A framework to quantify mass flow and assess food loss and waste in the US food supply chain

Human food activities are resource-intensive and contribute significantly to environmental problems. The U.S. uses a substantial portion of its resources for food production, and a significant amount of food is wasted. Reducing food loss and waste (FLW) is crucial for improving the efficiency of the food supply chain (FSC), enhancing sustainability, and improving food security. Existing studies on U.S. FLW quantification vary significantly in scope, data sources, and methodologies, leading to discrepancies in estimations. This study aims to provide a comprehensive and complete accounting of FLW in the U.S. by developing a consolidated framework for quantifying FLW, considering a life cycle approach and breaking down mass flows by FSC stages and major food commodity groups. The framework also quantifies FLW distribution across different management pathways to inform the development of improved FLW recycling and recovery practices. This study addresses the limitations of previous research by considering all five stages of the FSC (on-farm production, manufacturing, distribution, wholesale & retail, and consumption), incorporating both edible and inedible portions of FLW, and explicitly defining the scope of waste factors. The comprehensive analysis will provide a baseline for designing effective strategies to reduce, recycle, and recover FLW in the US.

Several studies have attempted to quantify U.S. FLW, but significant discrepancies exist due to varying definitions of FLW, data sources, scope, supply chain coverage, and accounting methods. Many studies focused only on the warehouse and retail (W&R) and consumption stages, excluded inedible parts, and did not quantify FLW management. Other studies included more stages but used waste factors with undefined scopes and did not compare results to actual mass flows. This lack of a complete and comprehensive accounting of U.S. FLW highlights the need for this study's consolidated framework, which addresses these limitations.

This study quantifies mass flow and FLW generation along five U.S. FSC stages using Mass Flow Analysis (MFA). Food is defined as any product intended for human consumption, excluding sodas and alcoholic beverages. The study focuses on agricultural materials and excludes other inputs. FLW is defined as food intended for human consumption that leaves the FSC, including edible and inedible parts, regardless of destination. Food products are categorized into 10 commodity groups. Nine FLW management pathways are considered, aggregated into four groups based on the EPA's Food Recovery Hierarchy: food donation, recycling (animal feed, industrial uses, composting, land application), and disposal (landfill, incineration, wastewater treatment). The MFA uses a mass balance approach when all inputs and outputs are known, and coefficients from various sources are used when only partial data are available. Two calculation pathways are employed, depending on data availability. Data sources primarily include USDA's Annual Agricultural Statistics, supplementary reports, and Loss Adjusted Food Availability (LAFA) data. An uncertainty analysis is conducted using a Monte Carlo simulation with 10,000 runs to determine the 95% confidence intervals for FLW generation.

The study estimates a total FLW of 335.4 million metric tonnes (MMT) from the U.S. food supply chain in 2016. Water evaporation during manufacturing accounted for 19% (63.7 MMT) of FLW. Recycling and recovery accounted for 55.3% (185.5 MMT), while landfill, incineration, or wastewater treatment accounted for 23.1% (77.3 MMT). Manufacturing was the largest contributor to total FLW (61%), primarily due to the separation of edible from inedible parts. However, manufacturing also had the highest recycling rate. The consumption stage accounted for 57% of FLW disposed of through landfill, incineration, or wastewater treatment. High-demand, perishable products (vegetables, fruits, dairy, grains, meat & poultry) accounted for 67% of food waste. Animal feeding was the largest FLW management pathway (116.6 MMT), largely contributed by manufacturing. Food donation represented only a small portion (2.4%, 8.1 MMT) of total FLW, with only 7.8 MMT of donated food consumed. Uncertainty analysis indicates relatively small confidence intervals for most FLW estimations.

The findings highlight the inefficiency of the U.S. food system, despite a significant portion of FLW being recycled or recovered. The low rate of food donation and high reliance on landfill/incineration/wastewater treatment underscore the need for improvement. The high FLW from perishable products in later stages emphasizes the need for strategies to extend shelf-life and optimize storage and transportation. The significant amount of manufacturing FLW, while mostly unavoidable or recycled, indicates opportunities for reducing FLW at downstream stages through advanced packaging or smaller packaging. Consumer behavior plays a critical role, and standardized labeling and consumer education are necessary to reduce premature food disposal. Government policies, such as incentives for food donation and infrastructure investments, can significantly improve FLW management.

This study provides a comprehensive quantification of FLW in the U.S. food system, revealing key contributors and management pathways. The findings emphasize the need for a multifaceted approach involving technological advancements, policy interventions, and consumer education to improve the efficiency and sustainability of the food system. Future research could focus on developing more precise estimations for specific stages or commodities, exploring the effectiveness of various policy interventions, and investigating the dynamics of consumer behavior and its influence on FLW generation.

The study acknowledges limitations related to data availability, particularly for on-farm FLW and distinguishing between food services and households. The use of waste factors from various sources also introduces uncertainty, although this is partially addressed by uncertainty analysis. The study assumes a uniform distribution in its Monte Carlo simulation, which may not perfectly reflect the real-world distribution of uncertainties. Some FLW management data were limited; for instance, there is no reliable data on on-farm animal-based FLW, and the combined estimates for the distribution and W&R stages.