Robbie Andrew

scholar orcid twitter
CICERO Center for International Climate ResearchHome   |   About me   |   Contact me

A history of estimating global CO2 emissions: Part I

First published: October 2018

Jump to section: Högbom 1894 | Arrhenius 1908 | Callendar 1938 | Plass 1956 | Revelle & Suess 1957 | Revelle et al. 1965 | Baxter & Walton 1970 | Broecker et al. 1971 | Keeling 1973 | Rotty 1981 | Marland & Rotty 1984 | Andres et al. 1999 | Boden et al. 2017 | Data download


In this piece I will be visiting various estimates of global emissions of carbon dioxide, in chronological order according to when they were made. The focus here is on these estimates rather than on the development of climate science.

Scientists have been estimating global emissions for more than a hundred years, for a variety of reasons. The scientific understanding of the global carbon cycle has underlined these efforts, and continues to this day, including in the annual publication Global Carbon Budget.

In the earliest years the carbon cycle of interest was the long-term balancing of geological, extraterrestrial, and natural-system inputs and outputs, primarily to understand the ice ages.

It was already well known in the scientific world by the late 19th century that CO2 in the atmosphere warmed the planet, and that changing levels would lead to changed surface temperatures.

While geological interest continues, interest in human perturbation of the carbon cycle has grown as our emissions have grown and can no longer be considered negligible, and this perturbation is of key importance in understanding climate change: how current and future climate change relates to our emissions.

The easiest way to summarise all of the estimates discussed in this article is the following animated GIF. Each of these landmark publications will be discussed here in turn. Click on any image for a larger view.

Arvid Högbom, 1894
Novelty: Likely first global estimate

In the early 1890s, Swedish geochemist Arvid Högbom presented some of his thoughts on the carbon cycle in nature to the Swedish Chemical Society. In this work he presented what was possibly the first ever estimate of the amount of carbon dioxide emitted into the atmosphere from the combustion of fossil fuels.

Högbom's article "On the probability of secular changes in the level of atmospheric CO2" (original title: "Om sannolikheten för sekulära förändringar i atmosfärens kolsyrehalt") was published in 1894 in the Svensk Kemisk Tidskrift (Swedish Chemistry Journal). Here 'secular' means 'over a period of about a hundred years'.

Here's the relevant text (p. 171, my translation):

"Current global hard coal production is in round numbers 500 million tonnes per annum, or 1 tonne per km2 of the Earth's surface. Transformed to CO2 this amount of coal represents approximately a thousandth part of the air's total CO2. This is equivalent to a layer of limestone of 0.003 mm over the entire surface of the Earth, or 3 mm over Sweden, or, expressed as a volume, 1.5 km3 of limestone."

Unfortunately Högbom didn't actually say how many tonnes of CO2 this was, but we can work it out directly from the information supplied:

Firstly, he says 500 Mt is the same as one tonne per km2, meaning he's using an approximate surface area for the Earth of 500 million km2 (it's actually about 510 million km2).

Then he says this is equivalent to 0.003 mm thickness of limestone (CaCO3) over 500 million km2 surface, or 1.5 km3. And 1.5 km3 is 1.5×1015 cm3, which, with a density of about 2.72 g/cm3, is about 4.1×109 tonnes. The CO2 'content' of this mass is roughly 44/100 (ratio of the molar masses), i.e. 0.44×4.1×109=1.8×109 tonnes. This is also consistent with his 'thousandth part'.

So he estimated about 1.8 Gt of CO2 were emitted at about the time he was writing. Given the starting point of 500 Mt of coal, 1800/500=3.6 tons CO2 per ton coal, which is about the ratio of the molar masses of CO2 to C (3.664), such that he appears to have assumed that coal was equal or close to 100% carbon (it's not, as we'll see later).

This gives us the first point on our graph, below, but note that here I plot in units of gigatonnes (Gt) carbon, rather than Gt carbon dioxide, so we end up at about 500 Mt again. And while this estimate is only for coal, use of natural gas at the time was almost zero, and oil was still in its infancy: almost all fossil energy was from coal.

The reason Högbom compared to limestone in this way was because he believed these emissions from coal combustion would be completely compensated for by mineralisation processes: formation of limestone and other carbonates taking CO2 out of the atmosphere. That is, his conclusion was that short-term emissions from burning fossil fuels would not affect the natural carbon balance. His interest was in changes in the carbon cycle over geological periods.

Arrhenius was the first to translate the passage I've quoted from Högbom into English, bringing it to a wider audience (including Chamberlain), in an extended excerpt of Högbom's article in his famous 1896 piece "On the Influence of Carbonic Acid in the Air upon the Temperature of the Ground". However, Arrhenius' translation is a little loose in places, even omitting one clause entirely, so I've retranslated from the original. With regard to units, Högbom uses the Swedish word 'ton', which is ambiguous, but since he otherwise in his article uses only metric units (cm, km, grams), I will assume here he meant tonne. The difference is nevertheless minor given the approximate nature of his analysis. Lastly, while Högbom specifically refers to hard coal ("stenkol"), it's not certain this specificity was warranted, and the production estimate might have included lignite ("brunkol").

Svante Arrhenius, 1906
Novelty: First time series

It was Svante Arrhenius who, inspired by Högbom's lecture and further discussions at the Stockholm Physics Society, worked further on the question of global warming. Arrhenius initially accepted Högbom's main conclusion, namely that in the short-term the carbon cycle was in balance, given that industrial emissions were relatively low. But he nevertheless performed the thought experiment asking what would happen if those emissions were to further increase.

In 1906, Arrhenius' book "Worlds in the Making: the Evolution of the Universe" was published (the original in Swedish, with the English translation following in 1908).

In this work he had the following to say (p. 54):

the annual combustion of coal ... has now (1904) risen to about 900 million tons and is rapidly increasing

It was this rapid increase, emissions nearly doubling in under 15 years, that led Arrhenius to continue:

we ... recognize that the slight percentage of carbonic acid in the atmosphere may by the advances of industry be changed to a noticeable degree in the course of a few centuries.

In a footnote, Arrhenius supplies the first time series of estimated global CO2 emissions:

It amounted in 1890 to 510 million tons; in 1894, to 550; in 1899, to 690; and in 1904, to 890 million tons.

And for the first time we see a remarkable acceleration in the figure below.

Popular Mechanics, 1912

In March 1912 the American popular science magazine Popular Mechanics published an article commenting on the unusual weather of the year before. In this piece, written by Francis Molena, an estimate is made of global CO2 emissions:

"In the United States about 500,000,000 tons of coal were mined in 1911. Suppose four times this amount were mined and burned in the whole world. When this amount of coal is burned, 7,000,000,000 tons of carbon dioxide are put into the atmosphere."

(Hat tip to @tsrandall.)

The caption for the figure above was reproduced verbatim in a small newspaper in New Zealand later the following year:

Guy Callendar, 1938
Novelty: Possibly first estimate of total cumulative emissions

Callendar, an English steam technologist, published "The Artificial Production Of Carbon Dioxide And Its Influence On Temperature" in 1938. In his abstract he states (p. 223):

"By fuel combustion man has added about 150,000 million tons of carbon dioxide to the air during the past half century."

Unfortunately there's neither repetition nor support for this estimate in the article, but the current estimate for that period is about 135,000 million tonnes (from Boden et al.).

In the main text, Callendar reports the "artificial production at present" as 4500 million tons (p. 224). In all cases where it is unambiguous, units are metric, so I will assume that these are metric tons.

Elsewhere he reports twice that the "annual net addition of CO2 to the air" was 4300 million tons, claiming that (p. 224):

"there is no geological evidence to show that the net offtake of the gas is more than a small fraction of the quantity produced from fuel."

This conclusion is opposite to that of Högbom, who suggested it likely that all fossil emissions were absorbed in the medium term. Later evidence, and current science, shows that about half of the CO2 we emit is absorbed by oceans and the terrestrial biosphere (plants and soils). Global warming is largely driven by the increased amount of CO2 in the atmosphere, absorbing outgoing infra-red radiation, so this 'partitioning' is critical to understanding the changing climate.

As with Arrhenius, Callendar believed that global warming would be beneficial, in Callendar's words delaying indefinitely "the return of the deadly glaciers" of the ice ages (p. 236).

Gilbert Plass, 1956

In 1956, Canada-born physicist Gilbert Plass published an article in the American Journal of Physics entitled "Effect of Carbon Dioxide Variations on Climate" in which he explored Earth's carbon balance and climate sensitivity.

Plass states (p. 379):

"The combustion of fossil fuels is adding 6×109 tons per year of carbon dioxide to the atmosphere at the present time."

There is no supporting evidence nor sources provided for this claim, but as we'll see it fits well with later estimates for the same time.

He then mentions other human activities that release carbon, without quantification (p. 379):

"In addition such activities as the clearance of forests, the drainage and cultivation of lands, and industrial processes such as lime burning and fermentation release additional amounts of carbon dioxide that are not included in the foregoing estimate. This is a large enough contribution to upset the carbon dioxide balance and to increase the amount in the atmosphere appreciably."

Roger Revelle and Hans Suess, 1957
Novelty: First long time series

Geologist Revelle and physical chemist Suess were interested in the fate of the carbon dioxide added to the atmosphere by human activities. While Callendar had believed that most of it remained in the atmosphere, Revelle and Suess came to a different conclusion, publishing their analysis as "Carbon Dioxide Exchange Between Atmosphere and Ocean and the Question of an Increase of Atmospheric CO2, during the Past Decades".

By exploring the ratios of the three naturally occurring carbon isotopes found in both wood and marine material (Suess was one of the founders of radiocarbon dating), Revelle and Suess concluded that the oceans had absorbed 'most' of the carbon released into the atmosphere by burning fossil fuels. Fossil fuels, having been buried for very long periods, have no 14C, which, generated by cosmic radiation in the atmosphere, is unstable with a half-life of about 5730 years.

On the way to this conclusion, they estimate emissions from fossil-fuel combustion per decade. The methods are not given, but reference is elsewhere given to the United Nations report "World requirements of energy, 1975–2000" presented at the International Conference on Peaceful Uses of Atomic Energy, held in Geneva in 1955, and this is most likely the source of the energy data.

This UN energy demand forecast out to 2000 appears to have spurred interest in emissions of CO2.

The emissions estimates they published (shown in the figure below) lie very close to previous estimates, and therefore likely assume very high carbon contents of the fuels.

Revelle and Suess also make an estimate of the carbon emitted due to clearing of forests over the preceding 100 years (p. 26):

"The increase in arable lands of about 4×1016 cm2 since the middle of the 19th century has resulted in a corresponding decrease of forest land of about 10%. Assuming that 70% of all the soil carbon is in forests (probably a considerable over-estimate), and that cultivation reduced this by 70%, the total decrease in soil carbon would correspond to 9×1016 gms of CO2, which is 4% of that in the atmosphere. At least four-fifths of this amount should have been transferred to the ocean."

Where 9×1016 gms is 90 billion tonnes (Gt). Current data from the Global Carbon Project gives land-use change emissions over the period 1850–1950 as 400 Gt CO2. It is curious that Revelle and Suess have focussed on soil carbon without mentioning the carbon lost from the trees themselves. (Recent research suggests that about 133 Gt of carbon – equivalent to about 490 Gt CO2 – has been lost from soils since the dawn of agriculture 12000 years ago.)

Based on their analysis, Revelle and Suess conclude:

"In contemplating the probably large increase in CO2 production by fossil fuel combustion in coming decades we conclude that a total increase of 20 to 40 % in atmospheric CO2 can be anticipated. This should certainly be adequate to allow a determination of the effects, if any, of changes in atmospheric carbon dioxide on weather and climate throughout the earth."

The increase in concentration of atmospheric CO2 recently passed 40%, and certainly seems to be 'adequate' in allowing us to determine the effects on climate.

Roger Revelle et al., 1965
Novelty: First to describe clearly data sources and methods

The 1965 report by the President's Science Advisory Committee Panel on Environmental Pollution included discussion of CO2 from fossil fuels as a potential global problem, the first authoritative US government mention of the issue.

Along with co-authors Wallace Broecker, David Keeling, Harmon Craig, and Joseph Smagorinsky, Revelle wrote in the Appendix "Atmospheric Carbon Dioxide" that:

"Within a few short centuries, we are returning to the air a significant part of the carbon that was slowly extracted by plants and buried in the sediments during half a billion years."

In presenting their case, Revelle and colleagues tabulated estimates of the world production of fossil fuels and emission CO2 from combustion of those fuels, both annually for 1950–1962 and decadal from the 1860s.

For the first time in the collected works I'm presenting, the authors present both their sources and the method for calculating emissions from the energy data.

"Sources: From 1860 to 1949, United Nations, Department of Economic and Social Affairs: "World Energy Requirements". Proceedings of the International Conference on the Peaceful Uses of Atomic Energy, Vol. 1, pp. 3-33, 1956. For [1950-62], United Nations, World Energy Supplies. Statistical papers, Series J, United Nations, New York."

Assumed carbon content: coal=75%; lignite=45%; liquid hydrocarbons=86%; natural gas=70%.

(While several volumes of the named proceedings are available online, I have been unable to find volume 1.)

While their carbon content for lignite is too high, the other factors are not far off global averages used in later studies. Lignite is a poor-quality coal, with high content of both moisture and 'ash' (minerals that provide no energy and result in ash), and these must be accounted for when assessing carbon content.

Roger Revelle has been quoted as saying that our emissions of CO2 are "man's greatest geophysical experiment" (Broecker et al., 1971, p.322)

Murdoch Baxter and Alan Walton, 1970
Novelty: First annual time series, first cement emissions

In 1970, Baxter and Walton, both chemists at the University of Glasgow, published "A Theoretical Approach to the Suess Effect", that is, the observation that the fraction of isotope 14C in the atmosphere was declining due to the addition of fossil carbon.

To drive their method, the authors updated previous estimates of fossil carbon emissions, going back to the source data, and theirs is the first presentation of annual fossil carbon emissions.

Unfortunately, in the process of improving the estimates, they used an even higher factor for the carbon content of lignite: 73%. They cite Clarke (1924), who presents a table showing three different estimates of the carbon content of lignite from three different studies (p. 766). These are clearly marked (on the preceding page, 765) as being reduced to an ash-free and water-free standard. Baxter & Walton appear to have incorrectly applied this factor to the mass of lignite, including ash and water. "In lignites the moisture is usually very high, for these coals are peculiarly hygroscopic" (p. 765). The unweighted average of the three studies is 72.95, which is most likely where Baxter & Walton's 73% comes from.

Importantly, Baxter and Walton may also have been the first to estimate CO2 emissions from global production of cement, which involves the calcination of fossil carbonates, particularly limestone.

The emissions data are only presented in graphical form in the paper, with reference to Baxter's PhD thesis for data in tabulated form. This thesis is housed at the Glasgow University Library, and is not yet digitised (but following my enquiry they have kindly bumped it up the list). So here I have used an online tool to extract the data from the figures, and obviously there is some error involved in this process, compounded by the fact the original is plotted on a log scale.

Wallace Broecker, Yuan-Hui Li and Tsung-Hung Peng, 1971
Novelty: First explicit mention of assumption of complete oxidation

In 1971, Broecker, Li, and Peng published "Carbon Dioxide—Man's Unseen Artifact" as a chapter in the book "Impingement of Man on the Oceans," and most of the article is devoted to ocean uptake of carbon. But for the purposes of their analysis they present decadal estimates of global emissions of CO2.

It is by now more common practice to indicate sources and methods, and Broecker et al. give us the following information (p. 288):

"A compilation of world fossil fuel production made by the United Nations, Department of Economic and Social Affairs (1956) and the 1950–1959 World Energy Supplies Statistical Papers, Series J, has generally been adopted as the basis for CO2 generation estimates...."

They assume (p. 288):

"the mean carbon content of coal to be 70 per cent, of lignite 42 per cent, of liquid hydrocarbon 85 per cent, and of natural gas 70 per cent, and that all the carbon in these fuels has been converted to CO2 gas"

Again the lignite factor is too high, while the factors for the other fuels are reasonable as global averages.

This is the first time in the works collected here that the assumption of complete oxidation of all fossil fuels has been made explicit. The validity of this assumption was first addressed by Keeling in 1973, the next study on our list.

David Keeling, 1973
Novelty: First fully documented, first estimates of fraction oxidised, first uncertainty assessment

Dissatisfied with the lack of documentation accompanying previous estimates of global fossil emissions, Keeling published his ground-breaking "Industrial Production of Carbon Dioxide from Fossil Fuels and Limestone" in 1973 in the journal Tellus.

Keeling stepped up by several orders of magnitude the level of detail and effort involved in assembling estimates of global CO2 emissions, and in the process corrected assumptions made by earlier scholars.

"The world average fractions of carbon in coal and lignite, estimated from calorific data, are found to be lower than previously assumed. When account is taken of handling losses and partial diversion to produce petrochemicals, road asphalt, and other non-fuels, the calculated CO2 emissions are further reduced by several percent even after allowing that most unburned materials eventually oxidize to CO2 in the environment."

Having gone back to original sources, the carbon contents by fuel type Keeling settled on were: Coal 70%, Lignite 28%, Liquids 84%, Natural gas 70%. Finally lignite's carbon content is represented correctly, allowing for both ash and moisture content.

Note in the figure below how much lower emissions from coal combustion are than Baxter and Walton's estimates: Keeling's much lower carbon contents, both for hard coal and lignite, make a substantial difference.

Keeling continued previous scholars' use of the UN's "World Energy Supplies" for energy data reported in physical terms. But he now introduced factors for how much of each energy type was actually oxidised (carbon converted to CO2), where all previous estimates had assumed that all carbon in fossil fuels was oxidised.

"All these published studies assumed 100% conversion of raw material to CO2."

In Keeling's analysis, unoxidised carbon included natural gas lost in handling and transport, particulate (soot) emissions from coal and oil combustion, hydrocarbon loss from oil, and 6% of oil used for non-energy purposes.

While Keeling's source data started in 1860, he suggested that earlier emissions could be estimated by assuming exponential growth, deriving an equation for this purpose.

Keeling was also the first to include estimates of uncertainty in his data and emissions estimates, further lifting the scientific rigour of his analysis above previous work.

Furthermore, Keeling was perhaps the first to call into question China's reported coal consumption:

It is questionable whether any already industrialized country could triple its coal production and consumption within three years, even under the stimulus of a "Great Leap Forward" as China's economic exertion from 1956 to 1960 has been called.

The question of data reliability in China continues to plague researchers.

Concluding, Keeling made some important observations:

"Considerable effort was required to make even this first step towards obtaining accurate values of the global emission of CO2. Much more work will be required if the presently estimated error limits of 13 % are to be substantially reduced."

"Ultimately [carbon contents of fuels, losses, and diversions] should be computed country by country from reliable inland studies of currently determined fuel compositions and emission factors for carbon dioxide, carbon monoxide, hydrocarbons, and air-borne particles."

Ralph Rotty, 1973, 1981
Novelty: First estimates of flaring emissions, first comparison of multiple energy datasets

Ralph Rotty, "both a meteorologist and a mechanical engineer" (p. 31), had already been estimating global CO2 emissions before he joined the Institute for Energy Analysis at Oak Ridge in 1974. Indeed, Keeling in his 1973 article acknowledged considerable assistance from Rotty.

In 1973, Rotty published a companion paper to Keeling's study, "Commentary on and extension of calculative procedure for CO2 production" in the same journal, Tellus.

In this 1973 paper, Rotty pointed out two additional sources of energy data – one from the UN and another from the US Bureau of Mines – but showed that all three sources were in agreement, as long as care was taken with differences in definitions.

Rotty also calculated emissions from flaring of natural gas in oil production, being the first to do so. These data were not included in previous estimates because the gas is never sold on the market, and is therefore not part of energy supply. The following year he published an article specifically on flaring, extrapolating flaring emissions back to 1935 using regression analysis of crude oil production.

Because of data limitations Rotty's estimates included both gas actually flared (combusted to form CO2), and gas vented (released directly to the atmosphere). By including vented gas the assumption is effectively made that this gas, largely methane, soon oxidises to CO2. Recognition of this assumption is important when calculating radiative forcing and global warming using these data and the remaining datasets in this Part I, all of which make the same assumption.

In later years Rotty published occasional updates to his estimates of global emissions, including in 1981 a presentation "Data for Global CO2 Production from Fossil Fuels and Cement" at a carbon cycle workshop in California.

This brief paper included a full tabulation of global emissions from 1950 through 1978, which are shown in the figure below. His methods were the same as Keeling's.

Rotty, 1983
Novelty: First estimates of global distribution of emissions

Thus far energy production data had been used to estimate global emissions, initially because they were the only data available, and later because they were seen as more accurate than consumption data at the global level.

Production data are collated from economic statistics as the physical amount of each fuel type (and cement) produced in each country.

Meanwhile, estimating actual consumption in each country relied on what is called the 'apparent consumption' approach. Rather than consumers reporting their fuel consumption, apparent consumption determines market supply in each country by adjusting production statistics for international trade, use in international bunkers (which are not assigned to countries), and changes in stocks.

Generating fuel consumption data in this way relied on adding and subtracting less certain data, which introduced additional uncertainty, but it was nevertheless the best method to estimate country-level consumption.

In addition, oil used as an energy source in refineries must be carefully accounted for.

But in 1983 Rotty published "Distribution of and Changes in Industrial Carbon Dioxide Production", and for the 'distribution' aspect he wanted to show where on the planet emissions were generated. For that, country-level data were necessary, and they had to be consumption based: where the fuels were combusted, as opposed to where they were produced.

Rotty then became the first to produce estimates of CO2 emissions by country, although because his interest was in helping climate modellers, his results are aggregated to latitudinal bands.

After a brief re-examination, Rotty used the same factors as Keeling in 1973.

Gregg Marland and Ralph Rotty, 1984
Novelty: Full re-examination of methods

In 1984, Marland and Rotty, both then at the Institute for Energy Analysis at Oak Ridge, published "Carbon Dioxide Emissions from Fossil Fuels: a Procedure for Estimation and Results for 1950–1982."

In this work Marland and Rotty completely re-examined Keeling's 1973 methodology and factors used to convert fossil fuel data to estimates of emissions.

The resulting factors were not significantly different from those of Keeling, and were still constant over time, an assumption that appeared valid given the available observational data.

The method developed by Marland and Rotty for estimates of global CO2 emissions was used at CDIAC right through until 2017, when funding was finally removed by the Department of Energy.

The figure below demonstrates that only coal emissions appear to be lower than those estimated by Rotty in 1979. This is largely because the UN had subsequently begun to publish energy data in energy terms (rather than only in tonnes or cubic metres), using country-specific data on carbon contents of fuels rather than the global averages that scholars had relied on previously.

The importance of these new data from the UN cannot be overstated. Scientists' previous estimates of global emissions were based on broad assumptions that gave ballpark figures and reasonable trends. Having country-level energy data in energy units not only led to substantially more robust global estimates, but also opened the way for country-level estimates.

Bob Andres et al., 1999
Novelty: First estimates before 1860, both global and national

Andres and colleagues first presented estimates of global emissions back to 1751 at a conference in 1993, later publishing "Carbon Dioxide Emissions from Fossil-Fuel Use, 1751–1950" in 1999. In this work they made use of historical data on country-level energy production and energy trade to create a long time series of both global and country-level estimates.

While it doesn't show on the graph here, Andres et al. were the first to both extend the time series of global emissions back to 1751 and also produced per-country emissions over that period.

Historical energy production data were taken from Etemad & Luciani's "World Energy Production 1800-1985", while energy trade data were taken from Mitchell's series "International Historical Statistics." According to Andres et al., footnotes in Etemad and Luciani allowed extension of the time series before 1800 to 1751.

In fact, the only data in 1751 were UK coal production, and the source used by Etemad & Luciani, Pollard's A New Estimate of British Coal Production, 1750–1850 repeatedly indicates that the time series start in 1750.

The methods applied to these data were identical to those used by Boden et al. 1995.

Tom Boden, Bob Andres and Gregg Marland, 1999–2017
Novelty: Annually updated series

From 1999 to 2017, Boden, Andres, and Marland at CDIAC published "Global, Regional, and National Fossil-Fuel CO2 Emissions", with emissions estimates from 1751.

These estimates were updated annually as new data were available from the UN. The methodology remained unchanged during this period.

Editions from 2010 through 2017 are still available online, currently hosted by Berkeley Labs.

The method for estimating global emissions was unchanged from that of Marland and Rotty 1984 (see above), while emissions prior to 1950 were directly from Andres et al 1999 (see above).

National estimates were calculated based on the method described by Marland and Boden 1993 (which is likely very similar to that of Marland and Rotty, but I haven't got my hands on the book yet!). In that work, Marland and Boden also made the first extrapolation of global emissions to 'last year' using BP energy data, which for many years have been released several months before any other source, albeit with less detail.

In 2016 it was announced that the Department of Defense would be withdrawing funding for CDIAC. The 2017 edition was the last from Oak Ridge, and at this point it's unclear whether efforts to relocate to another institution will be successful.

Coming soonish: Part II

Thus far I've only discussed methods developed up until the early 1990s (that includes Boden et al. 2017, which is still based on those methods, but with updated input data). Since then, various efforts have emerged using more specific methods, accounting emissions not only by country but by sector (transport, etc.), and including estimates for many more climate forcers than CO2. These will be the subject of Part II of this History of estimating global CO2 emissions.

Data included in this article
Revelle & Suess, 1957:RevelleSuess1957.xlsx
Revelle et al., 1965:RevelleEtal1965.xlsx
Baxter & Walton, 1970:BaxterWalton1970.xlsx
Broecker et al., 1971:BroeckerEtal1971.xlsx
Keeling, 1973:Keeling1973.xlsx
Rotty, 1981:Rotty1981.xlsx
Marland & Rotty, 1984:MarlandRotty1984.xlsx
Andres et al., 1999:Currently hosted at LBL (included in Boden et al. 2017)
Boden et al., 2017:Currently hosted at LBL
All CO2 estimates:EarlyEstimates1_data.csv

Högbom, A., 1894. Om sannolikheten för sekulära förändringar i atmosfärens kolsyrehalt. Svensk Kemisk Tidskrift VI, 169-176.

Arrhenius, S., 1896. On the Influence of Carbonic Acid in the Air upon the Temperature of the Ground. Philosophical Magazine and Journal of Science 41, 237-276.

Arrhenius, S., 1908. Worlds in the Making: the Evolution of the Universe (Translated by Dr. H. Borns). Harper & Brothers Publishers, New York and London.

Callendar, G.S., 1938. The artificial production of carbon dioxide and its influence on temperature. Quarterly Journal of the Royal Meteorological Society 64, 223-240.

Plass, G.N., 1956. The Carbon Dioxide Theory of Climatic Change. Tellus 8, 140-154. DOI: 10.3402/tellusa.v8i2.8969.

Revelle, R., Suess, H.E., 1957. Carbon Dioxide Exchange Between Atmosphere and Ocean and the Question of an Increase of Atmospheric CO2 during the Past Decades. Tellus 9, 18-27. DOI: 10.1111/j.2153-3490.1957.tb01849.x.

Revelle, R., Broecker, W., Craig, H., Keeling, C.D., Smagorinsky, J., 1965. Atmospheric Carbon Dioxide, in: The Environmental Pollution Panel, P.s.S.A.C. (Ed.), Restoring the Quality of Our Environment, pp. 111-133.

Baxter, M.S., Walton, A., 1970. A Theoretical Approach to the Suess Effect. Proceedings of the Royal Society of London. A. Mathematical and Physical Sciences 318, 213-230. DOI: 10.1098/rspa.1970.0141.

Broecker, W.S., Li, Y.-H., Peng, T.-H., 1971. Carbon Dioxide-Man's Unseen Artifact, Impingement of Man on the Oceans. John Wiley & Sons, Inc.

Keeling, C.D., 1973. Industrial production of carbon dioxide from fossil fuels and limestone. Tellus 25, 174-198. DOI: 10.3402/tellusa.v25i2.9652.

Rotty, R.M., 1973. Commentary on and extension of calculative procedure for CO2 production. Tellus 25, 508-517. DOI: doi:10.1111/j.2153-3490.1973.tb00635.x.

Rotty, R.M., 1981. Data for global CO2 production from fossil fuels and cement, SCOPE Carbon Cycle Workshop, La Jolla, California.

Marland, G., Rotty, R.M., 1984. Carbon dioxide emissions from fossil fuels: a procedure for estimation and results for 1950-1982. Tellus 36B, 232-261.

Marland, G., Boden, T., 1993. The Magnitude and Distribution of Fossil-Fuel-Related Carbon Releases. Springer Berlin Heidelberg, Berlin, Heidelberg, pp. 117-138.

Andres, R.J., Fielding, D.J., Marland, G., Boden, T.A., Kumar, N., Kearney, A.T., 1999. Carbon dioxide emissions from fossil-fuel use, 1751-1950. Tellus B 51, 759-765. DOI: 10.1034/j.1600-0889.1999.t01-3-00002.x.

Boden, T.A., Andres, R.J., Marland, G., 2017. Global, Regional, and National Fossil-Fuel CO2 Emissions. Carbon Dioxide Information Analysis Center, Oak Ridge National Laboratory, U.S. Department of Energy, Oak Ridge, Tenn., U.S.A.

Global CO2 emissions from cement production

Updated: Cement production has grown at an even faster pace than use of fossil fuels since 2000, and CO2 emissions resulting from this production are increasing in step. This new article makes the most thorough estimate to date of these emissions.Learn more »

Predicting atmospheric CO2 concentration

In early 2016, we predicted that the annual rise in carbon dioxide concentration at Mauna Loa at the end of that year. How did we do?Learn more »

Global Carbon Budget Figures

Every year the Global Carbon Project publishes a number of figures demonstrating the latest global carbon budget, and these are freely available for use in a number of formats.Learn more »

Global Carbon Budget 2018

The Global Carbon Project has released its 2018 edition of the world’s carbon budget, including historical emissions by country back to 1751. This multidisciplinary and international effort provides a set of consistent supporting data for further analysis.Learn more »