Climate impacts of alternative beef production systems depend on the functional unit used: Weight or monetary value

Significance

This paper presented a meta-analysis of life cycle assessments of greenhouse gas (GHG) emissions from beef cattle production in the United States and Canada. Our study highlights the implications of using carbon intensities measured by weight vs. economic activity for comparisons across differentiated beef cattle products. While a positive association was found between the proportion of lifespan on grassland and the conventional weight-based metrics, grass-finished beef was found to have lower GHG emission per economic activity than feedlot-finished beef. Systems that incorporated soil carbon sequestration were shown to lead to the greatest reduction in GHG emissions. Our study emphasized the need for more inclusive assessments of livestock-related soil carbon sequestration that account for regional differences and grazing management effects.

Abstract

Beef production has been identified as a significant source of anthropogenic greenhouse gas (GHG) emissions in the agricultural sector. United States and Canada account for about a quarter of the world’s beef supply. To compare the GHG emission contributions of alternative beef production systems, we conducted a meta-analysis of 32 studies that were conducted between 2001 and 2023. Results indicated that GHG emissions from beef production in North America varied almost fourfold from 10.2 to 37.6 with an average of 21.4 kg CO₂e/kg carcass weight (CW). Studies that considered soil C sequestration (C-seq) reported the highest mitigation potential in GHG emissions (80%), followed by growth enhancement technology (16%), diet modification (6%), and grazing management improvement (7%). Our study highlights the implications of using carbon intensity per economic activity (i.e., GHG emissions per monetary unit), compared to the more common metric of intensity on per weight of product basis (GHG emissions per kg CW) for comparisons across differentiated beef cattle products. While a positive association was found between the proportion of lifespan on grassland and the conventional weight-based indicator, grass-finished beef was found to have lower carbon intensity per economic activity than feedlot-finished beef. Our study emphasizes the need to incorporate land use and management effects and soil C-seq as fundamental aspects of beef GHG emissions and mitigation assessments.

Elevated atmospheric greenhouse gas (GHG) concentration is leading to numerous deleterious climatic changes, including higher average temperature and more frequent extreme weather conditions, and has been identified as a major threat to biodiversity and human wellbeing (1, 2). To address these threats, 124 countries have pledged to attain carbon neutrality by 2050 to 2060 (1). In accordance with this goal, efforts have been made globally to identify major GHG emission sources and potential strategies to enhance GHG mitigation and sequestration.

Globally, 20 to 25% of anthropogenic GHG emissions have been attributed to crop and livestock production and related land use changes (3, 4). Within the agricultural sector, GHG emissions from animal-based food production, especially beef, have received increasing attention in GHG emission reduction policy initiatives (5–8). Given the ongoing debate about the contribution of beef cattle production to anthropogenic GHG emissions, evaluating emissions from different beef cattle production systems and identifying mitigation strategies are critical (9–13).

An expanding body of literature has used life cycle assessment to evaluate the GHG emissions of beef production, and some review papers have synthesized and compared these studies on a global scale (14–16). GHG emissions and mitigation opportunities from beef production vary across the world due to climate, vegetation cover, land use, and socioeconomic differences (14, 16, 17). For example, African nomadic livestock systems differ markedly from US feedlot production with respect to biophysical conditions, livestock diets, and finishing time (18). Even in developed countries, beef operations differ; for example, compared to North America, European operations are generally much smaller and more integrated with crop production (19).

We conducted a meta-analysis of peer-reviewed publications that assess GHG emissions in the United States and Canada. As of 2023, the number of beef cattle in these two countries was 89.3 million and 5.5 million, respectively, which collectively contributed to nearly one-fourth of the world’s total beef supply (United States, 21%; Canada, 2%) (20–22). Published assessments of GHG emissions from beef cattle production in the western hemisphere have been conducted mainly in these two countries. This geographic focus also minimizes the effects of biophysical and cultural differences that dramatically affect GHG emissions on a global scale.

In the United States and Canada, 37% and 29% of total anthropogenic methane emissions are from livestock sector, respectively, of which a vast majority has been attributed to beef and dairy beef cattle (23). Both countries have initiated GHG emission mitigation efforts in the agricultural sector. Following the Global Methane Pledge in 2021, the US Methane Emissions Reduction Action Plan was released, which includes methane emission reductions in agriculture as exemplified by the Greener Beef cattle Initiative (24). Similarly, the Canadian Agricultural Partnership, launched in 2018, aims to mitigate GHG emissions while enhancing climate resilience within the agricultural sector (25).

In the United States and Canada, beef cattle are produced through either feedlot-finishing or grass-finishing system, with distinct phases of production (Fig. 1). Most beef cattle are finished in a feedlot, with grass-finished beef accounting for less than 2% of the total beef volume (26). The cow–calf phase, which generates weaned calves at 4 to 8 mo, is the common source of beef cattle for both feedlot- and grass-finishing beef production systems. While young animals, including male dairy calves, can directly enter the feedlot when weaned as “calf-feds,” most are placed into a stocker operation where they consume mostly grass-based diets, or alternatively in a backgrounding operation where they are fed a mixed ration of grass and grains in confinements before being moved to a feedlot as “yearlings” (27–29). Diets of feedlot animals consist mainly of high-energy grains combined with small portions of forage, while weaners in grass-finishing systems are primarily fed pasture grass diets until they reach marketable size, which results in lower average daily weight gains and longer finishing times than in feedlots (28).

Fig. 1.

Flow chart of grass-finishing and feedlot-finishing beef production systems in the United States and Canada.

To analyze factors affecting GHG emissions and mitigation potentials from alternative beef cattle production systems and practices across the United States and Canada, we synthesized 32 previous studies in these two countries that included 115 beef cattle production scenarios, of which eight studies included 21 scenarios that considered soil C sequestration (C-seq). Compared to the previous reviews, we include a more complete set of studies in our study region, removing the restriction of containing at least two contrasting scenarios per study (14, 16), and the cradle to farm gate restriction (14, 15).

Most of the GHG emission comparisons within the agricultural sector are made on a per weight basis, i.e., GHG emissions associated with one kilogram of product. For example, life cycle GHG emissions for wheat and beef production in the United States were estimated to be 1.1 ± 0.13 and 31 ± 8.1 kg CO₂e per kg, respectively (30). While GHG emissions per unit weight provide a valid measure to compare across agricultural product of the same type and quality, the logic of using the weight-based comparison for differentiated agricultural products is weak as those products are incommensurable even under the same weight of product.

Carbon (C) intensity per economic activity, which measures GHG emissions per monetary unit (MU), provides an indicator that compares GHG emission growth with economic growth (31). On a global scale, C intensity of gross world product declined from 0.35 to 0.24 kg C/$ at a rate of 1.3% per year from 1970 to 2000, yet it starts to increase at 0.3% per year after 2000 (31). It is, therefore, important to identify sectors with relatively high C intensity per economic activity and prioritize them for mitigation programs.

Compared to conventional feedlot-finished beef, grass-finished beef receives premium prices that vary from 48 to 193% depending on cuts (26, 32). To compare climate impacts of different beef production systems, we used two indicators: 1) GHG emissions with commonly used weight as a functional unit; 2) GHG emissions with monetary value of beef as a functional unit, which is also referred to as carbon intensities per economic activity. To better understand the climate impacts of beef cattle productions in North America and inform policy priorities and strategies aimed at reducing GHG emissions, we addressed the following three main objectives: 1) compare GHG emissions per weight and per MU across feedlot- and grass-finishing beef production systems; 2) examine mitigation potentials of alternative beef production scenarios; and 3) identify factors that affect GHG emissions and mitigation potentials.

Results

Summary of Reviewed Studies.

The breakdown of the scenarios by finishing type and by region is presented in SI Appendix, Table S1, which includes 32 studies incorporating 115 scenarios. Of these scenarios, 81 (70%) and 34 (30%) assessed GHG emissions of beef cattle production systems in the United States and Canada, respectively. Of all these scenarios, 16 (14%) analyzed GHG emissions of grass-finishing operations, and 99 (86%) represented feedlot-finishing operations, including 76 scenarios with beef cattle only, four with dairy beef cattle only, and 19 comprising a mixture of beef and dairy beef cattle. Eight of the studies (21 scenarios) considered the role of C-seq in the Midwest, Northeast, Northern Plains, and Southeast of the United States and Canada (SI Appendix, Table S1).

Among the beef cattle production scenarios in the United States, 22 (19%) did not specify the states where the studies were conducted and, therefore, could not be regionally categorized. Of the scenarios that did specify location, 70% were concentrated in the Southwest (29%), Southern Plains (21%), and Northern Plains (20%). Studies about beef cattle GHG emissions have increased dramatically since 2011 (SI Appendix, Fig. S1), with only seven scenarios during 2001 to 2010 (including only one grass-finishing scenario), and 108 scenarios during 2011 to 2023 (including 15 grass-finishing scenarios).

Of all 115 scenarios in our study, nine (8%) excluded the cow–calf phase in their analysis (33, 34), four (4%) excluded the stocker/backgrounding phase (35), and 12 (10%) considered neither of these two phases (36, 37). The remaining 90 scenarios (78%) included all production phases from birth to slaughter. Studies in which weaned calves were placed directly in a feedlot (38–40) were categorized as covering all production phases because no production phase is skipped from the GHG emission analysis.

Summary Statistics of Beef Production Systems and GHG Emissions.

Of the 90 scenarios that included all production phases, 19 considered C-seq. The inclusion of C-seq altered the distributions of GHG emissions per weight and per economic activity (Fig. 2). On average, scenarios that considered C-seq show significantly lower GHG emissions than scenarios that did not consider such sequestration (7.7 vs. 21.4 kg CO₂e/kg CW, or 4.9 vs. 19.6 kg CO₂e/MU, P < 0.05), where CW represents carcass weight (CW) and MU represents the monetary unit described in the Materials and Methods section (SI Appendix, Table S2). Among the 90 scenarios that either omit soil C from their assessment or assume soil carbon equilibrium, GHG emissions based on CW range from 10.2 (feedlot-dairy only) (41) to 37.6 CO₂e/CW (grass-finishing) (38), while GHG emission values among the 19 scenarios that considered C-seq range from −94.7 (feedlot-beef only) (42) to 24.4 CO₂e/CW (feedlot-beef only) (38). In contrast, GHG emissions based on MU range from 9.7 (grass-finishing) (43) to 31.0 CO₂e/MU (feedlot-beef only) (30) for scenarios that did not account for C-seq and from −94.7 to 24.4 CO₂e/MU for scenarios that accounted for C-seq (SI Appendix, Table S2).

Fig. 2.

Comparison of CW-based and MU-based GHG emissions between scenarios that either did not (blue) or did (beige) consider C-seq. Note: MU is the economic value of 1 kg CW of feedlot-finishing beef. The horizontal line and × within each box are the median and mean values, respectively; the Bottom and Top lines the 25th and 75th percentiles; the whiskers (vertical lines) indicate variability that is 1.5 times the interquartile range outside the 25th and 75th percentiles. Mean, minimum, and maximum values are provided in SI Appendix, Table S2.

GHG emissions and beef production system characteristics for all scenarios are described in Table 1, while those for scenarios with all production phases are in SI Appendix, Table S3. Of the 115 scenarios, 65 provide days to finishing, with an average of 561 d ranging from 382 (feedlot-beef only) (36) to 903 d (grass-finishing) (33). Additionally, 58 (50%) scenarios provided information to infer the proportion of days consuming grass, defined as the days spent on grassland (consuming a forage-based diet), divided by the total days from birth to slaughter. On average, beef cattle spent 67% of their lifespan on grassland ranging from 26% (feedlot-dairy only) (41) to 100% (grass-finishing scenarios). Finishing weight was available for 77 scenarios, averaged at 580 kg, ranging from 385 (grass-finishing) (43) to 725 kg (feedlot-beef only) (39).

Table 1.

Definitions and summary statistics of cattle GHG emissions and other variables from the selected literature

Variable	Definition	N	Mean	Min	Max
GHG emission_CW	GHG emission kg CO2e/kg CW	115	18.54	3.5	37.6
GHG emission_MU	GHG emission kg CO2e/(kg CW × MU)	115	16.81	3	31
Publication year	Year of study publication	115	2016	2001	2023
MitiP_CW	CW-based GHG emission mitigation potential	84	0.19	0	0.65
MitiP_MU	MU-based GHG emission mitigation potential	84	0.16	0	0.58
Slaughter weight	Cattle finishing weight	77	580	385	725
Days to slaughter	Total number of days between birth and slaughter	65	561	382	903
Proportion on grass	Number of days on grassland relative to total days to slaughter	58	0.67	0.26	1
Grass-finishing	Grass-finishing (1 = yes, 0 = no)	115	14%	0	1
Beef	Only beef cattle in the feedlot phase (1 = yes, 0 = no)	115	80%	0	1
Mix	Mixed beef and dairy cattle in feedlot phase (1 = yes, 0 = no)	115	17%	0	1
Dairy	Only dairy cattle in feedlot phase (1 = yes, 0 = no)	115	3%	0	1
Cow–calf phase	Cow–calf phase included (1 = yes, 0 = no)	115	82%	0	1
Stocker phase	Backgrounding/stocker phase included (1 = yes, 0 = no)	115	86%	0	1
All production phases	All production phases included (1 = yes, 0 = no)	115	78%	0	1
Regional comparison	Regional comparison of GHG emissions (1 = yes, 0 = no)	115	23%	0	1
System comparison I	Feedlot-finishing systems comparison (1 = yes, 0 = no)	115	11%	0	1
System comparison II	Feedlot- vs. grass-finishing systems comparison (1 = yes, 0 = no)	115	28%	0	1
GET	Growth Enhancing Technology comparison (1 = yes, 0 = no)	115	22%	0	1
Diet	Dietary comparison (1 = yes, 0 = no)	115	21%	0	1
Grazing	Grazing management comparison (1 = yes, 0 = no)	115	3%	0	1
Midwest	US Midwest (1 = yes, 0 = no)	115	10%	0	1
Northeast	US Northeast (1 = yes, 0 = no)	115	3%	0	1
Northern Plains	US Northern Plains (1 = yes, 0 = no)	115	11%	0	1
Northwest	US Northwest (1 = yes, 0 = no)	115	3%	0	1
Southeast	US Southeast (1 = yes, 0 = no)	115	3%	0	1
Southern Plains	US Southern Plains (1 = yes, 0 = no)	115	12%	0	1
Southwest	US Southwest (1 = yes, 0 = no)	115	17%	0	1
All US	US (no specific region specified) (1 = yes, 0 = no)	115	13%	0	1
Canada	Canada (1 = yes, 0 = no)	115	30%	0	1

Note: Summary statistics are based on the 115 included scenarios. Summary statistics for scenarios that include all production phases, including live weight (LW) based GHG emission variables, are provided in SI Appendix, Table S2.

GHG Emissions and Beef Cattle Production Systems.

Comparative results of GHG emissions per CW and per MU across different beef production systems are presented in Fig. 3. On a weight basis, GHG emissions of grass-finishing beef ranked highest (27.5, P < 0.05) and feedlot-dairy only ranks lowest (11.1, P < 0.05), while beef-only and mixed feedlots are intermediate (21.7 vs. 19.1) with no significant difference between them (Fig. 3 and SI Appendix, Table S4). On an economic activity basis, however, GHG emissions of the grass-finishing and dairy-only feedlot systems ranked lowest (12.5 vs. 11.1) with no statistically significant differences, while those for beef-only and mixed feedlots were significantly higher (P < 0.05) (Fig. 3 and SI Appendix, Table S4). Note that in this manuscript, GHG emission values based on both functional units are the same for the feedlot-finishing scenarios with the relative MU value for feedlot beef being one.

Fig. 3.

Comparison of CW-based and MU-based GHG emissions among four different cattle production systems. Note: Only the scenarios that include all cattle production phases without accounting for C-seq were included in the analysis. MU is the economic value of 1 kg CW of feedlot-finishing beef. The horizontal line and × within each box are the median and mean values; the Bottom and Top lines the 25th and 75th percentiles; the whiskers (vertical lines) indicate variability that is 1.5 times the interquartile range outside the 25th and 75th percentiles.

In terms of beef production system characteristics, the grass-finishing system had the highest percentage of time on grazing (99.5%), followed by beef-only feedlot system (58.9%) and dairy-only feedlot system (28.8%) (SI Appendix, Table S4). Given that the time for beef cattle to reach slaughter weight is highly affected by the proportion of time on grazing vs. high energy feed, grass-finishing had a significantly longer finishing period (735 d) than the three feedlot-finishing systems (516, 462, and 454 d for beef-only, dairy-only and mixture, respectively). Conversely, grass-finishing generated a significantly lower finishing weight (488 kg) than feedlot-beef only (610 kg) (P < 0.05) while feedlot-mixture and feedlot-dairy only fell in between (546 vs. 571 kg) with no statistical difference.

GHG Emission and Beef Production Variables.

GHG emission is positively correlated with days to finishing and percentage of days grazing on grassland, with the latter two also being positively correlated (0.46, 0.72, and 0.79, respectively, P < 0.01) (SI Appendix, Table S5). While GHG emission per CW was also positively correlated with grass-finishing, it was negatively correlated with mixed and dairy-only feedlots (0.49, −0.26, and −0.48, respectively, P < 0.05), which implies that fattening more dairy beef cattle in feedlots could reduce GHG emissions per CW. By contrast, grass-finishing has less climate impact (−0.58, P < 0.01) using a value-based emission indicator, which accounted for the price premiums of the grass-finishing beef. Finally, year is positively correlated with grass-finishing and days to finishing but negatively correlated with slaughter weight (0.33, 0,45, and −0.33, respectively, P < 0.05), indicating that grass-finishing, associated with longer finishing periods and lower slaughter weights, has attracted increased attention in recent years (SI Appendix, Fig. S1 and Table S5).

GHG Mitigation Potential.

To better understand GHG mitigation potential of beef production, we categorize the contrasting scenarios in different studies into the following comparison types: 1) scenarios that did or did not account for soil C-seq; 2) different regions (Region); 3) different production systems that included either a) feedlot-finishing systems with varying feedlot entry time and days in feedlot (System comparison I), and b) grass- vs. feedlot-finishing systems (System comparison II); 4) use growth enhancement technology (GET) or not; 5) scenarios with different diets (Diet); 6) scenarios with and without dairy beef cattle, which were further divided into dairy-only (Dairy) and dairy mixture (Mix); and 7) different grazing management strategies (Grazing) (SI Appendix, Table S6). Of the 25 studies that involve such scenario comparisons, the most frequently compared scenarios are C-seq (32%), System comparison II (32%), and GET (28%), followed by Region (16%), System comparison I (16%), Diet (12%), Mix (12%), Dairy (8%), and Grazing (4%).

On average, scenarios that considered C-seq have GHG emission mitigation potential values of more than 80%, while mitigation potentials of scenarios that did not account for C-seq are less than 20% (P < 0.05; Fig. 4 and SI Appendix, Table S2). Fig. 5 demonstrates GHG emission differences or mitigation potentials across various comparison types excluding the scenarios that accounted for C-seq. On average, studies that involve Dairy, System comparison II, and Region comparisons generated the highest GHG emission differences across scenarios, followed by Mix and GET (57%, 33%, 24%, 22, and 16%, respectively), while the remaining three comparison types (System Comparison I, Diet, and Grazing) demonstrated GHG emission differences of less than 10%. Overall, we found comparisons that focused on underlying production variations (e.g., different production regions, dairy cattle entering beef production system, and contrasting finishing systems) generate greater differences in GHG emissions than comparisons focused on within-system interventions (e.g., GET, altering diets, and grazing management practices).

Fig. 4.

Comparison of mitigation potentials of alternative production scenarios for CW-based and MU-based GHG emissions between scenarios that did not consider (blue) or consider (beige) C-seq. Note: The horizontal line and × within each box are the median and mean values; the Bottom and Top lines the 25th and 75th percentiles; the whiskers (vertical lines) indicate variability that is 1.5 times the interquartile range outside the 25th and 75th percentiles. Mean, minimum, and maximum values are provided in SI Appendix, Table S2.

Fig. 5.

GHG emission differences or mitigation potentials for different comparison types in CW-based and MU-based GHG emissions across different comparison types. Note: Only scenarios that did not account for C-seq were included in the analysis. The horizontal line and × within each box are the median and mean values; the Bottom and Top lines the 25th and 75th percentiles; the whiskers (vertical lines) indicate variability that is 1.5 times the interquartile range outside the 25th and 75th percentiles.

Except for the C-seq comparison, eight studies have comparison scenarios that differed in more than one aspect. For example, the study by Stackhouse et al. (41) contains scenarios that differed in both production systems and GET usage. Excluding such studies, the remaining single-aspect comparison demonstrated similar results as Fig. 5, which indicated the robustness of our findings.

Factors Affecting GHG Emissions.

Ordinary Least Square regression results for factors that affect weight- and value-based GHG emissions are presented in Table 2. While GHG emission results vary across the studies due to differences in assumptions, our regression models that incorporate production systems, production phases, geographic regions, and year of the study as explanatory variables explain 65% and 83% of the variation in GHG emissions by CW (adjusted R² in Models I and III, respectively), and 69% and 86% of the variation in emissions by MU (adjusted R² in Models II and IV, respectively).

Table 2.

Ordinary least square regression results of factors influencing CW- and MU-based GHG emissions

	Scenarios with all production phases		All scenarios including those with partial production phases
	CW-based emissions	MU-based emissions	CW-based emissions	MU-based emissions
	Model I	Model II	Model III	Model IV
Intercept	719.82***	430.18**	807.33**	651.10***
	(181.11)	(166.38)	(180.13)	(153.94)
Grass-finishing	6.57***	−9.13***	6.36***	−7.02***
	(1.07)	(0.98)	(0.98)	(0.84)
Mix	−2.92***	−3.00***	−2.70***	−2.48***
	(0.87)	(0.80)	(0.90)	(0.77)
Dairy	−11.89***	−11.62***	−10.82***	−10.93***
	(1.62)	(1.49)	(1.74)	(1.48)
Cow–calf phase			10.14***	7.71***
			(1.12)	(0.96)
Stocker phase			5.80***	7.34***
			(1.31)	(1.12)
Midwest	1.76	0.60	1.31	1.41
	(1.14)	(1.05)	(1.08)	(0.92)
Northeast	4.39***	2.80*	3.98**	2.79*
	(1.66)	(1.52)	(1.76)	(1.51)
Northern Plains	−0.58	−0.49	−0.90	−0.61
	(1.02)	(0.94)	(1.08)	(0.93)
Northwest	0.09	−0.53	−0.18	−0.02
	(1.55)	(1.42)	(1.65)	(1.41)
Southeast	8.83***	6.92***	8.76***	7.01***
	(1.82)	(1.67)	(1.94)	(1.66)
Southern Plains	−0.36	−1.48	1.11	0.03
	(1.21)	(1.11)	(1.10)	(0.94)
Southwest	−0.03	−0.09	−1.83*	−0.66
	(1.02)	(0.93)	(1.02)	(0.87)
All US	−0.39	−0.13	−1.60	−0.83
	(1.16)	(1.06)	(1.11)	(0.95)
Year published	−0.35***	−0.20**	−0.40***	−0.32***
	(0.09)	(0.08)	(0.09)	(0.08)
N	90	90	115	115
R2	0.69	0.73	0.85	0.88
Adjusted R2	0.65	0.69	0.83	0.86

^*, ^**, and ^*** represent P < 0.10, P < 0.05, and P < 0.01, respectively. SE are presented in the parentheses under each of the estimated mean values. All variables are defined in Table 1.

Compared with beef-only feedlot, dairy-only feedlot reduced weight-based GHG emission by 10.82 to 11.89 kg CO₂e/kg CW and value-based emission by 10.93 to 11.62 kg CO₂e/MU, whereas mixed feedlot reduced weight-based GHG emission by 2.70 to 2.92 kg CO₂e/kg CW and value-based GHG emission by 2.48 to 3.00 kg CO₂e/MU. By comparison, while grass-finishing increased weight-based GHG emission by 6.36 to 6.57 kg CO₂e/kg CW (Model I & III), it was found to reduce value-based emission by 7.02 to 9.13 kg CO₂e/MU, which is a more than three times greater than the value-based emission reduction in mixed feedlots (Model II & IV). Comparing the studies with all production phases with those that omit some of those phases, studies that included the cow–calf phase had, on average, 10.14 kg CO₂e/kg CW greater emissions, and studies that included the stocker/backgrounding phase reported significantly higher emissions, by 5.80 kg CO₂e/kg CW (Model III).

Compared with the studies in Canada, those in the United States that did not differentiate regions, and those in the Midwest, Northern Plains, Southern Plains, and Northwest regions all reported similar GHG emission values (Table 2). By contrast, studies in the Northeast and Southeast regions reported significantly higher GHG emission levels than the Canadian studies across all four models, while the Southwest region had lower GHG emissions but in Model I only.

The regression results also indicated that the studies published in the more recent years of our 2001 to 2023 literature review, reported significantly lower GHG emissions in beef cattle production than earlier publications; weight- and value-based GHG emission values for beef cattle decreased by 0.35 to 0.40 kg CO₂e/kg CW and 0.20 to 0.42 kg CO₂e/MU, respectively, for each later year of publication.

Factors Affecting Mitigation Potentials.

Ordinary Least Square regression results for factors that affect mitigation potentials of GHG emissions are presented in Table 3. The regression analyses were conducted for the scenarios that included all production phases excluding C-seq and incorporated various aspects of production system comparisons, geographic region, and publication year as explanatory variables. The regression models explain 83% and 88% of the variations in mitigation potentials for GHG emissions, per CW and MU, respectively (adjusted R² in Models V and VI).

Table 3.

Ordinary least square regression results of factors influencing GHG emission mitigation potentials

	Studies with all production phases
	CW-based emissions (Model V)	MU-based emissions (Model VI)
Intercept	−0.02	19.56**
	(6.82)	(7.23)
Regional comparison	0.18***	0.18***
	(0.03)	(0.04)
System comparison I	0.01	-0.03
	(0.03)	(0.03)
System comparison II—grass	−0.11***	0.23***
	(0.04)	(0.03)
System comparison II—feedlot	0.22***	−0.08**
	(0.02)	(0.03)
Growth enhancing technology (GET)	0.03	0.04
	(0.03)	(0.03)
Diet	0.06*	0.03
	(0.03)	(0.04)
Mix	0.09***	0.08**
	(0.02)	(0.03)
Dairy	0.45***	0.38***
	(0.04)	(0.04)
Grazing management	0.04	0.06
	(0.05)	(0.06)
Midwest	0.00	0.04
	(0.03)	(0.04)
Northeast	−0.03	−0.02
	(0.04)	(0.05)
Northern Plains	0.04	0.09***
	(0.03)	(0.03)
Northwest	0.06	0.09*
	(0.04)	(0.05)
Southeast	−0.22***	−0.17**
	(0.06)	(0.07)
Southern Plains	−0.07*	−0.02
	(0.03)	(0.04)
Southwest	0.04*	0.06*
	(0.03)	(0.03)
All US	−0.04	−0.01
	(0.04)	(0.05)
Year published	0.00	−0.01***
	(0.00)	(0.00)
N	65	65
R2	0.91	0.88
Adjusted R2	0.88	0.83

^*, ^**, and ^*** represent P < 0.10, P < 0.05, and P < 0.01, respectively. SE are presented in the parentheses under each of the estimated values. System comparison I refers to alternative feedlot-fishing system comparisons, while system comparison II refers to feedlot- vs. grass-finishing system comparisons, which is divided into feedlot-finishing scenarios (System comparison II—feedlot) and grass-finishing scenarios (System comparison II—grass). All other variables are defined in Table 1.

Studies involving dairy-only, grass-finishing, and different regions in their comparisons reported the highest GHG mitigation potentials, or greatest differences in GHG emissions (45%, 22%, and 18%, respectively, Model V). Comparatively, studies that included diet modifications and feedlot-mixture comparisons provide lower mitigation potentials (6% and 9%, respectively). When controlling for the other influencing factors, studies that included system comparison I, GET, and grazing management comparisons generated similar mitigation potentials with studies that do not include such comparisons. Similar findings are observed regarding mitigation potentials for value-based emissions.

Among studies that included grass- vs. feedlot-finishing comparisons, grass-finishing scenarios generated lower mitigation potential (−11%) but other scenarios generated higher mitigation potential (22%) for the weight-based GHG emissions (Model V). The converse results were found regarding value-based emissions (Model VI), which is consistent with the findings that grass-finishing systems have the highest weight-based GHG emissions but lowest value-based emissions (Fig. 3).

Compared with the Canadian studies, US studies that did not distinguish regions, and those in the Midwest and Northeast regions reported similar GHG emission mitigation potentials. Higher GHG emissions mitigation potentials per CW and MU were reported in the Southwest region (4% and 6%, respectively), while significantly lower GHG emission mitigation potentials per CW and MU were reported in the Southeast region (−22% and −17%, respectively). The regression results indicate that there was no difference in weight-based GHG emission mitigation potential in beef cattle production between 2001 and 2023, but that value-based emission mitigation potential decreased by an average of 1% per year of publication.

Discussion

In global efforts to reduce the deleterious effects of elevated atmospheric carbon, animal-based food production, especially from beef cattle, has been identified as a significant source of GHG emissions (4, 6, 8, 11). However, the estimates of GHG emissions from different beef cattle production systems have varied widely. Our meta-analysis of 32 studies, which covered 119 beef production scenarios in the United States and Canada, identifies several important considerations for mitigating beef GHG emissions in North America. Overall, our study found that grass-finishing beef produced greater weight-based but lower value-based GHG emissions than feedlot-finishing beef. On one hand, this apparent disparity is explained by the finding that grass-fed beef cattle take longer than feedlot-finished beef cattle to reach slaughter weights during which time they emit GHGs but, on the other hand, they command a 121% higher price than feedlot-finished animals, which result in lower value-based GHG emissions. Importantly, scenarios that considered C-seq found significantly lower GHG emissions from beef production than those that did not do so, yet only about a quarter of the studies in our meta-analysis accounted for C-seq in their GHG emissions estimates. The implications of these findings are teased out in the following discussion.

Role of Cropland and Grazing Management Practices in Beef Cattle GHG Emissions.

When comparing GHG emissions from grass- and feedlot-finished beef production, GHG emissions from the production of crops, notably maize, which form the basis of beef cattle feeds in feedlots, and the effects of grazing management should be considered because land management practices can significantly influence beef cattle-based GHG emissions (44). For example, conservation crop production practices can reduce GHG emissions and increase soil C-seq (45–48). Despite the importance of cropland management on the GHG emissions of feedlot feeding, tillage practices were reported in only 12 of the 32 papers included in our meta-analysis (19, 37–39, 42, 49–55), and only one study considered GHG emissions from soil erosion resulting from poor crop cultivation practices (36).

Grazing management practices could affect forage quality and, therefore, the values of methane conversion factor and GHG emissions of beef cattle. For example, Chiavegato et al. (56) found that higher stocking rate, lower stocking density, and proper rest period can reduce the methane conversion factor from 6.4 to 3.8% for lactating beef cows. However, almost all the studies in our meta-analysis adopted the default methane conversion value of 6.5% (57) for grazing beef cattle without distinguishing grazing management practices. One exception is Tichenor et al. (58), who found that, compared to the default conversion value of 6.5%, a seasonally weighted average conversion value of 5.5% led to a 9% lower global warming potential (GWP) of grass-finishing beef. Accordingly, grazing management that leads to higher quality forage may substantially decrease the global warming impact of grass-finishing beef production.

Despite the potential positive effect of improved grazing management on beef cattle GHG emissions, only seven of the 32 studies in our meta-analysis identified their grazing management practices, such as management intensive/adaptive multipaddock/short-duration, high intensity grazing (36, 38, 49, 58–60) or minimal rotational grazing (34), and only one compared GHG emission differences across grazing management practices (61). To accurately assess the GHG emissions of beef production, it is necessary for future studies to account for the cropping and grazing management effects.

Soil C-seq.

Grasslands cover about 40.5% of the Earth’s terrestrial surface (excluding Greenland and Antarctica) and have been estimated to store about one-third of the global terrestrial carbon stocks (62). In North America, Mollisols are the most extensive soil order, accounting for about 21.5% of the land surface; they are deep soils with high organic matter, and nutrient-rich surface horizons (63). If managed properly, such grassland soils have the potential to store vast amounts of carbon and mitigate most of the GHG emissions from livestock sector (64). However, these soils are subject to structural collapse and massive soil erosion when deprived of continuous ground cover as in simple crop rotation fields that applies excessive amounts of fertilizers to boost crop production (44).

Therefore, keeping rangelands intact plays the most important role in maximizing soil C stocks (65). In comparison to long-term grazing removal, moderate cattle grazing on semiarid grassland ecosystems may not affect long-term soil C-seq (66). Furthermore, improving grassland management could play a key role in increasing C-seq (65, 67–69). For example, Wang et al. (70) found that the C-seq potential varies between 0.3 and 3.5 Mg ha⁻¹ y⁻¹ when converting continuous grazing to multipaddock grazing in US Southern Great Plains. Additionally, conversion of cropland to grassland was found to result in an important carbon sink. For example, in the Southeastern United States, conversion of degraded row-crop fields to intensively managed grassland led to C accumulation of 8.0 Mg ha⁻¹ y⁻¹ (71). Such results have led some to conclude that application of intensified grazing management and land use conversion can potentially offset GHG emissions or even enable beef production to become a net carbon sink (36, 42, 44, 72).

Soil C-seq rates depend on soil types, precipitation, grazing, and land management legacies (65). While land converted from row crop production to grazing land has been shown to have an initially high C-seq rate, such rate tends to diminish over time as soils approach new equilibrium level (42, 73). Based on a literature review, Conant et al. (74) found that C-seq rates from different types of management improvements ranged from 0.11 to 3.04 Mg C ha⁻¹ y⁻¹.

Of the eight studies in our meta-analysis that incorporated soil carbon scenarios, two focused on feedlot-finishing systems (39, 42), one on the grass-finishing system (60), and the remaining five studies compared feedlot- vs. grass-finishing systems, among which two studies found lower net GHG emissions in feedlot- than grass-finishing (58, 59), whereas the other three found the opposite outcome (36, 38, 43). Due to the limited focus on C-seq in beef cattle production analysis, most C-seq data used by studies in our meta-analysis applied previously published average sequestration values of 0.41 (75) or 0.39 to 0.46 Mg C ha⁻¹ y⁻¹ (76) without accounting for local circumstances. Only two studies included in our meta-analysis analyzed on-farm soil organic carbon change. One study found that, over 4 y, adaptive multipaddock grazing sequestered 3.59 Mg C ha⁻¹ y⁻¹ (36), whereas the other study found that, over 20 y, multispecies pasture rotation on previously degraded cropland sequestered 2.29 Mg C ha⁻¹ y⁻¹ (60). The limited information about livestock-related soil C-seq underscores the urgent need for broader estimation of such metrics under varying land management history, grazing management approaches, and locations under varying soil and climate conditions.

Implications of Evaluating GHG Emissions by Economic Activity.

To improve production efficiency, feedlots commonly incorporate growth-enhancing hormones to improve beef cattle performance and antibiotics to minimize liver abscesses associated with long-term consumption of grain-based diets by beef cattle (77, 78). It has been estimated that eliminating the use of growth hormones in feedlots could lead to 9.8% greater GHG emissions per kg in beef cattle (79). Yet, over the last two decades, increasing human health and animal welfare concerns over feedlot meat production have led to a rapid growth in demand and increased willingness to pay a sizeable premium for grass-finishing beef (26, 28, 80, 81). As a result of this trend, in addition to weight-based GHG emissions, we also compared different beef production systems using the value-based indicator. When price premiums are accounted for, our results indicate that grass-finishing beef has a lower climate impact than feedlot-finishing beef.

Nevertheless, shifting entirely to grass-finishing beef production faces numerous barriers. For example, due to the slower fattening rate and lower finishing weight, such a shift would require an increase in the cattle population to maintain beef supply, yet the expansion of grass-finishing beef production would be constrained by the availability of grassland and marginal cropland (82). Taking this resource constraint into account, an entire shift to grass-finishing beef would support only about 60% of current beef production in the United States (83). Furthermore, if not managed properly, an accelerated shift toward grass-finishing could intensify overgrazing, thereby increasing grassland degradation and soil erosion. Therefore, a broader adoption of regenerative grassland management practices is essential to enhance grassland productivity and carrying capacity (82).

Additionally, a reduction in beef supply due to a shift away from feedlot-finishing would likely lead to an increase in beef prices. In developed countries, where overconsumption of protein has been associated with health-related problems (84), an increase in beef prices could curtail overconsumption and reduce GHG emissions per economic activity. Alternatively, without a concurrent reduction in demand, a reduction in beef supply in the United States and Canada could increase beef imports. This could cause significant environmental leakage if those imports were to come from places with active deforestation for cattle production or otherwise high GHG emissions, or other environmental impacts. Moreover, beef supply reduction from the leading beef exporting countries such as the United States and Canada may limit protein intake and intensify malnutrition problems in some developing countries. If there is a paradigm shift toward production of more grass-finished beef, the growth could be targeted to areas where cropland is becoming increasingly marginal, e.g., some areas overlying the Ogallala aquifer that will be unsuitable for crop production as the aquifer becomes depleted (85).

Climate Challenges and Regional Sustainability in Beef Production.

Given that climate change projections predict more frequent, severe, and prevalent droughts, irrigation plays an important role in the sustainable production of beef cattle in many regions of North America. In regions, such as the Southern Plains, the long-term sustainability of feedlot-finishing is being undermined as the aquifers used to irrigate the grain crops used for feedlot diets are being depleted at an accelerated rate (43). Additionally, pasture irrigation is often required to sustain grass-finishing beef operations, especially in regions where drought occurs frequently (34, 86). Yet of the studies incorporated in our meta-analysis, only ten considered the irrigation status of cropland (27, 34, 43, 51–55, 79, 86), and seven mentioned the irrigation status of pasture. Given that the groundwater resources needed to irrigate the grain-based feed crops used in feedlots are likely to become increasingly depleted in many regions of North America, long-term regional sustainability should be addressed when comparing the environmental efficacy of alternative beef cattle production systems.

Temporal Change in Beef-Related GHG Emissions.

Our meta-analysis of beef cattle studies in the United States and Canada during 2001 to 2023 indicates a downward trend in GHG emissions over time. When converting different types of GHG emissions to carbon dioxide equivalent, all studies in our sample used the 100-y GWP conversion factors; however, the conversion values have changed considerably during our study period. For example, the GWP for methane has been adjusted upward from 21 to 34, with most papers using intermediate values. By contrast, the nitrous oxide conversion value has been adjusted downward, from 310 to 265. Because methane constitutes a much higher proportion of GHG emissions than nitrous oxide, the preceding changes in these conversion factors tend to increase total anthropogenic GHG emissions in carbon dioxide equivalents (53). Despite the changes in GWP metrics tend to shift the emissions upward, a downward trend in beef GHG emissions during 2001 to 2023, which suggests that there has been an increase in beef production efficiency over the past two decades.

Concluding Summary.

The results of our North American meta-analysis have implications for measuring and mitigating the carbon footprint of beef production, which is of interest to policymakers, producers, and researchers. First, our study highlights the implications of using value-based functional units in addition to weight-based functional units when comparing GHG emissions of differentiated agricultural products. Second, we pointed out the necessity to account for cropping and grazing management effects in assessing GHG emissions of alternative beef production systems, which is lacking in most studies. Third, long-term regional sustainability should be evaluated when comparing the environmental efficacy of alternative beef cattle production systems, given that the groundwater water widely used to irrigate the grain-based feedlot diets is becoming increasingly depleted in many regions throughout North America. Fourth, our findings showed that soil C-seq has the highest mitigation potential for beef GHG emissions and emphasize the need to better quantify soil C-seq under different grazing management practices and in different regions.

Materials and Methods

Selection of Studies.

We reviewed peer-reviewed journal articles about the climate impacts of beef cattle production published between 2000 and 2023. A search for relevant publications was conducted in Google Scholar, Science Direct, and Wiley Library using the following keywords: beef cattle, beef production, beef cattle production, carbon emission, climate impact, GHG emission, life cycle assessment, and Canada, North America, and the United States. Using those keywords, we found three highly relevant recent review papers (14–16) that provide a global synthesis of beef cattle production GHG emissions. Excluding overlapping papers, these three reviews included 76 individual studies and they provided the basis for locating original research on GHG emissions of beef cattle production in North America. We then conducted a backward and forward citation search of the initially identified papers.

Among the initially screened papers, we determined whether the studies should be included in the final sample by checking their titles, abstracts, and keywords. All of the studies included in our meta-analysis meet both of the following criteria: 1) focus on the GHG emission effects of beef production systems and 2) focus on beef cattle production in the United States or Canada. Ultimately, we identified 17 studies that were covered by the review papers, and 15 additional studies excluded from the review papers because they did not meet their inclusion criteria or were published more recently. These 32 studies included 115 beef production scenarios that met our search criteria, of which 21 scenarios considered soil C-seq. Among three studies that compared historical and current scenarios (19, 27, 51), we only retained scenarios that assessed GHG emissions under the prevailing beef production technologies, as the historical scenarios were decades prior to the estimation time and could be biased.

Functional Units and Conversion Method.

The functional units used in the beef cattle production life cycle assessments that met our evaluation criteria were all based on CW, live weight (LW), or retail cuts. Among the 32 studies that met our inclusion criteria, eight included only LW (19, 35, 38, 49, 51, 87–89), 16 included only CW (27, 30, 33, 34, 36, 43, 52–54, 58–60, 79, 86, 90), and six included both LW and CW as the functional units (39–42, 50, 61). The remaining two studies used retail cuts as the functional unit (55, 91)^*. Except one study (55), all studies on GHG emissions of beef production excluded production stages beyond the farm gate, such as packing, retail, and restaurant use; therefore, to ensure consistency in our meta-analysis, we focus only on within-farm beef cattle production stages.

GHG emission estimates that used CW and LW can be compared because CW = LW × DP, where DP represents dressing percentage, which ranged from 50.0 to 62.5% among the grass-finishing scenarios (33, 34); and from 55.0 to 63.8% among the feedlot scenarios (59, 79). To avoid result discrepancies caused by DP assumption variations, we derived both CW- and LW-based emission values for all studies using the conversion methods discussed below.

For studies that included DP values, we used those values to convert CW- to LW-based emissions, and vice versa. For studies that did not provide a DP value for feedlot scenarios (49, 51, 61, 87, 88), we used 59% as the DP value, which is a weighted average derived from three studies that assumed that DP values are 62% for finished beef cattle and 50% for cull cows and that 76% of beef comes from finished beef cattle and 24% from cull animals (53, 55, 91). Retail cut was converted to LW or CW-based on the assumption that 29% of the LW is consumer benefit (55) and that retail cuts are, on average, 53.5% of CW.

During 2014 to 2021, the premiums paid for grass- over feedlot-finished beef differed across cuts and ranged between 48% (flank steak) and 193% (filet mignon) (26). To account for value differences between feedlot- and grass-finished beef, we adopted the intensity indicator based on economic activity, which captures GHG emissions resulting from the production of one MU of beef (31), as an alternative indicator to compare the climate impacts of alternative beef production systems. For CW-based functional units, the average price of feedlot-finished beef was set as the baseline MU, and the MU of grass-finished beef was derived from the average premium above this baseline [(0.48 + 1.93)/2 + 1 = 2.21].

To compare the climate impacts of different beef production systems, two indicators were compared: 1) GHG emissions per weight unit; and 2) GHG emissions per MU, or C intensity per economic activity. We assumed that one kg of feedlot- and grass-finished are valued at one and 2.21 MU(s), respectively. Given that these two GHG emission estimates were each highly correlated for CW and LW (98% and 99%, respectively, P < 0.01), the CW-based indicators can be substituted for LW-based indicators without loss of information. Therefore, for brevity, we presented only the CW-based results in the main manuscript and included LW-based GHG emission results in SI Appendix.

GHG Mitigation Potential.

To eliminate the impact of assumption differences across different studies, we conducted our intrastudy analysis of GHG mitigation potential by calculating the percentage differences of each scenario in the study relative to the highest-value scenario in the same study. This applied only to the studies that incorporated multiple beef production scenarios. No GHG mitigation potential was calculated for studies with only one beef production scenario, except for two studies from the northwestern United States, for which we used the estimated GHG emission of locally produced beef (87) as a baseline for that of regionally produced beef (88).

Mitigation potentials, or differences in GHG emissions across scenarios, were estimated using the GHG emissions per CW (MitiP_CW) and GHG emissions per MU (MitiP_MU). For example, we calculated the GHG mitigation potential of scenario X as

$$MitiP\_CW=\frac{CW\_X\_CW\_max}{CW\_max}*100\%,$$ [1]

where CW_max is the emission value for the highest CW-based GHG emission scenario in the given study, and CW_X is the CW-based GHG emission value for scenario X. No mitigation potential is calculated for the highest emission scenario, which is treated as the baseline. Due to the almost perfect correlation between the CW- and LW-based emission indicators, as reported above, we did not calculate the mitigation potential for the LW-based indicators.

Comparisons of GHG Emissions.

Of the 32 studies in our meta-analysis, eight considered C-seq by accounting for changes in soil organic carbon in their estimates (36, 38, 39, 42, 43, 58–60). GHG emissions were compared between scenarios that considered and did not consider soil C-seq to estimate the potential effect of soil C-seq on the climate impact of beef production.

Additionally, GHG emissions were compared across different beef production systems that were broadly categorized as either feedlot- or grass-finishing. Feedlot-finishing was then divided into three subcategories based on breed of calves: 1) feedlot–beef only included only beef breed calves; 2) feedlot–dairy only included only dairy breed calves; and 3) feedlot–mixed systems included beef and dairy breed calves (SI Appendix, Table S1). Duncan’s multiple range tests were conducted with different letters indicating means differ at a significance level of 5%.

Beef Production Regions and Study Years.

To investigate potential regional effects of GHG emissions in beef production, we further divided US beef production scenarios into seven regional categories according to Rotz et al. (53), which are Midwest, Northeast, Northern Plains, Northwest, Southeast, Southern Plains, and Southwest states. The study with states that spanned across two regions (e.g., ref. 91) was placed in both regions. US-based study scenarios that did not specify states or regions were included in the “All US” category. Similarly, as most studies in Canada did not specify region, they were all placed in the “All Canada” category. Besides the regional variables, we also included year of publication as a variable to analyze the time trend of beef GHG emissions; we used publication year because the assessment year was unavailable in most cases.

Ordinary Least Square Regression Models.

Regression models were developed to identify factors affecting GHG emissions per CW and per MU. Explanatory variables included production systems (grass-finishing, beef-only, mixed, and dairy-only feedlot systems), regions in the United States, and publication year of the study. The beef-only feedlot category and Canada serve as control variables. Among the regression models, Models I and II incorporated the 90 scenarios that included all production phases in their estimations, while Models III and IV used all 115 scenarios, including the 25 scenarios that do not contain cow–calf or stocker phases, or both, so we could ascertain the effects of these two phases on GHG emissions of beef production. Variables with limited observations, i.e., days to slaughter, proportion of time spent on grassland, and fishing weight (Table 1), were excluded from the regression analyses, but their associations can be inferred from the correlation table (SI Appendix, Table S5).

The two regression models that aim to identify factors affecting GHG emission mitigation potentials (Models V and VI) incorporated explanatory variables in three categories, which were comparison type, region in the United States, and publication year. Comparison types are defined in Table 1. We further divided the scenarios with feedlot- vs. grass-finishing system comparisons into feedlot-finishing scenarios (System comparison II—feedlot) and grass-finishing scenarios (System comparison II—grass) due to the differences in GHG emissions between feedlot- and grass-finishing systems.

Scenarios that consider C-seq were not included in the regression models for the following reasons: 1) the C-seq values used in the literature are mostly arbitrary and hypothetical; 2) scenarios with C-seq have the same beef production systems and technologies as their counterpart without C-seq. Therefore, our mitigation models aim to identify factors affecting GHG emission mitigation potentials without C-seq, while the effect of such comparison (with and without C-seq) are demonstrated in Figs. 2 and 4.

Disclaimer.

The findings and conclusions in this report are those of the authors and should not be construed to represent any official USDA or US Government determination or policy.

Supplementary Material

Appendix 01 (PDF)

Dataset S01 (XLSX)

Data, Materials, and Software Availability

All data are included in the manuscript and/or supporting information.