Systematic Review Of Greenhouse Gas Emissions For Different Fresh Food Categories

Abstract

This paper presents the results of a systematic literature review of greenhouse gas emissions for different food categories from life cycle assessment (LCA) studies, to enable streamline calculations that could inform dietary choice. The motivation for completing the paper was the inadequate synthesis of food greenhouse gas emissions available in the public domain. The paper reviewed 369 published studies that provided 1,718 global warming potential (GWP) values for 168 varieties of fresh produce. A meta-analysis of the LCA studies was completed for the following categories: fresh vegetables (root vegetables, brassica, leaves and stems); fresh fruits, (pepo, hesperidium, true berries, pomes, aggregates fruits and drupes); staples (grains, legumes, nuts, seeds and rice); dairy (almond/coconut milk, soy milk, dairy milk, butter and cheese); non-ruminant livestock (chicken, fish, pork); and ruminant livestock (lamb and beef). The meta-analysis indicates a clear greenhouse gas hierarchy emerging across the food categories, with grains, fruit and vegetables having the lowest impact and meat from ruminants having the highest impact. The meta-analysis presents the median, mean, standard deviation, upper and lower quartile, minimum and maximum results for each food category. The resultant data enables streamline calculations of the global warming potential of human diets, and is illustrated by a short case study of an Australian family’s weekly shop. The database is provided in the Appendix as a resource for practitioners. The paper concludes with recommendations for future LCA studies to focus upon with respect to content and approach.

Generated Summary

This meta-analysis investigates the greenhouse gas (GHG) emissions associated with various food categories derived from 369 published Life Cycle Assessment (LCA) studies. The primary aim is to synthesize existing data to enable streamlined calculations for dietary choices, thereby informing consumers about the environmental impacts of different food products. The study employs a systematic literature review methodology, adhering to the PRISMA Statement protocol to minimize bias and enhance scientific validity. Data were collated under broad category headings, including fresh fruits, vegetables, staples, dairy, non-ruminant livestock, and ruminant livestock. Furthermore, information on the LCA methods, geographical locations, and unique product descriptors was compiled. The study focuses on converting the recorded Global Warming Potential (GWP) values into a common functional unit to facilitate comparisons across different food types and production systems. The meta-analysis aims to produce a database that enables the streamlined accounting for individual meals, diets, catering organizations, and nations.

Key Findings & Statistics

The meta-analysis examined 369 published LCA studies, providing 1,718 GWP values for fresh produce. The data included 192 journal papers, 80 conference papers, 64 reports, and 29 web-based EPDs. The majority of GWP values (58%) were from the last five years. The study converted each GWP value into a common functional unit of kg CO2-eq /kg bone free meat (BFM) or produce, at the regional distribution centre (RDC).

• The lowest median GWP values were for field-grown vegetables (0.37 kg CO2-eq/kg), field-grown fruit (0.42 kg CO2-eq/kg), cereals (except rice) and pulses (0.50-0.51 kg CO2-eq/kg).
• Slightly higher values for tree nuts were found (1.20 kg CO2-eq/kg).
• Rice had the highest impact of the plant-based field-grown crops (2.55 kg CO2-eq/kg).
• Non-ruminant livestock had medium GWP values in fish (3.49 kg kg CO2-eq/kg BFM), chicken (3.65 kg CO2-eq/kg BFM) and pork (5.77 kg CO2-eq/kg BFM).
• Dairy products (cheese) and butter also shared a medium GWP values.
• Ruminant livestock in lamb (25.58 kg CO2-eq/kg BFM) and beef (26.61 kg CO2-eq/kg BFM) had the highest median GWP values.
• The meta-analysis identified a clear greenhouse gas hierarchy emerging across the food categories, with grains, fruit, and vegetables having the lowest impact and meat from ruminants having the highest impact.

Table 4 Summary of GWP values (kg CO2-eq/kg produce or bone free meat) across broad food categories:

• Vegetables (all field-grown vegetable): Median: 0.37, Mean: 0.47, Stdev: 0.39, Min: 0.04, Max: 2.54, Q1: 0.19, Q3: 0.60, No. of Studies: 33, No. of GWP values: 140.
• Fruits (all field grown fruit): Median: 0.42, Mean: 0.50, Stdev: 0.32, Min: 0.08, Max: 1.78, Q1: 0.28, Q3: 0.63, No. of Studies: 77, No. of GWP values: 250.
• Cereals: Median: 0.50, Mean: 0.53, Stdev: 0.22, Min: 0.11, Max: 1.38, Q1: 0.38, Q3: 0.63, No. of Studies: 31, No. of GWP values: 90.
• Legumes and Pulses: Median: 0.51, Mean: 0.66, Stdev: 0.45, Min: 0.15, Max: 2.46, Q1: 0.36, Q3: 0.83, No. of Studies: 16, No. of GWP values: 51.
• Passive greenhouse fruit and vegetable: Median: 1.10, Mean: 1.02, Stdev: 0.49, Min: 0.32, Max: 1.94, Q1: 0.54, Q3: 1.35, No. of Studies: 5, No. of GWP values: 15.
• Tree nuts combined: Median: 1.20, Mean: 1.42, Stdev: 0.93, Min: 0.43, Max: 3.77, Q1: 0.61, Q3: 2.13, No. of Studies: 7, No. of GWP values: 21.
• Milk world average: Median: 1.29, Mean: 1.39, Stdev: 0.58, Min: 0.54, Max: 7.50, Q1: 1.14, Q3: 1.50, No. of Studies: 77, No. of GWP values: 262.
• Heated greenhouse fruit and vegetable: Median: 2.13, Mean: 2.81, Stdev: 1.61, Min: 0.84, Max: 7.4, Q1: 1.74, Q3: 3.7, No. of Studies: 18, No. of GWP values: 53.
• Rice: Median: 2.55, Mean: 2.66, Stdev: 1.29, Min: 0.66, Max: 5.69, Q1: 1.64, Q3: 3.08, No. of Studies: 12, No. of GWP values: 27.
• Eggs: Median: 3.46, Mean: 3.39, Stdev: 1.21, Min: 1.30, Max: 6.00, Q1: 2.45, Q3: 4.05, No. of Studies: 19, No. of GWP values: 38.
• Fish: all species combined: Median: 3.49, Mean: 4.41, Stdev: 3.62, Min: 0.78, Max: 20.86, Q1: 1.99, Q3: 5.16, No. of Studies: 47, No. of GWP values: 148.
• Chicken: Median: 3.65, Mean: 4.12, Stdev: 1.72, Min: 1.06, Max: 9.98, Q1: 2.77, Q3: 5.31, No. of Studies: 29, No. of GWP values: 95.
• Cream: Median: 5.64, Mean: 5.32, Stdev: 1.62, Min: 2.10, Max: 7.92, Q1: 3.82, Q3: 7.14, No. of Studies: 3, No. of GWP values: 4.
• Pork: world average: Median: 5.77, Mean: 5.85, Stdev: 1.63, Min: 3.20, Max: 11.86, Q1: 4.50, Q3: 6.59, No. of Studies: 38, No. of GWP values: 130.
• Prawns/shrimp: Median: 7.80, Mean: 14.85, Stdev: 12.37, Min: 5.25, Max: 38.00, Q1: 6.76, Q3: 20.20, No. of Studies: 7, No. of GWP values: 11.
• Cheese: Median: 8.55, Mean: 8.86, Stdev: 2.07, Min: 5.33, Max: 16.35, Q1: 7.79, Q3: 9.58, No. of Studies: 22, No. of GWP values: 38.
• Butter: Median: 9.25, Mean: 11.52, Stdev: 7.37, Min: 3.70, Max: 25.00, Q1: 7.28, Q3: 12.41, No. of Studies: 4, No. of GWP values: 8.
• Lamb: world average: Median: 25.58, Mean: 27.91, Stdev: 11.93, Min: 10.05, Max: 56.70, Q1: 17.61, Q3: 33.85, No. of Studies: 22, No. of GWP values: 56.
• Beef: world average: Median: 26.61, Mean: 28.73, Stdev: 12.47, Min: 10.74, Max: 109.5, Q1: 22.26, Q3: 31.57, No. of Studies: 49, No. of GWP values: 165.

Table 5 provides the GWP values for individual foods from lowest to highest median values, with a breakdown of the median, mean, standard deviation, and range for each food category.

• The non-ruminant livestock category analysed LCA studies including fish, poultry and pork. 108 LCA studies were reviewed resulting in 446 GWP values.
• The analysis of fish was further broken down between different species of fish. Within the results of specific species of fish, pilchards, pollock, carp, herring and mackerel presented low GWP values.
• Chicken (3.65 kg CO2-eq/kg BFM) and eggs (3.46 kg CO2-eq/kg eggs) displayed similar median values.
• The world average for beef was 26.61 kg CO2-eq/kg BFM. When geographic locations were compared for beef, South American beef had the highest median greenhouse gas profile with 34.10 kg CO2-eq/kg BFM. Australian beef had the lowest median greenhouse gas profile with 22.88 kg CO2-eq/kg BFM.

Table 6, 7, 8, 9 show GWP values of vegetables, fruits, staples, fish, poultry, pork, lamb, beef, and dairy. These tables list the median and mean GWP values and standard deviations.

Other Important Findings

• The meta-analysis confirms the GWP hierarchy between broad food categories.
• There is an unequal representation of food types in LCA studies.
• Variations in data within each category may be attributable to different LCA approaches.

Limitations Noted in the Document

• The use of common functional units in food LCAs would make it easier to compare reports, avoid misrepresentation and strengthen the validity of comparisons.
• The study acknowledges that consumption, and end-of-life management of food and packaging will alter results.
• The meta-analysis is limited by the variety of methods and the varying quality of data from different studies.
• The study acknowledges that consumption, and end-of-life management of food and packaging will alter results. Of particular note is the 30% of purchased food that is not eaten, with a potential causal relationship to packaging design (see for example Wikström et al. 2014). This remains outside the scope of the study.

Conclusion

The meta-analysis identified a clear hierarchy within the GWP values, with grains, fruits, and vegetables having the lowest impact, while meat from ruminants had the highest. The median results could be used with confidence to provide a streamlined estimate of the impact of ingredients for dietary choice or menu planning, particularly by selecting food from differing categories. This is illustrated by a short case study in section 4.3 of an Australian family’s weekly shop. The findings highlight the importance of the GWP values within the fruits and vegetables, seeds, and nuts compared to livestock. It also stresses the need for greater consistency in reporting in environmental product declarations and developing product category rules to improve comparability in food LCAs. Future LCA studies should focus on underrepresented food types, common functional units, and standardized system boundaries to improve comparability and inform dietary choices for environmental purposes. The authors’ position is that the variation in farming methods and conditions has a more significant impact on the presented GWP values than methodological choices or publication type. The meta-analysis provides a clear message and presents generalizable findings that should not be dismissed because of methodological limitations, contributing to the development of streamlined tools to rapidly calculate the GWP of differing diets. The primary recommendation is the need to expand future LCA studies in order to accurately represent underrepresented food types as the market shifts towards sustainable protein sources. The results from the three scenarios show a large variation in the median GWP per week. Substituting ruminant meat (beef and lamb) for non-ruminant meat (kangaroo, duck and rabbit), and selecting an alternate fish species in pollock (Table 13B) produces an estimate 30% reduction in GWP in relation to the median weekly shop in Fig 11. An alternate diet that attempts to match the recommended weekly protein intake via a plant and fish based diet (Table 13C) produces an estimate 52% reduction in GWP related to the median weekly shop shown in Fig 11. The main takeaway is that consumer choice has a significant impact on the GWP.

DOI

10.1016/j.jclepro.2016.04.082