Prospects of carbon labelling – a life cycle point of view

Journal of Cleaner Production 72 (2014) 76e88

Contents lists available at SciVerse ScienceDirect

Journal of Cleaner Production journal homepage: www.elsevier.com/locate/jclepro

Prospects of carbon labelling e a life cycle point of view M.Q.B. Tan a, R.B.H. Tan a, H.H. Khoo b, * a b

National University of Singapore, Department of Chemical and Biomolecular Engineering, 4 Engineering Drive 4, Singapore 117576, Singapore Institute of Chemical and Engineering Sciences (ICES), 1 Pesek Road, Jurong Island, Singapore 627833, Singapore

a r t i c l e i n f o

a b s t r a c t

Article history: Received 2 May 2012 Received in revised form 24 September 2012 Accepted 24 September 2012 Available online 22 October 2012

The total carbon footprint (CF) of the following life cycle food imports was compared for the prospects of carbon labelling; one study on beef from Canada, three from the U.S.A., and one each from Japan and Brazil; three studies on pork from Australia and Canada; three case studies on chicken from Brazil and Finland; rice from Thailand; and finally, two investigations of potatoes from the UK and Australia. The CF results on average were: beef (32.0 kg CO2-eq/kg or 100 kg CO2-eq/kg protein), pork (4.5 kg CO2/kg or about 18 kg CO2-eq/kg protein), chicken (2.9 kg CO2-eq/kg or about 10 kg CO2-eq/kg protein); and for rice and potatoes, 3.0 and 0.43 kg CO2-eq/kg respectively. Per 1000 kcal they are 2.31 and 0.56 kg CO2-eq respectively. While land use is widely acknowledged as a source/sink for carbon emissions, the allocation of CO2 amounts associated with deforestation was complex and difficult to quantify; and hence omitted from the life cycle CF analysis. It was highlighted that the results are not strictly comparable in absolute terms, but serve the purpose of shedding light on the environmental issues in a food production chain. A standardized approach would definitely be a useful GHG accounting tool to provide an indicator for carbon labelling schemes. Factors influencing carbon labelling schemes in Singapore were raised and discussed. From a survey conducted, 76% responded positively on having carbon labels. Ó 2012 Elsevier Ltd. All rights reserved.

Keywords: Life cycle Food GHG Carbon label Carbon footprint Singapore

1. Introduction There is growing scientific evidence and consensus about climate change and its link to human activities (IPCC, 2007). This concern is compounded by a global rapid growing population, with increased demand for energy, materials, minerals, and food products. In order to feed growing populations, the world’s agricultural sector faces great pressures to increase its output. In the World Population Prospectus published by the United Nations in 2005, it was projected that the world’s population is set to grow by 76 million people annually, with 95% of this growth taking place in developing countries (United Nations, 2005). At the same time, developing countries are growing in affluence, resulting in a greater demand for high-value foodstuffs such as meats. The livestock sector has a significant impact on the environment in various ways. It has recently come to light that animal agriculture plays a greater and more significant role in its contribution to global warming (e.g., Kramer et al., 1999; Fiala, 2008). Policies aimed at sustainable living and consumption patterns are increasingly focussing on this challenge, drawing largely from the insights gained through the carbon

* Corresponding author. Tel.: þ65 6796 7341; fax: þ65 6267 8835. E-mail address: [email protected] (H.H. Khoo). 0959-6526/$ e see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.jclepro.2012.09.035

footprint of various food products (McAlpine et al., 2009). Livestock are a major source of land pollution as they emit organic matter, nutrients, pathogens and drugs into the soil, which then seep into lakes and rivers (De Vries and De Boer, 2011). Greenhouse gases are emitted either directly from the animals or indirectly through waste. Furthermore, in order to create farmland for the production of livestock, large expanses of forest land has had to be cleared, resulting in the destruction of natural habitats and a loss of carbon dioxide consuming forests (Cederberg et al., 2011). According to the Food and Agricultural Organization of the United Nations (FAO, 2006) the three most common greenhouse gases e carbon dioxide (CO2), methane (CH4) and nitrous oxides (N2O) e are generated by livestock; that is, 9% of all anthropogenic CO2 emissions, 37% of anthropogenic CH4 emissions, and 65% of anthropogenic N2O emissions. The majority of N2O produced is a result of manure, while CH4 is produced mainly from enteric fermentation, which is the process by which carbohydrates are digested in ruminant animals (Ogino et al., 2007). Additionally, the environmental impact of growing grain for animal feed is extremely intense; three quarters of all water-quality problems in rivers and streams in the U.S.A. are due to the agricultural industry (Bittman, 2008). The world has seen large increases in the demand for meat, shown by changes in the annual per capita consumption of meat,

which doubled between 1980 and 2002 from 14 kg to 28 kg. While the demand for meat is expected to stay relatively stable in developed countries, that for developing countries is expected to grow exponentially, reaching 37 kg per capita per year in 2030 (FAO, 2006). 1.1. Food consumption in Singapore According to the Agri-Food and Veterinary Authority of Singapore (AVA, 2012), the aggregate annual meat consumption for Singaporeans in 2010 was 61.3 kg per capita. The figures, displayed in Table 1, are far higher than the global average, but lower than that in other developed countries such as the U.S.A., which had an annual meat consumption of 90.9 kg in 2007 (American Meat Institute, 2009). Being a developed country, Singapore’s per capita consumption of meat has stabilized over the years. The trend is illustrated in Fig. 1. Of the three major meat types, chicken is consumed most abundantly, followed by pork and then beef. Even though the consumption of beef has not declined significantly in Singapore, as it has in the United States where per capita beef consumption is down 25% from 1980, Singaporeans still consume much less beef than their counterparts in other developed countries. For example, despite the decline in red meat consumption, the average American is estimated to have consumed 26 kg of beef in 2011 (Reuters, 2011), compared to 4.2 kg in Singapore in 2010. 2. Carbon labels Carbon labelling schemes, though voluntary, have been introduced in countries such as the United Kingdom, The Netherlands, and Japan (Gössling et al., 2011; Saunders et al., 2011), with significant participation from companies that don’t want to be seen as lagging behind their rivals in cornering the green market. For example, Gadema and Oglethorpe (2011) explored the correlation between the accessibility of information and the degree of carbon footprint score on consumers’ choices in selecting food products. They concluded that there exists a connection between the two to guide customer choices and therefore a targeted and welldocumented carbon labelling strategy is essential. Carbon labels are expected to provide consumers with the opportunity to make informed choices, especially where the need to reduce carbon emissions has become a global concern. In another example, Upham et al. (2010) conducted a study of stakeholder and public perceptions of grocery carbon labelling in the UK. Similar to other parts of the world, it was highlighted that the prospects of carbon labels e based on carbon footprint or CF information e are still in the early stage of development. Despite the challenges faced by its implementation, the study concluded that such methods paved the way forward in helping to reduce the carbon intensity of high volume grocery products and could have a substantial role to play in meeting carbon reduction targets. Such examples initiated by many countries have become the driving force for other nations to emulate. Strategies and plans to develop methods to measure greenhouse gas (GHG) emission data e also known as carbon ‘footprinting’ e are becoming widespread. The purpose of this information is to be used on product labels so that consumers’ awareness relating to the environmental impacts involved in the production of products can be enhanced. This paper aims to explore the prospects of carbon labelling of raw food products in Singapore. Before carbon labels can be

Per Capita Consumption / kg

M.Q.B. Tan et al. / Journal of Cleaner Production 72 (2014) 76e88

77

Annual trend in per capita meat consumption in Singapore 40 30 Chicken (kg)

20

Pork (kg)

10

Beef (kg) 0 2000

2002

2004

2006

2008

2010

Year

Fig. 1. Annual trend in per capita meat consumption in Singapore from 2000 to 2010 (AVA, 2012).

adopted, the carbon footprint of the top three most commonly consumed meats in Singapore e chicken, pork, and beef e and two commonly consumed raw staple foods e rice and potatoes e will first be investigated using a life cycle approach. The general steps for the research approach are illustrated in Fig. 2. 3. Life cycle greenhouse gas analysis In order to generate a complete GHG emissions profile generated from a production chain, a life cycle approach was used to analyze the GHG emissions associated with a product. Similar to Life Cycle Assessment (LCA), which is recognized as a systematic and comprehensive environmental impact accounting tool (Mouron et al., 2006), the life cycle greenhouse gas analysis approach identified all activities involved in the production, consumption and retirement/disposal of a product, analyses the relevant energy and material requirements, and evaluated the emissions from the product’s life cycle. The results of the life cycle greenhouse gas analysis can then be used for the GHG emission labels, known as carbon labels, which allow consumers to make more informed consumption decisions (Ruviaro et al., 2011; Gössling et al., 2011). The results of the life cycle-based CF study can be used to influence consumer behaviour through initiatives such as carbon labelling of products. While Tesco has abandoned its carbon

Table 1 Annual per capita consumption of meat in Singapore in 2010 (AVA, 2012). Meat

Chicken

Pork

Beef

Duck

Mutton

Consumption (kg)

32.3

20.2

4.2

2.8

1.8

Fig. 2. Stages involved in investigating carbon labelling prospects based on life cycle approach.

78

M.Q.B. Tan et al. / Journal of Cleaner Production 72 (2014) 76e88

per kg carcass weight, which refers to the meat product after slaughter and evisceration, and for some animals (e.g., pigs), the removal of hair and genitalia. To further make comparisons across meats, the carbon footprint per mass of protein will also be presented. For the comparison of non-meat food items, per kg product and per 1000 kcal will be displayed.

labelling programme due to a lack of critical mass in the market and the sheer amount of work involved in quantifying carbon emissions, other companies have found success in carbon labelling. One such example is Max Burgers in Sweden, which started putting carbon labels on its burgers in 2008 and saw a jump in the sales of vegetarian burgers over beef burgers. Additionally, their carbon labelling initiative was a marketing success, boosting customer loyalty and allowing the burger chain to double its market share in Sweden (Van Gilder Cooke, 2012). These results show that carbon labelling is potentially viable, although the Tesco experience would have to be examined in greater detail to improve the chances of success in future.

3.1.1. Comparison of beef production Studies on the carbon footprint of beef production were taken from four countries: Beauchemin et al. (2010) for beef production in Canada; Pelletier et al. (2010) for cowecalf farming system in Upper Midwestern region, U.S.A.; Ogino et al. (2004, 2007) for calf, cow, beef system from Japan; and Cederberg et al. (2009) for cow grazing in Brazil. The generic system boundary for all four beef production cases is illustrated in Fig. 3. Beef production systems typically consist of two main stages e the cowecalf stage and the beef fattening stage. The investigation carried out by Beauchemin et al. (2010) typified the animal husbandry system and cropping practices widely used in western Canada. The farm consisted of a beef production operation, a cropping operation and native prairie pasture for grazing. The assessment was conducted over a period of 8 years period, starting with young calves, continued through meat production cycles, and ended with the slaughter of matured cows. During the calves’ maturity stage, they were fed a high forage diet. At the final weight of 605 kg they were sent to the slaughterhouse for food processing. The total GHG emissions from the herd were calculated by summing lifetime emissions for the 120 cows, 4 bulls and all feedlot cattle over the 8 year production cycle. In the second case, Pelletier et al. (2010) compared the cumulative energy use, GHG emissions and other 12pollution associated with three beef production strategies found in the Upper Midwestern region, U.S.A. The cowe calf herd is maintained on legume frostseeded pasture forage and hay. After reaching an average weight of 636 kg the cows were sent

3.1. Food imported to Singapore Due to a lack of resources and land constraints, Singapore imports most of its food products from overseas. Therefore, instead of performing an assessment for a specific agricultural sector, the goal and scope is focused on compiling and comparing five different food products (three meats and two staples) imported to Singapore from various countries. The system boundary for all food products is from cradle-to-gate, where the final finished produce is transported to Singapore by shipment. Generally, system boundaries of whole farm models of food production systems are defined to assess GHG emissions from “cradle-to-farm gate”. Since five different food types are being investigated, the system boundary will be displayed separately for each food type e starting from farms and ending with delivered food item to Singapore. The choice of functional unit of GHG emissions has important implications for the interpretation of results. Most cases describe emissions as total CO2-eq per kg product basis, or total CO2-eq per kg liveweight or carcass weight for livestock production systems. In this paper, the comparison basis across meat products is selected as

INPUT

OUTPUT

Fertilizers Pesticides Fuel, Electricity

CO2 CH4 N2O

FEED

Feed Production and Transport

FEED

Feedlot Feedlot Transport of to Finishing phase

Backgrounding

Abattoir

Shipment to Singapore

Per kg carcass

Grazing Grazing

Waste Management CALVES

MANURE

MATURE COWS

FROZEN BEEF

Fig. 3. Life cycle stages of beef production from cradle to gate (“gate” relates to per kg beef carcass delivered to Singapore).

M.Q.B. Tan et al. / Journal of Cleaner Production 72 (2014) 76e88

to slaughter. Field-level GHG emissions of CO2, N2O, and other pollutants relating to nitrogen fertilizer application were calculated. In Japan (Ogino et al., 2004, 2007), calves were produced at an interval of every 14 months with a gestation period of 9 months and lactation period of 5 months. The calves were fattened from 8 months of age till 28 months of age and marketed. The results highlighted that in particular, N2O derived from nitrogen in cattle waste contributes to global warming in the process of Japanese beef production systems. Finally, Cederberg et al. (2009) studied beef production from a farming system with continuous grazing in Brazil. The average calving interval was 21 months and the average age at slaughter was 4 years, and average carcass weight was 200 kg. Calculations of emissions of N2O from cattle manure were done based on data on the nitrogen excreted by livestock. 3.1.2. Comparison of pork production For pork, the case studies were taken from Australia (Wiedemann et al., 2010) and Canada (Verge et al., 2009). The generic system boundary for pork production is illustrated in Fig. 4. Wiedemann et al. (2010) performed an investigation on two pork supply chains in eastern Australia, located in Queensland and New South Wales regions. In Australia’s pork production farms, weaner pigs are weaned at 3 weeks, transferred into nursery housing for another 3 weeks, and then transferred to the grow-out housing at 6 weeks. A matured pig’s liveweight is averaged as 96.9 kg. The authors identified that the major resource use and environmental issues were associated with energy use and GHG emissions for both Australian pork supply chains. In the second case Verge et al. (2009) calculated the GHG emissions from the Canadian pork industry during the period of 1981e2001. In this study, the system consisted of weaners and finishers that were fed a ration consisting of grains and oilseed crops. They were housed in regular grow-out housing facilities. Greenhouse gas emissions of CO2, CH4 and N2O from animals,

79

farming facilities and the crops used as animal feed were included in the analysis. 3.1.3. Comparison of chicken production The results of the assessments of poultry farms were extracted from Katajajuuri et al. (2006) based on a broiler system in Finland; and Da Silva et al. (2010) based on a traditional industrial chicken farming system in Brazil. The generic system boundary for chicken production is illustrated in Fig. 5. A broiler chicken production network in Helsinki, Finland, consisting 20 farms, was studied by Katajajuuri et al. (2006) using farm data. The study included feed production and fertilization, transport of feed and chicks to broiler housing, the heating of broiler housing, and slaughtering. Greenhouse gas emissions from electricity consumption at the farm were calculated using average Finnish grid data. Da Silva et al. (2010) analyzed four farming systems from cradle to farm gate of chicken production. The inventory system was adopted from typical farming practices found in Brazil. The life cycle stages included crop production as feed, and ends at the slaughterhouse. 3.1.4. Comparison of rice and potatoes production Since Singapore imports the vast majority of rice from Thailand, only the study from Thailand was considered (Kasmaprapruet et al., 2009). In the system boundary, fields are plowed by a diesel powered tractor, followed by manual seeding. After a certain stage, the seedlings are replanted in prepared paddy fields. Ripe paddy plants are harvested and then sent to the mill for further processing. The husk and bran of the rice plants are removed by de-husking and milling. The final products are milled rice, husk, bran layer and broken rice. The system boundary starts at the seeding stage and ends at the finished rice product. Although Singapore imports its potatoes mainly from Indonesia and the U.S.A., detailed studies on the cultivation of potatoes from these countries were not available. In this article, it is assumed that the cultivation of potatoes in these countries is similar to those in

INPUT

OUTPUT

Fertilizers Pesticides Fuel, Electricity

CO2 CH4 N2O

Feed Production and Transport

Grow-out Housing Breeder Pig Unit

Transport to Finishing phase

Abattoir

Shipment to Singapore

Per kg carcass

Deep Litter Shed

Waste Management WEANED PIGS

MANURE

MATURE PIGS

FROZEN PORK

Fig. 4. Life cycle stages of pork production from cradle to gate (“gate” relates to per kg pork carcass delivered to Singapore).

80

M.Q.B. Tan et al. / Journal of Cleaner Production 72 (2014) 76e88

INPUT

OUTPUT

Fertilizers CO2

Pesticides Fuel, Electricity

CH4 N2O

Feed Production and Transport

FEED

Broiler or Chicken Production

Waste Management

MATURE CHICKENS

Abattoir

Shipment to Singapore

Per kg carcass

FROZEN CHICKENS

MANURE

Fig. 5. Life cycle stages of chicken production from cradle to gate (“gate” relates to per kg chicken carcass delivered to Singapore).

the UK, and a study from UK was used (Lillywhite et al., 2007). In the UK, potato seedbeds are prepared using ploughs and power harrows. Seed potatoes are planted, fertilized using nitrogen fertilizer and sprayed with fungicide and insecticide. Flailing is carried out to remove potato tops before they are harvested. The system boundary covers the cradle to farm gate per kg potatoes (Lillywhite et al., 2007). The system boundaries for rice and potatoes production are illustrated in Figs. 6 and 7 respectively. 4. Results and discussions The detailed compiled data for beef, pork and poultry are presented as Tables 1AeC respectively in the Appendix A. 4.1. Total carbon footprint of beef A breakdown of the total CF per kg carcass beef from Canada, U.S.A., Japan and Brazil is displayed in Fig. 8. The total CF in terms of kg CO2-eq/kg carcass weight, the results are: 26.81 (Canada); 36.03, 33.31 and 38.47 (U.S.A.); 26.31 (Japan); and 30.65 (Brazil). The percentage contributions to the total CF (as an average of the six case studies) are illustrated in Fig. 9. From Figs. 8 and 9, it is observed that the main contributor to GHG emissions from the production of beef is from enteric fermentation, which contributes to 36e61% of the total CF. Other significant contributors are from manure management and feed production. Due to different farming practices across various countries (e.g., Beauchemin et al., 2010; Pelletier et al., 2010; Ogino et al., 2004, 2007; Cederberg et al., 2009), the feed production stage displays some variability, with the least significant impact coming from Canada. Cattle that are fed roughage, such as hay and rice straw, have higher enteric methane emissions than those that are fed a greater proportion of concentrates, such as corn and wheat bran (Ogino et al., 2007). In the U.S.A. cattle farming system,

feedlot-finished cattle were fed mainly concentrates, which contributed to relatively lower global warming potential as compared to grazing practices (Pelletier et al., 2010). Out of the four different cattle production practices, feedlot-fed finishing systems turned out to have the lowest global warming impacts, while pasture finishing systems had the greatest impact. Feed production for feedlot systems is highly energy-intensive as it requires fuel for machinery, pesticide and fertilizer production and application, crop processing and transportation (Pelletier et al., 2010). Compared to cattle farms in the U.S.A., Japan beef production resulted in a total lower carbon footprint. The lower GHG emissions may be attributed to the practice of using a feedlot system for cows and calves in Japanese farms, instead of grazing systems. In addition to that, the feeding length of cattle during the finishing stage has an influence on the levels of greenhouse gas emissions. By shortening the feeding length by 2 months e from 28 to 26 months e the GHG emissions of cattle farming can be potentially reduced from 20.6 to 19.7 kg CO2-eq per kg of carcass weight (Ogino et al., 2004). In comparison to the farming stages for beef production, the meat processing activities and transport of products contribute insignificantly to GHG emissions, taking up an average of 5.4% and 3.5% respectively. Having cattle farms near or next to wheat or rice farms has the benefit of reducing GHG emissions in more than one way. Transport emissions are reduced for gathering feed for cattle, and wheat and rice straw can be collected in exchange for compost from cow manure. In this manner, the environmental impact of the cowecalf system could be reduced by 22.2e27.4%, while that for the feed production and transport stage could be reduced by 19.1e21.0% (Ogino et al., 2007). 4.2. Total carbon footprint of pork A breakdown of the total CF per kg carcass pork from Australia and Canada is displayed in Fig. 10. The total CF is 5.70 and 3.27 kg

M.Q.B. Tan et al. / Journal of Cleaner Production 72 (2014) 76e88

81

INPUT

OUTPUT

Fertilizers Pesticides Fuel, Electricity

CO2 CH4 N2O Replanting seedlings

Seedlings grow

Seeding

Growing

Harvesting

RIPE CROPS

SEEDLINGS

Transport to Singapore

Dehusking, Milling, Packing

Drying

Transportation of paddy plants to mill

MILLED RICE

Per kg product Fig. 6. Life cycle stages of rice production from cradle to gate (“gate” relates to per kg rice delivered to Singapore).

CO2/kg carcass weight for Australia; and 3.92 kg CO2/kg carcass weight for Canada. The percentage contributions to the total CF (averaged from the three cases) are illustrated in Fig. 11. It can be observed from Figs. 10 and 11 that manure management forms the largest contributor of GHG emissions. This farming

stage contributes between 25% and 64% of the total CF scores. The high level of GHG emissions generated is due to the way that manure is handled. In the Northern system of Australia’s pork production system, waste is directed into effluent treatment ponds, which emit a large amount of methane. For the Southern system,

INPUT

OUTPUT

Fertilizers Pesticides Fuel, Electricity

CO2 CH4 N2O

Preparation of seedbeds

Seeding

Flailing

Harvesting

RIPE POTATOES

Transport to Singapore

Packing

PACKED POTATOES

Per kg product Fig. 7. Life cycle stages of potatoes production from cradle to gate (“gate” relates to per kg potatoes delivered to Singapore).

82

M.Q.B. Tan et al. / Journal of Cleaner Production 72 (2014) 76e88

Total CF in kg CO2-eq per kg carcass weight of beef 40 Not Classified 35 Transport

kg CO2-eq

30 25

Meat Processing 20 Manure Management

15 10

Enteric Fermentation 5 Feed Production

0

Fig. 8. Total CF from different beef production systems (per kg carcass weight beef).

only manure from the breeder facility is sent to the effluent treatment ponds. Methane emissions from the effluent treatment ponds can be mitigated by covering the ponds and flaring the methane, which converts methane into biogenic carbon dioxide that has no net global warming potential (Wiedemann et al., 2010). An alternative measure is to keep storage tanks as cold as possible as the low temperatures would reduce bacteria activity. The liquid manure could also be used as fertilizer, or digested to produce biogas which can then be used for heating or power production (Verge et al., 2009). Additionally, the type of housing has an indirect impact on the GHG emissions from swine manure management. A better way to reduce the carbon footprint of pork production would be to use the deep litter housing system in place of conventional slatted floor housing which sends manure to effluent treatment plants. The former, which is used in the finishing stage of the Southern Australia pork production system, results in significant reductions in GWP. It is also observed from Figs. 10 and 11 that other significant contributors to the CF results are pig farming and feed production, which makes up 19% and 20.6% respectively on average. However, these two components do not show as much variability in absolute terms over different farming systems compared to manure management. Hence, carbon mitigation efforts should be directed

Percentage contributors of GHG emissions in beef production 5.4%

3.5%

at optimizing the manure management strategies. The contribution of GHG for pork meat processing and transportation is slightly higher than in the case for beef (refer to Figs. 8 and 9), which is an average of 10.1% and 5.6% respectively. 4.3. Total carbon footprint of chicken A breakdown of the total CF per kg carcass chicken from Finland and Brazil is displayed in Fig. 12. The total CF is 3.52 kg CO2/kg carcass weight (Finland) and 2.32e2.87 kg CO2/kg carcass weight for Brazil. The percentage contributions to the total CF (averaged from the three cases) are illustrated in Fig. 13. As observed from Figs. 12 and 13, the main contributors of GHG emissions in the production of chickens are feed production and poultry farming. They contributed between 37e53% and 19e39% of the total CF results, respectively. In Brazil, where chicken manure is used as fertilizer, emissions from feed production were reduced (Da Silva et al., 2010). Thus, in order to reduce the GHG emissions from the production of chickens, more efficient methods of producing feed can be employed. It can be highlighted that significant GHG emissions from chicken production in Finland are mainly due to increased energy use for heating (Katajajuuri et al., 2006). As the case of the other meat products, chicken processing contributes to the least GHG emissions, with a percentage contribution of 7.2% on average. Unlike the case for beef and pork, the transport component contributes more significantly to GHG emissions at 18.1% on average. This demonstrates that the total CF of poultry products imported from neighbouring countries can be reduced substantially by shorter distance transportation. 4.4. Total CF results comparing three meat products

26.4% 19.2%

Feed Production Enteric Fermentation Manure Management Meat Processing Transport

44.7% Fig. 9. Percentage GHG contributions in the life cycle stages of beef production (averaged from total CF).

A comparison across the different meat products e beef, pork and chicken e is displayed in Fig. 14 (per kg carcass) and Fig. 15 (per kg protein). From Figs. 14 and 15, it is highlighted that beef production emits far more CO2-eq units than that for pork or chicken production, whichever functional unit is used. This is due to the high emissions from enteric fermentation, classified in the above graphs as livestock production. While manure management formed the major part of the carbon footprint of pork production, it has an even higher contribution, in absolute terms, to GHG emissions from beef

M.Q.B. Tan et al. / Journal of Cleaner Production 72 (2014) 76e88

83

Total CF in kg CO2-eq per kg carcass weight of pork 6.0

5.0

Not Classified

kg-CO2-eq

4.0

Transport Meat Processing

3.0

Manure Management 2.0

Pig Production Feed Production

1.0

0.0 Australia (Northern)

Australia (Southern)

Canada

Fig. 10. Total CF from different pork production systems (per kg carcass weight pork).

production. In beef production, manure management released an average of 6.27 kg CO2-eq/kg carcass weight while that for pork released 2.24 kg CO2-eq/kg carcass weight. Taking the upper-bound of GWP for each category, on a per kg of protein basis, beef has a global warming potential that is 437% more than that for pork and 715% more than that for chicken. Thus, the consumption of beef has a much greater environmental impact than that for pork and chicken. For most meats, transport is not a major contributor of GHG emissions, taking up an average of 3.5% and 5.6% for beef and pork respectively. For chicken production, the percentage contribution of transport is much higher at 18%; this is because chicken production in itself is less GHG-intensive than the production of other meats. Thus, the GHG emissions associated with the choice of purchasing chicken at a grocery store can be reduced by choosing those that are imported from neighbouring countries, whereas that for pork and beef, options of meat purchase can be made by choosing those with less carbon-intensive farming practices. 4.5. Total carbon footprint of rice and potatoes The detailed compiled inventory data and results for rice and potatoes are presented in Table 1D in the Appendix A. The total CF of rice is 3.0 kg CO2-eq/kg rice and the percentage contribution to

Percentage contributors of GHG emissions in pork production

5.6%

the total CF is illustrated in Fig. 16. It was observed that the main contributor from the production of rice was the cultivation stage due to the use of fertilizers during the stage of rice cultivation (about 51%). The next greatest contributor is enteric emissions, as methane is produced by the anaerobic decomposition of organic matter in the flooded paddy fields, which contributed to 42% of the total CF. Compared to other types of crops, the transportation of rice does not contribute significantly to the GHG emissions of the product. This is because the amount of pollution from transport pales in comparison to those coming from the high emissions due to rice cultivation. Furthermore, the transportation distance between Thailand and Singapore is relatively minimal compared to the shipment distance of food products imported from other countries such as Europe and U.S.A. Therefore, for the case of rice, the GHG emissions from transportation are hardly noticeable. As for potatoes, the total CF is 0.47 (UK) and 0.39 (Australia) in kg CO2-eq/kg. A breakdown of the total CF results for potatoes production from Australia and UK is displayed in Fig. 17, and the average contribution to the total CF is illustrated in Fig. 18. In comparison to the meat sector (cattle and swine), the production of staples such as grains and potatoes is considerably less carbon-intensive. However, depending on the cultivation process, certain crops such as rice can emit large amounts of GHG, especially methane from paddy fields (Fig. 16). It should be highlighted, however, that the major source of emissions from the production of staple foods is transportation. The transportation of potatoes form a significant portion of the GHG emissions from the life cycle stages of potatoes (as shown in Fig. 17); this implies that by importing potatoes from a nearby location, the GHG emissions from the life cycle of the potato can be drastically reduced. 5. Further analysis

10.1%

20.6% Feed Production

5.1. Meat production and GHG emissions

Pig Production Manure Management

19.0% 44.7%

Meat Processing Transport

Fig. 11. Percentage GHG contributions in the life cycle stages of pork production (averaged from total CF).

The choice of consumers in dietary options can have an influence in environmental impacts (De Vries and De Boer, 2011). Fig. 19 presents (a) the comparison between the percentage contribution of meat consumed against the total CF of beef, pork and chicken; and (b) the prices of meat products vs. CF totals. As shown in Fig. 19(a), the relative contribution of the different types of meat to the annual per capita consumption in Singapore is compared against the associated GHG emissions. Although beef comprised only 7.4% of the total mass of meat consumed by

84

M.Q.B. Tan et al. / Journal of Cleaner Production 72 (2014) 76e88

Total CF in kg CO2-eq per kg carcass weight of chicken 4.0 3.5

kg CO 2-eq

3.0 2.5

Transport

2.0

Not Classified Meat Processing

1.5

Chicken Production

1.0

Feed Production

0.5 0.0

Finland

Brazil (southern system)

Brazil (central-west system)

Fig. 12. Total CF from different chicken production systems (per kg carcass weight chicken).

Singaporeans, it turned out to be the most significant in terms of associated GHG emissions. Singaporeans consume 4.2 kg of beef annually on a per capita basis (AVA, 2012). Based on a population size of 5 million people (Statistics Singapore, 2011), this will amount up to about 672 million kg of CO2-eq each year. It is also observed from Fig. 19(b) that while there is a vast difference in the GHG emissions per kg of meat produced between beef and the other two meats, this difference is not matched by variations in price. This could be an indication that the environmental costs of beef production are not factored into its retail price. In contrast, the difference in the price of pork and chicken does show a closer correlation to the difference in their environmental impacts. Due to high grain subsidies in some countries such as the U.S.A., the actual resource-intensity of producing meats is distorted as the livestock consumes large amounts of this subsidized feed (Bittman, 2008). As the beef industry is a major beneficiary of these grain subsidies, it is likely that beef is over-produced and over-consumed, which is of great detriment to the environment. 5.2. Staples and GHG emissions In terms of total CF per calorie of staples, the total CF translates from 3.0 kg CO2-eq/kg to 2.31 kg CO2-eq per 1000 kcal of rice; and

from 0.43 kg CO2-eq/kg (on average) to 0.56 kg CO2-eq per 1000 kcal of potatoes. When comparing the total CF per kg basis for both, rice production emits seven times more CO2-eq than potatoes. Half of the emissions from rice production are due to the release of CH4 from paddy fields during cultivation, while the bulk of emissions from potato production comes from fertilizer application. It would be difficult however, to encourage consumers in Singapore decrease their carbon-intensity of food choices by shifting consumption from rice to potatoes, since rice is a traditional staple food in many Asian countries, including Singapore. 5.3. Limitation of carbon footprint results While the results of the studies have shed some light on the GHG emissions associated with the life cycle production of five common food products consumed in Singapore, it is important to note several caveats that come with this information. LCA studies are modelled with system boundaries that are accompanied by many different parameters and hence the absolute values of the results bear some uncertainty. As it is extremely difficult to directly measure the GHG emissions from a farm, most studies were conducted using datasets which use mean emissions from the different stages (e.g., Beauchemin et al., 2010; Pelletier et al., 2010; Cederberg et al., 2009; Ogino et al., 2004, 2007). The results of each case study were highly influenced by the farming practices

Percentage contributors of GHG emissions in chicken production Total CF in kg CO2-eq per carcass weight 35

18.1%

47.2%

Feed Production

25

Chicken Production

20

Meat Processing Transport

27.6%

Transport kg CO2-eq

7.2%

30

Meat Processing Manure Management

15

Livestock Production 10

Feed Production

5 0 BEEF

Fig. 13. Percentage GHG contributions in the life cycle stages of chicken production (averaged from total CF).

PORK

CHICKEN

Fig. 14. Comparison of total CF results (averaged) per kg of carcass of beef, pork and chicken.

M.Q.B. Tan et al. / Journal of Cleaner Production 72 (2014) 76e88

Total CF in kg CO2-eq per kg protein 120 100

kg CO2-eq

80

Transport Meat Processing

60

Manure Management Livestock Production

40

Feed Production

20 0 BEEF

PORK

CHICKEN

Fig. 15. Comparison of total CF results (averaged) per kg protein of beef, pork and chicken.

adopted, along with the GHG emission factors that are used. We thus suggest that a common measure for determining the GHG factor, and the CO2-eq total, be used by applying the IPCC Fourth Assessment Report: Climate Change 2007 (IPCC, 2007). While comparing the carbon footprint of the five food items, care was taken to ensure that the system boundaries of the case studies were the same for each product and that the relevant activities were included, along with the associated GHG emissions in the production chains. Despite this caution, it should be highlighted that some differences for staples were inevitable. For example, the reports on the production of pork included GHG emissions from the activities involved in setting up housing for swine (Wiedemann et al., 2010) while reports for beef production excluded this activity from their analysis. While it is unlikely that this difference would cause any significant changes to the results and conclusions of the studies, it should be noted that these differences do exist. Although the studies are not strictly comparable in absolute terms, they are useful in shedding light on the environmental issues faced in a food production chain. 5.4. Land use Finally, the problem of land use change was not considered in the studies. It should be highlighted that in all the 15 studies investigated e six studies of beef; three of pork; three of chicken; one of rice; and two of potatoes e the sequestration of CO2 during the cultivation of fodder and crops, or carbon emissions during land clearing, were not accounted for. According to Phetteplace et al.

(2001), 0.12 tonnes of carbon could be sequestered per hectare per year by the improvement of pastures during the cowecalf phase of beef production, while 0.4 tonnes of carbon could be sequestered per hectare per year if pastures were converted to management-intensive grazing. This results in grass-finished beef having a carbon footprint that is 15% less than that for feedlotfinished beef. This is because the improved feed systems would result in a reduction of 1.8 kg CO2-eq for feedlot-finished beef, but a reduction of 8.2 kg CO2-eq for beef finished on intensively grazed improved pastures (Pelletier et al., 2010). In another study on U.S.A. beef cowecalf through finishing production systems, it was found that, by using a finishing system of intensive grazing rather than direct grazing or feedlot, about twice the amount of carbon dioxide could be sequestered, resulting in a lower carbon footprint for intensively grazed cattle (Johnson et al., 2003). On the downside, cattle grazing have been reported to be a major driver for global deforestation and significant carbon emissions; since livestock farming has been reported to be the cause of large-scale clearing of forest areas (McAlpine et al., 2009). This aspect is not included sufficiently in many studies of animal food products. Most ruminants (cattle, pigs and poultry), are associated with changes in land use change and are reported to be a major driver of deforestation (Nepstad et al., 2006). If land use change or deforestation were to be included in estimates of life cycle GHG emissions, the total CF of meats, especially beef, is expected to dramatically increase. According to Cederberg et al. (2011), the carbon footprint of beef produced on newly deforested land is estimated at more than an astounding 700 kg CO2-eq per kg carcass weight if direct land use emissions are annualized over 20 years. While land use change is widely acknowledged as a source of GHG emissions, the allocation of CO2 emission associated with deforestation, however, is complicated (Müller-Wenk and Brandão, 2010). This aspect in the food production chain is difficult to quantify and hence omitted from the carbon footprint analysis. 6. Prospects of carbon labelling and GHG emissions mitigation initiatives Carbon labelling e the use of labels bearing information about the GHG emissions associated with the production of a good e has been suggested by several authors as a means for the mitigation of GHG emissions as it would serve as education on climate change (Upham et al., 2010; Young et al., 2010). According to Boardman (2008), such a policy could have one or a few of the following effects: -

Contribution of CF results (in kg CO2-eq) for rice production 0.06 0.06

0.01

0.04

Enteric emissions Cultivation Harvesting 1.26

Seeding and milling Rice husk combustion Others

1.52

Transport

Total CF: 3.0 kg CO2-eq/kg rice NOTE: GHG emissions from rice husk combustion are negligible and omitted from the chart

Fig. 16. Contribution to the total CF of rice production.

85

-

To provide information so as to enable customers to choose a less carbon-intensive product To provide a way for organizations to publicly commit to reducing the embodied carbon in their products To encourage retailers to only put up less carbon-intensive products for sale

In this paper, the potential impact of carbon labelling on shifting consumer behaviour will be discussed. The prospects of shifting behaviours of retailers are not within the scope of the study. While consumers in Singapore are not able to directly control how livestock is produced or how crops are cultivated, importers have the options of choosing food items that are less carbon intensive. However, financial incentives for importers to choose ‘greener’ products are lacking in the country. As end-users, consumers have the power to shift their demand between substitute products. Thus, if consumers were to have greater demand for meats and staples that bear lower CF labels, and willing to pay

86

M.Q.B. Tan et al. / Journal of Cleaner Production 72 (2014) 76e88

Total CF in kg CO 2 -eq per potatoes 0.50

Transport 0.45

Not classified

0.40

kg CO 2-eq

0.35

Emissions from machines

0.30

Emissions from fossil fuel use

0.25 0.20

Emissions from post-harvest processing

0.15

Emissions from electricity use

0.10

Emissions from soil N2O

0.05

Emissions from agrochemicals

0.00

United Kingdom

Australia

Fig. 17. Total CF from different potatoes production systems (per kg product).

a premium for them, Singapore can reduce its food-related GHG emissions. In order to do this, however, consumers have to be aware of the negative externalities associated with consuming products with high CF scores. It may be possible for carbon labels to encourage consumers to change their dietary preferences, especially for environmentally conscious citizens. 6.1. Factors restricting the implementation of carbon labels First of all, the success of carbon labelling is restricted by how the general public perceives the numbers stated on carbon labels. In order for the carbon label to have any benefit on consumer behaviour, it must be easily understood. The CF totals for the same food product can vary substantially depending on farming practices as well as the country where the food items are produced. For example, the production of 1 kg of beef can result in the emission of between 26.8 and 38.5 kg CO2-eq. In pork production, a change in the manure management practices could potentially reduce the associated emissions for 1 kg of pork from 5.7 to 3.3 kg CO2-eq. As such, the number that is printed on the carbon label would be confusing for consumers, since the general public is not expected to be well-versed in LCA concepts and calculations. This confusion is expected to be increased manyfold if land use emissions are factored in.

Percentage contributors of GHG emissions in potatoes production 4%

4%

Emissions from agrochemicals 11%

Emissions from soil N2O 18%

Emissions from electricity use 22%

Emissions from post-harvest processing

Emissions from fossil fuel use 41%

Emissions from machines

NOTE: GHG emissions from transportation are negligible and omitted from the chart

Fig. 18. Percentage GHG contributions in the life cycle stages of potatoes production (averaged from total CF).

Secondly, as can be seen from Tesco’s experience with carbon labelling, life cycle greenhouse gas analysis is a tedious process which requires a large amount of resources. Thus, for carbon labelling to take hold, more efficient methods of data collection should be developed. Finally, by using a single indicator e carbon footprint or CF e to represent a complex reality, essential information about a product is left out. For example, carbon footprinting excludes other environmental sustainability indicators such as acid rain, eutrophication and water resources. Using the carbon footprint indicator as the sole basis for making consumption choices could, therefore, be restricted to global warming prevention but does not include other environmental impacts. 6.2. Factors encouraging the implementation of carbon labels While there have been scattered efforts to raise awareness about climate change and the role that food plays in GHG emissions, there exists a serious lack of awareness in many parts of the world, including Singapore. Therefore, carbon labels could serve as an educational tool and help consumers make informed choices (Ruviaro et al., 2011; Gössling et al., 2011). As with the case of nutritional labels, the factor of time will play a major role for people to understand the information placed on carbon labels, and may eventually shift their purchasing behaviours. A classical example in Singapore is the Green Label, which is a labelling scheme administered by the Singapore Environment Council (2012) to provide a seal of endorsement for products which claim to be environmentally friendly. It has been reported that products that are certified by the scheme have ‘become more marketable and readily accepted by consumers or business associates when making a purchase’ (Singapore Environment Council, 2012). Furthermore, Max Burger’s success in carbon labelling holds many lessons for other companies looking to follow suit. Just as Tesco cited the lack of a critical mass of products carrying carbon labels as a reason for the failure of its carbon labelling programme, Max Burger’s programme was successful precisely because its products were substitutes for each other, and the carbon labels allowed consumers to easily take into account the carbon emissions associated with each burger when making their purchasing decisions. This shows that the decisions of consumers can be influenced by carbon labels, but there needs to be enough products carrying

M.Q.B. Tan et al. / Journal of Cleaner Production 72 (2014) 76e88

Differences in Meat Prices and GHG Emissions

Percentage comparison between kg of meat consumed and GHG 35

emissions 7.41%

90% 80% 70%

42.61%

35.63%

60% 50% 27.60%

40% 30%

25 20

15.27 15

11.25 10

56.97%

5.86

20% 29.79%

10%

4.49

5

2.90

0

0% kg consumed Chicken

a

31.97

30 SGD/kg and kg CO2-eq

100%

87

Meat Prices

kg CO2-eq Pork

Beef

Beef

Comparison between the percentage contribution of meat

consumed vs. the CF

Pork

kg CO2-eq/kg meat

Chicken

b Comparison of meat prices and associated GHG emissions

Fig. 19. Comparisons of (a) percentage meat consumed and (b) prices against kg CO2-eq total/kg.

the labels in order for consumers to make that comparison in the first place. 6.3. Survey from consumers in Singapore A poll conducted by the National Climate Change Secretariat in Singapore on climate change issues showed that 86% ‘felt a sense of responsibility in dealing with climate change’, while 74% were ‘concerned about it’. Another 58% felt that Singapore should act to mitigate the effects of climate change even if the costs were high (National Climate Change Secretariat, 2012). A similar survey, specifically on carbon labels, was undertaken by the authors in early 2012. The survey was conducted online, using the Google Documents online tool, and distributed via email and social media such as facebook and twitter. Of the 194 survey respondents, who replied over a one month period, 54% were female and 46% were male. The vast majority e 78% e belong to the 15e35 age group, while 20% were over 50 years old. The remaining 2% of respondents were from the 36 to 50 age group. 76% of respondents indicated that they would make a conscious effort to check the labels before making purchasing decisions, if carbon labelling were to be implemented. When asked about current schemes such as the Green Label scheme, which applies to most products except for food, drinks and pharmaceuticals, and the Energy Label scheme for electrical appliances, 68% of respondents said that they choose to buy certified-green products by looking at the green or energy label, while only a small percentage e 13% e said that they were unaware of such schemes. These results are encouraging for Singapore’s fight against climate change, as it shows that Singaporeans see themselves as part of the solution to mitigating climate change. 6.3.1. Limitations of study Both organic food consumers and purchasers of carbon-labelled food products are associated with values such as altruism, ecologymindedness, and universalism. As such, some lessons from organic food labelling can be applied to carbon labelling (Roos and Tjarnemo, 2011). An investigation conducted by Aertsens et al. (2009) concluded that organic food buyers are positively and significantly correlated with their organic dietary lifestyles. However, an attitudeebehaviour gap exists in other types of

labelling schemes e having the intention to purchase did not necessarily result in an actual purchase. In a survey conducted by Magnusson et al. (2001), only 4e10 percent indicated that they were very likely to buy organic products despite displaying positive attitudes towards environmentally friendly consumption. A possible cause for this discrepancy is that, even though the majority of consumers recognize the environment labels, they may not know or understand what the labels indicate. Furthermore, environment labelled products are perceived to be more expensive. Since food purchases are guided largely by habit, consumers might not change their purchasing decisions even if the labelled products are cheaper, as this information would be overlooked (Leire and Thidelle, 2005). Additionally, price, quality, convenience and brand familiarity are still the most important criteria in making purchasing decisions, resulting in labelled products being ignored (Vermeir and Verbeke, 2006). This is not surprising, given that the buying of food is a low involvement process where one would likely apply an automated rather than a reasoned cognitive process. One of the suggestions to decrease the attitudeebehaviour gap is to increase the availability of labelled products while improving awareness and knowledge regarding labelling programmes (Aertsens et al., 2009). 7. Concluding remarks In order to encourage people to make more informed choices when making grocery purchases, it is imperative that they have access to environmental performance information of products. Carbon labels serve the purpose of informing consumers about how purchasing different products within a substitutable range would have differing impacts on the environment. However, before any forms of carbon labels can be implemented, a method has to be employed to analyze and measure the life cycle emissions generated from the production of products. The carbon footprint (CF) of food exports from the following was compared: beef from Canada, three from the U.S.A., and two more from Japan and Brazil respectively; three studies of pork from northern and southern Australia and Canada; three case studies of chicken from south and central Brazil and Finland; one of rice from Thailand; and finally, two investigations of potatoes from UK and

88

M.Q.B. Tan et al. / Journal of Cleaner Production 72 (2014) 76e88

Australia. From these studies, there is a general consensus that beef is the most carbon-intensive meat product (average 32.0 kg CO2-eq/kg carcass weight), followed by pork (4.5 kg CO2-eq/kg carcass weight), and finally, chicken (average 2.9 kg CO2-eq/kg carcass weight). For staples, rice has a total CF that is roughly seven times that of potatoes when weight is used as a functional unit (3.0 kg CO2-eq/kg rice vs. 0.43 kg CO2-eq/kg potatoes). The choice of functional unit of GHG emissions has important implications for the interpretation of results (e.g., per carcass weight, per protein or per 1000 kcal). The quantitative investigations of the different systems and results from each case are not straightforward due to different farming practices within a defined system. However, a life cycle approach remains as an important GHG accounting tool for assessing the CF of products and for providing quantified information for use in carbon labels. We suggest that carbon footprint calculations must be standardized to avoid giving misleading information to policy makers, retailers, and consumers. In conclusion, consumers are now demanding more information about how their consumption patterns have an impact on the environment (Young et al., 2010; Upham et al., 2010). In order to keep up with the competition, it is in the interest of producers and retailers to provide guidance to their customers on the impact of their purchases on GHG emissions. Thus, despite the difficulties in accounting and reporting accurate emission figures, the implementation of carbon labelling is likely to have positive effects on the environment. Apart from the preliminary survey conducted, further investigations should be made on Singapore’s consumer patterns to determine in greater detail the nation’s response towards low carbon emission products. Raising awareness regarding the rationale behind carbon labelling schemes will also help encourage the nation’s fight against climate change. Appendix A. Supplementary material Supplementary material associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.jclepro. 2012.09.035. References Aertsens, J., Verbeke, W., Mondelaers, K., Huylenbroeck, G.V., 2009. Personal determinants of organic food consumption: a review. Br. Food J., 1140e1167. AVA, 2012. Agri-food and Veterinary Authority of Singapore e Statistics. http:// www.ava.gov.sg/Publications/Statistics/. American Meat Institute, 2009. http://www.meatami.com/. Beauchemin, K.A., Janzen, H.H., Shannan, M.L., McAllister, T.A., McGinn, S.M., 2010. Life cycle assessment of greenhouse gas emissions from beef production in western Canada: a case study. Agric. Syst. 103, 371e379. Bittman, M., 27 January 2008. Rethinking the meat-guzzler. The New York Times. Boardman, B., 2008. Carbon labelling: too complex or will it transform our buying? Significance, 168e171. Cederberg, C., Meyer, D., Flysio, A., 2009. Life Cycle Inventory of Greenhouse Gas Emissions and Use of Land and Energy in Brazilian Beef Production. The Swedish Institute for Food and Biotechnology. Cederberg, C., Persson, U.M., Neovius, K., Molander, S., Clift, R., 2011. Including carbon emissions from deforestation in the carbon footprint of Brazilian beef. Environ. Sci. Technol. 45, 1773e1779. Da Silva, V.P., van der Werf, H., Soares, S.R., 2010. LCA of French and Brazilian broiler poultry production systems. In: Notarnicola, B., Settanni, E., Tassielli, G., Giungato, P. (Eds.), Proceedings of the 7th Intl. Conference on LCA in the Agrifood Sector. Bari, Italy, pp. 475e480. De Vries, M., De Boer, I.J.M., 2011. Comparing environmental impacts for livestock products: a review of life cycle assessments. Livestock Sci. 128, 1e11. FAO, 2006. Livestock’s Long Shadow: Environmental Issues and Options. Food and Agriculture Organization of the United Nations, Rome. Fiala, N., 2008. Meeting the demand: an estimation of potential future greenhouse gas emissions from meat production. Ecol. Econ. 67, 412e419. Gadema, Z., Oglethorpe, D., 2011. The use and usefulness of carbon labelling food: a policy perspective from a survey of UK supermarket shoppers. Food Policy 36, 815e822. Gössling, S., Garrod, B., Aall, C., Hille, J., Peters, P., 2011. Food management in tourism: reducing tourism’s carbon ‘foodprint’. Tourism Manage. 32, 534e543.

IPCC, 2007. IPCC Fourth Assessment Report: Climate Change 2007. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA. Johnson, D.E., Phetteplace, H.W., Seidl, A.F., Schneider, U.A., McCarl, B.A., 2003. Management variations for US beef production systems: effects on greenhouse gas emissions and profitability. In: Proceedings of the 3rd Intl. Methane & Nitrous Oxide Mitigation Conf., Beijing, China, pp. 953e961. Kasmaprapruet, S., Paengjuntuek, W., Saikhwan, P., Phungrassami, H., 2009. Life cycle assessment of milled rice production: case study in Thailand. Eur. J. Sci. Res. 30, 195e203. Katajajuuri, J.M., Grönroos, J., Usva, K., Virtanen, Y.E., Sipilä, Venäläinen, E., Kurppa, S., Tanskanen, T.M., Virtanen, H., 2006. Broilerin Fileesuikaleiden Tuotannon Ymparistovaikutukset ja Kehittamismahdollisuudet (Environmental Impacts and Improvement Options of Sliced Broiler Fillet Production). MTT Agrifood Research, Finland. Kramer, K., Moll, H.C., Nonhebel, S., Wilting, H.C., 1999. Greenhouse gas emissions related to Dutch food consumption. Energy Policy 27, 203e216. Leire, C., Thidelle, A., 2005. Product-related environmental information to guide consumer purchases e a review and analysis of research on perceptions, understanding and use among Nordic consumers. J. Clean. Prod., 1061e1070. Lillywhite, R., Chandler, D., Grant, W., Lewis, K., Firth, C., Schmutz, U., Halpin, D., 2007. Environmental Footprint and Sustainability of Horticulture Including Potatoes e a Comparison with Other Agricultural Sectors. Department for Environment, Food and Rural Affairs (DEFRA), UK. Magnusson, M.K., Arvola, A., Hursti, U.-K.K., Aberg, L., Sjoden, P.-O., 2001. Attitudes towards organic foods among Swedish consumers. Br. Food J., 209e226. McAlpine, C.A., Etter, A., Fearnside, P.M., Seabrook, L., Laurance, W.F., 2009. Increasing world consumption of beef as a driver of regional and global change: a call for policy action based on evidence from Queensland (Australia), Colombia and Brazil. Glob. Environ. Change 19, 21e33. Mouron, P., Scholz, R.W., Nemecek, T., Weber, O., 2006. Life cycle management on Swiss fruit farms: relating environmental and income indicators for applegrowing. Ecol. Econ. 58, 561e578. Müller-Wenk, R., Brandão, M., 2010. Climatic impact of land use in LCA e carbon transfers between vegetation/soil and air. Int. J. LCA 15, 172e182. National Climate Change Secretariat, 2012. Climate Change & Singapore: Challenges. Opportunities. Partnerships. National Climate Change Secretariat, Singapore. Nepstad, D.C., Sticker, C.M., Almeida, O.T., 2006. Globalization of the Amazon soy and beef industries: opportunities for conservation. Conserv. Biol. 20, 1595e1603. Ogino, A., Kaku, K., Osada, T., Shimada, K., 2004. Environmental impacts of the Japanese beef-fattening system with different feeding lengths as evaluated by a life-cycle assessment method. Anim. Sci. 82, 2115e2122. Ogino, A., Orito, H., Shimada, K., Hirooka, H., 2007. Evaluating environmental impacts of the Japanese beef cowecalf system by the life cycle assessment method. Anim. Sci. J. 78, 424e432. Pelletier, N., Pirog, R., Rasmussen, R., 2010. Comparative life cycle environmental impacts of three beef production strategies in the Upper Midwestern United States. Agric. Syst. 103, 380e389. Phetteplace, H.W., Johnson, D.E., Seidl, A.F., 2001. Greenhouse gas emissions from simulated beef and dairy livestock systems in the United States. Nutr. Cycl. Agroecosyst. 60, 99e102. Reuters, 22 December 2011. U.S. beef consumption continues to decline. Chicago Tribune. http://www.chicagotribune.com/business/breaking/chi-usbeef-consumption-continues-decline-20111222,0,708055.story. Roos, E., Tjarnemo, H., 2011. Challenges of carbon labelling of food products: a consumer research perspective. Br. Food J., 982e996. Ruviaro, C.F., Gianezini, M., Brandão, F.S., Winck, C.A., Dewes, H., 2011. Life cycle assessment in Brazilian agriculture facing worldwide trends. J. Clean. Prod., 1e16. Saunders, C.M., Guenther, M., Tait, P.R., Kaye-Blake, W., Saunders, J., Miller, S., Abell, W., 2011. Consumer attitudes towards sustainability attributes on food labels in the UK and Japan. In: 85th Annual Conference, April 18e20, 2011, Warwick University, Coventry, UK. Agricultural Economics Society. Singapore Environment Council, 2012. Green Label. http://www.greenlabel.sg/ about. Statistics Singapore, 2011. Key Annual Indicators. Government of Singapore. http:// www.singstat.gov.sg/. United Nations, 2005. World Population Prospectus. UN Department of Economic and Social Affairs, New York. Upham, P., Dendler, L., Bleda, M., 2010. Carbon labelling of grocery products: public perceptions and potential emissions reductions. J. Clean. Prod. 19, 348e355. Van Gilder Cooke, S., 2012. Why going green can mean Big Money for fast-food chains. Time: World. http://www.time.com/time/world/article/ 0,8599,2111372,00.html#ixzz27MoBSVHm. Verge, X.P., Dyer, J.A., Desardins, R.L., Worth, D., 2009. Greenhouse gas emissions from the Canadian pork industry. Livestock Sci. 121, 92e101. Vermeir, I., Verbeke, W., 2006. Sustainable food consumption: exploring the consumer ‘attitudeebehavioral intention’ gap. J. Agric. Environ. Ethics, 169e194. Wiedemann, S., McGaham, E., Grist, S., Grant, T., 2010. Environmental Assessment of Two Pork Supply Chains Using Life Cycle Assessment. Rural Industries Research and Development Corporation. RIRDC Publication No 09/176, Australian Government. Young, W., Hwang, K., McDonald, S., Oates, C.J., 2010. Sustainable consumption: green consumer behavior when purchasing products. Sustain. Dev. 1, 20e31.