL FRAMEWORK
!
2.1 The concept of Sustainable Development
The realization of the connection between the long pursued economic growth, social wellbeing and health of the natural environment, reached the international political agenda at the United Nations World Commission on Environment and Development in 1989 (UNWCED, 1989). The official definition of sustainable development was presented as
“development that meets the needs and aspirations of the present without compromising the ability to meet those of the future.” Sustainability understands the world as a system composed of three inherently connected pillars: the economic, social and environmental (Pierantoni, 2004; United Nations General Assembly, 2014). Disregarding one of the pillars when seeking a sustainable socio-economic development, compromises the whole system’s sustainability (Kahuthu, 2006). The new Sustainable Development Goals (SDG) for 2015- 2030 (United Nations General Assembly, 2014), emphasize this interacting multidimensional nature, and calls for a socially inclusive and environmentally sustainable economic development. Despite global attempts and negotiations to agree on the new SDG, the difficulty in defining the term “sustainability” has lead to a variety of approaches to describe and evaluate the advancement towards it (Giovannini, 2004).
2.2 The Social-Ecological Systems approach to Sustainable Development
This study discusses current and future sustainability issues in the Chinese aquaculture sector, by adopting a social-ecological approach. This approach understands the world as an integrated system of human societies and ecosystems, referred to as social-ecological systems (SES) (Folke et al., 2010). Thus, the three pillars of sustainability are understood through the hierarchical nested circles model (Carter and Moir, 2012) (Figure 1): The environment is the basic life-support system for societies, and its stable functioning is a requisite for their success (Rockström et al., 2009). Society is represented at the centre of the model, an integrated part of ecosystem functioning, capable of shaping global patterns and processes (Steffen et al., 2011). Subordinated to these two realms is the economy, understood as services to society that steer our interaction with the environment (Carter and Moir, 2012; Griggs et al., 2013).
Figure!1.!The!nested!circles!model!of!sustainability!(Image!source:!Rockström!and!Klum,!2012).!
Recognizing the Chinese aquaculture production system as a SES, where aquaculture’s production system is embedded and interrelated to the health of the ecosystems, helps better understand its dependence and the impacts on the limited freshwater resources. Defining characteristics of SES are their complexity and adaptive nature i.e. complex adaptive systems (Gunderson and Holling, 2002). This translates into non-linear behaviours, interacting across time and space scales, and leading to system dynamics characterized by drastic changes, and behaviours which are hard to predict (Folke et al., 2010; Levin, 1998). e.g. uncertain market fluctuations or flood and drought events. Hence, the SES approach questions previous linear approaches to sustainability (Peterson et al., 2003). To achieve sustainable goals, the system can be managed for resilience, the capacity of a system to absorb shocks while maintaining its identity and function (Walker et al., 2002). This approach guides the governance of uncertainties and risks of detrimental impacts to society and the environment (Brand, 2008).
Desirable states can be maintained through re-organization and adaptation, while non- desirable states can be transformed (Walker and Salt, 2006).
2.3 Water footprint accounting as a sustainability evaluation method
The Footprint accounting was introduced to assess human activities’ environmental consequences (Giovannini, 2004; Tyedmers and Pelletier, 2007). The first suggestion to quantify water use towards a unit of reference was made within the Life Cycle Assessment (LCA) framework by Guinée et al. (1993). Hoekstra (2003) introduced the Water Footprint concept to illustrate the use of freshwater through human consumption, demonstrating the links between global trade and resource management (Hoekstra, 2009). The water footprint assessment developed by the WFN defines a water footprint as the direct and indirect freshwater consumption and pollution associated with products, consumers, businesses or nations (Hoekstra et al., 2011). Resource consumption is quantified combining the commonly addressed blue water component (i.e. consumption of surface or ground “liquid water” that flows through rivers, aquifers, groundwater, wetlands, damns, etc.) with Falkenmark’s green water concept i.e. the proportion of rainwater consumed by vegetation through evapotranspiration of the soil moisture (Falkenmark, 1995, 2003; Rockström, 2001;
Falkenmark and Rockström, 2006). An additional grey water component is also included to quantify water pollution i.e. the volume of freshwater consumed to dilute a pollutant to accepted water quality standards or natural background levels.
The WFN relates to other similar approaches that measure the environmental impacts of human demands. For example, the Life Cycle Assessment (LCA) is an internationally standardized methodology that measures the environmental efficiency of products or services over their entire life cycle (i.e. “from cradle to grave”) (Owens, 2002). Extractions and emissions are ultimately grouped into a limited number of impact categories (UNEP, 2015).
However, up until the recent publication of the ISO14046:2014 on the principles, very little attention was put on freshwater resource use (Koehler, 2008), commonly featured as a simple indicator of water input, obviating necessary details to understand and manage its sustainability (Owens, 2002).
3. CASE STUDY – China
!
3.1 China – the global aquaculture giant
Investments in research and development, economic incentives, long aquaculture farming tradition, and established fish eating habits, have spurred the Chinese global leadership as an aquaculture producer (NBSO, 2010; Hishamunda and Subasinghe, 2003). China was responsible for 62% of the global aquatic animal products in 2012, ten times more than any other nation (Figure 2) (FAO, 2012). Concurrently, China is also the leading export country of fishery commodities worldwide, with exports estimated at 18.2 billion US$ in 2012 (i.e.
more than double the second-ranked nation, Norway, with 8.9 billion US$), and with an average export rate of 21.6% between 2009 and 2012 (FAO, 2012). The export is mostly comprised of farmed species (Yu, 2012 cited as in Zhang et al., 2013). China is also the third largest importer of fishery products, driven by its cheap processing costs (Zhang, 2005), and shifting consumption preferences e.g. towards high-value aquatic species such as trout and shrimp (FAO, 2012). China reports its aquaculture production by body of water i.e. ponds, lakes, reservoirs, river/ditches, rice-fish systems and other, and by farming system i.e. cages, pens and other (NBSC, 2012). The rest of the production is reported as marine aquaculture but this also includes brackish water culture. This production takes places either off-shore, along the coast in ponds or in cages at different depths, and also in rafts. The bulk of Chinese aquaculture is sourced from the freshwater environment i.e. 73% of the farming area was situated in freshwater environments in 2014 (NASO, 2014), and 64% of the aquatic animal production by weight was freshwater (FishStat, 2012). Freshwater aquaculture is dominated by carps (i.e. 68% of freshwater production by weight), representing over 90% of the global production (FAO, 2012). Other commercially important species groups include tilapia (6% of freshwater production) and catfish (4% of freshwater production) (Fish Stat, 2012).
a)! !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
b)! !
Figure! 2.! a)! Total! aquaculture! production! in! 2012! by! weight! (%)! comparing! China! with! the! five!
continents,! b)! the! top! 13! countries! in! aquaculture! production! of! fish,! crustaceans,! molluscs,! etc.!
(FishStat,!2012).!
3.2 The sustainability of China’s aquaculture reliance on scarce freshwater resources
The fast expansion of aquaculture and its predominant freshwater-based production has led to a high competition for space and use of limited freshwater resources with other developing and expanding Chinese industries (especially agriculture), resulting in an overall intensification of aquaculture practices i.e. increase in output per land area (NBSO, 2010).
China’s consumption of its already scarce water resources has increased 5.8 times over the last 60 years (Xiao-jun et al., 2012), ascending to three times more than the average global per capita consumption (Chinese Academy of Sciences, 2007 cited as in Liu and Wu Yang, 2012). At the same time, flood and drought events have become more frequent and severe, due to climate change and hydrological cycle alterations (IPCC, 2007). The Chinese
62!
27!
5! 4! 2! 0!
0!
10!
20!
30!
40!
50!
60!
70!
China! Asia!(excl.!
China)! Americas!!! Europe!!! Africa!!! Oceania!!!
%!Global!produc-on!of!aqua-c! animals! !
Con-nent!
61.7!
6.3! 4.6! 4.6! 2.6! 2.0! 1.9! 1.6! 1.5! 1.3! 1.2! 1.1! 1.0!
0.0!
10.0!
20.0!
30.0!
40.0!
50.0!
60.0!
70.0!
%!Global!produc-on!of!aqua-c! animals!
Country!
government is seeking to achieve a sustainable water use and management by 2020 through its Water Pollution Action Plan, to address the large generated economic losses of its uncertain supply (Xiao-jun et al., 2012), which can incur in ecological and human well-being impacts (United Nations General Assembly, 2014). The most common strategy has focused on engineering solutions to ensure water supply (Xiao-jun et al., 2012). For example, the South-North Water Transfer Scheme is the world largest water diversion project. However, the adequacy and sustainability of such strategies to deal with the constantly changing socio- economic context and the nevertheless persistent water supply uncertainty have been contested (Jiang, 2009). Thus, the approach is shifting towards the management of water demand (Xiao-jun et al., 2012). The modernizing trend to improve parts of the aquaculture production chain such as increased productivities and improved feed efficiencies and compositions, has the potential to lower water consumption. This would especially be the case if the sustainable values of traditional aquaculture are preserved at the same time (WWF, 2015).
Improved water resource management is a key requisite to help alleviate the vulnerability associated with water scarcity (Jiang, 2009), and for moving towards more sustainable practices in aquaculture, avoiding the compromise between food and water security goals.
Thus, a better understanding of the aquatic food production social-ecological system of China will provide insights to guide the sustainable development of the sector as well as providing an example to other countries with aquaculture potential.
4. AIM & RESEARCH QUESTIONS
!
The main aim of this thesis is to perform a preliminary estimation of the water footprint of Chinese aquaculture by identifying key methodological aspects and variables specifically related to water footprint accounting in aquaculture. The applicability of these variables will be discussed, together with how the water footprint of Chinese aquaculture relates to water scarcity and sustainable water consumption at various scales.
Specific research questions relate to 1) what are the most water-consuming processes within an aquaculture production system? 2) what production systems and intensities of production have the largest water footprint? 3) what are the main advantages and limitations of the water footprinting methodology to inform about sustainability issues in aquaculture?
5. METHODS
In this study a water footprint was understood as the appropriation, consumption or loss of water from the available surface and groundwater bodies, through evaporation, return to another catchment area, to the sea, or incorporated into a product (Hoekstra et al., 2011) (refer to Appendix II for a list of terms and their definitions). The accounting methodology was based upon the manual developed by Hoekstra et al. (2011) (Figure 3), taking a LCA approach with a simplified goal and scope definition. This means including a reduced number of stages in the life cycle of aquaculture production system. Additional information needed to capture the specificity of water usage in different aquaculture farming systems, were collected from available literature and personal observation in the field.
The freshwater footprint of a product includes the water consumption at the production unit, the farm (i.e. direct water footprint) and supply chain (i.e. indirect water footprint) (Hoekstra et al., 2011). The WFN specifies three spatio-temporal levels of detail: Global, national/regional and local. In this study, net water consumption was estimated as much as possible at a regional scale. This scale can also be defined at the river-catchment area when geographically explicit data are available. The data were temporally analysed using annual and/or multi-annual averages.
5.1 The directwater footprint of Chinese aquaculture
The “direct water use” in this study will only refer to the blue water footprint of culture operations at the farms. Even though water pollution was not considered in this study due to time, data and methodological constraints, this is not to say that freshwater pollution is not important. For example, aquaculture systems may render freshwater useless for other purposes through e.g. eutrophication or salinization (Cao et al., 2007). The green water footprint at the farms was neither accounted for, considering the green water footprint of phytoplankton and other present plants as negligible.
! Figure!3.!!Different!types!and!components!of!the!water!footprint!assessment!(Hoekstra!et#al.,!2011).!
!
The farming systems were classified as semi-closed and open water systems:
5.1.1 Semi-closed water systems
Semi-closed water systems refer to man-made water embankments (i.e. ponds) of varying sizes used as production units for the stocking of aquaculture species. Water is obtained from natural sources such as rainfall, rivers, springs, etc. (Tidwell, 2012). Within freshwater aquaculture, China reports statistics for ponds and rice-fish systems. However, 71% of the total freshwater production in 2011 took place in ponds, hence, these were the focus on the study. Ponds can also be filled with brackish water, and brackish water ponds were included to illustrate the different methodological considerations to be taken compared to freshwater ponds. Also, the production of high-value species farmed in the brackish water environments, such as shrimp and prawns, is increasing (Pemsl and Bose, 2008; FAO, 2012).
The main water outflows found in aquaculture ponds (Boyd, 1985; Nath and Bolte, 1998;
Boyd and Gross, 2000; Braaten and Flaherty, 2000; Verdegem et al., 2006; Chapagain and Hoekstra, 2011; Hall et al., 2011) are illustrated in Figure 4 and discussed below.
Figure!4.!Schematic!representation!of!the!main!processes!through!which!water!can!be!consumed!or!
relocated!in!an!aquaculture!pond.!!
a. Evaporation from water surfaces
!
• Freshwater ponds:
Annual mean estimated evaporation values (meters year-1) from open-water were obtained using the STEAM model (Wang-Erlandsson et al., 2014) for the years 1999-2011, at a resolution of 1.5º. These values were plotted using the software QGIS (Quantum Geographical Information System) v2.6. From the Natural Earth public map database (http://www.naturalearthdata.com) a global vector layer of the Chinese internal administrative boundaries i.e. provinces and states, v.3.0.0, was added at a scale of 1:107. Within each Chinese province, zonal statistics were used to calculate the average evaporation per province (meters year-1).!The total pond area was multiplied by these evaporation rates for the year 2011 (NBSC, 2012) to obtain the total water volume evaporated (m3). This volume was then divided by the total pond production to obtain the evaporation water footprint (m3 tonne-1) per province.
• Brackish water ponds:
Brackish and marine water do not provide many of the functions of freshwater, and are therefore out of scope for most water footprints. Consequently, accounting for the evaporation of freshwater from brackish water ponds was considered irrelevant.
b. Infiltration losses
!
In this study, percolation and lateral seepage were treated under the same assumptions and referred to as infiltration losses (Verdegem and Bosma, 2009). Percolation is defined as the rate of vertical movement of water below the water table. Lateral seepage is defined as the horizontal movement of subsurface water (Huang et al., 2003). The different processes of seepage and percolation are difficult to measure separately, and are often studied as one variable (Wickham and Singh, 1978 cited as in Bouman et al., 1994).
Water infiltration is not considered water loss or consumption at the river catchment scale (Bouman et al., 2007). This argument is supported by other authors, such as Huang et al., (2003), who holds that infiltration losses from a paddy rice field recharge the groundwater, and surface water reservoirs, and shouldn’t be considered irrigation losses when managing water resources. Otherwise freshwater resources are easily double counted (e.g. both as seepage and as extracted groundwater). Because the study relies on annual data, the time lag between the consumption of water and its return to the catchment scale could not be captured.
Following the example of Chapagain and Hoekstra (2011) for consistency with other water footprint studies, infiltration losses will be accounted as a separate water consumption process that will later be discussed in comparison to the final blue water footprint.
• Freshwater ponds:
An estimation of the total volume of infiltrated water consumed in freshwater ponds was done using Equation 1.
Infiltration m! =Pond!area! m! ∗ Infiltration!rate! m
day ∗ Farming!cycle!(days) Production!(tonne)
Equation!1.!Equation!used!for!the!calculation!of!the!freshwater!consumption!through!of!infiltration!
loss!in!Chinese!aquaculture!ponds.!
!
A low infiltration rate of 2.5 mm/day was assumed to remain on the conservative side!(Yoo and Boyd, 1994). The length and number of farming cycles varies spatio-temporally and background data from Henriksson et al. (2014) on carp and tilapia ponds was used as a proxy to estimate the number of days ponds are used for freshwater species grow out. Carp and tilapia represent 46% of China’s freshwater production (FishStat, 2012). These data were collected in the provinces of Guangdong and Hainan in 2009. The average of different species of carps and their different grow out periods was taken. Results were extrapolated nationally, which for illustration purposes was considered acceptable.
The Chinese yearbooks do not report the total freshwater pond production desegregated into the different species groups cultured in the ponds, as well as the freshwater pond area dedicated to each of them. The reason for this is the large spatio-temporal overlap of the freshwater culturing techniques, e.g. almost all tilapia are farmed in polyculture systems together with carps and/or other species (Wang et al., 2014); making it next to impossible to allocate specific areas to specific species. Allocating specific evaporation and infiltration rates to different freshwater species groups could therefore not be done. Marine aquaculture annual statistics are, however, reported disaggregated and even if not relevant for direct freshwater usage, it provides a structure for linking indirect freshwater usage to species groups.
• Brackish water ponds:
The infiltration of brackish water does not contribute to freshwater consumption, unless freshwater is damaged through salinization of e.g. groundwater reservoirs, making the freshwater undrinkable. The lack of high-resolution data on location, common practice and brackish water initial salinity makes it difficult to estimate this water consumption. Hence, it was not part of the study.
c. Water drainage and exchange
!
• Freshwater ponds:
Although the freshwater pond water volume can be left to dry out i.e. through evaporation or infiltration, it is more often drained after harvest e.g. through drainage or irrigation canals, dumped in rivers and streams, etc. meaning the water still remains within the same catchment
area. Because of this, water is considered as not being consumed within the catchment area.
Moreover, the interruption of water flows due to water retention in ponds is considered a negligible water loss at this scale.
• Brackish water ponds:
The drainage of brackish water from ponds as well as its exchange can contribute to freshwater consumption through the salinization of freshwater. When brackish and fresh waters are mixed, the freshwater fraction becomes unsuitable for ground and surface water recharge, irrigation or household consumption among others uses. Thus, it is a net consumption of freshwater at the catchment scale. High-level inland ponds in China do commonly carry out this practice (Henriksson, 2015 pers. comm. Feb). Low-lying coastal brackish water ponds with daily tidal water exchange have a negligible freshwater use (Verdegem and Bosma, 2009). Using the background data in Henriksson et al. (2014), the sources and total volume of freshwater consumed to dilute the coastal saline water in inland brackish water ponds were identified and calculated.
5.1.2 Open water systems
Open freshwater systems refer to natural bodies of water that are stocked for commercial purposes, where oxygenation, waste removal and temperature control rely on natural ecological processes (Tidwell, 2012). Open water systems of interest to this study include lakes, reservoirs and rivers. In these systems, the evaporation water footprint cannot be allocated to the pens and cages, since that open water surface area already existed prior to the farming practices. Only the following direct water uses can be attributed to aquaculture:
- Water pollution: This component is outside the scope of this study.
- Water incorporated into the product’s biomass: For example, if 60-90% of the product’s biomass e.g. fish and crustaceans, is water (Boyd, 1990 cited as in Boyd & McNevin, 2014; Murray and Burt, 2001), translating into a water footprint of only 0.6-0.9 m3 tonne1. Therefore, it was considered negligible in this study.
5.2 Indirect water footprint of Chinese aquaculture
“Indirect water use” refers to the water footprint associated with the supply chain. In this study, the three water footprint components i.e. the green, blue and grey water footprints, are associated with the production of goods used in the farms i.e. aquaculture feeds.
To calculate the water footprint of aquaculture feeds, the database of Mekonnen and Hoekstra (2011) was used to obtain the green, blue and grey water footprint values of the raw agricultural commodity from which ingredients are derived. The production of feed for species across the three environments i.e. marine, brackish water and fresh water, were taken into account. Additional water usage for the processing, refinement, transport and consumption, of these ingredients was not included in the study. For some species this may be significant but for others it might not, and this is most likely the case for most other foods as well.
The final water footprint of feeds is influenced by the following parameters (Mekonnen and Hoekstra, 2010; 2010b):
!
a. Feed composition
!
The production of plant-based ingredients incurs in larger water footprints than animal-based ingredients, such as fishmeal and fish oil (Verdegem et al., 2006). At the same time, because of escalating costs and diminishing supplies of fishmeal, fish-based proteins are expected to gradually be substituted by plant protein concentrates (Tacon et al., 2011; Troell et al., 2014).
Hence plant-based ingredients were the focus of this study, considering the fishmeal and fish oil water consumption as negligible. Due to a lack of statistics on farm-made feed use, only the Chinese demand of commercial feeds was analysed using background data from Tacon et al. (2011).
b. Origin of ingredients
!
The production origin of the ingredients influences the final freshwater footprint through the different climatic and agricultural conditions found in different production locations (Mekonnen and Hoekstra, 2010). FAO’s “crops and livestock product trade matrix” database
(http://faostat3.fao.org/download/T/TM/E) was used to identify the countries from where the feed ingredients were sourced, using “China mainland” and the latest year (2011). FAO reports imports of ingredient grouping them in larger categories. The categories encompassing the ingredients demanded in the Chinese aquaculture were analysed. The top countries constituting 95% of the imports were evaluated. The specific water footprint of raw agricultural commodities was then averaged across the different identified countries of import. The water footprint of the remaining 5% was obtained using a global average.
c. Water footprint allocation method
!
For processes where more than one product is yielded (e.g. wheat milling yields wheat flour and wheat bran), the water footprint can be divided differently depending on what methodology is being applied. These include allocation based upon the relative weight proportion of the raw product (i.e. by mass), allocation based on economic values, or based on caloric content. In this study, both a mass and economic allocation were evaluated. In order to calculate the mass allocation factors, all products resulting from a process need to be identified e.g. soybean leads to soybean meal, oil and lecitin, as well as their share of the total mass e.g. soybean meal is 79% of a soybean (Equation 2). These values were obtained from the background data in Tacon et al. (2011) and supplemented by literature search (see Table A1 in Appendix I for details on the calculation). To calculate the allocation factors for economic values, the ingredient prices were obtained from Henriksson et al. (2014) and online searches for current market prices. These prices could then be multiplied by the previously calculated share of the total mass (Equation 3).
Mass and economic allocation factors were multiplied by the green, blue and grey water footprints of the raw agricultural commodities obtained from Mekonnen and Hoekstra (2011).
!"##!!""#$!%&#'! = (%!!"##)! (%!!"##)!
!!!!
Equation! 2.! Mass! allocation! calculation! of! the! freshwater! footprint! of! feeds! used! in! Chinese!
aquaculture!in!2010.!
!"#$#%&"!!""#$!%!"#! = (!"#$% ∗ %!!"##)! (!"#$% ∗ %!!"##)!
!!!!
Equation! 3.! Economic! allocation! calculation! of! the! freshwater! footprint! of! feeds! used! in! Chinese!
aquaculture!in!2010.!
d. Economic Feed Conversion ratio (eFCR)
!
The eFCR refers to the unit of weight of feed consumed per unit of weight of saleable product (Tverdal and Larssen, 2006), ultimately influencing the total quantity of feed required by each species group to grow to its harvest weight. The data are reported in Tacon et al. (2011) (see Table A2 in Appendix I). To convert the species group’s diet water footprint into a footprint per tonne of animals produced, the eFCR was multiplied by the diet’s overall water footprint.
6. RESULTS
!
6.1 Directwater footprint: The water footprint of Chinese fresh and brackish water aquaculture ponds
!
6.1.1 Evaporation water footprint in freshwater ponds
The evaporation water footprint of freshwater ponds in China is independent of what species is being farmed in the ponds, although densities impact on footprint per biomass. If aggregated at a provincial level (Figure 5a), aquaculture performed in the large northern and north-western provinces of Inner Mongol, Qinghai and, Xizang had the highest water footprints, ranging from 4621 m3 tonne-1 in Inner Mongol to 16621 m3 tonne-1 in Qinghai.
This is largely due to the low yields ranging from 0.9 to 3.2 tonne ha-1 (Figure 5b), as annual evaporation rates do not vary considerably (Fig 6).
The south-eastern and north-eastern coastal provinces of Guangdong, Fujian, and Liaoning had the lowest water footprints (1067 to 1386 m3 tonne-1), largely due to their higher yields (ranging from 10.9 to 13.1 tonne ha-1).Although the province of Guangdong ranks third in terms of total volume of water evaporated, with an annual evaporation rate of 1.51 m year-1, it also had the largest provincial contribution to the annual aquaculture production (17.4%) (for details on the provincial area (ha), production (tonnes), volume of water evaporated (m3) and yield (tonne ha-1), refer to Appendix I Table A3). Hebei and Shanxi follow Guangdong, Fujian, and Liaoning with slightly lower yields (9.2 and 10.3 tonne ha-1) and higher water footprints (1487 and 1398 m3 tonne-1 respectively). Evaporation rates remained similar (1.37 and 1.44 m year-1 respectively) but again, production areas had an order of magnitude difference (28266 ha and 2032 ha respectively).
a)
b)
Figure!5.!a)!The!evaporation!component!of!the!blue!water!footprint!of!freshwater!ponds!and!b)!the!
freshwater! pond’s! yield! of! Chinese! aquaculture,! averaged! at! the! province! level.
! Figure! 6.! Average! annual! evaporation! rates! in! the! Chinese! provinces! for! the! years! 1999! to! 2011!
(Wang[Erlandsson!et#al.,!2014).!
! 6.1.2 Infiltration losses in freshwater ponds
The farming cycle length (days year-1) was found to be an average of 206 (199 days for carps and 212 days for Tilapia). Accordingly the total infiltration water volume in freshwater ponds was calculated to be 724 m3 tonne-1.
6.1.3 Dilution losses in brackish water ponds
From the total number of shrimp brackish water ponds in Guangdong surveyed by Henriksson et al. (2014), 65% used freshwater sources to dilute the brackish or marine water of their ponds. 58% derived this freshwater from pumping groundwater while 7% sourced it from rivers or streams. The average river or stream freshwater footprint was 4345 ± 7149 m3 tonne-1, while the average ground freshwater footprint was 5034 ± 9225 m3 tonne-1, showing the large variation within the samples.
6.2 Indirect water footprint: The water footprint of Chinese aquaculture feeds
!
6.2.1 Quantity and composition of plant-based aquaculture feeds
The 11.8 million tonnes of feed demanded by Chinese aquaculture in 2010 was composed of 25 plant-based ingredients (Tacon et al., 2011) (Table 1). From these, only four ingredients accounted for 77% of the total quantity used, i.e. rapeseed/canola meal, soybean meal, wheat and maize.
Table!1.!Ingredients!used!in!Chinese!aquaculture!feeds!in!2010.!
!
6.2.2 The origin of the ingredients
The ingredient groups “Cassava & products”, “Peas” and “Soybean” were to a large extent imported (Table 2), meanwhile the remainder of ingredients were assumed to have been produced domestically. More than 95% of imported cassava, peas and soybean could be linked to specific countries of origin.
Table!2.!Quantity!used!and!imports!of!key!ingredient!groups!found!in!aquaculture!feeds!in!China.!In!
red! are! the! ingredient! groups! with! considerably! high! imported! quantities!
(http://faostat3.fao.org/download/T/TP/E).! Countries! from! where! ≥! 95%! of! the! ingredient! groups!
were!imported!from!in!2011!are!identified!(http://faostat3.fao.org/download/T/TM/E).!
6.2.3 The water footprint of feed ingredients
The final water footprint of producing the feed ingredients depends on the allocation methodology used (Figure 7). There is some similarity between the results of both allocation methodologies (see Appendix I Table A4 for a detailed table of the water footprint values of feed ingredients):
- The largest water footprint corresponds to green water with a substantial variation between the allocation methods (1540 ± 017 m3 tonne-1 and 1462 ± 527 m3 tonne-1 using economic and mass allocation respectively). Mass allocation showed sunflower seed meal as the major water consumer with 2504 m3 tonne-1, followed by all the soy-, pea- and wheat-based ingredients. Using economic allocation, only soybean oil and lecitin
(3939 and 3646 m3 tonne-1) show higher green water footprints, followed by sunflower seed meal, rapeseed/canola oil, the pea-based ingredients and the wheat-based ingredients except for wheat bran.
- The second largest water footprint is the grey. Using mass allocation, cotton seed meal has the largest grey water footprint, followed by the pea-based ingredients, sunflower seed meal, lupin kernel meal and rapeseed/canola oil. Using economic allocation, the same ingredients had the largest water footprint. The ranking, however, changed slightly. Rapeseed/canola oil showed the largest grey water footprint, followed by sunflower seed meal, pea-based ingredients, lupin kernel meal and cotton seed meal.
Overall, the grey water footprint is one order of magnitude smaller than the green, and the variation in both samples does not differ notably (i.e. 337 ± 207 m3 tonne-1 and 308
± 250 m3 tonne-1, using mass and economic allocation respectively).
- The blue water footprint showed high variation and did not differ notably using mass or economic allocation (i.e. 179 ± 199 m3 tonne-1 and 167 ± 203 m3 tonne-1, using mass and economic allocation respectively), and very similar ingredients are found at the top.
Using mass allocation, the wheat-based ingredients had the highest water consumption, 480 m3 tonne-1, and sunflower seed meal, cotton seed meal and the rice-based ingredients were in descending order lower. Using economic allocation, sunflower seed meal had the highest water footprint, 552 m3 tonne-1, followed by the wheat-based ingredients, broken rice and cotton seed meal.
a)
b)
Figure! 7.! The! green,! blue! and! grey! water! footprint! of! producing! the! ingredients! using! a)! mass!
allocation!and!b)!economic!allocation.!Wheat,!starch!(wheat!starch)!and!wheat!flour!were!assume!to!
be!the!same.!The!same!applies!to!wheat!by[products!and!wheat!bran.!!
!
6.2.4 The water footprint of different species groups
Different species groups consume different types of feed at different quantities, driving the overall freshwater footprints associated with their diets (Figure 8). The final water footprint of producing the different species groups using both allocation methodologies is presented in Figure 9.
• When using economic allocation, the water footprint of eels is among the highest compared to the other species groups. Eels’ diet is 93% composed of starch and soybean meal (Figure 8a). Wheat starch is among the most water consuming ingredients, especially it’s blue water footprint. Soybean meal shows a closer to average water footprint, especially its green water footprint component, two orders of magnitude larger than its blue and grey components. However, when using mass allocation, the water footprint is considerably lower.
• Trout showed the most varied diet: Sunflower seed meal only accounts for 9% of the diet but has consistently a large green, blue and grey water footprint. Soybean oil (22%) has a proportionately larger green water footprint, especially when using economic allocation, and this heavily impacts on trout’s water footprint. At the same time, its grey water component is among the lowest. Wheat (13%), assumed to be the same as wheat starch, soybean meal (21%) and corn gluten meal (8%). When using economic allocation, trout remains among the species groups with highest water footprints. However, trout has also the lowest eFCR (1.3), which greatly aids them in lowering its water footprint.
• Freshwater prawns have a diets where 50% of their plant-based feed is composed of rice bran, an ingredient which has among the lowest or second to lowest water footprint, mostly driven by a proportionately lower green water footprint as well as the lowest grey water footprint. The other 50% is wheat bran, which shows a lower water footprint when using economic allocation. On the contrary, it is with economic allocation that the water footprint of prawns notably increases.
!
Figure!1a.!Plant!ingredients!found!in!the!diets!of!the!different!species!groups!farmed!in!Chinese!aquaculture!for!the!year!2010.!Background!data!
from!Tacon!et#al.!(2011).!
!
!
Figure!8b.!Plant!ingredients!found!in!the!diets!of!the!different!species!groups!farmed!in!Chinese!aquaculture!for!the!year!2010.!Background!data!
from!Tacon!et#al.!(2011).!!!
