The Repurposing Life

The Countries Historically Responsible For Climate Change

In this post, we look at the countries historically responsible for climate change based on their CO₂ emissions from years 1850–2021. This analysis includes CO₂ emissions from land use and forestry, in addition to those from fossil fuels.

Historical responsibility for climate change is at the heart of debates over climate justice. History matters because the cumulative amount of carbon dioxide (CO₂) emitted since the start of the industrial revolution is closely tied to the 1.2°C of warming that has already occurred.

In total, humans have pumped around 2,500bn tonnes of CO₂ (GtCO₂) into the atmosphere since 1850, leaving less than 500GtCO₂ (14%) of remaining carbon budget to stay below the 1.5°C target for global temperatures. This means that the world has by now collectively burned through 86% of the carbon budget for a 50–50 probability of staying below this target, or 89% of the budget for a two–thirds probability.

The video below shows cumulative CO₂ emissions from fossil fuels, land use and forestry between 1850 and 2021 by nation in million tonnes.

Ranking In first place is the U.S., which has released more than 509GtCO₂ since 1850 and is responsible for the largest share of historical emissions, 20% of the global total. China is a distant second, with 11%, followed by Russia (7%), Brazil (5%) and Indonesia (4%). The latter two are among the top 10 largest historical emitters based on CO₂ from land use.

Climate Justice

Meanwhile, large post–colonial European nations, such as Germany and the UK, account for 4% and 3% of the global total, respectively, not including overseas emissions under colonial rule. These national totals are based on territorial CO₂ emissions, reflecting where the emissions take place.

In addition, this analysis looks at the impact of consumption–based emissions accounting in order to reflect trade–in carbon–intensive goods and services. Data on such accounts is only available for recent decades, even though trade has influenced national totals throughout history.

This analysis then explores the figures in relation to population, where countries like of China and India fall in the ranks. Because per–capita rankings depend strongly on the methodology used—unlike overall cumulative emissions—these figures do not relate directly to warming.

Finally, this post presents a detailed explanation of the qualitative and quantitative data behind this analysis, where it comes from and how it was used, including assumptions, uncertainty and changing national borders.

Why Cumulative CO₂ Matters

There is a direct, linear relationship between the total amount of CO₂ released by human activity and the level of the Earth’s surface warming. Moreover, the timing of a tonne of CO₂ emitted has a limited impact on the amount of warming it will ultimately cause.

This means CO₂ emissions from hundreds of years ago continue to contribute to the Earth’s warming—and current warming is determined by the cumulative total of CO₂ emissions over time. This is the scientific basis for the “carbon budget”, the total amount of CO₂ that can be emitted to stay below any given limit on global temperatures.

The link between cumulative emissions and global warming is measured by the “transient climate response to cumulative emissions” or TCRE, estimated by the Intergovernmental Panel on Climate Change (IPCC) Sixth Assessment Report to be 1.65°C per 1,000bn tonnes of carbon (0.45°C per 1,000GtCO₂).

This analysis shows that humans have emitted some 2,504GtCO₂ into the atmosphere since 1850, a figure that aligns with those presented by the IPCC’s Report (pdf) and the Global Carbon Project (GCP) 2020 report (pdf), an international effort to quantify carbon emissions and sinks each year.

Based on the TCRE, those cumulative CO₂ emissions correspond to an increase in warming of 1.13°C—and yearly global temperatures reaching around 1.2°C (in 2020) above pre–industrial levels.

This post does not consider emissions of non–CO₂ greenhouse gases or aerosols, which are predominantly short–lived and do not accumulate over time in the same way as CO₂. The warming impact of non–CO₂ gases is roughly balanced by the cooling from aerosols.

The chart below shows how rapidly global CO₂ emissions have risen over the past 70 years. It also highlights the split between CO₂ emissions from fossil fuels and cement (grey), compared with those from land use, land use change and forestry (green) in billions of tonnes between 1850 and 2021.

At a global level, emissions from land use and forestry have remained relatively consistent over the past two centuries. They amounted to around 3GtCO₂ in 1850 and stand at roughly 6GtCO₂ today, despite huge shifts in regional patterns of deforestation over time. A visible spike in 1997 was caused by widespread forest fires in Indonesia and parts of Asia, subsequently described as an unprecedented ecological disaster.

In contrast, fossil fuel emissions have doubled over the past 30 years, quadrupled over the last 60 years and risen nearly twelve–fold over the past century. The 0.2GtCO₂ released in 1850 amounts to just half of 1% of the roughly 37GtCO₂ that was emitted in 2021.

While the large majority of CO₂ emissions today are from burning fossil fuels, human activity, such as deforestation, has made a significant contribution to the cumulative total. Between 1850 and 2021 land–use change and forestry added some 786GtCO₂, nearly ⅓ of the cumulative total, the remaining ⅔ (1,718GtCO₂) are from fossil fuels and cement.

As emissions have increased, the carbon budget has been used up at an exponential rate. Since 1850, half the cumulative total has been released in the past 40 years alone. Beginning in 2022 if annual emissions remain at current levels, the remaining 1.5°C budget (for a 50% probability) will be used up within 10 years—and only seven years for a two–thirds probability.

National Responsibility for Historical Emissions

As for assigning national responsibility for burning through the carbon budget in the context of climate justice debates, this analysis primarily looks at cumulative national territorial emissions, since this is how the data is available. Responsibility is about dealing with the impact of climate change to date—as well as who should make the most effort to prevent further warming.

Cumulative national allocations assign “responsibility” for historical emissions to the modern–day country that occupies the territory where emissions took place in the past. This means that shifting territorial ownership and the unification and dissolution of countries are a complication.

On this basis—and including all human sources of CO₂—the video above shows the countries most responsible for historical emissions as they accumulate from 1850–2021. Each bar is marked with a contemporary national flag and represents a country’s cumulative emissions over time and is color–coded by region of the world, represented in the map at the top–right corner.

in the bottom–right corner are the year and size of the remaining carbon budget for the 1.5°C target as it is depleted over time. The history of national CO₂ emissions is also a history of development, the changing positions in the rankings relate to many factors.

In the early decades of the timeline, global CO₂ emissions were dominated by land–use change and forestry. During this time, the largest emitters, the U.S., Russia and China, were geographically extensive nations cutting down their temperate forests for agricultural land and for fuel.

In the U.S., waves of settlers spread across the continent from east to west, following what they called “manifest destiny”, clearing land for farming as they went.

Simultaneously, European countries like France, Germany and—especially—the UK (which had cleared their land for farming before 1850) begin to rise up the ranks as they become industrialized by the use of coal. These countries have significantly reduced their emissions in recent decades, yet they remain among the most important contributors to historical warming today.

The rainforest nations of Brazil and Indonesia were deforested back in the late 19^th and early 20^th centuries by settlers growing rubber, tobacco and other cash crops. But deforestation grew exponentially beginning around 1950 for cattle ranching, logging and palm–oil plantations.

The U.S. remains in first place for its cumulative CO₂ emissions throughout 1850–2021, since its development began first with the use of coal, and then with the inventions of the motor car. By the end of 2021, the U.S. emitted more than 509GtCO₂ since 1850. This is by far the largest share (20.3% of the global total) and is associated with 0.2°C of warming.

The chart below breaks down each country’s cumulative total into emissions from fossil fuel–use (grey) or from land–use change and forestry (green).

In second place is China, with 11.4% of cumulative CO₂ emissions to date and around 0.1°C of warming. China has had high land–related emissions throughout its history, but the main cause of its current place is its rapid, coal–fueled economic boom since 2000.

China’s CO₂ output has more than tripled since 2000, overtaking the U.S. in becoming the world’s largest annual emitter, responsible for around 25% of the current yearly total. Read further on China’s pre–industrial coal use.

Russia is third, with some 6.9% of global cumulative CO₂ emissions, followed by Brazil (4.5%) and Indonesia (4.1%). In the chart above, the latter pair are in the top 10 as a result of their emissions from deforestation, despite low totals from the use of fossil fuels.

Germany, in sixth place with 3.5% of cumulative emissions due to its coal–reliant energy industry, illustrates how some countries’ land sectors have become cumulative CO₂ sinks rather than sources, as trees have returned to previously deforested areas.

India is in seventh place with 3.4% of the cumulative total—just above the UK—as a result of a higher contribution from land–use change and forestry. Japan with 2.7% and Canada with 2.6%, finish the list of the top 10 historical largest contributors.

International transport emissions from aviation and shipping would rank 11^th in the list if they were to be viewed as a nation, but these are always excluded from national data. It is worth noting that the data used for this post is based on the scientific approach to accounting for land–use emissions, which differs from official data submitted to the UN. The difference relates to what is counted as a “human” versus “natural” source or sink of CO₂.

Cumulative Consumption Emissions

Some countries have reduced their territorial emissions at home, but continue to rely on imported high–carbon goods. Consumption–based emissions accounts assign responsibility to those that use the products and services rendered with fossil energy. This methodology reduces the total for major exporters, like China.

Even though international trade in these has taken place throughout history, gathering consumption emissions data relies on detailed trade tables that are only available from 1990 on.

However, it is still possible to examine the impact of traded CO₂ on countries’ cumulative emissions, as shown by the 20 largest contributors in the chart below. The grey bars show cumulative national emissions, the light grey areas indicate CO₂ associated with exports and the red areas indicate emissions linked to imports.

The top 19 countries based on cumulative consumption emissions are the same as the top 19 countries based on a territorial basis. None of the top 10 shift position in the ranks despite now having a much larger CO₂ footprint than their territorial total. This larger footprint adds to the share of responsibility of wealthy nations. 

The U.S. and Japan each gain 0.3 percentage points of the global cumulative total, Germany and the UK add 0.2 points each, and China and Russia’s shares drop by 1.1 and 0.5 points respectively.

The consumption accounting used in this chart includes only CO₂ from fossil fuels and cement, which is why Brazil and Indonesia’s totals barely change. Consumption–based data prior to 1990 is excluded from this analysis because it is unavailable. In the 19^th century, the UK exported large volumes of energy– and carbon–intensive goods.

According to a 2017 paper, other industrializing nations, such as the U.S. and Germany, did the same. In 1890, nearly 20% of UK energy use related to exported goods, meaning a consumption accounting basis a similar proportion of its CO₂ emissions would have been allocated overseas.

Consumption–based accounting does not resolve the question of responsibility for emissions, however, given that both sides of a trade relationship are likely to gain financially. In a modern context, only one side of that relationship has full responsibility over the CO₂–emitted—regardless of historical colonial rule.

A third approach, related to “scope 3 emissions”, is to make fossil–fuel producers and fossil–fuel exporters like Australia responsible for the CO₂ released with the burning of their coal, oil or gas. However, national emissions data on a production basis are not yet available and could risk double–counting of CO₂ produced in one place and used elsewhere.

Cumulative Per–Capita Emissions

The idea of national responsibility runs into issues like the unequal size, wealth and carbon–intensity of current populations, as well as those of previous generations. These issues apply both within and between countries. Countries themselves are rather arbitrary human constructs, resulting from accidents of history, geography and politics.

One way to rectify these issues is to normalize countries’ contributions to cumulative CO₂ emissions according to their relative populations. Based on the mathematical modelling of climate systems at the University of Exeter, per–capita data is not relevant to the climate, unlike cumulative historical emissions that relate to current warming. What matters for the atmosphere and the climate is cumulative CO₂ emissions.

One example of this is that small countries with high per–capita emissions are still relatively unimportant for warming overall. This analysis approaches the question of accounting for relative population sizes in two different ways:

1. The first approach takes a country’s cumulative emissions in each year and divides it by the number of people living in the country at the time, implicitly assigning responsibility for the past to those alive today.
2. The second approach takes a country’s per–capita emissions in each year and adds them up over time. This approach gives equal weight to the per–capita emissions of the populations of the past and of the present.

Rank	Country	Cumulative Emissions Per Population 2021, tCO2	Rank	Country	Cumulative Per Capita Emissions, tCO2
1	Canada	1,751	1	New Zealand	5,764
2	U.S.	1,547	2	Canada	4,772
3	Estonia	1,394	3	Australia	4,013
4	Australia	1,388	4	U.S.	3,820
5	Trinidad & Tobago	1,187	5	Argentina	3,382
6	Russia	1,181	6	Qatar	3,340
7	Kazakhstan	1,121	7	Gabon	2,764
8	UK	1,100	8	Malaysia	2,342
9	Germany	1,059	9	Republic of Congo	2,276
10	Belgium	1,053	10	Nicaragua	2,187
11	Finland	1,052	11	Paraguay	2,111
12	Czech Republic	1,016	12	Kazakhstan	2,067
13	New Zealand	962	13	Zambia	1,966
14	Belarus	961	14	Panama	1,948
15	Ukraine	922	15	Côte d’Ivoire	1,943
16	Lithuania	899	16	Costa Rica	1,932
17	Qatar	792	17	Bolivia	1,881
18	Denmark	781	18	Kuwait	1,855
19	Sweden	776	19	Trinidad & Tobago	1,842
20	Paraguay	732	20	United Arab Emirates	1,834

The table above lists the cumulative emissions 1850 through 2021 for the top 20 countries weighted by population (left), versus per–capita (right). The ranking excludes countries with a population of less than 1 million people.

Several of the top 10 countries for cumulative emissions overall, namely China, India, Brazil and Indonesia, are absent from the table. These countries have made large contributions to global cumulative emissions, but they also have big populations, making their impact per person much smaller. Those same four countries account for 42% of the world’s population, but only 23% of cumulative emissions for years 1850–2021.

The remainder of the top 10 countries, namely the U.S., Russia, Germany, the UK, Japan and Canada, account for 10% of the world’s population, but 39% of cumulative emissions. In the table above Canada ranks in first place, followed by the U.S., Estonia, Australia, Trinidad and Tobago, and Russia.

For the larger countries on this list, their rank reflects combinations of high deforestation rates during the 19^th and mid–20^th centuries—when populations were smaller—as well as their high per–capita fossil fuel use in recent decades.

For others, such as Estonia, which has long relied on oil sands for most of its energy needs has had high annual per–capita emissions. The Estonian government has pledged to phase out oil sands production by 2040. As a former Soviet state, Estonia’s emissions before 1991 are estimated according to its share of the USSR’s total at that time.

The Caribbean island nation of Trinidad and Tobago has only 1.4 million people, but ranks high due to its large oil and gas industry, which also feeds a large chemicals sector. As for the cumulative per–capita ranking (table right) the list changes, although it still ranks Canada, Australia and the U.S. high on the list.

New Zealand ranks at the top because of extensive deforestation during the 19^th century, when its native Kauri forest was cleared for timber. At that time the country’s small population had high annual per–capita emissions. By 1900, the cumulative total was around two–thirds of the total emitted by present day.

Other countries ranked by emissions from deforestation include Gabon, Malaysia and the Republic of Congo, as well as several South American nations. In terms of assigning “responsibility” for these emissions, there arise issues related to colonization and the extraction of natural resources by foreign settlers.

Methodologies Explained

#1 Fossil Data

Scientists have been making estimates of global CO₂ emissions for more than a century. In 1894, Swedish geochemist Arvid Högbom made the earliest attempt. A translation by the Center for International Climate Research (CICERO) in Norway, explains the estimate: current global hard coal production is in round numbers 500m tonnes per annum, or 1 tonne per km² of the Earth’s surface. Transformed to CO₂ this amount of coal represents 1/1,000^th of the air’s total CO₂.

Högbom’s account put global CO₂ emissions from coal burning at about 1.8GtCO₂ in 1890, an estimate that was close to a contemporary estimate of emissions from coal at that time, about 1.3GtCO₂. Svante Arrhenis followed Högbom with his seminal 1896 work that was the first to predict that changing atmospheric CO₂ levels could substantially alter the Earth’s temperature. Since then, scientists have developed their own time series estimating CO₂ emissions from the burning of fossil fuels, which agree within a few percent.

The data for this article is drawn from the estimates of national historical CO₂ emissions from fossil fuels and cement production, developed by the CDIAC in the U.S. The CDIAC figures (1750 through present day) are maintained and updated by the Appalachian Energy Center at Appalachian State University.

The historical fossil CO₂ estimates are based on a methodology developed by 1984 and since refined. It uses records of fossil fuel production, trade and use that includes estimates of the amount of CO₂ released when coal, oil or gas is burned, by weight. A sophisticated version (pdf) of this initial approach is still used to estimate contemporary emissions today.

CO₂ emissions are seldom measured, rather they are estimates from the best data available on the amount of fossil fuel produced and how it was used. Because fossil–fuel CO₂ emissions are connected with energy, a tracked commodity with a critical role in economic activity, there is plenty of data that can be used for estimating emissions.

There is data on fossil–fuel use and processing going back to 1751. Constructing estimates for those early years can be accomplished because early on there were only a few countries burning fossil fuels and the rate of their growth is puts the majority of global emissions during the most recent decades.

China, with a population of some 400 million people even in 1850, is recorded as having zero emissions from fossil fuel burning until around the turn of the 20^th century. This is because China has been using coal for thousands of years, one account suggests it burned hundreds of thousands of tonnes a year to make iron as early as the 11^th century.

This coal use was said to be localized due to the high cost of transport, and the Mongol invasion collapsed some iron hubs. A 2004 article (pdf) claims China remained predominantly reliant on wood fuel, causing widespread deforestation. By 1900, European countries were mainly energized by coal—but energy use in rural China during the last year of the Qing dynasty (1911) was the same as it had been 100 or 500 years earlier, making China’s coal use hard to quantify before 1900.

The Center for History and Economics at Harvard University has compiled and hosts (pdf) another data set of global historical energy use used in this article that supports the figures given by CDIAC.

#2 Industrial Baseline

The data analyzed in this post begins in 1850, because this coincides with the IPCC definition of the pre–industrial baseline period (1850–1900), and because data on national emissions from land use and forestry is not available before 1850.

Based on the CDIAC figures, only a handful of countries were emitting significant CO₂ from fossil–fuel burning before 1850, and continued to have miniscule totals well into the 20^th century. Accounting beginning in 1850 excludes just 3.8GtCO₂ of fossil fuel emissions released from 1750–1850, which is 0.2% of the total emitted during all of 1750–2021. Nearly three–quarters (2.8GtCO₂) of emissions before 1850 was from the UK, adding data from 1750 would add 0.1 percentage points to the UK’s share of global cumulative emissions.

The CDIAC data set is also used by the GCP, which has been aggregated with other useful information by Our World in Data (OWID). This analysis takes fossil emissions data through to 2019 from the OWID compilation. This analysis estimates emissions in 2020 and 2021 using the real–time figures published by Carbon Monitor (CM), which contains data for major economies and the world in aggregate.

Data for fossil CO₂ emissions from international transport are reported separately by one of the GCP’s  collaborators. International transport emissions halved in 2020 due to COVID before returning to 2019 levels in 2021. GCP’s data set is also the source for consumption–based emissions, which run from 1990 onwards. Population data comes from OWID and Gapminder.

#3 Changing Borders

Territorial changes and the unification or disintegration of national entities affects historical division of emissions. The CDIAC data takes into account changing national boundaries over time. E.g.: responsibility for emissions from the coal– and mineral–rich region of Alsace–Lorraine switches between France and Germany, following contemporary borders.

Similarly, emissions from the area that is now Pakistan are reported under India’s total prior to the country’s partition in 1947, followed by Bangladesh splitting from Pakistan in 1971. The breakup of the former Soviet Union or the former Yugoslavia—or the combining of North and South Vietnam or East and West Germany—leave data that allows for reconstruction.

The treatment of countries within supranational entities, such as the Austro–Hungarian or Ottoman empires, creates the potential for double–counting. Differently from CDIAC, the GCP aggregates and disaggregates national emissions according to modern geographic entities, joining East and West Germany into a single unit.

Whereas CDIAC reports emissions from Czechoslovakia as a single country until its separation into the Czech Republic and Slovakia in 1991, the GCP reports figures for the two countries throughout. Based on the shares of emissions from the Czech Republic and Slovakia projected backwards through time from their split in 1991.

GCP uses this same approach to countries in the former Soviet Union, whereas CDIAC reports data for the USSR from 1830–1991 and for the independent states thereafter. This analysis uses the GCP reporting of national emissions, rather than those used by CDIAC.

#4 Land–Use, Land–Use Change and Forestry

Estimated national CO₂ emissions from land use, land–use change and forestry (LULUCF) are the average of two data sources: Houghton and Nassikas (2017) hereafter HN and Hansis et al (2015) hereafter BLUE.

Updated versions of these datasets, covering 1850–2019 were shared by the director of the department of geography at the Ludwig–Maximillians University Munich. Both datasets derive from “bookkeeping models”, which record changes in soil and above–ground carbon stocks over time, based on aggregate levels of land–use change.

The senior scientist emeritus at the Woodwell Climate Research Center explains the calculating of the annual emissions from land–use change from two kinds of data: the first reconstructs areas of croplands, pastures, forests and other lands. The second (the carbon data) is how much carbon is in the vegetation and soils of different ecosystems and how those stocks change as a result of land–use change and forestry.

The bookkeeping model shows how much carbon is lost or gained when land use changes as a result of human activity, which is based on knowing the annual changes in the carbon stocks of a hectare of land undergoing some kind of land use change. For example, clearing a forest for cropland, or planting a forest on open land.

The data on carbon stocks and their changes from land use are taken from ecological and forestry sources. The two LULUCF datasets contain differences at global and national levels, explored in a 2021 joint paper.

HN aggregates the different underlying land–use data at the national level, whereas BLUE is spatially explicit, allowing for tracking shifting cultivation that may affect carbon stocks across a wider area, even if the area of farmland stays the same. HN and BLUE estimate carbon stocks for each type of land use, as well as the share of stocks that quickly decompose differently.

As with the estimates of fossil CO₂ emissions, the LULUCF figures change as they go back in time. This results from incomplete data. Rates of change in land use were lower further back in the past than in the last 60 years. The accuracy of global land use and forestry emissions amounts to around plus or minus 2.5GtCO₂ per year, similarly to that for fossil fuels. However, in relative terms, the accuracy amounts to ±50% of the estimated LULUCF total.

This analysis assumes land–use emissions in the most recent years were unchanged since the most recent estimate.

A third dataset on LULUCF emissions is the “OSCAR” time series (pdf) averages together HN and BLUE figures for the annual Global Carbon Budget analysis. OSCAR is reported at regional rather than country level, and therefore not used in the national historical emissions analysis. The OSCAR data falls roughly in the middle of the other two data sets. The cumulative global total for LULUCF differs from the three–way average used by GCP by less than 2%.

As with fossil fuels, this analysis starts in 1850, excluding some CO₂ emissions related to pre–industrial land–use change, predominantly forest clearance. A 2012 paper (pdf) explores regional land–use change emissions during the 1,000 year pre–industrial period from 800–1850.

In Europe, there was a large uptick of emissions due to widespread forest clearance until the Black Death, followed by another wave of deforestation during the Renaissance. Global land–use change emissions overall were dominated by China and the region of South Asia made up of India. These pre–industrial CO₂ emissions increase Asia’s share of current warming by 2–3 percentage points, while reducing North America and Europe’s share by a similar amount.

Sources:
Analysis: Which countries are historically responsible for climate change?
https://www.carbonbrief.org/analysis-which-countries-are-historically-responsible-for-climate-change/
by Simon Evans on 5 October 2021
Data Sources:
Global Carbon Project
https://www.globalcarbonproject.org/
Carbon Dioxide Information Analysis Center (CDIAC)
https://cdiac.ess-dive.lbl.gov/
Our World in Data
https://ourworldindata.org/
Carbon Monitor
https://carbonmonitor.org/
Houghton and Nassikas (2017)
https://agupubs.onlinelibrary.wiley.com/doi/abs/10.1002/2016GB005546
Hansis et al (2015)
https://agupubs.onlinelibrary.wiley.com/doi/10.1002/2014GB004997