Carbon footprint as environmental performance indicator for the ...
See discussions, stats, and author profiles for this publication at: http://www.researchgate.net/publication/230770765
Carbon footprint as environmental performance indicator for the manufacturing industry ARTICLE in CIRP ANNALS - MANUFACTURING TECHNOLOGY · DECEMBER 2010 Impact Factor: 2.54 · DOI: 10.1016/j.cirp.2010.03.008
CITATIONS
READS
27
161
3 AUTHORS: Alexis Laurent
Stig Irving Olsen
Technical University of Denmark
Technical University of Denmark
30 PUBLICATIONS 265 CITATIONS
46 PUBLICATIONS 904 CITATIONS
SEE PROFILE
SEE PROFILE
Michael Zwicky Hauschild Technical University of Denmark 173 PUBLICATIONS 4,689 CITATIONS SEE PROFILE
Available from: Alexis Laurent Retrieved on: 24 September 2015
CIRP Annals - Manufacturing Technology 59 (2010) 37–40
Contents lists available at ScienceDirect
CIRP Annals - Manufacturing Technology jou rnal homep age : ht t p: // ees .e lse vi er. com/ci rp/ def a ult . asp
Carbon footprint as environmental performance indicator for the manufacturing industry A. Laurent, S.I. Olsen, M.Z. Hauschild (2)* Department of Management Engineering, Technical University of Denmark (DTU), Lyngby, Denmark
A R T I C L E I N F O
A B S T R A C T
Keywords: Manufacturing Lifecycle Carbon footprint
With the current focus on our climate change impacts, the embodied CO2 emission or ‘‘Carbon footprint’’ is often used as an environmental performance indicator for our products or production activities. The ability of carbon footprint to represent other types of impact like human toxicity, and hence the overall environmental impact is investigated based on life cycle assessments of several materials of major relevance to manufacturing industries. The dependence of the carbon footprint on the assumed scenarios for generation of thermal and electrical energy in the life cycle of the materials is analyzed, and the appropriateness of carbon footprint as an overall indicator of the environmental performance is discussed. ß 2010 CIRP.
1. Background
2. Comparison procedure
In the global endeavour to meet the international commitments to reduction of greenhouse gas emissions, many companies integrate environmental issues into their management systems, with potential effects in their entire production chains. Several tools and metrics have been developed to measure the environmental impact of a product in the life cycle perspective of the whole product chain. A metric that has gained prominence in recent times is the carbon footprint (CFP), which quantifies the climate change impact of greenhouse gas (GHG) emissions in a life cycle perspective [1]. However, more encompassing approaches exist, one of the most prominent being a proper Life Cycle Assessment (LCA). Like the CFP, an LCA focuses on a product system, comprising all the processes related to a product or a service—from the cradle to the grave. In contrast to the CFP, an LCA assesses all the environmental impacts of the system, not just the contributions to climate change [2]. Considering the multi-faceted nature of environmental impacts from production systems, one may contest the ability of carbon footprint to represent the overall environmental performance of a product. To investigate the legitimacy of CFP as indicator of environmental impacts more broadly, a comparative analysis has been performed of CFP and life cycle impacts on human health, from the production of a range of metals, chemicals, plastics and textiles—all materials that are central for our manufacturing industry.
In LCA the inventory analysis quantifies the ‘‘elementary flows’’ of the product system in the form of inputs from the environment without prior human transformations and outputs to the environment without further human transformations. In the Life Cycle Impact Assessment (LCIA), this information is translated via a characterization step and aggregated to environmental impact indicator results related to human health, natural environment and resource depletion [3]. After the characterization, impact category results are expressed in different metrics and can hence not be compared across impact categories. Therefore, a ‘‘normalization’’ is performed, translating all impacts to a common unit by calculating ‘‘the magnitude of the impact indicator results [characterization] relative to some reference information (socalled normalization references)’’ [4]. The normalization reference applied in this study is the annual contribution of an average person to each impact, and the resulting common unit for all impact categories is the Person Equivalent (PE). The CFP typically considers the six GHGs identified in the Kyoto Protocol, i.e. CO2, CH4, N2O, SF6, HFCs and PFCs. The normalization reference for the CFP was calculated based on the global per capita emission data for these GHGs in 2004 applying the latest set of global warming potential (GWP) factors, released by the IPPC as characterization factors [5]. Human toxic impacts (non-cancer effects) (HTI) from releases of toxic substances to the environment were assessed using characterization factors calculated with the USEtoxTM-model (www.usetox.org). This model was developed as part of the UNEP-SETAC’s Life Cycle Initiative and provides recommended factors for a global assessment of human toxicity [6]. To ensure consistency in the comparison to CFP, the normalization reference for the HTI category was calculated for the same emission year as for global warming, i.e. 2004.
* Corresponding author. E-mail address: [email protected] (M.Z. Hauschild). 0007-8506/$ – see front matter ß 2010 CIRP. doi:10.1016/j.cirp.2010.03.008
38
A. Laurent et al. / CIRP Annals - Manufacturing Technology 59 (2010) 37–40
3. Results and discussion Both the CFP and the HTI were calculated with the LCA software SimaProTM using the inventory data for production of materials available in the Ecoinvent database [7], and compared after normalization. In the comparison, the materials were divided into four categories: metals, chemicals, plastics and textiles. Results are shown for each category in Fig. 1. 3.1. CFP versus HTI With some variations the HTI scores dominate the CFPs for most of the materials shown in Fig. 1, typically with 1 order of magnitude for plastics, textiles and paper materials and up to 2 orders of magnitude for metals and chemicals. For metals, the relative differences vary around a factor of 10 for aluminium, steel and nickel, 10–30 for the production of solder, whereas factors of ca. 30–70 are found for the production of zinc. Copper shows more disparate results depending on the type of production considered (e.g. combined metal production, primary/secondary copper) with differences varying from 5 to 70 while the inclusion of Cu in alloys such as bronze or brass results in factors of up to 150. Overall, this means that the CFP contributes only a minor part to the material’s full normalized environmental impact. Nevertheless, there may well be a methodological bias in the normalization of the CFP and HTI results as discussed in Section 4, explaining at least part of the dominance of the HTI scores. Whether CFP gives a small or large contribution to the total environmental impact depends on this, and it is therefore also of interest to decide whether the CFP results are able to predict the human toxic impacts, i.e. whether the plotted data points show proportionality between HTI and CFP results. This may be the case when the two types of impacts are caused by the same activities in the life cycle (for aluminium in Fig. 1a the production of aluminium hydroxide), and then the CFP may serve as a good proxy of overall environmental impact. On the other hand, it is clear from Fig. 1, that for most of the materials, no such proportionality is observed. For the same material the ratio between the two impact scores may vary up to two orders of magnitude between different data sets and then CFP is not a good proxy of the human toxic impacts. 3.2. Influence of energy production scenarios A large share of the environmental impacts from the life cycle of material production comes from the generation of thermal or electrical energy used in the life cycle processes. Thermal energy is generally produced by combustion of (fossil) fuels but electricity production involves technologies of widely differing environmental impact profiles ranging from renewable energy technologies and nuclear power with limited carbon footprint to coal fired power with a high CFP per produced kWh. The ratio between CFP and HTI and hence the ability of CFP to predict overall environmental impact may thus depend on the electricity scenario that is used when producing the material, and this will typically differ between countries and regions. Furthermore, a shift in Table 1 Electricity supply source apportionment in the baseline scenario and the two alternative scenarios. Fig. 1. The normalized HTI scores plotted against normalized carbon footprint scores (both in PE/kg) for a) metal production (the category ‘‘Others’’ includes, among others, productions of lead, zinc, solder), b) chemical production (both organics and inorganics), c) plastic production (the category ‘‘Others’’ includes other thermoplastics, elastomers and biopolymers) and d) textile and paper production. Each dot represents the impacts of one material viewed in a life cycle perspective. The inserted line represents an equal magnitude for the two normalized impact scores. Clusters located above this line represent materials for which the normalized HTI is higher than the normalized CFP.
Hard coal Lignite Oil Natural gas Nuclear Hydropower Wind Others (PV, cogen.) a
Baseline scenario
Scenario 1 Nat. gas
Scenario 2Wind
16% 14% 4% 16% 32% 12% 2% 1%
8% 7% 0% 28% 30% 10% 7% 3%
– – – – – – ca. 100% –
Totals not exactly equal to 100% due to approximations and roundings.
A. Laurent et al. / CIRP Annals - Manufacturing Technology 59 (2010) 37–40
energy technology is foreseeable as policy focus shifts towards outfacing fossil fuels and GHG emissions. To explore the importance of the technology mix applied in the modelling for thermal and electrical energy production, two alternative scenarios, were built and applied to three of the materials: aluminium, copper (coupled production with nickel), and carbon monoxide. The first energy scenario, named ‘‘nat. gas’’, represents foreseeable changes in the electricity technology mix replacing coal by natural gas—see Table 1. The second scenario, named ‘‘wind’’, represents the immediate consequence of an extreme shift of the present energy supply sources of a factory to wind turbines as exclusive electricity source in order to investigate the sensitivity of the results to more extreme (but still possible) electricity scenarios. In both alternative scenarios, natural gas entirely substitutes oil and coal in the production of thermal energy.
39
For all three materials both the HTIs and the CFP decrease in both alternative scenarios, but Fig. 2 clearly identifies three patterns. The first pattern occurs when the toxic impacts do not stem primarily from energy production but from other processes in the life cycle of the material (typically disposal of materials during production). This is observed for the production of aluminium, where the changes in the energy generation scenario has little influence on the HTI, even when all electricity is produced by wind power plants (see Fig. 2a). In contrast, the CFP is strongly reduced, and in such a case, the ratio between human toxicity and carbon footprint results depends strongly on the choice of the energy generation scenario. The second pattern is observed for the case of copper production (coupled with nickel). Here, the main cause of HTI is emissions from the energy generation. While the change of
Fig. 2. The sensitivity of carbon footprint and human toxicity to changes in energy generation from the baseline scenario (used in Fig. 1) to one of the two alternative energy scenarios for productions of a) aluminium, b) copper and c) carbon monoxide.
40
A. Laurent et al. / CIRP Annals - Manufacturing Technology 59 (2010) 37–40
energy scenario influences the results for both indicators, the HTI drops off more than CFP, even when electricity is supplied entirely by wind power plants (see Fig. 2b). This is due to the change to use of natural gas for thermal heat production. While, the ratio between CFP and HTI for production of 1 MJ electricity decreases by more than half from the baseline scenario to wind, the trend is reversed when the Nat. gas scenario is used instead of the baseline scenario, and the HTI is reduced more than the CFP due to very limited emissions of toxic metals in the life cycle of natural gas. For materials where thermal energy requirements prevail over electricity requirements, the switch of heat source to natural gas thus causes the impacts on human health to decrease at a higher rate than the impacts on climate change. The third pattern is observed when the production requires more electricity than thermal energy, and there are no major sources of HTI in the life cycle, apart from the use of energy. This pattern is illustrated by the carbon monoxide example in Fig. 2c. The shift away from oil/coal/lignite as energy carriers in the baseline scenario results in a higher decline of impacts on human health than of impacts on climate change, which is similar to the second pattern illustrated in Fig. 2b. The dominance of electricity supply over thermal energy becomes obvious when the baseline scenario is switched to the wind scenario and CFP drops at a higher pace than HTI, thus enlarging the gap between the two impact categories. The three patterns observed in Fig. 2 show the strong influence of the energy scenario on the ability of CFP to represent HTI. When energy production is the dominating source of both CFP and HTI, the change in the ratio between CFP and HTI, for a replacement of the baseline scenario by the alternative electricity scenarios, depends on the required amount of thermal energy relative to the electricity supply and will thus vary on a case-by-case basis. On the other hand, when human toxicity stems from non-energy-related processes (e.g. waste disposal during production), reductions in GHG emissions are decoupled from toxic emissions and the ratio between CFP and HTI will decrease when the electricity production is changed to less fossil fuel-based technologies. 4. Uncertainties The quantification of the global warming potentials of greenhouse gases is based on considerable scientific consensus work [5]. The resulting global warming potentials have not changed considerably over the last 15 years and are presumably associated with high reliability, ensuring high accuracy in the obtained results. The same goes for the inventory of GHG emissions which covers a very restricted number of gases and for the most important gas among them, CO2, is based on stoichiometric calculations form the used amount of fuel. The inventory, the CFP results and the normalized values in Fig. 1 are thus deemed reliable for this impact category. The situation is different with respect to the assessment of HTI scores. Here, the emissions, inventory for both materials and normalization reference is associated with considerable uncertainties due to the number of potentially contributing substances and the need to measure emissions for each of these. Also the factors that are used to characterise the human health impacts of a substance can be rather uncertain [6]. The HTI scores are in general dominated by metal emissions for most of the materials which indicates that metals are likely to be inventorized with much more precision than organic substances. This bias in the inventory data for the materials may be somewhat counteracted by a similar bias in the inventory data underlying the normalization reference for HTI [8]. Considering the relatively higher certainty of the CFP
results, this may help explain the observed domination of HTI results over CFP results. 5. Conclusions and recommendations For most categories of materials analyzed here, CFP scores turned out to be dominated by HTI scores representing the impacts on human health. Biases in the normalization applied to bring the two impacts on a common scale for the comparison may explain some of this dominance, but except for a few materials there was no strong proportionality between the two impacts. CFP has thus been found to be a poor representative of this other category of environmental impact. Only in cases where the main sources were the same for both types of impact there was a proportionality that meant that CFP could be used to represent toxic impacts as well. Investigating the dependence on the technologies and fuels used to provide the thermal and electrical energy in the life cycle of the produced materials different patterns were observed. Overall there was a strong dependence on the applied energy technologies for both types of impact, and this is unfortunate for the general use of CFP to represent human health impacts, but potentially also other types of environmental impact. Electricity production scenarios show large variations between geographical regions, and although a frequent use of coal as fuel is observed, there are notable examples of countries that base their energy supply on large shares of nuclear power (e.g. France and Sweden) or hydro power (e.g. Norway). The inventory of the toxic emissions is reasonable for electricity production for which regulations and control of fluegas emission exist, but for other types of processes in the life cycle of the materials, it must be expected to be much more deficient than the inventory of GHGs. It is likely that a satisfactory inventory of all chemical emissions with contributions to HTI would make the deviations between the CFP and the HTI even larger than what was observed here. In summary, the large variations observed among materials— both the type of material itself and the type of production—and the poor correlation between CFP and HTI show that carbon footprint cannot be taken to represent the overall environmental impact, let alone the human health impacts in the choice of materials for more sustainable production. The applicability of carbon footprint as indicator of environmental sustainability in the design and manufacture of products thus needs to be documented on a case-by-case basis. References [1] Finkbeiner M (2009) Carbon Footprinting—Opportunities and Threats. International Journal of Life Cycle Assessment 14:91–94. [2] Hauschild MZ (2005) Assessing Environmental Impacts in a Life-cycle Perspective. Environmental Science & Technology 39(4):81A–88A. [3] Finnveden G, Hauschild MZ, Ekvall T, Guine´e J, Heijungs R, Hellweg S, Koehler A, Pennington D, Suh S (2009) Recent developments in Life Cycle Assessment. Journal of Environmental Management 91:1–21. [4] ISO. (2006) ISO 14044 International Standard. Environmental Management–Life Cycle Assessment—Requirements and Guidelines, International Organisation for Standardization, Geneva, Switzerland. [5] IPCC, 2007, Climate Change 2007: The Physical Science Basis, in Solomon S., Qin D., Manning M., Chen Z., Marquis M., Averyt K.B., Tignor M., Miller H.L. (Eds.), Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA, 996 pp. [6] Rosenbaum RK, Bachmann TM, Gold LS, Huijbregts AJ, Jolliet O, Juraske R, Koehler A, Larsen HF, MacLeod M, Margni M, McKone TE, Payet J, Schuhmacher M, Van de Meent D, Hauschild MZ (2008) USEtox—the UNEP-SETAC toxicity model: recommended characterisation factors for human toxicity and freshwater ecotoxicity in life cycle impact assessment. International Journal of Life Cycle Assessment 13(7):532–546. [7] Ecoinvent Centre, 2007, Ecoinvent Data, Swiss Centre for Life Cycle Inventories, Du¨bendorf. www.ecoinvent.ch. [8] Heijungs R, Guine´e JB, Kleijn R, Rovers V (2007) Bias in Normalization: Causes, Consequences, Detection and Remedies. International Journal of LCA 12(4):211– 216.
Carbon footprint as environmental performance indicator for the manufacturing industry ARTICLE in CIRP ANNALS - MANUFACTURING TECHNOLOGY · DECEMBER 2010 Impact Factor: 2.54 · DOI: 10.1016/j.cirp.2010.03.008
CITATIONS
READS
27
161
3 AUTHORS: Alexis Laurent
Stig Irving Olsen
Technical University of Denmark
Technical University of Denmark
30 PUBLICATIONS 265 CITATIONS
46 PUBLICATIONS 904 CITATIONS
SEE PROFILE
SEE PROFILE
Michael Zwicky Hauschild Technical University of Denmark 173 PUBLICATIONS 4,689 CITATIONS SEE PROFILE
Available from: Alexis Laurent Retrieved on: 24 September 2015
CIRP Annals - Manufacturing Technology 59 (2010) 37–40
Contents lists available at ScienceDirect
CIRP Annals - Manufacturing Technology jou rnal homep age : ht t p: // ees .e lse vi er. com/ci rp/ def a ult . asp
Carbon footprint as environmental performance indicator for the manufacturing industry A. Laurent, S.I. Olsen, M.Z. Hauschild (2)* Department of Management Engineering, Technical University of Denmark (DTU), Lyngby, Denmark
A R T I C L E I N F O
A B S T R A C T
Keywords: Manufacturing Lifecycle Carbon footprint
With the current focus on our climate change impacts, the embodied CO2 emission or ‘‘Carbon footprint’’ is often used as an environmental performance indicator for our products or production activities. The ability of carbon footprint to represent other types of impact like human toxicity, and hence the overall environmental impact is investigated based on life cycle assessments of several materials of major relevance to manufacturing industries. The dependence of the carbon footprint on the assumed scenarios for generation of thermal and electrical energy in the life cycle of the materials is analyzed, and the appropriateness of carbon footprint as an overall indicator of the environmental performance is discussed. ß 2010 CIRP.
1. Background
2. Comparison procedure
In the global endeavour to meet the international commitments to reduction of greenhouse gas emissions, many companies integrate environmental issues into their management systems, with potential effects in their entire production chains. Several tools and metrics have been developed to measure the environmental impact of a product in the life cycle perspective of the whole product chain. A metric that has gained prominence in recent times is the carbon footprint (CFP), which quantifies the climate change impact of greenhouse gas (GHG) emissions in a life cycle perspective [1]. However, more encompassing approaches exist, one of the most prominent being a proper Life Cycle Assessment (LCA). Like the CFP, an LCA focuses on a product system, comprising all the processes related to a product or a service—from the cradle to the grave. In contrast to the CFP, an LCA assesses all the environmental impacts of the system, not just the contributions to climate change [2]. Considering the multi-faceted nature of environmental impacts from production systems, one may contest the ability of carbon footprint to represent the overall environmental performance of a product. To investigate the legitimacy of CFP as indicator of environmental impacts more broadly, a comparative analysis has been performed of CFP and life cycle impacts on human health, from the production of a range of metals, chemicals, plastics and textiles—all materials that are central for our manufacturing industry.
In LCA the inventory analysis quantifies the ‘‘elementary flows’’ of the product system in the form of inputs from the environment without prior human transformations and outputs to the environment without further human transformations. In the Life Cycle Impact Assessment (LCIA), this information is translated via a characterization step and aggregated to environmental impact indicator results related to human health, natural environment and resource depletion [3]. After the characterization, impact category results are expressed in different metrics and can hence not be compared across impact categories. Therefore, a ‘‘normalization’’ is performed, translating all impacts to a common unit by calculating ‘‘the magnitude of the impact indicator results [characterization] relative to some reference information (socalled normalization references)’’ [4]. The normalization reference applied in this study is the annual contribution of an average person to each impact, and the resulting common unit for all impact categories is the Person Equivalent (PE). The CFP typically considers the six GHGs identified in the Kyoto Protocol, i.e. CO2, CH4, N2O, SF6, HFCs and PFCs. The normalization reference for the CFP was calculated based on the global per capita emission data for these GHGs in 2004 applying the latest set of global warming potential (GWP) factors, released by the IPPC as characterization factors [5]. Human toxic impacts (non-cancer effects) (HTI) from releases of toxic substances to the environment were assessed using characterization factors calculated with the USEtoxTM-model (www.usetox.org). This model was developed as part of the UNEP-SETAC’s Life Cycle Initiative and provides recommended factors for a global assessment of human toxicity [6]. To ensure consistency in the comparison to CFP, the normalization reference for the HTI category was calculated for the same emission year as for global warming, i.e. 2004.
* Corresponding author. E-mail address: [email protected] (M.Z. Hauschild). 0007-8506/$ – see front matter ß 2010 CIRP. doi:10.1016/j.cirp.2010.03.008
38
A. Laurent et al. / CIRP Annals - Manufacturing Technology 59 (2010) 37–40
3. Results and discussion Both the CFP and the HTI were calculated with the LCA software SimaProTM using the inventory data for production of materials available in the Ecoinvent database [7], and compared after normalization. In the comparison, the materials were divided into four categories: metals, chemicals, plastics and textiles. Results are shown for each category in Fig. 1. 3.1. CFP versus HTI With some variations the HTI scores dominate the CFPs for most of the materials shown in Fig. 1, typically with 1 order of magnitude for plastics, textiles and paper materials and up to 2 orders of magnitude for metals and chemicals. For metals, the relative differences vary around a factor of 10 for aluminium, steel and nickel, 10–30 for the production of solder, whereas factors of ca. 30–70 are found for the production of zinc. Copper shows more disparate results depending on the type of production considered (e.g. combined metal production, primary/secondary copper) with differences varying from 5 to 70 while the inclusion of Cu in alloys such as bronze or brass results in factors of up to 150. Overall, this means that the CFP contributes only a minor part to the material’s full normalized environmental impact. Nevertheless, there may well be a methodological bias in the normalization of the CFP and HTI results as discussed in Section 4, explaining at least part of the dominance of the HTI scores. Whether CFP gives a small or large contribution to the total environmental impact depends on this, and it is therefore also of interest to decide whether the CFP results are able to predict the human toxic impacts, i.e. whether the plotted data points show proportionality between HTI and CFP results. This may be the case when the two types of impacts are caused by the same activities in the life cycle (for aluminium in Fig. 1a the production of aluminium hydroxide), and then the CFP may serve as a good proxy of overall environmental impact. On the other hand, it is clear from Fig. 1, that for most of the materials, no such proportionality is observed. For the same material the ratio between the two impact scores may vary up to two orders of magnitude between different data sets and then CFP is not a good proxy of the human toxic impacts. 3.2. Influence of energy production scenarios A large share of the environmental impacts from the life cycle of material production comes from the generation of thermal or electrical energy used in the life cycle processes. Thermal energy is generally produced by combustion of (fossil) fuels but electricity production involves technologies of widely differing environmental impact profiles ranging from renewable energy technologies and nuclear power with limited carbon footprint to coal fired power with a high CFP per produced kWh. The ratio between CFP and HTI and hence the ability of CFP to predict overall environmental impact may thus depend on the electricity scenario that is used when producing the material, and this will typically differ between countries and regions. Furthermore, a shift in Table 1 Electricity supply source apportionment in the baseline scenario and the two alternative scenarios. Fig. 1. The normalized HTI scores plotted against normalized carbon footprint scores (both in PE/kg) for a) metal production (the category ‘‘Others’’ includes, among others, productions of lead, zinc, solder), b) chemical production (both organics and inorganics), c) plastic production (the category ‘‘Others’’ includes other thermoplastics, elastomers and biopolymers) and d) textile and paper production. Each dot represents the impacts of one material viewed in a life cycle perspective. The inserted line represents an equal magnitude for the two normalized impact scores. Clusters located above this line represent materials for which the normalized HTI is higher than the normalized CFP.
Hard coal Lignite Oil Natural gas Nuclear Hydropower Wind Others (PV, cogen.) a
Baseline scenario
Scenario 1 Nat. gas
Scenario 2Wind
16% 14% 4% 16% 32% 12% 2% 1%
8% 7% 0% 28% 30% 10% 7% 3%
– – – – – – ca. 100% –
Totals not exactly equal to 100% due to approximations and roundings.
A. Laurent et al. / CIRP Annals - Manufacturing Technology 59 (2010) 37–40
energy technology is foreseeable as policy focus shifts towards outfacing fossil fuels and GHG emissions. To explore the importance of the technology mix applied in the modelling for thermal and electrical energy production, two alternative scenarios, were built and applied to three of the materials: aluminium, copper (coupled production with nickel), and carbon monoxide. The first energy scenario, named ‘‘nat. gas’’, represents foreseeable changes in the electricity technology mix replacing coal by natural gas—see Table 1. The second scenario, named ‘‘wind’’, represents the immediate consequence of an extreme shift of the present energy supply sources of a factory to wind turbines as exclusive electricity source in order to investigate the sensitivity of the results to more extreme (but still possible) electricity scenarios. In both alternative scenarios, natural gas entirely substitutes oil and coal in the production of thermal energy.
39
For all three materials both the HTIs and the CFP decrease in both alternative scenarios, but Fig. 2 clearly identifies three patterns. The first pattern occurs when the toxic impacts do not stem primarily from energy production but from other processes in the life cycle of the material (typically disposal of materials during production). This is observed for the production of aluminium, where the changes in the energy generation scenario has little influence on the HTI, even when all electricity is produced by wind power plants (see Fig. 2a). In contrast, the CFP is strongly reduced, and in such a case, the ratio between human toxicity and carbon footprint results depends strongly on the choice of the energy generation scenario. The second pattern is observed for the case of copper production (coupled with nickel). Here, the main cause of HTI is emissions from the energy generation. While the change of
Fig. 2. The sensitivity of carbon footprint and human toxicity to changes in energy generation from the baseline scenario (used in Fig. 1) to one of the two alternative energy scenarios for productions of a) aluminium, b) copper and c) carbon monoxide.
40
A. Laurent et al. / CIRP Annals - Manufacturing Technology 59 (2010) 37–40
energy scenario influences the results for both indicators, the HTI drops off more than CFP, even when electricity is supplied entirely by wind power plants (see Fig. 2b). This is due to the change to use of natural gas for thermal heat production. While, the ratio between CFP and HTI for production of 1 MJ electricity decreases by more than half from the baseline scenario to wind, the trend is reversed when the Nat. gas scenario is used instead of the baseline scenario, and the HTI is reduced more than the CFP due to very limited emissions of toxic metals in the life cycle of natural gas. For materials where thermal energy requirements prevail over electricity requirements, the switch of heat source to natural gas thus causes the impacts on human health to decrease at a higher rate than the impacts on climate change. The third pattern is observed when the production requires more electricity than thermal energy, and there are no major sources of HTI in the life cycle, apart from the use of energy. This pattern is illustrated by the carbon monoxide example in Fig. 2c. The shift away from oil/coal/lignite as energy carriers in the baseline scenario results in a higher decline of impacts on human health than of impacts on climate change, which is similar to the second pattern illustrated in Fig. 2b. The dominance of electricity supply over thermal energy becomes obvious when the baseline scenario is switched to the wind scenario and CFP drops at a higher pace than HTI, thus enlarging the gap between the two impact categories. The three patterns observed in Fig. 2 show the strong influence of the energy scenario on the ability of CFP to represent HTI. When energy production is the dominating source of both CFP and HTI, the change in the ratio between CFP and HTI, for a replacement of the baseline scenario by the alternative electricity scenarios, depends on the required amount of thermal energy relative to the electricity supply and will thus vary on a case-by-case basis. On the other hand, when human toxicity stems from non-energy-related processes (e.g. waste disposal during production), reductions in GHG emissions are decoupled from toxic emissions and the ratio between CFP and HTI will decrease when the electricity production is changed to less fossil fuel-based technologies. 4. Uncertainties The quantification of the global warming potentials of greenhouse gases is based on considerable scientific consensus work [5]. The resulting global warming potentials have not changed considerably over the last 15 years and are presumably associated with high reliability, ensuring high accuracy in the obtained results. The same goes for the inventory of GHG emissions which covers a very restricted number of gases and for the most important gas among them, CO2, is based on stoichiometric calculations form the used amount of fuel. The inventory, the CFP results and the normalized values in Fig. 1 are thus deemed reliable for this impact category. The situation is different with respect to the assessment of HTI scores. Here, the emissions, inventory for both materials and normalization reference is associated with considerable uncertainties due to the number of potentially contributing substances and the need to measure emissions for each of these. Also the factors that are used to characterise the human health impacts of a substance can be rather uncertain [6]. The HTI scores are in general dominated by metal emissions for most of the materials which indicates that metals are likely to be inventorized with much more precision than organic substances. This bias in the inventory data for the materials may be somewhat counteracted by a similar bias in the inventory data underlying the normalization reference for HTI [8]. Considering the relatively higher certainty of the CFP
results, this may help explain the observed domination of HTI results over CFP results. 5. Conclusions and recommendations For most categories of materials analyzed here, CFP scores turned out to be dominated by HTI scores representing the impacts on human health. Biases in the normalization applied to bring the two impacts on a common scale for the comparison may explain some of this dominance, but except for a few materials there was no strong proportionality between the two impacts. CFP has thus been found to be a poor representative of this other category of environmental impact. Only in cases where the main sources were the same for both types of impact there was a proportionality that meant that CFP could be used to represent toxic impacts as well. Investigating the dependence on the technologies and fuels used to provide the thermal and electrical energy in the life cycle of the produced materials different patterns were observed. Overall there was a strong dependence on the applied energy technologies for both types of impact, and this is unfortunate for the general use of CFP to represent human health impacts, but potentially also other types of environmental impact. Electricity production scenarios show large variations between geographical regions, and although a frequent use of coal as fuel is observed, there are notable examples of countries that base their energy supply on large shares of nuclear power (e.g. France and Sweden) or hydro power (e.g. Norway). The inventory of the toxic emissions is reasonable for electricity production for which regulations and control of fluegas emission exist, but for other types of processes in the life cycle of the materials, it must be expected to be much more deficient than the inventory of GHGs. It is likely that a satisfactory inventory of all chemical emissions with contributions to HTI would make the deviations between the CFP and the HTI even larger than what was observed here. In summary, the large variations observed among materials— both the type of material itself and the type of production—and the poor correlation between CFP and HTI show that carbon footprint cannot be taken to represent the overall environmental impact, let alone the human health impacts in the choice of materials for more sustainable production. The applicability of carbon footprint as indicator of environmental sustainability in the design and manufacture of products thus needs to be documented on a case-by-case basis. References [1] Finkbeiner M (2009) Carbon Footprinting—Opportunities and Threats. International Journal of Life Cycle Assessment 14:91–94. [2] Hauschild MZ (2005) Assessing Environmental Impacts in a Life-cycle Perspective. Environmental Science & Technology 39(4):81A–88A. [3] Finnveden G, Hauschild MZ, Ekvall T, Guine´e J, Heijungs R, Hellweg S, Koehler A, Pennington D, Suh S (2009) Recent developments in Life Cycle Assessment. Journal of Environmental Management 91:1–21. [4] ISO. (2006) ISO 14044 International Standard. Environmental Management–Life Cycle Assessment—Requirements and Guidelines, International Organisation for Standardization, Geneva, Switzerland. [5] IPCC, 2007, Climate Change 2007: The Physical Science Basis, in Solomon S., Qin D., Manning M., Chen Z., Marquis M., Averyt K.B., Tignor M., Miller H.L. (Eds.), Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA, 996 pp. [6] Rosenbaum RK, Bachmann TM, Gold LS, Huijbregts AJ, Jolliet O, Juraske R, Koehler A, Larsen HF, MacLeod M, Margni M, McKone TE, Payet J, Schuhmacher M, Van de Meent D, Hauschild MZ (2008) USEtox—the UNEP-SETAC toxicity model: recommended characterisation factors for human toxicity and freshwater ecotoxicity in life cycle impact assessment. International Journal of Life Cycle Assessment 13(7):532–546. [7] Ecoinvent Centre, 2007, Ecoinvent Data, Swiss Centre for Life Cycle Inventories, Du¨bendorf. www.ecoinvent.ch. [8] Heijungs R, Guine´e JB, Kleijn R, Rovers V (2007) Bias in Normalization: Causes, Consequences, Detection and Remedies. International Journal of LCA 12(4):211– 216.
