Food systems contribute significantly to global greenhouse gas (GHG) emissions (21–37% of global anthropogenic GHGs). Current national emissions accounting primarily uses a production-based approach, considering only emissions within a country's borders. However, increasing international food trade means that consumption patterns significantly influence a nation's actual GHG contribution. This study addresses this knowledge gap by examining trade-adjusted agricultural emissions. The researchers aim to quantify the discrepancies between production-based and trade-adjusted accounting methods across various geographical scales and investigate how different specifications of trade adjustments affect the results. The importance lies in informing the development of more effective national emissions reduction targets that consider the global implications of food trade and consumption patterns. Existing studies on trade-adjusted emissions have limitations: some focus solely on producer or consumer perspectives, neglecting intermediary trading countries; others lack a broad overview of food-related agricultural emissions at the national level. This study aims to fill these gaps and enhance the accuracy of national GHG inventories.

Numerous studies have explored the contribution of food systems to climate change and the limitations of production-based emission accounting. Previous research has estimated the discrepancies between production-based and consumption-based (trade-adjusted) accounting for various sectors. However, a comprehensive analysis of trade-adjusted agricultural emissions at global, regional, and national levels remains scarce. Existing studies on trade-adjusted approaches in the agricultural sector often overlook the role of intermediary trading countries in emissions accounting. This paper builds upon prior research by providing a more comprehensive analysis that incorporates a wider range of food items and considers the diverse roles of producer, consumer, and intermediary nations.

The researchers utilized data from the FAOSTAT database (Food and Agriculture Organization Statistical Database) covering 1986-2017. This database provided comprehensive data on agricultural production, trade flows (including bilateral trade data from the Detailed Trade Matrix dataset), and emission intensities. The study employed a bilateral trade input-output (BTIO) approach to adjust agricultural production emissions for trade. Trade-adjusted agricultural emissions (TAE) were calculated as the sum of production-based emissions (PBE) plus import emissions minus export emissions. The analysis included three sensitivity analyses to assess the robustness of the main findings. The first sensitivity analysis replaced regional emission intensities with global ones for non-producer countries. The second analysis adopted a technology-adjusted approach suggested by Kander et al. (2015), accounting for differences in carbon efficiencies across countries. The third analysis incorporated emission intensities associated with both production and imports to account for re-exports. Additionally, the study examined agricultural land-use emissions using FAOSTAT data, adjusting them for trade flows. Data preprocessing involved converting processed food items to their primary equivalent mass using caloric content ratios from FAO and USDA sources. Emission intensities for a category called 'others' (containing various fruits, vegetables, and oilcrops) were calculated using a residual method, subtracting emission from the 14 main categories from the total. The study used three-year averages for most analyses unless otherwise specified.

Global trade-adjusted agricultural emissions (TAE) increased in absolute terms from 3.86 Gt CO₂/yr in 1987 to 5.02 Gt CO₂e/yr in 2015, mirroring the rise in global agricultural production. However, per capita TAE decreased during this period due to improvements in agricultural efficiency. Regional trends showed increases in most regions, except Europe, Oceania, and the Former Soviet Union, which experienced declines due to factors such as increased productivity and economic downturns. At the country level, major agricultural importers such as city-states and Middle Eastern countries showed substantial differences between TAE and PBE, with TAEs being significantly higher. Conversely, major agricultural exporters like Australia and New Zealand had lower TAEs than PBEs. Global emissions embodied in food exports and imports increased significantly over time. Ruminant meat and milk products were major contributors to embodied emissions in most regions, while paddy rice was dominant in Asia. In recent years, the 'others' food group (including soybeans) saw an increasing share of emissions, reflecting the growing demand for feed for animal-source food. The sensitivity analyses showed that the main findings were relatively robust across alternative specifications. Differences between the original approach and the sensitivity analyses were minor except for a few small, trade-dependent countries and when using the technology-adjusted approach. The analysis of agricultural land-use emissions revealed substantial differences between trade-adjusted and unadjusted values, particularly for major exporters (decreases) and importers (increases). The findings demonstrate the importance of considering trade adjustments to improve the accuracy and policy relevance of agricultural emissions accounting.

The findings highlight the significant impact of international trade on agricultural emissions, revealing substantial differences between production-based and consumption-based accounting. The large discrepancies, particularly for major importers and exporters, underscore the need to move beyond a solely production-based approach. The inclusion of intermediary trading countries in emissions accounting is crucial, as their role in emissions reduction should not be overlooked. The three sensitivity analyses support the robustness of the main findings. The study challenges the notion that wealthier countries always have higher per capita emissions; several low-income and emerging economies exhibited high per capita TAEs, largely attributed to consumption of emission-intensive food items or less efficient agricultural practices. The results have implications for national emission reduction targets under the Paris Agreement, highlighting the need to integrate trade-adjusted approaches for effective climate change mitigation. The study also points to the major contribution of animal-source foods, particularly ruminant meat and milk products, to emissions, supporting the argument for dietary shifts towards less emission-intensive foods.

This research demonstrates that adjusting agricultural emissions for trade is crucial for accurate climate change mitigation strategies. The significant discrepancies between production-based and trade-adjusted emissions highlight the limitations of solely production-based accounting in a globalized food system. The study's methodology, incorporating a BTIO approach and sensitivity analyses, provides a robust framework for improving national GHG inventories and informing policy decisions. Future research could explore the implications of trade adjustments for nutrient distribution and analyze the effects of specific trade agreements on embodied emissions.

The study's reliance on FAOSTAT data introduces certain limitations. The lack of detailed emission intensities for some food items, grouped under 'others,' might influence the accuracy of overall emissions estimates. The study mainly focuses on farm-gate emissions, excluding pre- and post-production processes like land-use change (though a sensitivity analysis addresses this partially) and transportation emissions. Assuming equal contributions of all items to land-use change for simplicity also limits the precision. The use of the BTIO approach, while transparent and suitable for bilateral agreements, might not capture the complexities of multi-regional trade flows as accurately as MRIO approaches.