Comparison of food carbon footprints using 20- and 100-year timeframes

1 Introduction

As anthropogenic activities continue to accelerate climate change at unprecedented rates, providing greater transparency in emissions calculations remains paramount for informing public and policy-making decisions. Global Warming Potential (GWP) quantifies the radiative forcing of a Greenhouse Gas (GHG) over a timeframe relative to carbon dioxide (CO₂). It is often used in life cycle assessments (LCAs) and other environmental calculations to convert GHG emissions into a standard unit of carbon dioxide equivalence. GWP over a 100-year period (GWP100) is commonly used as a metric in emission assessments, but using GWP100 alone can underrepresent impacts of gases with short-lived lifespans. Using GWP over a 20-year period (GWP20) in addition to GWP100 is especially useful in analysis of food system emissions as GWP20 can better account for climate impact of prevalent agricultural emissions such as methane (CH₄). However, food emission databases lack individual GWP20 derived emission data that could be essential in political and institutional food policy shifts.

The Intergovernmental Panel on Climate Change (IPCC), which developed the framework for GWP in 1990, acknowledged the practical limitations of GWP in their First Assessment Report after providing example figures for 20-, 100-, and 500-year time periods, stating that each time horizon “should not be considered as having any special significance” [Intergovernmental Panel on Climate Change (IPCC), 1990, p. 59]. However, the timeframe strongly influences these calculations, and in 1995, after the release of the IPCC's 1994 Special Report [Intergovernmental Panel on Climate Change (IPCC), 1994], the 100-year time frame was adopted collectively during the first Conference of the Parties in 1995 [United Nations Framework Convention on Climate Change (UNFCCC), 1995, p. 14]. While many have agreed that one time frame should not be presented alone and that the short-term impacts of certain GHGs on climate, environmental systems, and human health must be considered (Balcombe et al., 2018; Mar et al., 2022; McAuliffe et al., 2023) the majority of literature and policy reports have used and continue to use GWP100 alone (Heller and Keoleian, 2015; Clune et al., 2017; Poore and Nemecek, 2018; Springmann et al., 2018).

GWP is particularly important through the lens of food systems, which alone contribute to about one third of anthropogenic GHG emissions (Crippa et al., 2021). A business-as-usual food system will, on its own, emit more GHGs than the total allowable budget from all sectors to remain within the 1.5 degrees C warming level (Clark et al., 2020). CH₄ and nitrous oxide (N₂O) are significant components of this global agricultural footprint, accounted for by the metabolism of ruminant animals, including cows, goats, and sheep, as well as manure management processes. Both gases are significantly more potent than CO2 over their lifetimes; however, they persist for a shorter period.

Limited recent studies have presented dietary LCA data using both a 20-year and a 100-year time frame. Using Food Frequency Questionnaire data from 55,504 participants, Scarborough et al. (2023) compared the dietary carbon footprints of vegans, vegetarians, pescatarians, and meat-eaters under both GWP20 and GWP100. While the trends in carbon footprint between the different diets remained the same for the two time points, the absolute numbers changed, as did the gap between the groups. With the 100-year timeframe, the carbon footprint decreased by about 65% between medium meat consumption and a vegan diet and by 76% between a high-meat diet and a vegan diet. When the time frame of interest was shortened to 20-years, the data show a 71% decrease from medium-meat consumption and a vegan diet, and an 82% decrease from high-meat consumption to a vegan diet. Ivanovich et al. (2023) also compared carbon footprints for the food system under 100- and 20-year timeframes, estimating the climate impact of the entire global food system to be 13 GtCO₂e year⁻¹ with a 100-year timeframe and 22 GtCO₂e year⁻¹ with a 20-year timeframe. The paper also provides a breakdown of the contributions of the various gases to global warming for a series of foods, and it presents a comparison of the GWP20 and GWP100 values for 12 broad food categories in the supplemental figures. The footprint of ruminant meat shows the greatest difference between the two timeframes.

Even though the need for consideration of shorter timeframes in LCA calculations has been widely mentioned in the literature, GWP20-derived emission data for a wide-ranging database of foods are not readily available. The goals of this work are to: (1) calculate GWP20 CO₂-equivalent emissions for 16 broad food categories and 222 specific food items in a comprehensive database alongside GWP100 CO₂-equivalent values, utilizing known mass ratios between various GHGs; and (2) use both the GWP20 and GWP100 values in carbon footprint calculations of college dining hall meals as a case study with real-life implications.

2 Materials and methods

Ivanovich et al. (2023) built a comprehensive emissions inventory for individual food items based on life cycle analyses of individual food items using GWP100. The Supplemental Information from Ivanovich et al. (2023) provides a table of values for CO₂, CH₄, and N₂O in kg gas/kg food for these foods. We used these data to derive ratios between the three gases for every category of food items. Then, using commonly referenced databases (Heller and Keoleian, 2015; Clune et al., 2017) for total CO₂ equivalent emissions under GWP100 and applying both the ratios from Ivanovich et al. (2023) and the multipliers for GWP100, we calculated the amounts of all three gases (consistent with the given ratios) that led to the commonly referenced values, as the breakdown of various gases is not typically given in databases. Once the amounts of each gas were known, the calculation of carbon footprints in CO₂eq could be calculated using the standard equation and known multipliers for each time frame. The above methodology is demonstrated in the following workflow:

m_{C_*O*2}, m_{C_*H*4}, and m_N₂O are values noted for each designated food in Ivanovich et al., 2023:

mCO2=kg CO2kg foodmCH4=kg CH4kg foodmN2O=kg N2Okg food

Using the above, we solved for the ratio of CH₄ and N₂O to CO₂, named as r_*gas* below:

rCH4=mCH4mCO2rN2O=mN2OmCO2

The GWP equation over 100 years is shown below, where GWP100CH4-CO2eq, the 100 year warming potential of CH₄ (non-fossil), is the measure of how much energy 1 kg of CH₄ absorbs relative to 1 kg of CO₂ over 100 years, and GWP100N2O-CO2eq, the 100 year warming potential of N₂O, is the measure of how much energy 1 kg of N₂O absorbs relative to 1 kg of CO₂ over 100 years, and TotalC_O₂eqGWP100 is the estimated total equivalent amount of CO₂ (kgCO₂) produced per kg of a food based on each gas's impact on warming over a 100 year time frame.

GWP100CH4-CO2eq=27.0

From Forster et al. (2021), Table 7.15, CH₄-non fossil GWP-100.

GWP100N2O-CO2eq=273

From Forster et al. (2021), Table 7.15, N₂O GWP-100,

Total C_O₂eqGWP100 values for various foods were found from Heller and Keoleian (2015) and Clune et al. (2017) to get a broader selection of Total C_O₂eqGWP100 data for individual foods. Ivanovich et al. (2023) derived ratios from general food groups were matched up with their corresponding specific food from which the Total C_O₂eqGWP100 were taken from:

Total CO2eqGWP100 =mCO2+GWP100CH4-CO2eq×mCH4 + GWP100N2O-CO2eq×mN2O

From the earlier ratios, we substitute

mCH4=rCH4×mCO2mN2O=rN2O×mCO2

to solve for m_{C_*O*2},

mCO2=Total CO2eqGWP1001+GWP100CH4−CO2eq×rCH4+ GWP100N2O−CO2eq×rN2O

Then, using m_{C_*O*2}, m_{C_*H*4} and m_N₂O can be computed for each food.

With each individual component of the total emissions computed, we were able to plug m_{C_*O*2}, m_{C_*H*4}, and m_N₂O into both the GWP equations:

Total CO2eqGWP100 =mCO2+GWP100CH4-CO2eq×mCH4 + GWP100N2O-CO2eq×mN2OTotal CO2eqGWP20 =mCO2+GWP20CH4-CO2eq×mCH4 + GWP20N2O-CO2eq×mN2O

Where Total C_O₂eqGWP20 is the total kg of equivalent CO₂ released per kg of a food based on GWP20,

GWP20CH4-CO2eq=79.7

From Forster et al. (2021), Table 7.15, CH₄-non fossil GWP-20.

GWP20N2O-CO2eq=273

From Forster et al. (2021), Table 7.15, N₂O GWP-20.

Table 1 shows the applied above methodology for eight selected food items depicted in Figure 1. Additional calculations can be found in the Supplementary Information.

Table 1. Methodology applied to eight food items.

Expanded food items were compiled Heller and Keoleian (2015) and Clune et al. (2017) and used referenced ratios from eight corresponding food items from Ivanovich et al. (2023).

To contextualize GWP20- and GWP100-derived carbon-footprint estimates in an institutional setting, we selected 15 meals actually served across UCLA Dining Hall menus, chosen to span a range of protein and dietary categories (ruminant meat, poultry, pork, fish, dairy, egg, and plant-based) for comparison. Ingredient lists for each meal (all of which individually constitute one serving size) were obtained from UCLA Dining's published nutrition facts. Each recipe was split into its constituent ingredients, and the mass (g) of each ingredient was estimated using standard volume-to-mass conversions together with USDA FoodData Central entries (U.S. Department of Agriculture, 2019). Each ingredient was then matched to its corresponding food item in our combined GWP20/GWP100 database (Table 2) or, where no exact match existed, to the closest available food category, to calculate its individual CO₂eq contribution. These per-ingredient contributions were used to create the CO₂eq per meal reported in Figure 2, under both the GWP20 and GWP100 frameworks.

Table 2. GWP20 and GWP100 derived Total CO₂eq for individual foods, grouped by food category (kg CO₂eq/kg food).

3 Results

Total C_O₂eqGWP100, Total C_O₂eqGWP20, and corresponding emission components (CH₄−CO₂eq and N₂O−CO₂eq for GWP20 and GWP100) were determined for 222 individual food items. Note that as only the GWP of CH₄ (non-fossil) varies from a 20 to 100 year timeframe (N₂O retains the same value of 273 under GWP20 and GWP100), the differences generated by changing the time horizon are essentially driven by methane. Food items were also categorized into general food groups to demonstrate higher differences in emission changes between GWP20 and GWP100 derived values in different foods. Averages and deviations for Total C_O₂eqGWP20 and Total C_O₂eqGWP100 are shown across food groups in Table 3.

Table 3. GWP20 and GWP100 derived Total CO₂eq for different food categories in kg CO₂eq /kg food.

Standard deviation and coefficient of variation are not reported for Animal Fats (n = 1 thus statistically undefined) or Rice (n = 2 with identical values yielding insignificant 0).

Comparing total CO₂ equivalent emissions derived from GWP20 vs. GWP100 for different food categories reveals that the scale of difference is most prominent in ruminant food sources such as red meat, where Total C_O₂eqGWP20 is approximately 2.34-fold the value of Total C_O₂eqGWP100 (Table 3). In non-ruminant foods, the scale of difference is smaller—vegetables only see a 7% increase from Total C_O₂eqGWP100 to Total C_O₂eqGWP20. Diverse selected foods are highlighted in Figure 1 to demonstrate CH₄'s impact on differences between Total C_O₂eqGWP100 and Total C_O₂eqGWP20.

Total C_O₂eqGWP20, and Total C_O₂eqGWP100, for individual foods are shown in Table 2. Using CO₂, CH₄−CO₂eq, and N₂O−CO₂eq for GWP20 and GWP100 from our calculations, we applied this data to real meals served at a university dining hall to contextualize use of GWP derived emissions in an institutional setting. Figure 2 demonstrates that although the types of meals with the largest Total CO₂eq don't change, the magnitude of those emissions does change significantly for some meals. Generally, ruminant source foods increase their impact on an overall meal under GWP20 given CH₄'s greater near-term radiative forcing. This is exemplified in the larger share of emissions contributions in a meal under GWP20 by dairy when compared with non-ruminant animal proteins (e.g., Meat-poultry in the Chicken Katsu Sandwich). The ratio of dairy to plant-based ingredients in the Impossible Burger also increases under GWP20 with the former now accounting for closer to half of the Total CO₂eq of the meal compared with roughly one-third under GWP100.

4 Discussion

In a food policy setting, complementing GWP100 with GWP20 derived values could offer a simple and streamlined solution to portraying the near-term potency of CH₄, which is central to agricultural systems. Due to methane's high impact yet short atmospheric perturbation time, it offers an effective pathway of fast emission reductions. If implemented widely, the use of GWP20 has the potential to provide more urgency in future work impacting climate action and industry reform. For example, using GWP20 in addition to GWP100 to calculate food emissions could change whether a meal meets the threshold for low and high carbon footprint labels. This could directly influence consumer choice as interest in consumer-facing climate labels grows in the form of both front-of-packaging “climate scores” (European Parliament. Directorate General for Parliamentary Research Services, 2023) and university menus (Wolfson et al., 2025). Numerous studies have shown that effectively designed climate labels can enhance consumer awareness of the environmental impacts of foods and ultimately encourage climate-friendly choices (Camilleri et al., 2019; Slapø and Karevold, 2019; Crippa et al., 2021; Fresacher and Johnson, 2023; Beyer et al., 2024; Lenk et al., 2025).

At the wider scale, as cities and institutions begin to integrate climate targets within food policy, the proportional contribution of distinct ingredients will have substantial consequences on the overall climate impact of institutional procurement. Using GWP20 could potentially help inform the substitution of high footprint ingredients or production systems for replacements that are less environmentally detrimental. Similarly, the revision of subsidies to favor the production of less emissions intensive foods at regional and national levels could also potentially benefit from the use of GWP20 in highlighting methanogenic production systems, especially now when climate finance is yet to substantially penetrate agrifood systems (Galbiati et al., 2023) and can be guided toward more impactful frameworks.

It is critical to consider the role of positive feedback loops and tipping points in this discussion (Armstrong McKay et al., 2022). The latest status update of the planetary boundaries concept (Richardson et al., 2023) shows that we are currently exceeding six of the nine planetary boundaries, including the climate change boundary. Additionally, recent extension of the framework that considers the injustice resulting from the disproportionate burdens shouldered by some populations shows that we are currently at or beyond all of the boundaries (Rockström et al., 2023). In de-emphasizing the importance of potent short-lived gases like CH₄, the contribution to positive feedback loops and the instigation of tipping points (planetary boundaries) are mathematically overlooked. Incorporating GWP20 alongside GWP100 could aid in the accuracy and standardization of climate reporting in all affected communities and industries.

Several limitations should be considered when interpreting the results of this study. First, the total CO₂eq values used in our calculations come from two independently constructed literature meta-analyses—one aggregating 369 studies across 168 food varieties Clune et al., (2017) and one compiling GHG intensity values for approximately 100 foods from a range of approximately 25 published LCA sources (Heller and Keoleian, 2015). While (Clune et al. 2017) does record study-level system boundary and geographic information, neither source reports IPCC characterization factor version used in the GWP100 totals. Therefore, determining system boundaries, geography, or and production systems from the above studies with sufficient specificity to fully harmonize with the gas ratios we apply from a third, independent dataset (Ivanovich et al., 2023), was outside the scope of our ability within this study. Secondly, this workflow assumes that the relative contribution of CH4, N₂O, and CO₂ within a food category, as reported by Ivanovich et al. (2023), can be validly transferred to the specific food items and production systems represented in our other two source datasets. This assumption cannot be independently verified with the data available to us and may not hold where production systems, geography, or technology diverge substantially. Third, because the source GWP100 values may have been calculated using earlier IPCC characterization factors than the AR6 values we apply throughout, our back-calculated gas masses and resulting GWP20 values are best interpreted as internally consistent estimates rather than recovered physical emissions. Fourth, several food items lacked category-specific gas-ratio data and were instead assigned a proxy value from the most similar available food category, introducing additional uncertainty for those items. Fifth, we do not propagate uncertainty from the source datasets through our calculations, so the deviation statistics reported in Table 3 reflect only the spread of point-estimate values across food items within a category, not the underlying measurement or modeling uncertainty in each value. Sixth, our functional unit (kg CO₂eq per kg food) does not account for nutritional density or serving-size equivalence across food types, which affects comparability between, for example, leafy greens and animal proteins. Finally, the UCLA Dining Hall case study reflects the recipes, sourcing, and portioning practices of a single university dining program and may not generalize to other institutional or consumer contexts.

5 Conclusion

While applying multiple time frames in LCA calculations is recognized as important, data allowing calculations using GWP20 are not widely available. This paper presents both GWP20- and GWP100-derived carbon-footprint estimates, recharacterized from existing LCA-based emissions data using externally derived gas ratios, for broad food categories and specific food items. The carbon footprints of actual university dining hall meals are calculated using both time frames side by side. While all meals generally have higher footprints using GWP20, meals containing ruminant meat have a wider gap between the two timeframes. GWP20-derived food carbon-footprint estimates reflect near-term impacts and can result in increased potential for mitigation through education and consumer choice. Using GWP20-derived CO₂ emissions alongside GWP100 may better exemplify the proportion and magnitude of food impacts on the environment, with potential implications for agricultural production, commodity purchasing, subsidization, cultural focus, food policy, and institutional procurement if applied in future analyses.

Generative AI statement

The author(s) declared that Generative AI was used in the creation of this manuscript. During the revision of this manuscript, the author(s) used Claude (Sonnet 5; Anthropic, PBC) to cross check quantitative calculations and identify any possible quantitative or qualitative discrepancies. The author(s) carefully reviewed and verified any suggested discrepancies and updated the manuscript accordingly.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher's note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fsufs.2026.1717648/full#supplementary-material

Supplementary material Supplementary material at the source

Footnotes

Edited by: Isabella Nyambayo, Glyndwr University, United Kingdom

Reviewed by: Piotr Rzymski, Poznan University of Medical Sciences, Poland

Paschal Simon Milindi, University of Dar es Salaam, Tanzania