, Volume 94, Issue 1, pp 29–41

The human carbon budget: an estimate of the spatial distribution of metabolic carbon consumption and release in the United States

  • Tristram O. West
  • Gregg Marland
  • Nagendra Singh
  • Budhendra L. Bhaduri
  • Adam B. Roddy
Open Access

DOI: 10.1007/s10533-009-9306-z

Cite this article as:
West, T.O., Marland, G., Singh, N. et al. Biogeochemistry (2009) 94: 29. doi:10.1007/s10533-009-9306-z


Carbon dioxide is taken up by agricultural crops and released soon after during the consumption of agricultural commodities. The global net impact of this process on carbon flux to the atmosphere is negligible, but impact on the spatial distribution of carbon dioxide uptake and release across regions and continents is significant. To estimate the consumption and release of carbon by humans over the landscape, we developed a carbon budget for humans in the United States. The budget was derived from food commodity intake data for the US and from algorithms representing the metabolic processing of carbon by humans. Data on consumption, respiration, and waste of carbon by humans were distributed over the US using geospatial population data with a resolution of ~450 × 450 m. The average adult in the US contains about 21 kg C and consumes about 67 kg C year−1 which is balanced by the annual release of about 59 kg C as expired CO2, 7 kg C as feces and urine, and less than 1 kg C as flatus, sweat, and aromatic compounds. In 2000, an estimated 17.2 Tg C were consumed by the US population and 15.2 Tg C were expired to the atmosphere as CO2. Historically, carbon stock in the US human population has increased between 1790 and 2006 from 0.06 Tg to 5.37 Tg. Displacement and release of total harvested carbon per capita in the US is nearly 12% of per capita fossil fuel emissions. Humans are using, storing, and transporting carbon about the Earth’s surface. Inclusion of these carbon dynamics in regional carbon budgets can improve our understanding of carbon sources and sinks.


Agriculture Carbon dioxide Food consumption Horizontal carbon transfer Population Respiration 


Humans are heterotrophic organisms that ingest carbon in the form of complex carbohydrates and release carbon in the form of carbon dioxide, methane, and a mixture of solid, dissolved, and volatile waste products. As early as 1843 Justus Liebig recognized that humans were heated by the oxidation of carbon compounds and as a result released “carbonic acid” to the atmosphere (Leibig 1843). With a global population now exceeding 6.5 billion people, humans move large amounts of carbon about on the Earth’s surface: growing, transporting, and consuming agricultural and forest products. Movement of carbon for human consumption is particularly evident in developed countries where food is now grown in concentrated cropland regions and transported over sometimes large distances for consumption by humans in metropolitan areas and by animals in pasture and rangeland areas and on farms. These transfers of carbon are significant in our attempt to understand the spatial and temporal distribution of the global cycling of carbon (Ciais 2007; Denning et al. 2005). The objective of the current research is to balance the individual, human, metabolic carbon budget and to estimate the spatial and temporal distribution of carbon releases, particularly of carbon dioxide.

Carbon dioxide is taken from the atmosphere by agricultural crops through the process of photosynthesis. After crops are harvested, CO2 is released back to the atmosphere during the decomposition of crop residues and during the processing and consumption of agricultural commodities by humans and livestock. Released CO2 is taken up again by crops during the following year or during the next crop rotation. This cyclical process is considered to result in no net emission of CO2 to the atmosphere (West and Marland 2002a; EPA 2005; Penman et al. 2003). Carbon dioxide is also released to the atmosphere from the use of fossil fuels in agricultural machinery, the production of agricultural inputs such as fertilizers and insecticides, and during the tillage of agricultural soils (West and Marland 2002b); but our focus here is on the photosynthetic products and the ultimate release of the contained carbon.

While uptake and release of carbon may be approximately balanced for the full Earth and at the temporal scale of a year, the uptake and release of carbon can be very much separated in place and over the course of the year. Ciais et al. (2001) analyzed carbon uptake and release by agricultural products, including imports and exports, at a 1 × 1 degree (latitude/longitude) resolution. Ciais et al. (2001) concluded that the “horizontal displacement is far from negligible and [that] it has a large impact on the spatial (and temporal) distribution of carbon fluxes over large areas of the world…” In an analysis of global agricultural carbon displacement, Ciais (2007) found that CO2 emissions in densely populated regions can be 20% higher when considering CO2 losses from food consumption in addition to fossil-fuel emissions.

In completing a carbon budget for a single US city (Phoenix, Arizona), Koerner and Klopatek (2002) concluded that human activity, including transportation, produced more than 80% of carbon input to the atmosphere. They estimated that 61% of soil CO2 emissions came from agricultural lands and that human respiration alone contributed as much CO2 as five power plants. Koerner and Klopatek (2002) noted, as we have above, that respiration from humans is not a net source of CO2 to the atmosphere, but that humans have a dramatic effect on the carbon balance and the transfer of carbon between regions.

To balance regional carbon budgets it is thus necessary to understand and quantify the location of carbon sources and sinks, and to do this we need to know how humans redistribute carbon around the Earth and in time. The North American Carbon Program is intended to improve estimates of carbon fluxes by improving ground-based and atmospheric-based flux estimates (Denning et al. 2005) and to compare these two estimates for enhanced accuracy (Ogle et al. 2006). The human carbon budget is one component of ground-based estimates.

In this paper, we balance the human carbon budget and estimate the spatial distribution of net carbon release associated with the consumption of agricultural commodities by humans. We focus on the carbon budget, including both CO2 and CH4. Net emissions of CO2 and CH4 are provided in units of carbon and are also, where appropriate, combined in a standard carbon-equivalent (Ceq) unit that is calculated using the 100-year global warming potential of 23 for CH4 (Ramaswamy et al. 2001). Many components of the human carbon budget have not previously been calculated or documented. We document our calculations for carbon intake, expiration, excretion, other bodily releases of carbon, and the spatial distribution of these releases in the methods section of this paper. Additional details on these calculations are provided in the appendix. We use the final estimates to conduct an analysis of the US human carbon budget and report the main components of this budget in the results section.

Methods of carbon accounting

Our focus is on the United States. Methods have been developed for estimating carbon uptake by agricultural crops (Hicke et al. 2004; Hicke and Lobell 2004; Lobell et al. 2002; Prince et al. 2001). These methods can be used to estimate carbon uptake at the county level within the US. Release of this carbon occurs in processing plants, in landfills, and during the consumption of food products by humans and livestock. We focus here on the consumption and release of carbon by humans at a sub-county scale. We seek a mass balance for carbon uptake and release by humans and we consider the stock of carbon in humans. For an individual, the intake and output of carbon should approximately balance over the course of a year. We estimate the intake and output independently and compare the two results. Imbalance in the annual carbon budget, in the form of weight gain, is quantified and discussed.

Consumption of carbon by humans

Carbon consumption per capita was estimated by developing carbon coefficients for food commodities and then analyzing food consumption in the US. A list of food commodities and data on food consumption are available in the food commodity intake database (FCID), available from the National Technical Information Service (EPA 2000). The FCID is a compilation of survey data and includes intake of food commodities by age, sex, and race. Survey data collected for the FCID include 2 days of food intake for ~21,700 individuals, including 11,800 children from birth to 19 years of age. Appendix 1 details the data and assumptions used to arrive at estimates of C intake by humans.

We estimate mean carbon consumption with a 95% confidence interval for all ages at 52.96 ± 0.43 kg C capita−1 year−1. Carbon consumption differs significantly between males and females in all age groups with mean male and female intake across all age groups of 60.14 ± 0.70 and 45.56 ± 0.46 kg C capita−1 year−1, respectively. Data were analyzed in 5-year age groups commensurate with US population block-group data (US Census Bureau 1994) (see Fig. 7). Age groups that were not significantly different from one another were aggregated into revised age cohorts, and this revised grouping (Fig. 1) was used for the remainder of analyses on the human carbon budget. Average intake for young adults in the 15–39 year age group is 67.33 ± 1.21 kg C year−1. There was no significant difference for carbon intake among different races.
Fig. 1

Mean carbon consumption by age and gender in the US. Brackets illustrate the 95% confidence interval

Expiration of carbon by humans

Respiration is the process of inspiring and expiring air. Dry atmospheric air at standard temperature and pressure (STP = 0°C and 760 mm Hg) consists of 78.09% (by volume) N2, 20.95% O2, 0.03% (now approaching 0.04%) CO2, 0.93% Ar, and traces of other gases (Weast 1972). The body uses O2 and generates CO2, thereby expiring air with about 16.5% O2 and 3.5% CO2 (Johnson 1992; Keele et al. 1982; Piiper 1996; Thews 1983; Tuttle and Schottelius 1965). Average minute volume (VE) of air expired is about 6 L, consisting of a tidal volume (VT) of 500 mL per breath at about 12 breaths per minute (Ganong 2001; Johnson 1992; Keele et al. 1982). Mean CO2 released through human respiration at rest is therefore estimated at about 58.6 kg C year−1. This value represents a healthy young adult, and can increase by an order of magnitude during heavy exercise (Piiper 1996).

Data compiled by Altman et al. (1958) showed a significant difference in ventilation (inspiration and expiration) at rest among age classes and between males and females. These data also showed a significant difference in the percentage CO2 content of expired air among age classes, but not between males and females. A trend is evident in the efficiency of oxygen use and CO2 generation with increasing age, such that air intake increases while the percentage of CO2 in expired air decreases (Fig. 2). Combining volume of expiration with the concentration of expired CO2 indicates a trend of total expired CO2 over a gradient of age (Fig. 3) similar to that of food consumption (see Fig. 1). This trend supports our assumption of approximately equal carbon inputs and outputs for the annual human carbon budget.
Fig. 2

Volume of expired air versus carbon content of expired air in 32 subjects ranging from 9 to 85 years of age. Data compiled by Altman et al. (1958)

Fig. 3

Total carbon expiration for 32 subjects, ranging from 9 to 85 years of age. Total expired carbon is the multiple of the volume of expired air and the fractional carbon content (see Fig. 2). Data compiled by Altman et al. (1958)

Excretion and secretion of carbon by humans

In addition to expiring CO2 during the process of respiration, humans also release carbon as excreta, flatus, sweat, and aromatic compounds. Humans excrete, on average, 80–170 g feces and 1,000 g urine per capita per day (Orten and Neuhaus 1982). Carbon in flatus and sweat is a relatively small component of the human carbon budget, but is included here for completeness. Appendix 2 details our estimates of carbon excretion and secretion by humans. Our estimates are that average per capita loss of carbon by humans in the US is 4.62 kg C year−1 in feces, 1.93 kg C year−1 in urine, 0.0815 kg C year−1 in sweat, and 0.0293 kg C year−1 (0.1154 kg Ceq year−1) in flatus.

The treatment of solid or liquid wastes following their exit from the body influences whether the carbon is emitted to the atmosphere as CO2 or CH4. About 25% of domestic wastewater in the US is treated in septic systems (EPA 2005). The remainder of domestic wastewater is treated in central wastewater treatment facilities where the waste is decomposed by either aerobic or anaerobic processes. It has been estimated that 16.25% of wastewater is anaerobically digested (EPA 2005), and this estimate includes treatment in septic systems. As detailed in Appendix 2, our estimate is that total annual carbon release from human waste is 0.53 kg C capita−1 year−1 as CH4 and 6.02 kg C capita−1 year−1 as CO2, with another 0.11 kg C capita−1 year−1 as volatile and inorganic compounds in flatus and perspiration. Considering the global warming potential of CH4, total net Ceq emissions from human waste and subsequent waste treatment is 10.65 kg Ceq capita−1 year−1.

Our analysis of the fate of CO2 from human waste assumes that emissions occur in the same year that carbon is excreted from the body. The majority (75%) of human waste in the United States is processed at wastewater treatment plants, and these treatment plants increase the rate of decomposition in an effort to reduce the chemical oxygen demand of waste effluent. Remaining solid waste is dried and transported to hazardous waste landfills (Jack Graham, Engineer and Assistant Director of Water Quality, Maryville, TN, USA). Emissions of N2O also occur from processing the nitrogenous compounds in human waste at wastewater treatment facilities (Czepiel et al. 1995). In conducting an analysis of N2O emissions from a wastewater treatment plant in Durham, New Hampshire, Czepiel et al. (1995) estimated net per capita N2O emissions of 0.0032 kg N2O per year. The focus here is on the human carbon balance and these compounds are not considered further in spite of the fact that N2O is a greenhouse gas.

Carbon stock of humans

The uptake and loss of carbon by humans represent carbon flows. The total human population is a carbon stock, and this stock changes over time as the total US population increases in number and in mass per person. The human body consists of about 60% water, 15% protein, 6% minerals, and 19% fat (ICRP 1975; Going 1996). Almost all carbon in the body is found in fat, protein, and glycogen, with smaller amounts of carbon in bone mineral and lipids (Heymsfield et al. 1996). As detailed in Appendix 3, our estimate of carbon content in the average human adult is 20.7 kg C per capita. Average carbon content across all ages, ranging from 8 to 27 kg C per capita, is 17.3 kg C per capita.

The historical change in human carbon stock in the US is nearly proportional to the increase in population, with the addition of historical changes in average body mass. Average mass and height among adults increased 11 kg and 2.5 cm, respectively, between 1960 and 2002 (Ogden et al. 2004). Increase in the body mass index (BMI—a measure of mass in kilograms divided by height in meters) indicates that fat gain, not height increase, was the main contributor to the increase in average mass (Ogden et al. 2004). Weight gain and obesity are an increasing concern in the US and Blanck et al. (2006) report that obesity in the US has increased from 15.3% in 1995 to 23.9% in 2005. The mass estimates in this analysis are based on sample data from the FCID (EPA 2000), which were taken between 1994 and 1996. We corrected for changes in average mass across age categories since 1960, based on data presented by Ogden et al. (2004). Between 1790 and 2006, the US population increased nearly 80-fold from 3.9 million to 300 million (Fig. 4). In this same time period, the human carbon stock increased from 0.06 Tg to 5.37 Tg (teragram = million metric tonne), with half of this increase occurring since 1950. Had there not been the ~5% increase in average per capita weight between 1990 and 2000, with an additional 2% increase projected to 2006, the US human population would instead have been sequestering 4.84 Tg C in 2006.
Fig. 4

Carbon stock in the US human population from 1790 to 2000. This historical trend is based on census population data (US Census Bureau 1993, 2002) and an average 17.3 kg C per capita. The last point on the graph represents the estimated US population of 300 million on October 17, 2006

Geospatial distribution of human respiration

To estimate the release of carbon by humans we need to first estimate how the human population is spatially distributed. LandScan (Dobson et al. 2000, Bhaduri et al. 2007) is a grid-based population dataset that estimates where people are located around the world at a resolution of 30 arcseconds (~1 km × 1 km at the equator). LandScan combines data on population, road networks, digital terrain, schools, prisons, commercial and residential buildings, and nighttime lights with high-resolution satellite imagery to estimate where people are located during the day and during the evening. In our analysis, we use a US version of LandScan for year 2000 that uses block-group population data from the US Census Bureau and that estimates the ambient (i.e., integrated diurnal movements) location of humans at a resolution of 15 arcseconds (~450 m × 450 m). Estimates of carbon intake and release by humans are applied to unique age and sex categories within the block-group demographic data and then distributed within the LandScan model. The model does not distinguish variations that might occur over the course of a year so it does not capture any differences in respiration that might occur due to an annual variation in the level of human activity or of human distribution. Our computation distinguishes between human respiration and the discharge of other waste products but this cannot be captured by LandScan so the assumption is implicit that all of these emissions occur in the same place. LandScan allows for estimates, at the finest scale possible, of the direct release of human carbon into the atmosphere, the approximate location of net Ceq flux from humans (including the treatment of human waste), and the stocks of human carbon in the US.

Results and synthesis

The average adult (ages 15–39) in the US contains about 20.7 kg C. Our best estimate of the carbon budget for an average person in the US achieves reasonable balance between annual carbon intake and annual carbon output (Fig. 5). Our estimate is that the average young adult, between the ages of 15 and 39, takes in about 67.0 kg C year−1, and that this is balanced by release of about 65.5–58.6 kg C as expired CO2, 6.6 kg C as feces and urine, and considerably less than 1 kg C as flatus, sweat, and aromatic compounds (Fig. 5). Carbon that is excreted in the US is treated in private septic systems or municipal wastewater treatment facilities with the combination of these two processing methods releasing about 6.02 kg C as CO2, 0.53 kg C as CH4, and 0.11 kg C as other volatile and inorganic species. In our analysis, we associate emissions from waste with the same year in which they were excreted from the body.
Fig. 5

Schematic of consumption, metabolization, and post-waste processing of consumed carbon in the US. Solid lines represent transfer of solid carbon and double lines represent loss of gaseous carbon emissions. Excreta includes feces (5 kg C) and urine (2 kg C). Units are kg C per year for an average adult (ages 15–39), except for units of net Ceq emissions from wasterwater treatment facilities that includes a global warming potential (GWP) factor of 23 for CH4

Balancing the carbon budget for the average human adult involves estimates of total carbon consumption and of the fractional loss of carbon by different physiologic pathways. Given estimates of carbon consumption for different age-gender categories, the fractions of carbon loss for an average adult can be applied to individual age–gender categories (Table 1). Since the carbon budget is required to be balanced for each age–gender category, carbon consumed will equal the sum of carbon expired and the carbon excreted for each category. Total release of greenhouse gases is the sum of carbon expired and carbon released from the treatment of excreted carbon. The carbon stock per capita uses a model for total body carbon along with an estimated percentage of body fat for each age–gender category. Carbon consumption data are based on a survey of the US population; therefore, our estimates of human carbon stocks and flows represent the US population.
Table 1

Estimated carbon consumed, released, and stored by the US population

Age (gender)

Carbon consumed

Carbon expired

Carbon excreteda

Total release of CO2 and CH4

Estimated carbon stock


(kg C capita−1 year−1)

(kg Ceq capita−1 year−1b

(kg C capita−1)

<5 (M)






<5 (F)






5–9 (M)






5–9 (F)






10–14 (M)






10–14 (F)






15–39 (M)






15–39 (F)






40–59 (M)






40–59 (F)






60–74 (M)






60–74 (F)






























aExcreted carbon includes feces, urine, flatus, and secretion of sweat and odor

bCarbon-equivalent (Ceq) emissions include carbon expired and the CO2 and CH4 emissions from excreted carbon that is decomposed at wastewater treatment facilities and municipal solid waste facilities, using global warming potential values of 1 and 23 for both gases, respectively

We are able to estimate total human stocks and flows of carbon in the United States by combining our estimates of carbon stocks and flows for different age–gender categories with population data from LandScan (Fig. 6). In 2000, an estimated 17.2 Tg C were consumed by the US population and 15.2 Tg C were expired to the atmosphere as CO2. Including excretion and secretion of carbon, it is expected that the entire 17.2 Tg C was released to the atmosphere the same year. Increases in population and personal body mass result in some small annual increase in carbon stocks. The spatial distribution of carbon release by humans is nearly opposite the distribution of carbon uptake by agricultural crops (see Hicke et al. 2004).
Fig. 6

Carbon released directly by the US human population in 2000. Enhanced area represents carbon released in the metropolitan area of Chicago, Illinois, USA. Spatial resolution is commensurate with the 450 m × 450 m (20.25 ha) Landscan 2000 data

Figure 6 shows that the densely populated areas of the United States are strong sources of CO2 from human respiration. With average per capita consumption of agricultural carbon at 67 kg C year−1 and average per capita use of fossil-fuel carbon at 5430 kg C year−1 (Marland et al. 2006), emissions from human respiration are 1.2% of those from fossil-fuel use when averaged over the country. The fraction is likely to be somewhat larger in densely populated areas. The fraction is likely to be still larger in urban concentrations of developing countries where there is less automobile traffic. Our estimate includes only the agricultural carbon that humans directly consume and subsequently respire. If we consider total harvested carbon in the US (Ciais et al. 2007) divided by total US population, as is done with the estimate for fossil-fuel consumption per capita, we then compare agricultural carbon use of 627 kg C year−1 with fossil-fuel use of 5430 kg C year−1 and see that the displacement and release of total harvested carbon is nearly 12% of fossil-fuel emissions. Tracking carbon flows in all agricultural commodities would include consumption and use by humans and livestock, processing wastes, textile products, and imports and exports. We have documented the human component of agricultural carbon displacement here.


We thank Maithilee Kunda for her assistance with calculations of human respiration and Aarthy Sabesan for initial analyses on human food consumption. We kindly acknowledge support from the US Department of Energy Global Change Education Program for A. Roddy. This work was supported by the National Oceanic and Atmospheric Association, Human Dimensions of Climate Change Program. Additional resources were provided by the US Department of Energy, Office of Biological and Environmental Research, Carbon Dioxide Information Analysis Center. Oak Ridge National Laboratory is managed by UT-Battelle, LLC, for the US Department of Energy under contract DE-AC05-00OR22725.

Open Access

This article is distributed under the terms of the Creative Commons Attribution Noncommercial License which permits any noncommercial use, distribution, and reproduction in any medium, provided the original author(s) and source are credited.

Copyright information

© The Author(s) 2009

Authors and Affiliations

  • Tristram O. West
    • 1
  • Gregg Marland
    • 1
    • 2
  • Nagendra Singh
    • 3
  • Budhendra L. Bhaduri
    • 3
  • Adam B. Roddy
    • 4
    • 5
  1. 1.Environmental Sciences DivisionOak Ridge National LaboratoryOak RidgeUSA
  2. 2.International Institute for Applied Systems AnalysisLaxenburgAustria
  3. 3.Computational Sciences and Engineering DivisionOak Ridge National LaboratoryOak RidgeUSA
  4. 4.Swarthmore CollegeSwarthmoreUSA
  5. 5.Department of Integrative BiologyUniversity of California-BerkeleyBerkeleyUSA

Personalised recommendations