The Carbon Hoofprint of Cities is Shaped by Geography and Production in the Livestock Supply Chain

US cities’ meat emissions rival home energy use, driven by geography—not appetite—revealing big climate wins from diet and waste shifts.

A 2025 study reveals US cities' meat consumption generates 329 MtCO₂e annually—equivalent to all US at-home fossil fuel combustion—with per capita "carbon hoofprints" varying threefold (500-1,731 kgCO₂e) despite similar consumption levels, as surprising regional differences in meat production intensity create opportunities to halve emissions through reduced food waste and dietary shifts from beef to chicken.

Published by Benjamin P. Goldstein, Rylie E. O. Pelton, and colleagues from the University of Michigan and University of Minnesota, the research deploys the expanded FoodS3 (Food-System Supply-Chain Sustainability) model to map complete "meatsheds" for all 3,531 US cities by linking 3,143 feed and livestock-producing counties with urban consumption centers. Using spatially resolved carbon accounting combined with supply chain reconstruction, the researchers tracked commodity flows across multiple stages—from animal feed production through livestock raising and processing to final consumption—revealing that meat consumed in cities sources from geographically dispersed rural regions with dramatically varying GHG production intensities. The study demonstrates that traditional spend-based carbon accounting using national averages produces substantial overestimates and underestimates across cities, while the high-resolution approach identifies specific reduction strategies ranging from demand-side dietary changes (potentially reducing emissions 37-51%) to supply-side interventions like silvopastoral systems in targeted regions.

Key Findings: Geographic Production Intensity Drives Urban Carbon Hoofprints

Total US Cities' Hoofprint of 329 MtCO₂e Rivals National Energy Consumption

Residents in US cities annually consume 11.1 Mt of meat comprising 4.6 Mt of chicken, 3.7 Mt of beef, and 2.7 Mt of pork, with the three largest cities—New York City, Los Angeles, and Chicago—alone consuming 3.2 Mt. The aggregate carbon hoofprint of 329 MtCO₂e exceeds the territorial CO₂ emissions of the United Kingdom (305 Mt) and Italy (313 Mt), matching total US at-home fossil fuel combustion (334 Mt).

The total hoofprint of individual cities ranges from 1.68 ktCO₂e (Calais, Maine) to 18.1 MtCO₂e (New York City), with New York City's hoofprint alone exceeding combined GHG emissions of Maine and New Hampshire. The nationwide hoofprint equals 43% of GHGs generated from at-home fossil fuel combustion and electricity consumption across all US cities, increasing to more than 50% for 60 million Americans living in high-hoofprint cities.

Per Capita Hoofprints Vary Threefold Despite Consistent Consumption

The per capita carbon hoofprint averages 1,093 ± 237 kgCO₂e across US cities. Despite remarkably consistent per capita meat consumption of 78.6 ± 1.87 kg annually, hoofprints vary from 500 kgCO₂e per capita in Houghton, Michigan, to 1,731 kgCO₂e per capita in Richmond, Missouri—a factor of 3.5 difference. Per capita hoofprints are highest in parts of Texas, Oklahoma, and Missouri, and lowest in the upper Midwest, especially Michigan and Wisconsin.

This dramatic variation occurs because GHG intensities of most beef and pork supplied to the Midwest fall in the lowest two quintiles, whereas the opposite characterizes high-hoofprint regions. Consequently, meat consumption proves a poor proxy for hoofprint, with correlation between per capita consumption and hoofprint extremely weak (r = -0.04, P = 0.01): 868 cities have above-average consumption but below-average hoofprint, while 931 cities show the opposite pattern.

Carbon Intensity of Meat Strongly Correlates with Urban Hoofprints

The carbon intensity of meat production shows strong correlation with hoofprint (r = 0.99, P < 2 × 10⁻¹⁶), demonstrating that using GHG intensities based on national or regional averages would produce substantial overestimates and underestimates across US cities. This finding reveals that urban carbon accounting methods must move beyond the common practice of using national average carbon intensities of meat in their calculations.

Beef consistently comprises the largest hoofprint component at 73 ± 7% of the total, followed by pork and chicken, which is expected given beef's environmental intensity. However, considerable variation exists across cities, with the beef share ranging from 41% in Pine City, Minnesota, to 87% in Monte Vista, Colorado. Although pork typically ranks as the second largest component (range 5-26%; μ = 16 ± 5%), chicken supersedes pork for 386 cities (μ = 11% ± 3%).

GHG Intensity of Meat Varies Up to 15-Fold Across Production Regions

The GHG intensity of beef production varies by a factor of 4.3 across cities, increasing to 4.9 for chicken and 15 for pork. This wide range in meat-related emissions stems from several factors, chief among them being variability in emissions from feed driven by land-use change, yield differences, differing nitrogen fertilizer application rates, and related N₂O emissions.

These factors combine across supply chains to influence the GHG intensity of feed consumed in different regions, strongly affecting chicken and pork hoofprints since feed comprises a larger proportion of total GHGs for these meats. Pork production intensity is further compounded by differences in manure management, with some locations having negligible or negative manure emissions due to methane capture and energy use.

Meatsheds Span Hundreds of Counties and Thousands of Kilometers

Mapping beef supply chains for Atlanta, Los Angeles, and New York City demonstrates the vast spatial extent of urban meatsheds. Although cities typically source processed meat from a handful of counties, the meatshed can eventually encompass hundreds of counties spanning thousands of kilometers. Los Angeles sources beef from ten counties with processing facilities, which are supplied with livestock raised in 469 counties, which in turn rely on feed grown in 828 different counties.

Average distances for beef supply chains measure 920 ± 630 km from processing to city, 420 ± 150 km from feedlot to processing, and 430 ± 230 km from feed farmer to animal producer for the 20 largest US cities. Distances from processing to city exceed those from feedlot to processing (P = 0.0006) and from crop farm to feedlot (P = 0.0005), though considerable ranges exist within cities (for example, 140-2,180 km from processing to New York City).

Chicken and Pork Meatsheds Show Greater Geographic Concentration

Chicken and pork meatsheds demonstrate less geographic expansiveness than beef due to the vertical integration typifying US poultry and pork industries and high travel mortality rates for hogs and chickens. Average distance from broiler farms to processing measures 190 ± 370 km, shorter than distances from beef feedlots to processing (P = 0.0264).

Average distances from crop farm to broiler farm are also smaller and statistically distinguishable from beef for chicken (μ = 310 ± 200 km, P = 0.0184) and pork (μ = 140 ± 90 km, P < 0.0001). Pork processing occurs mainly in Iowa (31% of US processing), Illinois (10%), Minnesota (10%), and Missouri (8%), while most cities source processed chickens from the mid-Atlantic and Southeast. Most hog and chicken farmers are located within or near the US corn belt, whereas beef feedlots are closer to grazing lands and further from feed producers.

Unexpected Intensity Patterns Create Counter-Intuitive Dietary Pathways

The high resolution of FoodS3 reveals notable exceptions to conventional wisdom about meat emissions hierarchies. In contrast to most studies, the emissions intensity of chicken exceeds that of pork for some cities (for example, Monte Vista, Colorado), though this exception occurs for only ten cities. Similarly, ten instances exist where pork emissions intensity exceeds beef (for example, Calexico, California).

These cities consume high proportions of beef from culled dairy cattle rather than feedlots, combined with pork supplied from areas using low-tech manure management systems (for example, uncovered lagoons) and significant land-use change emissions from feed. The GHG intensity of beef production is lower for cities sourcing higher proportions from culled beef and dairy cattle versus feedlots, with cities near major dairy-producing areas like Wisconsin or southern Pennsylvania more likely to consume this lower-intensity beef.

Regional Hoofprint Differences Reflect Production Geography Complexity

Clear differences emerge between counties within modelled meatsheds, with low-GHG production intensity areas producing significant volumes of culled dairy cattle contrasting with those producing more feedlot beef using GHG-intensive feed. New York City, for example, sources meat from the Northeast and Midwest via processors in Pennsylvania, with substantial variation in production intensity within its beef meatshed.

Dissecting emissions by production stage reveals further intercity variation. Grazing generates 63 ± 14% of beef emissions on average but drops to 17% in San Diego where feedlot emissions predominate. Pork emissions split evenly between growing feed (μ = 52 ± 10%) and hog farming (μ = 46 ± 11%), while chicken emissions arise mostly from feed (μ = 70 ± 5%). However, these percentages vary considerably, with hog farming driving 10-80% of pork GHGs depending on location.

Why This Matters: Actionable Pathways for Urban Decarbonization

The Goldstein et al. study provides cities with unprecedented clarity on how to reduce their meat-related climate impacts through both consumption and production interventions. The finding that hoofprints vary threefold despite similar consumption reveals that traditional carbon accounting dramatically misrepresents cities' actual climate impacts from meat, potentially misdirecting mitigation efforts and investments.

For cities with mild climates and low-carbon energy grids, the hoofprint's relative importance increases dramatically—reaching 77% in San Diego and 81% in Los Angeles of emissions from at-home energy use. Conversely, the ratio drops in colder cities like Boston (24%) and Chicago (19%), and further still when combined with carbon-intensive grids and low hoofprints (5% in Springs, New York). This variability means that meat consumption represents an especially cost-effective decarbonization lever for specific urban contexts.

The four modeled policy scenarios reveal substantial reduction potential. Scenario 1 (halving edible food waste at retailers and consumers) reduces emissions by 16%, translating to less meat production needed to satisfy total consumption. Scenario 2 (substituting half of beef consumption evenly between pork and chicken) achieves 28% reduction. Scenario 3 (substituting half of beef with chicken only) produces 33% reduction. Scenario 4 (abstaining from meat once weekly, "meatless Monday") generates 14% reduction.

Combined implementation of halving food waste, substituting chicken for beef, and weekly meat abstention achieves a 51% reduction from baseline, cutting the total hoofprint by 123-142 MtCO₂e (37-43%). Reductions are largest for cities with GHG-intensive beef. Notably, these strategies preclude requiring overall meat consumption reduction, allow moderate beef consumption, and align with ongoing dietary shifts. Since substituting chicken for beef and cutting in-home food loss lower grocery expenditures, mitigating the hoofprint becomes less expensive and implementable more swiftly than some decarbonization strategies in other sectors like home energy retrofits.

Cities are increasingly promoting low-carbon diets, with New York City expanding availability of plant-based foods at city events and facilities while enlisting universities, companies, and other institutions as allies. The research suggests cities should adopt an expanded notion of urban food policy addressing economic and retail barriers limiting access to affordable, sustainable food. Educational campaigns in public spaces and schools tailored to social and cultural diversity could emphasize fiscal and health benefits of sustainable diets. More directly, cities can influence menus at hospitals and schools, promote plant-based products with local retailers, and support initiatives like urban farming linked to sustainable dietary shifts.

Production-Side Interventions and Urban-Rural Collaboration

While most work reducing GHG-associated meat production will occur in rural areas, cities can play a supporting role by identifying emissions hotspots in meat supply chains and deploying targeted production-based solutions. Beef silvopastoral systems could reduce the total hoofprint of US cities by 6% or considerably more for specific cities, with future research needed to model additional climate-friendly practices including alternative feeds, agrivoltaics, and bioenergy from manure to identify which strategies would prove most effective for different cities.

This spatial specificity could inform urban-rural collaborations to accelerate urban food-system decarbonization. Large US cities like Los Angeles have adopted measures addressing food-system GHGs through both production and consumption, with the research providing requisite knowledge for targeted interventions. By making opaque urban-rural linkages more transparent, the study clarifies how rural and urban livelihoods are intertwined and the collective responsibility for cities and rural communities to collaborate in the shared goal of reducing environmental impacts.

The FoodS3 modeling architecture can be deployed for other agricultural or non-agricultural commodities where supply and demand data exist at subnational scales. Forest products could be modeled using US Forest Inventory and Analysis statistics on timber harvests and mill outputs at county level, or metals using mining and mineral processing data from commercial and government sources. Incorporating consumption data, pollution inventories, and remote-sensing data could then link urban processes to environmental or social change elsewhere.

Methodological Advances and Remaining Uncertainties

The study overcomes two major obstacles that have hindered development of high-resolution urban-rural linkage mapping: scarce or proprietary data linking production to consumption, and limited spatially explicit data on environmental intensity of production. The FoodS3 model reconstructs meat supply chains for 93% of the US population using mass-balanced supply-demand modeling, combining supply chain models with spatial carbon accounting.

Monte Carlo analysis estimates a mean absolute percentage error of 26.5% across all cities from uncertainty in buyer-supplier relationships, dropping quickly with city size (below 20% for cities with populations above 125,000). Uncertainty from consumption estimates shows mean absolute percentage error from baseline ranging 0.02% to 5% for individual cities, below 1% for all cities.

However, uncertainty exists regarding estimated GHG reductions in scenarios. Reduced domestic demand might spur increased exports of US meat, though exports could substitute meat produced less efficiently elsewhere and curb land-use change associated with pastures and feed farming. Consequential models incorporating international trade could quantify net GHG impacts or other stressors and economic outcomes for farmers domestically and in importing countries. Future iterations could consider subnational origins and impacts of meat imported by the USA, or expand the framework to countries with comparable agricultural statistics.

Additional uncertainty from GHG inventories arises from several sources including variability in input and output parameters, model selection, and scenario assumptions and allocation decisions. Future work should overcome the computationally challenging task of quantifying this uncertainty at county level and propagating it through downstream stages, though the approach improves representativeness by integrating detailed, subnational production data and using tier 2 and 3 methodologies to reflect local environmental conditions rather than relying on national averages.

Sign up for our newsletter or connect with us on social media to stay up-to-date with our latest posts and permaculture inspiration.

Explore our inspiring series and posts:

• Foundation

• Manual 

• Perspective 

• Design 

• Reads

• Plants

• Soil

• Climate

Love the post? Share it with your circle, inspire your people.