Master’s Thesis, 60 ECTS
Social-ecological Resilience for Sustainable Development Master’s programme 2016/18, 120 ECTS
ENVIRONMENTAL IMPACTS OF CIRCULAR SCENARIOS FOR THE TEXTILE INDUSTRY
A PLANETARY BOUNDARIES-BASED LIFE CYCLE ASSESSMENT OF COTTON T-SHIRT
Piero Grilli
Stockholm Resilience Centre
Research for Biosphere Stewardship and Innovation
STOCKHOLM RESILIENCE CENTRE
Research on the Planetary Boundaries framework - Fashion within boundaries project
AUTHOR:
PIERO GRILLI grilli.piero@circeco.com
Stockholm Resilience Centre, Stockholm University
SUPERVISOR:
TIINA HÄYHÄ tiina.hayha@su.se
Stockholm Resilience Centre, Stockholm University
CO-SUPERVISOR:
RAFAEL LAURENTI rafael.laurenti@epfl.ch
École Polytechnique Fédérale de Lausanne
EXAMINER:
RAJIB SINHA rajib.sinha@abe.kth.se KTH Royal Institute of Technology
Abstract
Planetary Boundaries (PB) and Circular Economy (CE) are becoming the paradigm for sustainability.
There is an increasing interest to operationalise PB into a framework for businesses to maximise profitability within environmental limits. The context of the cotton textile industry makes a good setting for understanding the casual chain of connections between the socio-economic system expressed by extensive global supply chain of cotton, and its ecological interconnection with the Earth’s system that is put under pressure. For this study, life cycle assessment (LCA) is recognised as a suitable approach for measuring the linkages between those two systems. Results show that among all phases of the life cycle of a cotton t-shirt, the usage phase is the most impacting for most of environmental indicators, followed by the manufacturing and farming phase. Production or extraction of raw material as elements considered upstream in the supply chain are the predominant cause of impacts in this case study. Increasing circularity of the system yields to an improvement in environmental performance. However, the system remains largely unsustainable when taking into account the state of the Earth’s system, through the PB. When assessing sustainability through LCA, ecological references like PB, must be considered to understand absolute environmental sustainability a product system. This will reveal whether less impacting options in the system, are still deteriorating a state of the environment as a part of the Earth’s system, which needs to be the area of protection. In conclusion, linking planetary boundaries to life cycle assessment can help assess the absolute environmental sustainability, as opposed to relative sustainability, of a production system. Combining an assessment of the state of the environment (safe, critical/uncertain, at risk), and the assessment of environmental hotspots of the system under scrutiny, will determine where priority for goals and actions of improvements for environmental sustainability needs most attention.
TABLE OF CONTENT
1 INTRODUCTION _________________________________________________________ 1 2 THEORETICAL BACKGROUND ___________________________________________ 3
2.1 Social-Ecological System Perspective ____________________________________________________ 3 2.2 Planetary Boundaries framework ________________________________________________________ 4 2.3 Life cycle thinking ___________________________________________________________________ 5 2.4 PB-based LCA - where are we now? _____________________________________________________ 6 2.5 Circular economy – product and material circularity________________________________________ 11
3 CASE STUDY AND BACKGROUND INFORMATION ________________________ 13
3.1 Cotton production, manufacturing, use and disposal for the EU consumers ______________________ 13 3.2 Scenario analysis ___________________________________________________________________ 16
4 METHODOLOGY ________________________________________________________ 18
4.1 Goal and scope of the LCA ___________________________________________________________ 19 4.2 Functional unit _____________________________________________________________________ 20 4.3 System boundaries and expansion ______________________________________________________ 20 4.4 System processes and flows ___________________________________________________________ 21 4.5 Data quality and source ______________________________________________________________ 22 4.6 Data gap and limitation ______________________________________________________________ 24 4.7 Life Cycle Impact Assessment _________________________________________________________ 24
5 RESULTS _______________________________________________________________ 31
5.1 Results by each scenario _____________________________________________________________ 31 5.2 Results of comparison between scenarios ________________________________________________ 36 5.3 Results in relation to PB ______________________________________________________________ 38
6 DISCUSSION ____________________________________________________________ 44
6.1 Life cycle assessment: hotspots of environmental impact for a cotton t-shirt _____________________ 44 6.2 Scenario analysis: comparison of circular scenarios ________________________________________ 45 6.3 PB-based LCA: insights and the way ahead ______________________________________________ 46 6.4 Strengths and limitation of PB-based LCA methodology ____________________________________ 48
7 CONCLUSIONS __________________________________________________________ 50 REFERENCES ________________________________________________________________ 52 ANNEX I - THEORETICAL BACKGROUND _____________________________________ 56 ANNEX II - CASE STUDY DESCRIPTION AND BACKGROUND ___________________ 58 ANNEX III – METHODOLOGY _________________________________________________ 59 ANNEX IV – RAW DATA ______________________________________________________ 64 ANNEX V – FINAL ETHIC REVIEW ____________________________________________ 69
1 Introduction
Planetary Boundaries (PB) brings at the centre of the sustainability discussion global environmental dynamics, highlighting the non-linearity of the Earth system (Rockström et al. 2009; Steffen et al.
2015). The framework is becoming the ecological foundation for corporate sustainability (Antonini and Larrinaga 2017; Whiteman, Walker, and Perego 2013; Bjørn et al. 2017; Crépin and Folke 2014;
Clift et al. 2017). In practice, businesses are looking for ways of integrating PB-based targets that are more effective and scientifically sound to improve their environmental performance.
The operationalization of PB needs to have a systematic approach, which must be translated into sustainability targets and goals that are not limited to only some environmental indicators in a specific place at a specific time. Instead, it needs to considerate the functioning of the socio-economic system under scrutiny in relation to the Earth’s system as whole. In this context, an example of this approach is the study done by Bjørn et al. (2015) that has pointed out the importance of taking into account absolute ecological reference, such as PB, when assessing sustainability.
The thesis is part of the ‘Fashion within boundaries’ project, a collaboration between the Stockholm Resilience Centre, Ellen MacArthur Foundation, and H&M Group. Its overarching goal is to operationalise the PB framework for the textile sector, while understanding how the Circular Economy (CE) approach could help the textile industry stay within the planetary boundaries. To understand environmental impacts of the textile industry on planetary boundaries, a systemic approach looking at the impacts within the whole global supply chain is needed. With the same holistic approach, it is important to look at how CE principles applied to the supply chain can affect these environmental impacts. In this context, my research focuses on understanding:
How can a Planetary Boundaries-based Life Cycle Assessment be used to assess environmental sustainability of different circular scenarios for a cotton t-shirt?
This study will follow the path of previous studies on assessing absolute environmental sustainability through using the PB framework in combination with the LCA tool for understanding environmental sustainability in the textile context for the case chosen. FIGURE 1 below shows how the main research question is informed by four steps, which are made by several more specific sub-questions. This would be the approach on this study. Because of the diversity and complexity of the textile sector, the thesis focuses on a case study of a specific garment – an average white cotton t-shirt consumed in the European Union (EU). The EU has been selected as the focal area because of its emphasis and actions of the increasing role of circular economy and bio-economy in its policies (European
FIGURE 1 Research question diagram showing how the specific questions feed the research steps that feed each other in steps, and each concur to build the answer for the overarching research question.
2 Theoretical background
2.1 Social-Ecological System Perspective
In the thesis, the Driver-Pressure-State-Impact-Response (DPSIR) framework (OCDE 1993; EEA 1999) is been adopted to conceptually structure the research. FIGURE 2 shows interactions between the Earth’s natural system and the socio-economic system, adapted to the case of a life cycle of a t- shirt. The DPSIR framework is based on a systems approach to understand the interconnected socio- economic and ecological systems as a whole instead of looking at fragmented sectoral parts (Gari, Newton, and Icely 2015).
FIGURE 2 This figure wants to emphasise that impacts caused by resources use and emissions are the results of socio- economic activities, and their severity depends on the state of the environment, this casual chain is described by the DPSIR framework, figure adapted from (UNEP 2010).
The DPSIR framework illustrates how socio-economic drivers, in this case the making, use and disposal of a t-shirt, cause environmental pressure (emission and resource use) throughout all stages of the entire t-shirt life cycle. These pressures, for example chemical pollution to water, affect the state of the environment through changing the water quality and its chemical composition and causes impacts such as risen levels of water toxicity that potentially affects biodiversity and human health.
Eventually, problematic impacts will lead to social responses by policy makers and other actors,
2.2 Planetary Boundaries framework
Exponential growth of environmental damaging human activities, in particular over the last century (Steffen et al. 2011), are having increasing pressure on the Earth system, creating the risk of abrupt or irreversible environmental changes at the global level (Rockström et al. 2009). Human activities, including industrial processes like the making of a t-shirt, entail environmental impacts that can erode stability and quality of the local, regional and global environment.
The Planetary Boundaries (PB) framework (Rockström et al. 2009; Steffen et al. 2015) proposes precautionary maxima for human perturbation to the functioning of the Earth system. Breaching one or more boundaries may trigger abrupt, non-linear change to the local, regional and global environment that would be not favourable for human health and prosperity (TABLE 9 briefly describes the planetary boundary processes and their control variables). The framework is an approach to global sustainability and is used with the scope of defining a ‘safe operating space’ for human activities to develop and thrive without compromising the resilience of the Earth system (Rockström et al. 2009).
However, not all human-induced environmental impacts are actually producing effects directly at the global level, but rather producing effects at local and regional levels, whereas they can cascade and create global concerns. Local and regional impact have systemic impacts in the Earth system as PB processes interact with each other.
The PB framework is meant to define boundaries of the Earth system processes at the global level but many processes act and need to be scrutinised at the local and regional level because they have impacts and create regime shift at those levels before creating possible global impacts. The framework, in fact, distinguished boundaries that are directly related to processes at global scale, such as climate change, ocean acidification and stratospheric ozone depletion, and others that relate to spatially heterogeneous processes, such as landsystem change, freshwater use, atmospheric aerosol loading, biosphere integrity and biogeochemical flows. The difference lay in the way processes behave, showing cross-scale complexity. Some Earth system processes like climate change have thresholds that shows behave at the large scale, while some other like landsystem change have thresholds that shows behave at the regional scale. However, processes happening at different scales are connected to each other. Boundaries at regional scales provide the base for the resilience of the Earth system through their functions like for example acting as carbon sink or by regulating water, nutrients and other fluxes. On the other hand, global scale processes such as climate change impact regional and local processes.
Rockström et al. (2009) acknowledged that control variables for many processes are spatially heterogeneous, so there is a need of having control variable at the local or regional level but compatible with global level boundaries. Some Earth system processes such as climate change are
associated with known thresholds at the global scale, so they have a control variable that refer to the global situation. Some other instead, are recognised to show more heterogeneity and have only identifiable local or regional thresholds, so their control variable refer to a local or regional situation, and in those cases, the initial local and regional situation matter. This is the case for example of freshwater use where the depletion of the resource and so crossing of the threshold in a local or regional context would depend on the resource availability in the first place. Other processes, such as biogeochemical flows are tightly linked with regional or local conditions, but at the same time, the control variable needs to refer to a global resource stock situation (Rockström et al. 2009; Steffen et al. 2015). These are the reasons behind the development of a two-level set of control variable for some boundaries and why it is important to take into account scales of environmental impacts.
2.3 Life cycle thinking
Life cycle-based approach is employed to assess how human activities impacts the environment.
Understand production, consumption and disposal impacts patterns along a supply chain become important for identify possible solution and advance sustainability. Hotspots of potential environmental impacts represent areas of intervention for potential damage reduction. The strength of a life cycle thinking lies in:
• Bringing together a range of several environmental impacts giving a more holist approach, compared to footprints that usually focus on one specific environmental impact. This can help avoid shift of burdens;
• Cradle-to-grave or cradle-to-cradle approach allows to include all different parts of the system under investigation for understanding how they influence one another but also how they behave as a whole;
• Modelling the system as a conglomeration of interrelated and interdependent parts in an interdisciplinary way that includes the social and ecological parts that rather than be considered merely as sum of parts, it is an entity that expresses synergy and behaviours;
• Allowing an analysis of the propriety of the system unintelligible otherwise if considering only individual parts as the whole has greater meaning than the sum of its parts;
• Integrating resource use and emission over an entire life cycle into one analysis and so giving the whole picture of the impact related to a specific supply chain of a specific product or service;
• Allowing environmental impacts to connect to a specific good or service, or to a specific process or strategy of any system defined, facilitating the comparison of alternatives and helping identify
• Helping in determine which process is more sustainable overall helping optimizing individual processes or production system leading to an overall reduction of resources use and emissions.
A consumption-based approach goes in line with the social-ecological system prospective of the life cycle thinking. Socio and economic structures, such as international trade, are making more complex the casual chain between environmental impacts and their sources, while disconnecting responsibility and drivers. This is the case when potentially environmental impacts, related to the production of a good, are spatially separated from the consumption of that same good.
Consumption-based approach helps to accounts for all environmental impacts involved in the consumption of a commodity.
Analytical tools such as footprint indicators and life cycle assessment can truck environmental impact of specific product or processes helping to reconnect responsibility and to rebuild the casual chain between human activity and their impacts (Mozner 2013). The available tools, such as life cycle assessment and environmental footprints, can help to assess how socio-economic structures, such as international trade, are linked to pressure on planetary boundary processes (Häyhä et al. 2016).
2.4 PB-based LCA - where are we now?
Life Cycle Assessment (LCA) is a methodology for assessing impacts along a supply chain through the quantification of environmental impacts of a product’s entire life cycle. Each process over the life cycle, from extraction to disposal, create some type of pressure that reflect in emissions and resources use, which are expressed in scores of environmental impacts. These go to form impacts categories through characterization, which can relate to the state of the environment through normalization.
Characterization is a calculation on how a substance released in or taken away from the environment contribute to influence a specific category of impacts, like the magnitude that different GHGs have in impacting a specific environmental issue the global warming potential. When the impacts results are related to an existing state of the environment or target that is wanted to be achieved the result are normalised.
LCA environmental impacts indicators generally do not include a sustainable reference value like an ecological carrying capacity or a local or global boundary, but LCA environmental impacts indicators can be modified to refer to a sustainable level. This can be calculated either before or after environmental impact indicators but needs to be able to capture scale, spatial variation and to be expressed in metrics similar to those of LCA indicators (Bjørn and Hauschild 2015). The UNEP SETAC initiative (Frischknecht and Jolliet 2016) suggest that options for characterization, normalization and/or weighting can be well integrated with concept of planetary boundaries and carrying capacity for a distance-to-target approach. Study at the European level (Sala et al. 2016)
have also pointed out that this approach can gain relevant information regarding the magnitude of environmental impacts comparing to a reference state. This can both facilitate communication of results and effectively monitor progress towards sustainable consumption and production. In LCA, when comparing the sustainability of functionally equivalent products or service, indicators of performance are compared, and the product or the service with the lowest environmental impact is considered sustainable. This approach is considered relative environmental sustainability because it is the sustainability in relation to different alternatives. From relative sustainability perspective, a product or service appears sustainable if it is compared with worse alternatives. This is because it focuses on minimising environmental impacts without addressing if the product or service life cycle reflects also sustainable consumption and production patterns (Bjørn and Hauschild 2013).
TABLE 1 summarises the main studies that have attempted to integrate absolute environmental sustainability references, including the planetary boundaries, into the LCA framework. Heijungs, De Koning, and Guinée (2014) were the first ones discuss connecting the PB framework in the LCA.
The authors saw PB as theoretical reference for absolute environmental sustainability and that production and consumption should take place by minimising the limited resources available and emission possible. Great advancement was done by Bjørn and Hauschild (2015) with a carrying capacity-based normalization reference to be applied in the LCA. Carrying capacity was defined as
“the maximum sustained environmental intervention a natural system can withstand without experiencing negative changes in structure or functioning that are difficult or impossible to revert”
(Bjørn and Hauschild 2015). Björn et al. (2016) continued on the way of applying carrying capacity on LCA further advancing characterization factors, defining thresholds and control variables. Ryberg et al. (2016) study was the first to identify and discuss development and operationalization of a PB- based LCA. Ryberg et al. (2016) pointed out theoretical and technical challenges for a PB-based LCA. However, the study by Sandin, Peters, and Svanström (2015) was the first that used the PB framework to set targets for impact reduction at the product scale, and that applied the PB framework to an LCA case study. The most comprehensive study on global environmental impacts using PB in LCA as normalization factors was by Sala et al. (2016). Normalisation factors are used in LCA to estimate the significance of an impact associated to a product or a system in relation to a reference value (being a region, a country or the entire globe). The study explores existing, but also estimate new sets of normalization factors in the LCA context and assesses their feasibility in relation to PB.
Finally, the work done by Doka (2015; 2016) translate PB into a per-capita allowance, as the equitable annual allowance of environmental impact per person. This per-capita allowance of environmental burden, was included also in the study of Sala et al. (2016), and it is useful in applied case studies of
TABLE 1 Summary of studies done towards a PB-based LCA; studies are in order of advancement and the year of publication often does not correspond to the year of which the study, and their findings were available to others.
AUTHORS APPROACH CASE STUDY ADVANCEMENT CHALLEGES
Heijungs, De Koning, and Guinée (2014)
Maximising affluence (prosperity) within the planetary boundaries - Absolute sustainability reference based on PB PB are introduced as a reference impact that represent a target or a value, and consumption patterns are to check for best fit limits of environmental sustainability Backcasting LCA for exploring ways meet sustainability levels (PB) through adapted affluence (seen as consumption levels)
Input-output model for EU-27 (minimum consumption level represented by Bulgaria and maximum impact level represented by EU targets for 2020 and 2050 GHGs reductions)
PB is theoretically seen as reference for absolute environmental sustainability Maximising prosperity within the PB boundaries goes along with the concept of decoupling environmental degradation and economic growth
The backcasting approach works well when the environmental “allowance” in known
The study deals only with GHGs but in a multi-objective analysis many assumptions would not hold, and uncertainties would be several
The use of the IPAT equation applied gives a linear and static analysis, this is a greater issue because the study assumptions include static population, technological advancement, dematerialization, eco-efficiency
Bjørn and Hauschild (2015)
Carrying capacity-based normalization reference in LCA
Carrying capacity calculated from thresholds found in literature and expressed in metrics identical to midpoints impact indicators in LCA
The normalization reference expresses the carrying capacity of a reference region divided by its population and thus describe the annual person share of carrying capacity
Theoretical framework developed for normalization factors for Europe and the world to connect LCA impact indicators with carrying capacity references
Introduction and application of absolute environmental sustainability in a context Theoretical framework as PB with control variables used for normalization in LCA
Spatial generic approach - spatial variation increases relevance of the normalization reference for impacts that manifest at local and regional scales Uncertainty related to quantification of carrying capacity (may be overestimated in the study)
Location of the control variable in the impact pathway
Bjørn et al.
(2016)
Spatially differentiated characterization factors for terrestrial acidification
Carrying capacity as absolute environmental sustainability indicators Define of a control variable and threshold for acidification that describe structure and function of a natural system
Theoretical framework for LCA impact category terrestrial acidification
(worldwide model) Specific case study on US for terrestrial acidification
Strong realistic model for characterization in terrestrial acidification (carrying capacity for terrestrial acidification in a context)
Methodology for spatially resolved LCIA models with carrying capacity
Mathematical expression for calculating spatially resolved occupation of carrying capacity for any emissions based on LCA impact categories
New sets characterization factors must be calculated for each impact categories Advanced and time demanding methodology
Ryberg et al.
(2016)
Discuss operationalization of a PB-based LCA
Definition of theoretical and technical challenges for a PB-based LCA
The Holocene as a new area of protection when constructing a LCIA methodologies Spatial differentiation reflecting local or regional environmental heterogeneity when modelling impacts in LCIA
Identify match between PB control variables and damage indicators in the impact pathway expresses in the same metrics
Theoretical framework linking PB and LCA approaches
Applying the precautionary principle instead of best-estimate for defining areas of protection and environmental impacts
“allowance”
Further environmental constrains are introduced to LCA as the PB framework does not accept trade-offs between boundaries
The goal in LCA is expanded to seek the minimum total environmental impact without degrading the new area of protection
Need to ensure that all relevant environmental impacts are covered while avoid impact coverage to overlap Knowledge gap on the PB challenge to relate impacts categories to the new area of protection
Weighting impact “allowance” introduce the need for a value-based task
Precautionary approach in contrast with LCA as higher weight might implicitly be given to most uncertain boundaries although this might not reflect severity
Sandin, Peters, and Svanström (2015)
Using the PB framework to set targets for impact reduction at a product scale Identify the PB that relate to the impact categories of a product system
Interpret what the boundaries means for setting targets for the global market segment (global annual allowance) Application of ethical principles to further allocation - a value-based choice is used to allocate allowed impacts
Global market segment and product scale study LCA of five garments for modelling Swedish clothing consumption for several impact categories related to PB
Matching PB with appropriate impact categories
Methodology for environmental sustainability targets at the product scale Operationalise the PB framework in LCA in a context
Theorising and applying in practice the ethical dimension for allocation of environmental allowance
Assumptions on the value-based choice brings uncertainties
The allocation of ‘environmental allowance’ through ethical principles is highly normative and can be impractical
Sala et al.
(2016)
Assessing different sets of normalization factors for the global scale and EU-27 are compared and evaluated following different methodologies for assessing the level of environmental pressure and estimating impacts in LCA
Critical review on different PB-based normalization methodologies by exploring
EU and global consumption-based normalization factors that relate to PB
Comprehensive study on comparing existing and estimating new normalisation factors in relation to PB Normalization factors based on EU-27 basket of products, input/output approach and apparent consumption
Comparison of PB-based normalization factors also from other studies
Completeness of global inventories of emissions and resource use, as well as inventories for impact categories are affected by limited availability of data (e.g. many chemicals, their toxicity and their impact at different scales)
More robust quantification is needed to understand the extent to which PB are
Doka (2016) Combining LCI results with PB through the calculation of impact assessment method Calculation of a PB per-capita allowance as an equitable annual allowance of environmental burden per person
Theoretical framework focus on the PB per- capita allowance
Per capita limits calculation is useful for applied studies ‘planetary boundary allowance’
The aggregated score could be useful for traditional product LCA
No geographical or scale differentiation for environmental impacts
With the aggregated score, compensation of environmental burdens become mathematically possible, but this also imply that boundaries have equal importance
Highly uncertain and highly experimental
2.5 Circular economy – product and material circularity
Circular economy (CE) is the alternative archetype to the linear model, where sustainable developments are focused on minimising resource use and harmful emissions. It is conceived for implementing new production and consumption patterns that would help achieving sustainability and well-being for society at low or no environmental costs (Ghisellini, Cialani, and Ulgiati 2016; The Ellen MacArthur Foundation 2012). In many cases, the textile industry has made efforts to adapt to a circular model (Walter R. Stahel 2015), with the intention of lowering their environmental impact, which in turn impact human well-being. The industry has, in fact, strong feedback mechanisms with the ecological system that could allow reduction of pressure on PB and so there are several opportunities for a sustainable textile system through the circular transition (The Ellen Mac Arthur Foundation 2017).
Geissdoerfer et al. (2017) and Sauvé, Bernard, and Sloan (2016) argue that the framework is narrower and more effective in promoting business sustainability, as it has more distinctive path ways for its implementation. The circular model promotes the reducing, reusing, refurbishing and remanufacturing of materials increasing the efficiency of the resources throughput (The Ellen MacArthur Foundation 2012). This in turn provides shelter to business from price fluctuation and volatility (Lacy et al. 2014) and create virtuous cycles in the system that benefit the earth’s system too. At all phases across a supply chain, it is encouraged to use as fewer resources as possible. This lower the impact on ecosystem through keeping resources in circulation for as long as possible, which reduce the need of new resources and emissions. Large-scale of reuse and recycling of material, for example, could help reduce CO2 emissions as well as landscape and habitat degradation, which will in turn could help limit loss of biodiversity (European Parlimentary Research Service 2016).
Kirchherr, Reike, and Hekkert (2017) define 9 strategies for product and material circularity for the transition from a linear to a circular system. Strategies are grouped based on their focus area of improvement from smarter application of materials to extend material life that works on closing the loop while from extend product life to smarter product use that works on slowing the loop. FIGURE 3 below illustrate the 9 strategies also applied to the textile system under investigation in this study.
FIGURE 3 The 9R circular economy framework for garments from linear system to a circular system transition. The R are grouped based on their focus area of improvement from smarter application of materials to extend material life that works on closing the loop while from extend product life to smarter product use that works on slowing the loop. Adapted from (Potting et al. 2017).
3 Case study and background information
3.1 Cotton production, manufacturing, use and disposal for the EU consumers
For this study, an average supply chain of a cotton t-shirt has been considered where the final consumption takes place in the European Union (EU). Cotton is widely utilized as natural fiber for clothing in the EU, and globally over 82% of natural fibre consumption is cotton (FAO/ICAC 2013).
FIGURE 4 shows the share of producers, manufacturers and consumers countries of cotton globally.
Cotton garments imported in the EU are usually produced and manufactured elsewhere. Cotton production takes mainly place in India, China, the US and Pakistan while manufacturing takes mainly place in China, India, Pakistan and other South-Eastern Asian countries (OECD and FAO 2015). So, EU’s consumption is at the top among other countries despite producing and manufacturing such a low share of the world cotton.
FIGURE 4 Producers, manufacturers and consumers countries of cotton (USDA 2018; WTO 2017) (raw data in ANNEX II - Case study description and background).
Environmental impacts of the cotton supply chain are therefore spread over a multitude of countries as it is evident that the life cycle of most of cotton garments consumed in EU develop along a global supply chain. So, one garments can be responsible of emissions and resource use that spread from one side to the other of the globe, and affecting local, regional and global state of the environment.
During the cradle-to-gate path so from cultivation of cotton, through the manufacturing of the cotton material, till the final garment is finished, there is more than half amount of cotton material lost to the weight of the actual final cotton t-shirt (FIGURE 5). During the yarn production a large amount of cotton is diverted to other use as considered not valid for yarn and fabric making, the finishing processing and garment making are also phases where cotton in form of processing product does not end up as final product but become either some sort of by-product or waste (Jewell 2017).
FIGURE 5 Material loss in kg during the manufacturing of a cotton t-shirt from cotton fiber to final garment based on data from Jewell (2017).
Disposal after usage of garment also creates a great amount of inefficient use of material and products.
The gate-to-grave path for clothing in the EU (FIGURE 6) involves more than half of the total amount of cotton t-shirts being disposed to landfill. About 25% of the total amount of clothes disposed is incinerated mostly with energy recovery. About 20% of the total garments in EU is collected and sorted of which 80% is used for recycling fiber or reuse of garment. The recycling also includes down-cycling and the reuse mostly involve sending clothes to developing countries rather than be reused in EU (Beton et al. 2014).
FIGURE 6 End-of-life for garments in the EU, based on data from Beton et al. (2014).
3.2 Scenario analysis
Scenarios for this study, were chosen taking into consideration the actual path of textile in EU (FIGURE 6 described previously), while also trying to compare scenarios that differ the most in order to better remark environmental impact differences. For the textile industry, the idea was to oppose more circular scenarios represented by extended life span and reuse of product, to more linear scenarios represented by the most common cradle-to-grave approach. The scenario matrix below (FIGURE 7) is been built following the principles of the R strategies of the circular economy framework described in the FIGURE 3. The scenarios chosen are placed in the matrix in relation to the variables for the development of a circular system. This helps to understand what type of strategy they most represent.
FIGURE 7 Scenario matrix, defining scenarios from cradle-to-grave to cradle-to-cradle, the linear system is the starting point while the circular system is the arriving point that is achievable by closing and slowing the loop of flows of material and product. Several combinations are possible as each R strategy (R9 recover, R3 reuse, R1 rethink) can be placed in this scenario graph based on their level of circularity. The boxes in red represent the scenarios considered for this study.
The scenario building is therefore the result of an analysis of the sector, in particular in its end-of-life phases, but also based on what is often advocate by the circular framework for the textile industry (The Ellen Mac Arthur Foundation 2017). Accordingly, the different scenarios are defined as following:
LANDFILL – Baseline scenario where the t-shirt is after the use phase simply discarded and goes either into landfill. This scenario is characterised by a linear system flows of simply take-make- disposal and is the most common in scenario for textile in EU.
INCINERATION - Baseline scenario, which is also an actual textile product pathway in the EU. It can involve energy recovering, which is recognised as the last strategy for circularity by closing a small loop at the end of the supply chain. Despite of this, this scenario is far away from other strategies towards circularity and it is close to a cradle-to-grave approach.
REUSE RoW - Reuse in the rest of the world. The difference with the scenario above is the greater distance that the garment has to travel before being used again. This scenario is being considered mainly because most of garments in EU that are intended for reuse, are going to be reused outside the EU and so this scenario is more realistic.
REUSE EU - Reuse within the EU scenario. After use, the garment is collected, sorted and redistributed for donation or for being sold in another European country. The product is reintroduced into the system, and so decrease the need for the production of a new garment, which will involve new emissions and resource use along the supply chain. This scenario in the matrix lean more towards closing the loop rather than working on a leverage of slowing consumption.
LONG-LASTING DESIGN - Long-lasting design (LLD) is often advocated by circular economy frameworks (The Ellen Mac Arthur Foundation 2017). It represents a situation where a product is designed to last longer and/or there is a changing in consumption behaviour towards the refuse of getting a new product by satisfying the same function through a more intense use of the existing product. The scenario also follows a cradle-to-grave path like the previous one at the end, but since the product utilisation are extended, there would be more efficiency into the resource used and emission that relate to it, that is also why the scenario in the matrix lean more towards slowing the loop rather than closing the loop.
4 Methodology
Life Cycle Assessment (LCA) was employed to quantitatively evaluate the environmental performance of a cotton t-shirt throughout its life cycle, i.e., from cotton cultivation to manufacturing of the garment, to the distribution, use and eventually the disposal. The evaluation was done for environmental impacts of scenarios with the different path for reintegration and/or longer usage of the product into the system. Life Cycle Assessment is a standardised tool that quantifies the overall resources required and emissions associated to a product or a service allowing to identify activities that impact the most the ecosystem (Finkbeiner et al. 2006).
The LCA is carried out in an iterative manner following four steps (FIGURE 8):
1. Definition of goal and scope, which includes defining the main elements of the LCA like system boundaries, impact method and functional unit of the study.
2. Life cycle inventory (LCI) is a comprehensive list of relevant flows from and to the environment that relate to each process of the product’s system. This includes all resources used, in form of material, energy, water and land area, as well as all emissions to soil, water and air.
3. Life cycle impact assessment (LCIA) is where the flows from the LCI are related to potential environmental impacts such as climate change, eutrophication, acidification, toxicity and resources depletion as common examples. Classification, characterization and, weighting or normalization are activities included in this step.
4. Interpretation is the final step and involves drawing conclusions from the results of the LCI and the LCIA, including identification of significant issues and evaluation of the study also in term of its limitations.
FIGURE 8 Phases of LCA based on Finkbeiner et al. (2006). It is an iterative method where each phase is interdependent as will inform how the other will be completed, the double ended arrow represents this.
GaBi software (http://www.gabi-software.com) was employed to model the product’s system and to simulate scenarios of the system from a life cycle perspective. Some pre-existing aggregated and single processes came from databases associated with GaBi, while other processes were created based on literatures, and some other based on assumptions. The program allowed also to create new processes, to combine and to model them together in order to customised models of a product system.
4.1 Goal and scope of the LCA
The goal of the LCA in this study is to assess the environmental impacts and benefits of different scenarios of product in a textile supply chain in relation to the PB framework. Therefore, different types of scenarios were taken into account opposing circularity with non-circularity. Comparison between scenarios was done to estimate how closing the loop at the product level affect the supply chain emissions and use of resources in relation to PB. The intent is to determine if increasing circularity of a product level, through reuse or longer use, would result in a net environmental benefit, and if of what type. Information deriving from the analysis informs the potential environmental benefits of product reuse in a textile supply chain from a consumer perspective but in relation to global environmental challenges represented by the PB framework.
4.2 Functional unit
In LCA, the functional unit quantifies the function of the studied system and it provides a reference to which inputs and outputs can be related to. For comparative LCA in particular, the focus is on the function of the product, but it also needs to represent the physical unit of the function of the product system (Finkbeiner et al. 2006). This study investigates the environmental performance of circular scenarios and compares them to more linear scenarios through a functional unit that relates to the performance assessed by a product, one cotton t-shirt. This comprises a qualitative and quantitative description that makes the product’s service life the functional unit. One year of service life of a cotton t-shirt with 50 washing was the functional unit of the LCA. Important to take into account, is the amount of cotton used (in all its forms) to deliver one cotton t-shirt by each scenario.
The type of shirt is plain without any prints or pattern made by 100% conventional grown cotton, and weight on average 0.220 kg as finished garment. It is assumed that the user wears the t-shirt once or twice a week before washing and that the t-shirt completely finishes its service life in one year (after 50 washes). Although, it is more likely that a person uses 5 t-shirts over a period of 5 years, in this way annual impacts can better relate to the product. This is based on the study in the EU, done by Beton et al. (2014). Finally, no assumptions are made on the actual quality of the garment by the end of its service life, beyond the statistics about the number of reusable garments collected and if a garment is reusable, it has maintained a decent quality level.
4.3 System boundaries and expansion
A consumption-based approach is employed for this study, in order to be able to comprehensively take into account the total resource use and emission derived from the consumption of a cotton t-shirt, regardless of where and by whom this was produced, manufactured, used and disposed. Resource use and emissions, related to textile products of a global supply chain, are often fragmented across actors and countries, but the approach used by this study incorporates them all through looking into the entire life cycle of the product.
The LCA have cradle-to-grave baseline scenarios, spanning from raw material extraction to end-of- life treatment, and cradle-to-cradle scenarios that involve the reintroduction of the product into the supply chain through reuse. Life cycle of a cotton t-shirt includes fibre production, textile and garment manufacturing, distribution, use and end-of-life handling.
The geographical boundaries for use, disposal, collecting and sorting are limited to the EU, but for expanding the system to the all supply chain no geographical boundaries are considered. This allow to include agricultural production of the raw material, and all processes of manufacturing and distribution. In fact, many processes used for the building of the scenarios modelling in the LCA
software uses country or region’s specific data or global averages. This is discussed in the data section (4.5). TABLE 2 summarises what processes and assets are included and not in the building of the model with the LCA software.
TABLE 2 A summary of what is included in the study and what is not.
WITHIN SYSTEM BOUNDARY OUTSIDE SYSTEM BOUNDARY
• Extraction of resources
• Manufacture of materials, including fuels and chemicals used
• Electricity generation
• Garment sorting
• Transport from producing countries to manufacturing countries, from manufacturing countries to consumer countries, and after use transport to disposal or sorting facilities.
• Capital equipment, building and maintenance
• Packaging of the product for transport and sale
• Transport from growing site to ginning site
• Transport from manufacturing of fabric to cut and sew site
• Transport from seller to the consumer
• Softener of other product for textile care used during the usage phase
• Possible washing and drying in sorting facility for the reuse of textile
• Retails and warehouses energy consumption
4.4 System processes and flows
The pathway of a cotton t-shirt consumed in the EU under consideration for this study is described in the FIGURE 9 below. Process of cotton production are farming and ginning, t-shirt manufacturing starts with yarn production using the cotton fiber bales from cotton production, fabric production using the yarn (this study considers a knitted fabric), and so follow the finishing processes like dyeing and others, and finally the cut and sewing for the making of the final garment. The finished product is then distributed in the selling country (EU), used and disposed. This flow of material and product is the classical linear system flow is the most common for the EU case.
Circular linear flows, instead reintroduce the product or material to the linear supply chain closing the loop and so diminishing the need for new material or product. Instead of disposing the product after use, the t-shirt is collected, and sorted in what can be recycled or what can be reused and depending on this on which step of the supply chain the still valuable material or product is reintroduced.
FIGURE 9 Graphical representation of possible processes of material and product flows involved in the textiles system of a cotton t-shirt.
4.5 Data quality and source
The goal and the scope of the study create implication for data collection, influencing choices regarding system boundary, processes involved and environmental parameters to investigate, but most importantly was the functional unit that also greatly affected data collection.
The main implications deriving from these choices were, foremost, that many choices had to be country specific or account for the country where the process was taking place the most, so many data were based on general sector data or country averages where processes were taking place. Since, the system boundary is expanded to represent a global supply chain, the result is that for some
FARMING
GINNING
USAGE YARN PRODUCTION
DISPOSAL RETAIL GARMENT MAKING
SORTING OF GARMENTS
COLLECTION OF GARMENTS MECHANICAL RECYCLING
TRANSPORT (2) TRANSPORT (1)
WAREHOUSE FABRIC PRODUCTION
FINISHING PROCESSES
COTTON PRODUCTIONGARMENT MANUFACTURINGDISTRIBUTION
TRANSPORT from producer countries to manufacturer countries
TRANSPORT from manufacturer countries to user countries (EU)
Disposal as textile waste for landfill or incineration
RECYCLING
REPAIR REUSE
EXTENDED USE LINEAR SYSTEM FLOWS
CIRCULAR SYSTEM FLOWS
processes what is taken into account is the place where the process take place for the majority (e.g.
manufacturing takes place the most in China for EU import of textiles). However, data collection for textiles process have a high degree of generalization and representativeness. This means, also that processes commonly occurring for other textiles are plausibly used for this case study as there is no reason that suggest otherwise (e.g. the cotton farming process for pair jeans is the same as for a t-shirt and so on).
LCA data are typically non-specific and based on averages so environmental impacts are not actual measurements of impacts but rather potential impacts based on models. It is important to take this into account when discussing conclusion, because models built with LCA are not real specific system but a representation of it made on averages mostly. This is typical for attributional LCA for evaluating the average impact of the system per functional unit. The attributional approach allowed for this study to best isolate the investigated product’s system and better compared different scenarios. Following what discussed, it is understandable why the LCI was constructed with a mixture of generic data and some other more specifically tailored data for the specific case, both deriving from LCA databases and literatures. TABLE 3 summarises the data used source and type.
TABLE 3 Summary of data used source and type.
PHASES DATASET SOURCE COMMENTS
Agriculture production
Seeds to bales (including ginning)
GaBi database The data derive from a multi-output economic allocation approach as two valuable co-products (cotton fiber and seed)
Transport (1) Growing countries to manufacturing countries
(Beton et al. 2014; Jewell 2017; Wiese, Toporowski, and Zielke 2012)
Assumptions were made based on the literature see TABLE 11
Manufacturing Fiber to yarn GaBi database The data derive from a multi-output allocation approach as some by- product of manufacturing were considered too valuable to be excluded
Yarn to knit greige fabric GaBi database Knit greige fabric to finish
fabric
GaBi database
Garment making (cut and sew process)
GaBi database and Jewell (2017)
Process created using flows from the database and quantities from literature
Transport (2) Manufacturing countries to EU retails
(Beton et al. 2014; Schmidt et al. 2016; Wiese,
Assumptions were made based on the literature see TABLE 12
Usage Usage of garment in EU28 Usage (Beton et al. 2014) Washing powder impacts results (Medina et al. 2015)
Assumption were made on usage only based on the literature see TABLE 16 and TABLE 17 Reuse EU Collecting (Schmidt et al. 2016) Assumptions were made based on
the literature see TABLE 15 Sorting
Reuse RoW Collecting Sorting Transport after
disposal
Transport after disposal in each scenario
GaBi database (Schmidt et al. 2016)
Assumption were made based on the literature see TABLE 13
Incineration of textiles
EU-28 incineration of textiles GaBi database Data based on averages of waste incineration in EU28
Textiles to landfill EU-28 textiles on landfill GaBi database Data based on averages of waste to landfill in EU28
4.6 Data gap and limitation
Data search revealed scarcity in some data regarding some processes, but also, the fact that some data came in an aggregate form resulted in a limitation on the analysis. This also limited the possibility of exploring, more empirically, additional product and material circularity like mechanical recycling.
Several assumptions were, therefore, necessary in order to build the product’s system model on GaBi (ANNEX III -
Data source and assumptions).
Moreover, some flows between the product’s system and the ecosphere are not always quantifiable and often excluded, however, those can be described qualitatively. What was excluded, instead were flows that that have a high level of uncertainty if assumed. On the other hand, there were cases where no info was available and so legitimate proxies based on literature were used as best guesses.
4.7 Life Cycle Impact Assessment
During the life cycle impact assessment (LCIA), flows from the LCI are converted into impact categories that reflect potential impacts on human, environmental and resource scarcity, in form of environmental impact scores (Huijbregts et al. 2016). The three main activities form this step:
classification, characterization and weighting.
• Classification involves linking specific substances from the inventory analysis to specific environmental impacts that are represented by specific LCIA indicators.
• Characterization involves calculating results for the impacts categories using a conversion factor, basically how each substance in the LCI contribute to LCIA indicators. Basically, it considers impacts per unit of stressor.
• Weighting and normalization are optional activities. The first is used to determine importance of impacts in relation to a reference point, while the second allows to group impacts categories in endpoints results.
The ISO 14044 standard defines normalization as the “calculation of the magnitude of the impact indicator results relative to reference information” and weighting as the “conversion and possible aggregation of indicator results across impact categories using numerical factors based on value- choices” (Finkbeiner et al. 2006). As part of the LCIA phase, normalization and weighting are both considered as optional steps. Normalization step allows to relate impacts of a specific supply chain to reference environmental values associated to the system, that could be regional, national, continental or global system that the supply chain relate to throughout its processes.
In the next section, the LCIA framework used in this study is explained in more details.
4.7.1 The ReCiPe framework
The LCIA methodology chosen for this study is ReCiPe as it has a broad set of impact categories, 18 midpoints and 3 endpoints indicators (ANNEX III -
ReCiPe framework). This method was first developed by Goedkoop et al. (2009) and updated by M.
a. J. Huijbregts et al. (2016), and it provides characterization factors that relate mainly to the global scale. This is of high relevance for investigating a textile industry with a supply chain that span globally. Furthermore, environmental mechanisms included in the inventory are effectively modelled with ReCiPe LCIA results.
The ReCiPe framework uses two ways of deriving characterization factors, one at midpoint level located along the impact pathway, and one at endpoint level which correspond to the area of protection (ecosystem quality, resource scarcity and human health). The two ways are complementary, midpoints characterization has closer relation to the environmental flows in the LCI and so have relatively lower uncertainty, while endpoints characterization is more uncertain but informs more the environmental significance of the flows in the LCI (Goedkoop et al. 2009).
4.7.2 Further LCIA development for this study
In this study, state-of-the-art knowledge in PB-based LCA is applied to the case study. The PB framework is accepted as present (Steffen et al. 2015) despite the ongoing discussion on variables
the framework in TABLE 9). However, knowledge gaps understandably are going to affect the discussion of results.
The focus here is on bridging the global scale of the state of the Earth’s system and impacts that are affecting it, with the lower scales of the drivers. Taking on from the work done on the building of a PB-based LCA, described in the Theoretical Background section (2.4), weighting and normalization factors based on PBs (Doka 2016; Sala et al. 2016) have been applied to the results of the LCIA, in order to understand how, per functional unit, different scenarios influence the increase or decrease of impacts in relation to the state of the Earth’s system expressed by the PB.
Firstly, the Holocene state is being considered as the focus area of protection, based on that, the casual chain of the impact pathway is being built showing the connections of different environmental drivers, pressure, state and impacts in the ReCiPe framework (Huijbregts et al. 2016). This was important to find the points where the PB’s control variables can be measured. Processes involved in the impact pathway are divided into inventory, environmental mechanisms, midpoint and endpoint impact based on their location in the LCA process, i.e., before or after classification, characterization, weighting and normalization. Inventories include all pressures from the cotton textile system under investigation through the LCA. FIGURE 10 shows the causal chain of connections with the processes considered setting the Holocene state as area of protection, and the processes highlighted in red as the points where the control variables for the PB are measured.
Secondly, a method for downscaling the boundaries was needed in order to relate to the scale of a single consumer. A per-capita allowance boundary was used in relation to the relevant impact that the system taken into account have. Doka's (2015; 2016) work on PB allowance per person is being used for this task. The challenge was mainly to match ReCiPe LCIA results with what the boundary variable actually measures. Hence, the previous step of understanding where the boundary’s variable is located in the impact pathway was fundamental. It was important to make sure that variable and LCA results at different levels where measuring the same thing with the same unit. TABLE 4
summarise for each boundary how this is being assessed (Doka 2015; 2016) and how the method is compatible with ReCiPe LCIA results. It is important to point out also that, when PB is converted into per-capita allowance is being assumed a human population of 10 thousand million. This was considering population growth forecast in order to avoid continuous adjustment as population growth and for maintain a precautionary approach and so give time to policy maker to act (Doka 2015; 2016).
FIGURE 10 The causal chain of connections of different environmental drivers, states and impacts in the ReCiPe framework (Huijbregts et al. 2016) with focus on processes that involve the PB framework and so the Holocene state as area of protection. Processes are divided in inventory, environmental mechanism, midpoint and endpoint impact based on their
