A carbon footprint analysis in the textile supply chain

Abstract

This research work focuses on the application of life-cycle assessment methodology to determine the carbon footprint of different players involved in a supply chain of the textile sector. A case study of a product by a textile leader company was carried out. This study demonstrates that, in the textile chain, the main contribution to the greenhouse effect is provided by the electrical and thermal energy used and by the transportation (since different production phases are delocalised in a wide range that goes from South Africa, Italy, Romania and all around the world, from the distribution centre to the stores). The Monte Carlo analysis has been used in order to obtain, for each calculated impact, not only the average value but also the distribution curve of the results characterised by uncertainty parameters. Moreover, a sensitivity analysis was carried out to evaluate the impact of management choices such as:

• a change in the transportation modality, from aeroplane to boat;

• a combination of road and rail transportation; and

• a selection among suppliers that allows the firm to cut environmental impacts.

Keywords:

• life-cycle assessment
• carbon footprint
• textile sector
• environmental supply chain
• sensitivity analysis

1. Introduction

The ‘sustainable development’ philosophy has been gaining a growing interest from public institution, customers and companies that, to consolidate and increase their position in the market, have to necessarily deal with environmental issues. Through techniques and instruments of eco-efficiency such as design for environment, environmental management systems or life-cycle assessment (LCA), the economy and the ecology have to become synergic and not in contradiction as presented in many cases (see Halldorsson et al. 2009). Seuring and Müller (2008) proposed a literature review on sustainable supply chain management taking into account 191 papers, published from 1994 to 2007. They found out that frequently external triggers, which are placed on focal companies by governing agencies, customers and stakeholders, are proposed. Such pressure as well as incentives might lead to action by those companies.

This study was proposed by a leader company itself compelled by the increasing attention of the market to environmental issues. The main aim of the project was to evaluate the application of the LCA methodology in order to obtain a tool to support the company environmental policy. The first step was to define the object of the analysis: one of the basic items produced by the company was chosen as a case study. A simple wool sweater without buttons, laces, zippers or other accessories, with standard production processes for the yarn, weaving and finishing treatments, was taken into account. The result of this study was the evaluation of the carbon footprint of such product to assess its impact on the greenhouse effect. The audiences to which the results were addressed were the textile leader company that promoted the research and its suppliers. Moreover, other than for external communication, the purposes of the management in applying the LCA methodology are related to the product improvement, and to the support for strategic choices and benchmarking. Seuring and Müller (2008) identified two strategies followed by leader companies in order to create a sustainability supply chain management: ‘supplier management for risks and performance’ and ‘supply chain management for sustainable products’. In the first approach, one of the main concerns of the companies is a loss of reputation if related problems are raised, while the second approach requests the definition of life-cycle-based standards for the environmental and social performance of products that are then implemented throughout the supply chain.

The following sections present the results of this research work: in Section 2, the material and methods utilised to define the input and the output of the system and to perform the uncertainty analysis are described. In Section 3, the case study, the definition of the system boundaries, life-cycle inventory and impact analysis are presented. Sections 4 and 5 present the results of the sensitivity analysis and conclusions.

2. Material and methods

The application of the LCA methodology was carried out following the international standards and using the Simaprò software and the ecoinvent database (Pre Consultants 2006; Swiss Centre for Life Cycle Inventory 2009) applied to the production phases of the chosen item (wool sweater). In accordance with ISO 14041, the stage of life-cycle inventory assessment (LCIA) involves the collection of the data concerning the processes and various calculation procedures. The relationship between the item produced and the environment is defined.

The focus of this project was the calculation of carbon footprint measured in CO₂ equivalent. For this purpose, the International Panel of Climate Change (IPCC 2007) method was selected. This method developed by the IPCC allows to quantify the greenhouse effect by measuring the equivalent CO₂ and lists the climate change factors of IPCC with a time frame of 20, 100 and 500 years. In this work, a time frame of 100 years was considered. Normalisation and weighting are not a part of this method.

The English guidelines (PAS 2050 2008)4 were taken into account for the following aspects of the boundary definition:

• the modelling of the transportation phase to the stores was considered and an average transportation model was used;

• the exclusion of the capital goods from the analysis.

2.1 Related research works

Many companies focused their attention on environmental impact of their products. Levi Strauss & Co. (2009a, 2009b) commissioned a LCA of two of their core products. By taking a product life-cycle approach to their work, they were able to develop a set of strategies to address the greatest impacts of their business on the environment.

Wiedmann and Minx (2007) calculated that textiles and clothing are responsible for around 4% of the secondary carbon footprint of an individual in the developed world. The problem has been largely addressed at the European level, the EU COST Action 628 (Nieminena et al. 2007) was established to produce first-hand industrial environmental data of textiles in Europe, as well as to suggest tools for comparisons of present technologies and practices with cleaner applications, including the economic effects. LCA was used to set up criteria for an environmental product declaration for textile products. Unique, first-hand industrial data were collected from five European textile industries.

McCurry (2009) described a study about future trend in textile industries carried out from a Freedonia Group, Ohio-based consultancy firm. The study stated that the industries will pay attention to a host of environmental issues that aim to reduce their carbon and environmental footprints.

An ecological footprint study in the textile field has been proposed by Herva et al. (2008). They analysed a textile tailoring plant with the overall purpose of developing a tool for evaluating the environmental impact evolution due to the performance of the plant, as well as for comparing the environmental behaviour of different tailoring processes. Therefore, the selected data were those from the manufacturing work. Data were divided into three main categories: energy, resources and waste. The principal contribution to the final environmental factor (expressed in hectares of land) was the resources category, mainly due to the high value associated with the cloth. The consumed energy was the second contributor, while the waste category remained in the third place. The final outcomes were divided by the production rates to obtain a comparable relative index, easy to be interpreted by different stakeholders.

Regarding woollen products, Barber and Pellow (2006) presented a research that produced a detailed inventory of resource inputs for New Zealand Merino wool and assessed its total energy use profile relative also to other textiles. Results of this study show that New Zealand Merino fibre production and early-stage processing uses significantly less energy than synthetic fibres.

Brent (2004) proposed for a wool product a modified LCIA procedure, which is based on the protection of resource groups. A distance-to-target approach is used for the normalisation of midpoint categories, which focuses on the ambient quality and quantity objectives for four groups: air, water, land and mined abiotic resources. The case study establishes the importance of region specificity for life-cycle inventory (LCIs) and LCIAs.

In comparison to other works in the textile sector, this research carried out a carbon footprint analysis of woollen products involving all the supply chain stakeholders and evaluating the impact of new management choices, such as the transportation modality or the criteria for supplier selection.

2.2 Uncertainty analysis

Uncertainty is a measure of the ‘goodness’ of results. Since the LCA is a model, there can always be errors that cause a certain level of uncertainty in the results. The uncertainty may depend on several factors: poor correspondence between the software models and the reality and between geographic and temporal aspect, non-representative measurement samples, non-complete information, poor reliability of data, etc. (ISO 14040 2006). Traditionally, the LCA (inventory and impact assessment) is a deterministic model used for estimating the potential impacts associated with a product. However, the LCA's primary weakness lies in its improper treatment of the uncertainty resulting from the sparse and imprecise nature of available information and the simplified model assumptions (Lo et al. 2005). This fact is demonstrated by Jimenez-Gonzalez and Overcash (2000) who, comparing the LCI results for refinery products among several available databases, have shown that the variability of estimated emissions to the atmosphere, water-borne and solid waste are approximately 50–150, 1000 and 30%, respectively.

Variability and parameter uncertainty of unit process inputs and outputs, e.g. measurement uncertainties, process-specific variations, temporal variations, etc., are expressed in quantitative terms on the level of individual inputs and outputs of unit processes. This type of uncertainty has been treated consistently and in a quantified way within the ecoinvent project (Frischknecht and Rebitzer 2005). In this work, a lognormal distribution has been assumed for all unit processes of ecoinvent data. In fact, several reports in the field of risk assessment and impact pathway analysis have shown that the lognormal distribution seems to be a more realistic approximation for the variability in fate and effect factors than the normal distribution (Hofstetter 1998).

The Monte Carlo analysis has been used in order to obtain, for each calculated impact, not only the average value but also the distribution curve of the results characterised by uncertainty parameters.

The statistical principle of the Monte Carlo method consists in repeating calculation many times. Each time a random value is chosen for each flow, for example an emission or raw material input. In this work, the uncertainty analysis was based on 1000 calculations. According to Langevin et al. (2010), this number of iterations was a compromise solution between simulation time and precision of results. Uncertainty information can be derived with basic statistical methods from the distribution of the calculation results.

The values chosen in the Monte Carlo analysis are within a specified distribution. In this study, the uncertainty estimation for each emission was calculated using the following Equation (1) developed by Weidema and Wesnaes (1996). The pedigree matrix (see Table A1 in Appendix) was used to calculate the U _i terms shown in the following standard deviation (SD) equation:

where U ₁ is the uncertainty factor of reliability; U ₂, the uncertainty factor of completeness; U ₃, the uncertainty factor of temporal correlation; U ₄, the uncertainty factor of geographic correlation, U ₅, the uncertainty factor of other technological correlation, U ₆, the uncertainty factor of sample size; and U _b , the basic uncertainty factor (see Table A2 in Appendix).

In order to reduce uncertainty as much as possible, the maximum level of detail in the input data has been reached by obtaining the main data directly from the different parties involved in the process.

3. Case study

The LCA methodology has been used to define the carbon footprint of a wool sweater made by a leader company in the textile sector. The entire production chain has been examined and all the single contributions to the environmental impact have been evaluated. Hypothesis for improving the resources use and management has been proposed, again using the LCA approach.

The idea with which many decisions were made was to maintain a quite generic point of view since the main goal was to assess if the LCA methodology could be applied continuously and systematically to the entire garment collection.

3.1 System boundaries

It was decided to define the boundaries of the system starting from sheep breeding at the farms in South Africa, from which the greasy wool is obtained, then considering the other production phases, the distribution to the final stores all around the world and, finally, the user phase (washing and final disposal) is shown in Figure 1.

Figure 1 System boundaries.

Those boundaries comprehend all the phases related to the sweater production, distribution and use: wool scouring, dyeing, spinning, knitting, transportation to the distribution centres, and then to the selling stores, washing and final disposal.

All the packaging, especially the ones used for sending the sweater from the production site to the store, were considered.

Regarding the environmental boundaries, it was decided to consider the agricultural systems as part of the environment and this means, for example, that pesticides are viewed as emissions.

3.2 Functional unit

In order to maintain a general point of view and to consider an average sweater, all the different characteristics were included in the functional unit.

The sweater chosen as the case study had the following features:

• 100% Merino wool;

• four colours; and

• 2009 winter collection.

The medium weight method was used to calculate the average sweater. The parameters included in the average sweater, representative of the functional unit, were:

• Distribution size – each size assumes a different weight as shown in Table 1.

  Table 1 Percentage distribution for each size.

  	S	M	L	XL	XXL	XXXL	Total
  Distribution piece-size (%)	4.71	20.59	25.29	24.71	17.65	7.06	100
  Pieces for size (n)	141.18	617.65	758.82	741.18	529.41	211.76	3000
  Weight total for sizes (g)	32654.12	150798.53	195017.65	201911.29	151025.29	63131.29	794538.18

• Colour – each colour has a slightly different dyeing process. Based on the colour distribution for the examined sweater in the entire winter 2009 collection, it was possible to calculate the contribution of each colour in the medium sweater in terms of weight (Table 2).

  Table 2 Average sweater characteristics.

  Accessories
  Number	Description	Weight
  1	Main label (regular fit)	0.31 g
  1	Flag label	0.11 g
  1	Origin label	0.11 g
  1	Hang tag	1.83 g
  1	Safety pins (hang tag)	0.2 g
  1	Cord Ritza (hang tag)	0.3 g
  1	Care label	0.3 g
  2	Barcode	0.1 g
  1	Silk paper	4.25 g
  1	Polybag	15.65 g
  Colour	001	032	038	521	522	Total (kg)
  Weight distribution (%)	111.81	3.22	82.25	32.00	35.58	264.85

• Net weight – 264.85 g (without accessories).

• Colour and accessories.

3.3 Life-cycle inventory analysis (LCIA)

The collected data have been divided into two different groups: primary and secondary data are shown in Table 3.

Table 3 Data sources.

			Secondary data
Supplying company	Production processes	Primary data	Software model	BREF	ARPA
A	Sheep breeding	X	X
B	Scouring			X	X
C	Dyeing	X		X	X
D	Spinning	X	X		X
E	Trowing and vaporising	X			X
F	Knitwear factory	X	X

The primary data were collected directly from the companies involved in the different phases of the production process by a questionnaire. Secondary data were extracted from:

• The software models (yarn production bast fibres/IN U 2007).

• ‘BREF’ (best available technology reference document) of the European Community (2003). This document provides information about the best techniques for the textile sector.

• ‘Analysis of the production cycle in the textile and wool sector’ of the ARPA of Piemonte Region (Foddanu et al. 2005).

For the final phases of the life cycle, use and final disposal, data were obtained from the literature, market analyses, company data and ENEA database.

3.3.1 The production process

According to Byoungho (2004), the textile industry represents one of the most various and complicated productive processes of the entire manufacture system. Generally speaking, it is possible to distinguish two main different processes: mechanical (spinning and weaving) and chemical (washing, dyeing and finishing; see Goldbach et al. 2003).

The complete processing cycle of the analysed item involves several steps. First comes shearing, followed by sorting and grading, making yarn, dyeing, finishing, making fabrics, making up the sweater and distributing.

In synthesis, seven companies are involved in the production of the analysed item with 17 transportation phases for a total of more than 10,000 km, 5 KW h of electrical energy, 18 MJ of thermal energy, 60 g chemicals and detergents, 200 g of packaging and 350 g of waste.

Regarding the use of the sweater, data were obtained from company data and from ENEA (2003) guidelines. A medium life of 5 years with 15 washes per year was considered, 30°C of washing temperature, 10 litres of water for each cycle, 130 ml of chemicals5 (soap and conditioner).

For the final life phase, the final disposal, from the literature data, it was established that 49% was disposed of, 49% was burned and the last 2% was reused.

The transportation scheme for the distribution centres and the relative percentage are shown in Figure 2.

Figure 2 Scheme of the transportation from the warehouse to the DC.

3.4 Impact calculated by IPCC (2007)

By using IPCC (2007), it was possible to assess the contribution of each phase to the carbon footprint measured as CO₂ equivalent. Figure 3 shows that the total amount of CO₂ produced is 1.947 kg for the single item for the phases that go from breeding to final disposal.

Figure 3 Percentage of CO₂ in the different phases.

It is also evident that four phases give the main contribution to the CO₂ production:

1. The transportation from the DC to the stores (0.470 kg CO₂).

2. The sweater realisation (0.384 kg CO₂), since the material reaches the knitwear factory in Romania by truck (Euro 3), and it goes back to Carpi in the same way.

3. The breeding phase (0.376 kg CO₂).

4. The washing phase (0.280 kg CO₂).

3.4.1 Uncertainty assessment (IPCC 2007)

The Monte Carlo analysis was used to determine the uncertainty in the calculation of the CO₂ equivalent by setting the boundaries from breeding to final disposal. It calculated an average value of 1.947 kg with a SD of 0.179. The range goes from 1.614 to 2.331 kg with a 95% confidence interval (Figure 4).

Figure 4 Uncertainty analysis kg CO₂ eq 100 (IPCC 2007).

3.5 Life-cycle interpretation

In the research reported here, the LCI results were analysed and processed by means of a contribution analysis and an uncertainty analysis based on Monte Carlo simulations. The accuracy obtained using such method is acceptable since the range of the confidence interval is about 20% of the steady-state mean.

The research demonstrated that the CO₂ production related to the chosen sweater depends mostly on the complexity of the supply chain and on the distribution system.

The transportation is the main CO₂ generator in the sweater production process; in fact, the CO₂ equivalent produced in each phase depends mainly on the goods transportation modality and the travel length. Figure 5 shows the contribution of the transports (goods and packaging) in all the garment life cycle.

Figure 5 Contribution of transportation for each phase.

Figure 5 also shows that the packaging gives an important contribution of nearly 40% to the CO₂ emission, especially in the final phases of the life cycle.

For this reason, a model for designing a new packaging system has been created. New boxes of different sizes (small/large-sized) and less weight have been tested in order to insert more sweaters into a single box. However, those attempts were often refused from sales managers mostly for aesthetic reasons, but also because the solution proposed did not optimise the truck saturation. In fact, the results demonstrate that the packaging was already been conceived for minimising the volumes transported and that a new design provided irrelevant improvements ( < 1%). Moreover, the production and disposal of the packaging do not contribute to the CO₂ production.

The model of the transportation to the distribution centres represents a particularly interesting field for the leader company for two main reasons: it is under its direct control and it is a phase that involves all the goods produced by the company. This means that the model represents big volumes of items and, as a consequence, an improvement in this phase can affect widely the entire production chain.

The carbon footprint for each destination was calculated considering not only the impact related to the transportation and its modality but also the volume of the goods. The results attributed to the main impact, in terms of CO₂ equivalent produced, to the Carpi–USA destination are shown in Figure 6. The following graph clearly shows that a variation for the USA transportation modality can have the greatest impact in terms of CO₂ saving.

Figure 6 CO₂ equivalent for the different destinations.

Even if Australia and Japan have a similar volume, the goods travel on a freight ship for Australia, and aeroplane for Japan, and this justifies the greater impact of the last route.

4. Sensitivity analysis

A sensitivity analysis was carried out in order to evaluate the impact of management choices such as:

• a change in the transportation modality, from aeroplane to boat;

• a combination of road and rail transportation;

• a suppliers selection that allows the firm to cut down the environmental impacts; and

• a change in consumer behaviour.

4.1 Transport

It has already been demonstrated that the main impact in terms of CO₂ production is given by transportation. In this paragraph, the sensitivity of this parameter to the distance, volumes of goods and transportation means are presented.

The calculated impact of 1.947 kg CO₂ is a global average value, the real value for the single sweater depends strongly on the selling-point localisation. Figure 7 shows how the CO₂ equivalent produced depends on the transportation to the selling points but it does not give indications about the priorities to address in order to reduce it, since, for this purpose, an analysis of volumes is required.

Figure 7 Carbon footprint related to the selling-point location.

The carbon footprint produced during the transportation phase to the stores is about 34%; by changing the transportation modality (from road to rail and from plane to boat), it is possible to cut the CO₂ production by 20–30%.

The transportation of the goods to the distribution centres contributes for the 0.22 kg to the total carbon footprint. The contribution of each DC weighted over the volumes is shown in Figure 8.

Figure 8 Contribution of transportation modality related to the volumes of goods.

In order to reduce the CO₂ equivalent, two different approaches are required. First, it is necessary to look upon the destinations that give the higher contribution considering distances, volumes and modality. Second, the impact of a change in the transportation modality has to be modelled: from aeroplane to boat and from road to rail.

By changing these parameters, a total reduction of 0.18 kg that corresponds to 84% in this phase has been calculated.

4.2 The combination of road–rail transportation

In the research project, the transportation distance between Carpi and Wendlingen was analysed with an ‘accompanied combined transportation’ model. From Carpi, the trucks travel by road to the hub in Como, from which they reach Chiasso where they can be loaded directly on the trains and arrive in Basilea. From there, the trucks go back to the road until they arrive at the final destination in Wendlingen.

The difference, in terms of grams of CO₂ equivalent produced for each sweater, considered for the different transportation modalities is shown in Figure 9.

Figure 9 CO₂ production along the road Como–Wendlingen.

The adoption of the combined transportation methods would allow the company to reduce about 37% of the carbon footprint along this road. The total reduction (considering the entire process from sheep breeding to the stores) would be only 0.4%. This is due to the fact that the Como–Wendlingen road provides a small contribution to the global impact. Even if the benefit on the single sweater seems to be low, nevertheless, the improvement obtained with this operation on the total flow of goods transported by the leader company along this road is significant. Considering that the average gross weight of the sweater is 476 g and the annual flow is 500.000 kg, the benefit, in terms of CO₂ saving, is of 5.3 tonnes per year.

4.3 Suppliers selection

The selection of suppliers that guarantee good standards in terms of environmental impact with reference to the European Commission Standard (BREF) is a key factor in terms of CO₂ equivalent reduction. For the sweater production processes, an amount of 0.066 kg CO₂ per piece has been accounted for. Referring to the BREF, this is a fair low contribution and it is due to the company environmental awareness in the supplier selection.

From the study, it has been assessed that a high percentage of CO₂ derives from the transportation between suppliers. By hypothesising that all the suppliers (the dyeing mills, the spinning and the knitwear factory) are located in a nearest area, it is possible to save more than 0.4 kg of CO₂ for garment (28%). In Table 4, it is shown how the reduction of distances between suppliers could decrease the CO₂ produced.

Table 4 Benefit deriving from distance reduction.

Distance reduction	10%	20%	50%
Kg CO2	1.33	1.27	1.07
Benefit (%)	4.6	9.2	23.0

Moreover, the use of Euro 3 trucks to travel back and forth from the warehouse to the knitwear factory in Romania increases the effect. If Euro 4 trucks were used, the benefit would be 2.15% for pieces (about 3 tonnes for the annual production).

4.4 A change in consumer behaviour

The washing phase provides a great climate change impact (0.28 kg CO₂). A change in consumer behaviour could considerably decrease this impact. As an example, it should be possible to cut down the number of washing per year to less than the 15 times established in the ENEA study and used in this work to assess the impact of the final life phase. Moreover, the consumer could modify the washing temperature (assumed as 30°C in this research) changing considerably the climate change impact (Figure 10).

Figure 10 CO₂ production changing the washing temperature.

5. Discussion and conclusions

In the last few years, life-cycle thinking has been the focal point in the environmental policy development of the European Community standardised by the Integrated Product Policy. Many other standards are utilised in other countries.

In this scenario, LCA provides the scientific references for all the activities related to the Corporate Sustainability Report, which is a tool being used by more than 500 of the major companies all around the world to communicate their environmental policies to the market. The use of the LCA methodology allows the companies to evaluate and communicate the environmental impact of their processes and products.

With the introduction of more restrictive environmental standards, such as PAS 2050, the carbon footprint analysis could be one of the criteria to evaluate the suppliers from an environmental point of view and to improve the entire supply chain and to support the realisation of green products by the eco-design.

The aim of this paper was to assess the carbon footprint associated with a particular product of a leader textile company. The selected item has been utilised as a case study: average characteristics were chosen in order to define the input for the LCA model. This decision was made with the objective of performing a pilot project to assess the feasibility of the LCA methodology to be applied to the entire company production.

A strong support of a leader company was essential to overcome the reluctance of some stakeholders in providing data. According to Seuring and Muller (2004), inhumane working conditions or contaminations of the (local) environment could be frequently a problem in order to obtain information from suppliers.

With the use of LCA, it is possible to face the environmental analysis at different levels. In the proposed case study, the main aim was the carbon footprint evaluation but along the study, many other aspects were considered.

The LCA method could be a powerful tool to address the eco-efficiency promotion and to provide several benefits to the company. Particularly, the work developed in this case study allowed to make the customers aware that the product is environmentally sound; in fact, the results obtained were introduced in the marketing campaign. The study also allowed the company to improve production performances in terms of efficient managing and use of resources.

Moreover, the collaboration between several stakeholders made this work useful to increase the environmental consciousness of the employees and of the supply chain operator involved along the production process.

Acknowledgements

This research work was made possible due to the collaboration between the Knitwear and the IT & Logistic Departments of the company that, for privacy reasons, will be called GHB and the Department of Energy Studies of Marche Polytechnic University.

Additional information

Notes on contributors

M. Bevilacqua

1

G. Giacchetta

2

B. Marchetti

3

Notes

1. [email protected]

2. [email protected]

3. [email protected]

4. Where products are distributed to different points of sale (i.e. different locations within a country), emissions associated with transport will vary from store to store due to different transport requirements. Where this occurs, organisations should calculate the average release of GHGs associated with transporting the product based in the average distribution of the product within each country, unless more specific data is available (publicly available specification, PAS 2050 (2008), par. 6.4.6, note 4).

The GHG emissions arising from the production of capital goods used in the life cycle of the product shall be excluded from the assessment of the GHG emissions of the life cycle of the product (publicly available specification, PAS 2050 (2008), par. 6.4.3).

5. Data obtained from ENEA.

Appendix

Table A1 represents the pedigree matrix and Table A2 the basic uncertainty factor U _b values are defined in Pre Consultants (2006).

Table A1 Pedigree matrix.

Score	1	2	3	4	5
U 1, Reliability	Verified data based on measurements	Verified data partly based on assumptions or non-verified data based on measurements	Non-verified data partly based on qualified estimates	Qualified estimate (e.g. by industrial expert); data derived from theoretical information (stoichiometry, enthalpy, etc.)	Non-qualified estimate
	1.00	1.05	1.10	1.20	1.50
U 2, Completeness	Representative data from all sites relevant for the market considered over an adequate period to even out normal fluctuations	Representative data from >50% of the sites relevant for the market considered M over an adequate period to even out normal fluctuations	Representative data from only some sites (≪50%) relevant for the market considered or>50% of sites but from shorter periods	Representative data from only one site relevant for the market considered or some sites but from shorter periods	Representativeness unknown or data from a small number of sites and from shorter periods
	1.00	1.02	1.05	1.10	1.20
U 3, Temporal correlation	< 3 years of difference to our reference year (2000)	< 6 years of difference to our reference year (2000)	< 10 years of difference to our reference year (2000)	< 15 years of difference to our reference year (2000)	Age of data unknown or more than 15 years of difference to our reference year (2000)
	1.00	1.03	1.10	1.20	1.50
U 4, Geographical correlation	Data from area under study	Average data from larger area in which the area under study is included	Data from smaller area than area under study, or from similar area		Data from unknown or distinctly different areas (North America instead of Middle East, OECD-Europe instead of Russia)
	1.00	1.01	1.02		1.10
U 5, Further technological correlation	Data from enterprises, processes and materials under study (i.e. identical technology)		Data on related processes or materials but same technology, or Data from processes and materials under study but from different technologies	Data on related processes or materials but different technologies, or data on laboratory-scale processes and same technologies	Data on related processes or materials but on laboratory scale of different technologies
	1.00		1.20	1.50	2.00
U 6, Sample size	>100, continuous measurement, balance of purchased products	>20	>10, aggregated figure in environmental report	≥ 3	Unknown
	1.00	1.02	1.05	1.10	1.20

Table A2 Basic uncertainty factor.

Input/output group	U b	Input/output group	U b
Demand of		Emission to air of
Thermal energy	1.05	CO2	1.05
Electricity	1.05	SO2	1.05
Semi-finished products	1.05	Combustion: NO X , NMVOC total, methane, N2O and NH3	1.50
Working materials	1.05	Combustion: CO	5.00
Transport services	2.00	Combustion: individual hydrocarbons, TSM	1.50
Waste treatment services	1.05	Combustion: PM10	2.00
Infrastructure	3.00	Combustion: PM2.5	3.00
Resources		Combustion: polycyclic aromatic hydrocarbons (PAH)	3.00
Primary energy carriers	1.05	Combustion: heavy metals	5.00
Metals, salts	1.05	Process emissions: individual VOCs	2.00
Land use, occupation	1.50	Process emissions: CO2	1.05
Land use, transformation	2.00	Process emissions: TSM	1.50
Waste heat		Process emissions: PM10	2.00
Emission to air, water and soil	1.05	Process emissions: PM2.5	3.00
Emission to water of		From agriculture: CH4, NH3	1.20
BOD, COD, DOC, TOC	1.50	From agriculture: N2O, NO X	1.40
Inorganic compounds (NH4, PO4, NO3, Cl, Na, etc.)	1.50	Radionuclides (e.g. Radon-222)	3.00
Individual hydrocarbons, PAH	3.00	Process emissions: other inorganic emissions	1.50
Heavy metals	5.00	Emission to soil of
From agriculture: NO3, PO4	1.50	Oil, hydrocarbon total	1.50
From agriculture: heavy metals	1.80	Pesticides	1.20
From agriculture: pesticides	1.50	Heavy metals	1.50
Radionuclides	3.00	Radionuclides	3.00