1 Introduction

1.1 Global guidance on LCIA indicators

Phase 3 of the United Nations Environment Programme (UNEP)-Society of Environmental Toxicology and Chemistry (SETAC) Life Cycle Initiative (2012–2017) has launched a flagship project to provide global guidance and build consensus on environmental life cycle impact assessment (LCIA) indicators (see http://www.lifecycleinitiative.org/ under the Phase 3 activities for the full list of projects). Following initial scoping workshops in 2012 and 2013 and a stakeholder consultation, the effort has been focused in a first stage on impacts of global warming, fine particulate matter, water use impacts and land use impacts on biodiversity (Jolliet et al. 2014), plus cross-cutting issues and LCA-based footprints. A first Pellston workshop™ is planned for January 2016 to issue recommendations for these categories.Footnote 1 In a second stage, the project will address human toxicity, acidification, eutrophication, ecotoxicity and energy resources. It will also provide recommendations on how to integrate these individual indicators in a consistent framework, ensuring consistency of indicator selection process and assessment across impact categories. The deliverables will include a global guidance publication with a supporting web-based system that contains the finally selected LCA-based environmental impact category indicators and characterisation factors (for various regions), which will be available for viewing and download.

1.2 Objective of the present paper and outline

The aim of the present paper is twofold: (a) report the process and progress made and the preliminary results obtained in each impact category (see Sect. 2) and (b) present and document a worldwide rice LCA case study to be used to test and compare LCIA indicators across impact categories (see Sect. 3). Section 4 contains first conclusions from the work of the task forces and the rice LCA case study.

2 Progress and achievements per task force and impact category

2.1 Introduction

This section contains a description of the goals of the various task forces, the process and progress made since the start of this flagship project as well as first results achieved so far. First, the flagship project as a whole is described, followed by the task forces global warming, fine particulate matter, biodiversity impacts of land use, impacts of water use, cross-cutting issues and, finally, LCA-based footprint indicators.

2.2 General guidance and coordination (Rolf Frischknecht, Olivier Jolliet)

Goal

The goal of this flagship project is to run a global process aiming at global guidance and consensus building on a limited number of environmental indicators. The global process involves more than 100 world leading environmental and LCA scientists, organised in impact category specific task forces (TFs) and complemented by a cross-cutting issue TF to develop scientifically robust indicators suitable for a global consensus. The first phase of the consensus-finding process will end with a Pellston workshop™, a 1-week workshop that will take place in Valencia, Spain, from 24 to 29 January 2016, where invited experts and stakeholders aim to agree on recommended environmental indicators.

Process and progress

Following the paper presenting and scoping the LCIA guidance flagship project (Jolliet et al. 2014), coordination and continuous project monitoring was carried out involving all TF co-chairs by monthly teleconferences and standard forms enabling each TF to report technical content, timeline and memberships. The achievements made by each TF are described in more details in each subsection of the present paper. The first milestones in 2014 and 2015 have been the set-up of multiple LCIA guidance workshops in Basel (15–16 May 2014), Adelaide (15 September 2014) and Barcelona (7–8 May 2015). These workshops enabled each TF to get feedback from multiple stakeholders, to address specific critical cross cutting issues and key points for guidance on life cycle-based footprints.

Results

The Pellston workshop™ proposal as well as the steering committee composition has been approved by SETAC and by the International Life Cycle Board of the UNEP-SETAC Life Cycle Initiative. Guidance on Life Cycle-based footprinting will remain part of the flagship project but treated separate from the Pellston workshop™ that will focus on the LCIA indicator recommendations. The steering committee convened in a meeting at SETAC Barcelona for the members attending the congress, as well as via three teleconferences on 19.6, 23.6 and 9.7. Main outcome of the steering committee activity is an initial list of 48 foreseen participants to the Pellston workshop™ coming from 15 different countries, ensuring representation of the different continents, sectors and TFs. Finally, the development of a cross TF rice LCA case study has been finalised in collaboration with the particulate matter (PM) TF to facilitate impact category-specific model comparisons and testing. This rice LCA case study is further detailed in Sect. 3 of this paper.

2.3 Global warming (Annie Levasseur, Francesco Cherubini)

Goal

Global warming is considered as having a high environmental relevance among life cycle impact assessment (LCIA) categories. Global warming potential (GWP), i.e. the cumulative radiative forcing caused by a greenhouse gas (GHG) emission over an arbitrary time horizon (TH, e.g. 20, 100 or 500 years) relative to that of CO2, is the most common metric used in LCA to group the climate impacts from GHGs into CO2 equivalents, a mid-point indicator. As global warming causes impacts on both human health and biodiversity, some LCIA methods also provide end-point indicators to assess those impacts.

The goals of the global warming TF are to propose guidelines for the development of characterisation factors for the global warming impact category and to identify research needs in this area. Building on the Chapter 8 of the Intergovernmental Panel on Climate Change (IPCC) 5th Assessment Report for Working Group I (Myhre et al. 2013), a systematic analysis of available emission metrics and characterisation methods for the various climate forcing agents is first performed. Discussion involves data sources, modelling conditions and temporal/geographical scope among others.

Process and progress

The global warming TF is composed of 13 scientists having complementary expertise in climate modelling, climate metrics and/or global warming impact assessment in LCA. Two subtask forces have been established to work on specific issues, using the 5th Assessment Report of IPCC (IPCC 2013) as the basic source of knowledge: (1) well-mixed GHGs and near-term climate forcers (NTCFs) and (2) biogeophysical climate forcings from land use and land cover changes.

The global warming TF focused on drafting peer-reviewed articles presenting the outputs of its activities. Preliminary recommendations are given in a paper entitled: “Updating life cycle impact assessment from climate science: midpoint indicators”.

Chapter 8 of the last IPCC report states that “the most appropriate metric will depend on which aspects of global warming are most important to a particular application, and different climate policy goals may lead to different conclusions about what is the most suitable metric with which to implement that policy” (Myhre et al. 2013). LCA traditionally aggregates emissions of well-mixed GHGs to so-called “CO2 equivalents” using GWP100 as the default emission metric. GWP was introduced by the first IPCC assessment report in 1990 with illustrative purposes and does not embed any climate system responses or direct link to policy goals (Shine 2009). GWP100 was perceived to align well with the general principles of LCA, which privileges impacts integrated over time and space. An alternative to GWP shown in the 5th IPCC Assessment Report is the global temperature change potential (GTP), which is defined as the impact of a GHG emission pulse on global temperature at the chosen TH, again relative to CO2. Values of GTP for a TH of about 40 years are usually similar to those of GWP100. The choice between GWP100 and GTP100 can thus be seen as a preference to reduce global warming over the next few decades (GWP100) or in a longer period (GTP100). Another option is the integrated GTP (iGTP), which is based on the integrated impact on global temperature of an emission pulse and yields characterisation factors similar to GWP. In general, aggregating species with significant different atmospheric lifetimes to common units is challenging (Myhre et al. 2013). In these cases, the choice of the metric and the TH crucially matters for the importance given to forcers with a relatively short lifetime of up to a few decades, like CH4, compared to very long-lived forcers, like CO2.

Human activities perturb the climate system also through emissions of pollutants such as NOX, SOX, black carbon (BC) and others. These species have lifetimes shorter than the hemispheric mixing time and are usually called NTCFs. The atmospheric concentrations of NTCFs are very heterogeneous, and the resulting impact largely depends on the source region. The consideration of NTCFs in LCA should be cautious. Their impacts at local scale can differ from those averaged globally, some of them have cooling contributions (calling for a consistent treatment of negative characterisation factors across impact categories), and their impacts on climate have a confidence level that is lower than that for GHGs, especially in the cases in which aerosol-cloud interactions are important (Myhre et al. 2013). Changes in land use or land cover also cause impacts on climate. In addition to possible emissions of GHGs, changes in surface albedo (the ratio between reflected and incident solar radiation at the surface) have direct global climate implications, which are usually of opposite sign to those of CO2. Albedo changes can be converted to CO2 equivalents using either carbon equivalent factors or conventional emission metrics like the GWP or GTP (Caiazzo et al. 2014; Cherubini et al. 2013).

The Global Warming Task Force argues for an active consideration of these issues by the LCA community. The choice of one metric over another and the inclusion or exclusion of certain forcing agents can have large effects on the final LCA results, with the ultimate risks to identify mitigation options that are suboptimal, if not counterproductive.

Results

The TF concludes that the LCA methodology should gradually move towards the use of several sets of CFs based on different metrics and/or THs so that analysts can test the sensitivity of conclusions to such choices. A step-by-step approach is recommended consisting briefly in selecting a few metrics and THs, calculating impact scores and communicating about the meaning and impact of modelling choices if conclusions vary. Preliminary recommendations are as follows: (a) to use midpoint metrics which include climate-carbon feedbacks for CO2 and all well-mixed GHGs, (b) to quantify the impact including and excluding NTCFs to determine whether or not their inclusion changes conclusions and (c) provide uncertainty estimates with the characterisation factors.Footnote 2 In the TF paper, the pros and cons of each choice will be clearly presented and discussed so to ensure informed decisions by LCA practitioners.

During the plenary meeting in Barcelona, there were expressions of concerns for departing from cumulative metrics in LCA. In an LCA context, cumulative metrics like GWP or iGTP can ensure consistency across impact categories, while GTP is a special case as it is an instantaneous indicator. Future research can explore the feasibility within the LCA framework of better assessing the temporal and spatial heterogeneities of the various climate impact dynamics, for instance via time-distributed indicators (in terms of radiative forcing or temperature profiles) and subglobal impact categories (to be ideally addressed in combination with the regionalisation efforts in other environmental impact categories).

2.4 PM (Tom McKone, Peter Fantke)

Goal

Fine particulate matter (PM2.5) is one of the most important environmental factors contributing to the global human burden of disease. However, different models are currently used leading to inconsistent LCIA results reported for this category, thereby emphasizing the need for a clear guidance for practitioners. To address this need, a TF was initiated to build a health impact framework and factors for PM2.5 and its precursors. The TF will harmonise existing tools and methods and determine recommended factors for various source and exposure archetypes, incorporating human health effects from PM2.5 outdoor and indoor releases in LCIA.

Process and progress

The TF is subdivided into three subtask forces working on (1) indoor intake fractions, (2) outdoor intake fractions and (3) exposure-response functions, respectively. The common rice LCA case study and cross-cutting issues as well as a final framework for combining results and recommendations from all three groups are jointly assessed by all groups. The TF started with an expert workshop in Basel, Switzerland, in August 2013 to get initial guidance of how to address including health effects from PM2.5 exposures consistently into LCIA. The outcome of this initial guidance workshop was complemented by a literature review of the state of the knowledge in PM2.5 exposure assessment. Fantke et al. (2015) provide the workshop findings compiled into a set of preliminary recommendations for addressing PM2.5-related health effects in LCIA. A second workshop was held in Cincinnati in October 2014 to address cross-cutting issues between indoor intake fraction, outdoor intake fraction and exposure-response. TF progresses and results were, in addition, reported in several conferences at SETAC-Europe, SETAC North-America and International Society for Exposure Science (ISES) annual meetings.

Results

The indoor subgroup has reviewed and identified a set of factors (ventilation rate, room volume, occupancy, inhalation rate, time-activity patterns, particle penetration efficiency, deposition, filtration, phase changes and indoor chemistry and finally emission sources) influencing indoor inhalation intake of PM2.5 and an inventory of available approaches to be evaluated in a model comparison (Hodas et al. 2015). Indoor models are currently being compared for different archetypes (buildings with only mechanical air exchange, closed buildings with natural ventilation, open buildings with high-ventilation rate in tropical area and one-path vs. recirculated ventilation) and integrated with outdoor intake into a consistent and comprehensive framework to assess impacts of PM2.5 exposure indoors and outdoors. The exposure-response group performed the analysis of potential exposure-response curves, building on the work of Apte et al. (2015), and the exposure-response curves used in the global burden of disease (Burnett et al. 2014). Three possible exposure-response options have been considered, illustrated by a comparison between USA and China: derivative at regional working point, slope from regional working point to zero or fixed slope based on reference situation. The next step is to analyse the consequences of each of these options on the estimates of impacts on human health caused by PM2.5 emissions as reported in the different scenarios of the rice LCA case study (see Sect. 3).

2.5 Biodiversity impacts of land use (Ottar Michelsen, Assumpció Antón, Llorenç Milà i Cazals)

Goal

An intensive transformation of natural areas over the last decades has caused the loss of important environmental assets. Land use and, particularly, land use change from natural habitats to cropland and other human-dominated land cover types, as well as fragmentation of habitats, are among the main drivers of biodiversity loss and degradation of a broad range of ecosystem services. Considering the manufacturing and production activities of human society as some of the main drivers to land use and land use change, there is the need to focus on the prediction of potential impacts, such as biodiversity loss. No clear consensus exists yet on the use of (a) specific impact indicator(s) to quantify land use impacts on biodiversity within LCA. To address this gap, a TF on the quantification of land use impacts on biodiversity within LCA was initiated. This TF aims at global guidance and consensus regarding indicators and methods for the assessment of biodiversity impacts from land use (and, if possible, land use change) in LCA.

Process and progress

The TF on impacts on biodiversity from land use has developed an updated list of criteria for the evaluation of existing methods, which has been applied to 30 methods. Following a start-up workshop to frame the evaluation criteria in Basel (15 May 2014), multiple expert workshops and stakeholder consultations took place in 2014 (San Francisco, USA, 7 October 2014: focus on importance of evaluation criteria and choice of indicators; Brussels, Belgium, 18–19 November 2014: focus on evaluation criteria and usability of methods; Sao Paulo, Brazil, 12 November 2014: Stakeholder consultation on indicators for biodiversity in LCA). The TF is currently collaborating with the cross-cutting issues TF on defining ecosystem quality.

Results

The review and evaluation task has highlighted the diversity of approaches towards characterising biodiversity change (Teixeira et al. 2015). Currently, taking a conservative view, only two (three) methods could be operational for practitioners of LCA at a global scale: de Baan et al.(2013) which is further developed in Chaudhary et al. (2015) and Coelho and Michelsen (2014), but they may not be well suitable to catch specificity of rice production as well as different intensiveness of agricultural practices. In addition, there are several doubts about regional representativeness of biodiversity damage (see rice LCA case study in Sect. 3). A number of other models could be adapted to LCA with relatively little effort to provide either local or regional characterisation factors on a global scale, but there will be challenges in extracting the meaningful data from these models and combining them in a meaningful way to develop consensus-based characterisation factors at both the local and the regional levels (Teixeira et al. 2015). The experts highlighted the importance of the procedure established by the Life Cycle Initiative, including the early engagement of stakeholders in the consensus-building process. Experts agree that LCA may go beyond inventory data for land use (LU)/land use change (LUC) and relate elementary flows to their respective impacts on biodiversity. Local biodiversity loss due to LU/LUC typically requires a level of spatial resolution and farm-specific knowledge for which other non-LCA tools may be more accurate. It was also agreed that a good LCA indicator for biodiversity necessarily has to include geographical location, several aspects that depict the state of ecosystems at that location and a measure of intensity. Species richness is a good starting point for assessing biodiversity loss, an indicator often used by LCIA developers. However, complementary metrics are suggested to be considered in modelling, such as habitat configuration, including fragmentation and vulnerability, and/or intensity-based indicators (NPP/HANPP).

2.6 Water use (Anne-Marie Boulay, Stephan Pfister)

Goal

Since the consensus framework defining impact assessment of water use in LCA was published in 2010 (Bayart et al. 2010), several impact assessment methods were developed, assessing impacts on human health, ecosystem quality or resource depletion. Water is also gaining importance and attention as an area of concern and impact category, and stakeholders are often asked to communicate on their water use impacts in a single and simple metric. Among LCA practitioners, there is a need for a harmonised set of indicators assessing impacts from water use, both at the midpoint and endpoint levels, including one simple scarcity/stress-based metric providing an alternative to the more complex damage-oriented methods. The TF therefore focuses on the development of such a consensual stress-based metric in addition to harmonised damage-oriented impact pathways for human health and ecosystem quality.

Process and progress

The water use TF (WULCA) presently has three active subtask forces. The stress subtask force has collected input from more than 50 experts via workshops held on three continents: in Zurich, Switzerland, San Francisco, USA, and Tsukuba, Japan (Boulay et al. 2014), followed by an additional expert meeting held in April 2015 via teleconference, to provide preliminary recommendations (Boulay et al. 2015), which were presented in May 2015 at the SETAC Conference in Barcelona. The human health subgroup held one expert workshop in Barcelona, with representatives of Food and Agricultural Organization (FAO), World Resource Institute (WRI) and scientists and consultants. The main modelling challenges were discussed and will guide preliminary decisions on the modelling of this impact pathway. The ecosystem quality subgroup started framework discussions and plans to recommend methods in the next phase.

Results

The pre-recommended method by the stress subtask force is called AWaRe. This water use midpoint indicator is based on the relative Available WAter Remaining in a watershed, after the demand of humans and aquatic ecosystems has been met. It is first calculated as the water availability minus the demand (humans and aquatic ecosystems) per unit area of the considered region (m3 m−2 year−1). This ratio is then inverted and hence represents the area “virtually occupied” to cover the additional water consumption in a sustainable manner. In a second step, the value is normalised by the world average value in order to provide the regionally available water remaining relative to the world consumption-weighted average of this value. The indicator ranges from 0.1 to 100, with a value of 1 corresponding to the world average, and a value of 10, for example, represents a region where there is ten times more area required to provide a cubic metre of water sustainably than the world average. Or, in other terms, there is ten times less water available on the same area compared to the world average. The indicator is calculated at the subwatershed level and monthly time step and then aggregated to country and annual resolution. This aggregation can be done in different ways to better represent either an agricultural use or a domestic/industrial use, based on the time and region of water use. Characterisation factors on annual and country level for agricultural and non-agricultural use are therefore provided separately, as it implies different spatiotemporal aggregation. This method quantifies the potential of water deprivation, to either humans or ecosystems, and serves in calculating a water scarcity footprint according to ISO 14046, clause 5.4.5. The group is still also testing an alternative indicator combining demand-to-availability and area-to-availability ratios in view of the final recommendation. These two indicators are currently undergoing a stakeholder consultation. For the subgroup human health, no pre-recommendations are available yet, but it is agreed that Motoshita et al. (2014) will be used as a starting point and focus on additional methodological improvements to make the method more robust as well as add a domestic user deprivation module. More complex approaches have been analysed and discussed (improved economic models, differentiating water source and quality) but do not seem yet mature enough for this global consensus (limited time for complex development and low quality of available data).

2.7 Cross-cutting issues (Francesca Verones)

Goal

In addition to providing recommendations for specific LCIA category indicators, it is important that individual indicators are consistent and integrated in a comprehensive framework. The cross-cutting issues TF therefore aims to continue research and development on the LCIA framework and to provide guidance on issues relevant for all impact categories such as spatial differentiation, uncertainty, etc. It will produce a guidance document on how to best reach consensus, ensuring consistency of indicator selection process and assessment approaches across impact categories.

Process and progress

The TF started its work in January 2015, with ten active subtasks defined (framework, human health, ecosystem quality, resources, spatial aspects, normalisation, uncertainties, temporal aspects, relationships between social LCA and biophysical indicators and the overall consensus finding). The whole TF has had monthly teleconferences, where progress is presented and feedback is given. In between the teleconference, each subtask, led by different people, has worked on the relevant questions and problems. One physical meeting took place on 7 May, 2015, following the SETAC conference in Barcelona.

Results

Preliminary recommendations have been agreed upon for several subtasks, and main lines of thoughts have been developed further at the Barcelona workshop. For human health, it was agreed upon to maintain disability-adjusted life years (DALY) as endpoint unit, acknowledging that this contains already a (internationally well accepted) weighting. For ecosystem quality, the preliminary recommendation is to have “species disappearance” as endpoint. This implies selecting a potentially disappeared/affected fraction of species (PDF/PAF)-based metric. There are, however, different PDF/PAF-based approaches, thus the importance to ensure compatibility. Other endpoint units are also an option, as long as a conversion to PDF/PAF is possible. The main line of thought is that the species extinction that we aim to capture at the endpoint level should be quantifiable for any region of the world and should also account in some way for the vulnerability of the respective ecosystems and species assemblages. How such a vulnerability indicator should look like is an object of further investigation.

In most topics, multiple research activities are ongoing, inside and outside of the TF, and it is important to build framework and recommendations in a compatible way that may incorporate future developments. We therefore need to distinguish immediate recommendations from future ones: It might, for example, be advantageous in the future to split “ecosystem quality” into a more anthropocentric “ecosystem services” and an “ecosystem well-being” (ecosystems and biodiversity in their own right) category. However, such a recommendation cannot be made at this stage, due to the present lack of operational methodologies.

2.8 Footprint indicators (Brad Ridoutt)

Goal

The proposed goal of this activity is (a) to form a common understanding of the differences between footprints and existing LCA impact category indicators, (b) to formalise a universal definition of an LCA-based footprint and (c) to achieve consensus about acceptable ways of calculating LCA-based footprints which address area of concern in parallel with area of protection; outputs are in the form of guidance documentation intended for developers of LCA-based footprints. The scope of the work is not to define specific footprint indicators or to define a specific list of footprint indicators. The scope is to provide global guidance to enable the development of LCA-based footprints which address any area of concern (AoC)—defined now or in the future according to end-user demand. AoC is an environmental topic defined primarily by the interests of stakeholders and the community.

Process and progress

This TF was established in March 2014, and a statement on goal and scope was drafted that same month. Multiple teleconferences, presentations and meetings took place in particular in Basel (15/16 May 2014) and Adelaide (15 September 2014), and a first seed document was circulated for which more than 100 comments were received. Consensus was reached in resolving the comments on the seed document during a face-to-face TF meeting in San Francisco (8 October 2014) and various teleconferences and has led to an initial scientific paper (Ridoutt et al. 2015b). Liaison is also ensured with ISO TC207/SC3/WG6 which is developing a new international standard on footprint communications (ISO 14026). Key elements of the UNEP-SETAC Life Cycle Initiative’s work have already been adopted in the ISO 14026 Working Draft. Following a further face-to-face TF meeting in Barcelona (8 May 2015), a second paper has been accepted for publication which introduces the universal footprint definition and related terminology and includes a discussion of some of the modelling implications (Ridoutt et al. 2015a).

Results

The initial work undertaken by this TF involved forming a consensual understanding of the difference between footprint indicators and existing LCA impact category indicators. Without differentiation, footprints serve no additional purpose in LCA. In short, footprints have a primary orientation toward society and non-technical stakeholders and report on only selected topics of societal concern. On the other hand, LCA impact category indicators have a primary orientation toward stakeholders interested in a comprehensive evaluation of environmental performance and trade-offs.

While LCA reports and footprints may serve different purposes, the TF work has underscored the importance of LCA inventory and impact assessment modelling as the basis for footprint calculations. To avoid confusion and contradictions, footprints should provide guidance for decision-making which is coherent with LCA results of equivalent scope. For example, a water footprint should deliver results which are consistent with LCA indicator results pertaining to water use. This issue highlights the importance of this TF’s work. Currently, there is a proliferation of footprints, many of which report results which are inconsistent with those of LCA studies, leading to contradictory messages in the marketplace (Pfister and Ridoutt 2013; Ridoutt and Huang 2012; Ridoutt et al. 2012), and a reluctance to apply footprinting and LCA methods in many public policy and business contexts.

So far, the TF has reached consensus on four defining attributes that should characterise all footprint indicators: environmental relevance, accurate terminology, directional consistency and transparent documentation, as described by Ridoutt et al. (2015b). Recognising the differences in purpose and orientation between footprints and LCA study reports, the TF is now working towards consensus on technical issues relating to footprints which may deviate from conventional practices in LCA. One such issue is double counting. In LCA, emphasis is placed on avoiding double counting between indicators when reporting a profile of impact category indicators. However, when reporting a profile of footprints, some forms of overlap of impacts may be permissible in order to make individual footprints comparable between products. For example, if a product reported two footprints, a water footprint and a toxicity footprint, some water-related toxicity impacts might be reported in both. The plenary meeting in Barcelona emphasised that the shares of footprint contributions in combined footprints must be identified and declared. With this measure, potential overlaps can easily be identified by the users of the information. The important consideration with footprints, with their orientation toward society and non-technical stakeholders, is that they should be able to be interpreted without consulting technical study reports. These and other matters are to be discussed in a forthcoming scientific article.

2.9 How to ensure practicality of harmonised LCIA indicators

The TFs working on a harmonisation of environmental LCIA indicators will apply several of them in the impact assessment phase of a common LCA case study. This case study should represent a common LCA as performed today and thus contain the key elements to take into account when recommending one particular indicator. The following section described the case study developed within this flagship project.

3 Rice case study to test life cycle impact category indicators

3.1 Introduction and overview

The rice LCA case study serves as a test application for the environmental life cycle impact category indicators developed and recommended by the TFs of this UNEP SETAC Life Cycle Initiative flagship project.

The case study is not meant to be fully representative for rice production and consumption in the regions covered but should serve as a test environment and ensure the applicability of the harmonised life cycle impact category indicators developed within this flagship project in common LCA studies and background LCA databases.

The rice LCA case study deals with the supply chain of cooked white rice from the agricultural field to the cooking pot covering three different regions in three scenarios:

  • Urban China: rice production and processing in rural China, distribution and cooking in urban China

  • Rural India: rice production, processing, distribution and cooking in rural India

  • US-Switzerland: rice production and processing in the USA, distribution and cooking in Switzerland.

Each scenario covers the four life cycle stages: farm production, processing, distribution and cooking as shown in Fig. 1. The functional unit (FU) is 1 kg of cooked white rice consumed at home as most frequently selected in LCA studies of rice production systems (Blengini and Busto 2009; Brodt et al. 2014; Fusi et al. 2014; Hokazono and Hayashi 2012). The complete case study inventories are explained in detail in the Electronic Supplementary Material.

Fig. 1
figure 1

Product system and system boundary of the rice production, processing, distribution and cooking for the three scenarios. Country flags indicate the differences among the scenarios

Full size image

3.2 Life cycle stages

Farm production

The “farm production” stage includes all field operations for rice production from cradle to farm gate. The use and production of fertilisers, pesticides and diesel for field management practices are incorporated in addition to their direct emissions. Furthermore, direct land occupation and the quantity of water withdrawn and evaporated have been accounted for. It is assumed that straw is incorporated into the soil (soil improvement). So, paddy rice is the only product and no allocation is needed (in accordance with Blengini and Busto 2009; Brodt et al. 2014; Fusi et al. 2014).

Processing

The “processing” stage includes the transportation of paddy rice from the farm to the next rice processing plant, dehusking and milling (processing) of rice and primary packaging of the refined rice (internal plastic bag and external cardboard box).

During the processing of raw rice, rice husks and rice bran are generated as co-products. Rice husks are commonly used as bedding materials, animal feed or energy source for rice grain drying, for parboiling and for electricity generation.Footnote 3 Rice bran serves as an excellent binder and ingredient for animal feed.Footnote 4 Table 1 shows the prices and volume fractions of the co-products of white rice. This information is used to apply an economic allocation, in which white rice receives an allocation factor of 0.94, rice bran 0.05 and rice husk 0.01.

Table 1 Economic allocation based on prices and volume fractions of the co-products of white rice
Full size table

Distribution

Distribution includes both the transport of the refined rice from the processing plant to the point of sale and the transport by the consumer from the point of sale to home. Rice is commonly stored in bulk containers with a payload of 21.6 t each. Therefore, a lorry with a carrying capacity of more than 32 t is required.

Cooking

Cooking includes country-specific cooking operations at home. One kilogram of cooked rice requires 0.7 kg of parboiled white rice, 1.1 (China and Switzerland) and 1.7 L (India) of water and 5 to 10 g of table salt. The table salt is disregarded in any of the three scenarios and not part of the product system analysed.

3.3 Scenarios

Introduction

The following subsections include a description of the three scenarios of the rice case study. Life cycle inventory data stem from the ecoinvent data v2.2, from the Agri-footprint database and from working and scientific papers (Jungbluth 1997; Singh et al. 2014). All data were harmonised as far as possible and embedded into the ecoinvent data v2.2. Table 2 summarises the key characteristics of the three scenarios.

Table 2 Key characteristics of the three scenarios
Full size table

Urban China

The province of Hunan, the top rice production province in China, is chosen as the rice-producing area. It exports most of its surplus to other provinces. Most of Hunan’s arable land is farmed using modern techniques, including mechanised irrigation and chemical fertilisation.Footnote 5 It is estimated that most of the province’s cultivable land is devoted to paddies (wet-rice fields), a great many of which are flooded by river systems during most of the year2. The rice yield stated in Table 2 (6450 kg/ha) equals China’s average rice yield, which is among the highest in AsiaFootnote 6 (Blonk Agri-footprint 2014). According to this source, about 27 g methane is emitted per kilogram of rice produced. The average amount of water needed for Chinese rice production includes irrigation water withdrawn from water wells and river water run-off (Chapagain and Hoekstra 2011).

Once raw rice is harvested, it is transported to Chenzhou, a city in the province of Hunan by a lorry complying with an emission standard equivalent to EURO III, which was nationwide implemented in 2007.Footnote 7 After dehusking and milling, white rice is packed and transported to a supermarket in Shanghai covering a distance of 1300 km (see Table 2).

In China, it is common to buy rice in 10-kg bags. Hence, for this scenario, it is assumed that one bag of rice of 10 kg is bought per shopping trip in a supermarket in Shanghai and carried home by public bus (6 km both ways). In China, rice is mostly cooked in electric rice cookers. Besides 0.7 kg of white rice, 1.1 l of water is added for cooking.

Rural India

The State of West Bengal is India’s top rice production state and produces almost 16 % of India’s total rice.Footnote 8 Hence, West Bengal is the rice production area selected for this scenario. The rice yield in India amounts to 3500 kg/ha (OECD/FAO 2014), which is a bit more than half of China’s average rice yield (see Table 2). Therefore, more land is occupied to produce 1 kg of rice, and hence, more diesel for agricultural machines is assumed to be used (larger areas to cultivate). The amount of water withdrawn per kilogram rice depends on the cultivation method and irrigation regime applied (Shantappa et al. 2014). According to Chapagain and Hoekstra (2011), the irrigation water withdrawn from groundwater and surface water per kilogram paddy rice sums up to 1.435 m3 in India. Furthermore, biogenic methane emissions play an important role in the paddy rice production. In India, 30 g methane is emitted per kilogram of rice produced.Footnote 9

The harvested rice is transported by lorry to the relatively small city of Contai, in the State of West Bengal. Since 2000, also India started adopting emission and fuel regulations for light- and heavy-duty vehicles. Bharat Stage III is required since 2010 nationwide.Footnote 10 Even though the Bharat standard does not seem to include smoke threshold levels, it is regarded as equivalent to EURO III6.

After processing, the rice is transported to Kolkata, the capital of the State of West Bengal, which is only a 150 km drive away from Contai. In contrast to the urban China scenario, the consumer in India is assumed to use a bicycle to go to a supermarket located 1.5 km away from home and buys a 10-kg bag of white rice (see Table 2).

In rural areas of West Bengal, 74 % use firewood for cooking while in urban areas, only 14 % use firewood (TERI 2013). They use open three-stone stoves for indoor cooking with firewood. Cooking with firewood has a low efficiency, generates a high amount of particulate matter (PM) emissions and uses about 50 % more water than when cooking with an electric cookerFootnote 11 (Singh et al. 2014).

USA-Switzerland

Arkansas records the highest rice production in the USA.Footnote 12 This is why the third scenario focuses on that area for rice production with an average US rice yield of 7452 kg/ha (Nemecek et al. 2007). According to this source, about 41 g methane is emitted per kilogram of rice produced. The processing plant is assumed to be in Little Rock, Arkansas. The US EPA and California Emission Standards for heavy-duty vehicles applicable from 2007 are tougher and thus set lower acceptable threshold levels (measured in grams per brake horsepower-hour) with regard to NOX, PM and HC emissions than the EURO VI emission standard.Footnote 13 Taking the average age of a US lorry into account, a lorry with emission standard EURO V is considered representative (see Table 2).

Once processed and packed, white rice is transported to the Mississippi State Port (670 km from Little Rock), where it is reloaded into a transoceanic freight ship and shipped to the Port of Rotterdam in the Netherlands (around 8000 km). In Rotterdam, the rice is loaded onto a lorry complying with EURO V to supply a supermarket in Geneva (930 km from Rotterdam). The EURO V emission standard was introduced in Europe in 2008.Footnote 14

Since a passenger car, complying with EURO 5 and fuelled with diesel, is used to go to a supermarket 5 km away from home, a larger volume of groceries is purchased each time. In Switzerland, it is common to buy rice in 1-kg bags. Hence, one bag of rice of 1 kg is bought per shopping trip along with other groceries (17 kg). At home, the white rice is cooked on a gas stove fuelled with natural gas.

3.4 Selected case study results

The environmental impacts of cooked rice according to the three scenarios are quantified with the following four indicators relevant to the TFs:

  • Global warming (GWPs, 100-year integration (IPCC 2013, Table 8.A.1))

  • Fine PM

  • Land occupation (land competition (Guinée et al. 2001, Sect. 4.3.3)

  • Water withdrawal (gross water use, no scarcity weighting)

The indicators fine PM, land occupation and water withdrawal represent a non-characterised sum of PM2.5 emitted, square-metre-years of land occupied and m3 of water withdrawn, respectively.

Global warming

Global warming impacts are quantified in Fig. 2 using the GWPs (100-year integration time) published in the main table of the 5th Assessment Report (IPCC 2013). The TF also produced results, not shown here, for several modelling choices such as TH, with or without climate-carbon feedbacks, with and without NTCFs to test the sensitivity of LCA results to them. The production stage causes the largest share of GHG emissions in all three scenarios (see Fig. 2). India demonstrates the highest overall impact on global warming due to the amount of GHGs being emitted during cooking with firewood (76 % of it being non-sustainably harvested wood and, thus, giving rise for non-biogenic CO2 emissions). The higher GHG emissions of rice production in the USA is due to a more elaborated inventory “rice, at farm/USA” compared to the inventory of Chinese rice production (see Electronic Supplementary Material) and should thus not be over-interpreted. The impact of distribution is lowest in India due to the small distance from the processing plant to the supermarket and the use of the bicycle for shopping.

Fig. 2
figure 2

Global warming impact of 1 kg of cooked rice in the three scenarios, measured in kilogram CO2 equivalents (IPCC 2013, Table 8.A.1)

Full size image

PM (PM <2.5 μm)

Figure 3 shows the sum of particulates <2.5 μm emitted during the four life cycle stages of the three scenarios. Most PM emissions arise during cooking with firewood on a three-stone fireplace in India. Surprisingly, also cooking in the China scenario causes a substantial share of the PM emissions, which is due to their electricity being mostly generated from coal. The PM emissions of the “distribution” stage are highest for the USA/Switzerland scenario because of the overseas transportation. However, this is a minor contribution to the overall PM emissions.

Fig. 3
figure 3

Particulate matter <2.5 μm emitted per kilogram of cooked rice in the three scenarios, measured in kilogram

Full size image

Land occupation

Agriculture and, thus, the “rice production” stage occupy most land (see Fig. 4). Rice cultivation in India occupies more land than rice cultivation in USA and China because of its lower yield. Furthermore, the considerably high land demand in the “cooking” stage of India is due to the demand of firewood.

Fig. 4
figure 4

Land occupation per kilogram of cooked rice in the three scenarios, measured in square-metre-years land occupied

Full size image

Water withdrawal

Nearly 100 % of the water is withdrawn during the rice production stage from groundwater and surface water for irrigation as shown in Fig. 5. All further life cycle stages contribute negligible shares to the total amount of withdrawn water.

Fig. 5
figure 5

Water withdrawal per kilogram of cooked rice in the three scenarios, measured in cubic meter water withdrawn

Full size image

4 Conclusions

The topical TFs are well on track in identifying and harmonising relevant and applicable environmental LCIA indicators. While the global warming TF focuses on the most appropriate metrics among those published in the 5th Assessment Report of IPCC (IPCC 2013), the water use TF (WULCA) proposes a new scarcity-based midpoint water use indicator. The PM TF identified the best available model parameters to quantify human health effects caused by outdoor and indoor PM 2.5 emissions, listening to domain experts, and the LU TF identified species richness as a good starting point for a harmonised indicator and will explore complementary metrics such as vulnerability and habitat fragmentation. With local and regional variability, the LU TF faces a similar problem like the water use TF did in the past and does still today. The cross-cutting issues TF and the rice LCA case study help ensuring consistency across environmental indicators and applicability of the LCIA indicators proposed, respectively. The rice LCA case study with its multi-tier supply chains and generic processes will motivate the TF to provide generic as well as geographically and temporally differentiated harmonised indicators.

The results of the current work of the TF will be documented in white papers, some of which will be published in scientific journals. The white papers will be the relevant input for the 5-day Pellston workshop™, where delegates from the TF and further stakeholders will discuss and agree on best practice and harmonised LCIA indicators as well as an update on the general LCIA framework. Based on the progress made in the last 12 months and the current state of discussions, it is very likely that a relevant outcome will be produced during the Pellston workshop™ taking place in Valencia, Spain, from 24 to 29 January 2016.