image: Global carbon dioxide emissions from 1750 to 2022: the CDIAC-FF annual product
Credit: Michael Erb & Gregg Marland
Tracking Fossil Fuel Emissions
Understanding the sources and quantities of carbon dioxide (CO₂) released to the atmosphere is central to assessing past emissions and identifying potential reductions. Fossil fuel combustion is the principal source of human-caused CO₂ emissions, alongside emissions from natural gas flaring and venting and the manufacture of cement. Consistent estimates across countries and years help distinguish changes in emissions from changes in data coverage or accounting methods. The CDIAC-FF dataset provides annual estimates reaching back to 1750 and records emissions by fuel type and, for recent decades, economic sector.
Authors Michael Erb and Gregg Marland from the Research Institute for Environment, Energy and Economics at Appalachian State University updated the CDIAC-FF dataset through 2022. The research describes a data product, using United Nations Statistics Division energy data for fuel production, trade and consumption from 1950 onward, with earlier estimates drawn from historical sources. The authors estimate emissions using both a reference approach based on apparent national fuel use and a sectoral approach based on consumption by economic sector. The inventory covers fossil fuel use, natural gas flaring and venting, and cement production, with national and global estimates kept within consistent accounting boundaries over time.
A Dataset Updated Through 2022
The updated inventory estimates global emissions in 2022 at 10,287 metric megatons of carbon (Mt C), a 2.72% increase over 2021 and a 2.30% increase over 2019. These totals are reported as mass of carbon, rather than mass of CO₂, consistent with the CDIAC-FF dataset. Emissions from solid fuels and global per-capita emissions had largely stopped growing over the preceding decade, although solid-fuel emissions rose between 2020 and 2022. The update includes revisions to underlying energy data and methods, while retaining the dataset’s long historical coverage.
The authors identify China, India, Russia, Indonesia, Iran and Saudi Arabia among countries with substantial emissions growth over the decade from 2013 to 2022. In contrast, growth was small or negative in several other high-emitting nations, including the United States, Japan, Germany and the Republic of Korea. The regional record shows large emissions from North America over time, high emissions in Eastern Europe before the 1990s and rapid recent growth in eastern and southern Asia. Comparisons across regions require care because national borders and regional classifications have changed, including after the dissolution of the Soviet Union in 1991.
Sectoral Sources and Pandemic Changes
The sectoral estimates extend back to 1995 for all countries, reflecting the growing availability of detailed energy-consumption data. In 2022, electricity and heat generation accounted for 39.2% of global CO₂ emissions in the sectoral account, while road transport accounted for 14.5%. Manufacturing, construction and non-fuel industry contributed a further 16.9%. Between 1995 and 2022, emissions from electricity and heat plants recorded the largest increase by sector, rising by 1,789 Mt C, or 82%.
The pandemic year provides a comparison between preliminary estimates and a record based on more complete data. The updated inventory estimates that global CO₂ emissions in 2020 were 4.64% below 2019. Comparing 2020 with the mean of 2019 and 2021 shows the clearest decline in aviation, while household and non-energy-use emissions remained approximately normal. Most countries experienced lower emissions in 2020, but emissions in 2019 and 2021 were comparable in most countries, indicating that the decrease was generally short-lived.
Methods, Revisions and Uncertainty
In the reference approach, national fuel use is estimated from domestic production and imports, minus exports, stock changes and fuels loaded for international commerce. Emissions are calculated using fuel quantities, carbon content and the fraction oxidized; emissions from international shipping and aviation fuels are included in global totals but not assigned to individual countries. The sectoral approach estimates fuel use in specific economic activities and uses product-specific carbon contents. Because these approaches use different data and methods, their estimates need not match: for 2022, the global reference total was 10,287 Mt C, compared with 10,149 Mt C in the sectoral account.
A methodological update revises how carbon stored in non-energy uses of liquid fossil fuels is treated. For years from 1990 onward, the calculation uses storage fractions derived from U.S. Environmental Protection Agency estimates; for earlier years, stored carbon is prescribed as 6.7% of liquid fuel use. The authors note that storage factors vary by country and over time and carry substantial uncertainty. Other refinements include changes to biofuel accounting, solid-fuel energy conversions, bunker-fuel oxidation fractions, cement data and population aggregation; excluding the non-energy-use change, these modifications altered national reference-approach totals by no more than 0.2% in any year since 1751 when the previous product was rerun.
Uses and Limits of the Record
The updated dataset supports comparisons of historical and contemporary emissions by country, fuel group and economic sector. Its consistent coverage from 1750 makes it possible to examine changes over long periods, while the sectoral records identify the economic activities responsible for a large share of recent emissions. The 2022 results point to electricity and heat generation, manufacturing and construction, and road transport as major sources in the global sectoral account. The authors present these patterns as evidence that emissions profiles differ among countries and that sectoral data can inform consideration of potential reductions.
The authors caution that uncertainty varies among countries and is difficult to quantify across the many national estimates. They cite a prior estimate of ±8.4% (2 sigma) uncertainty for the recent global total, while national uncertainty can vary widely; comparisons among datasets also depend on harmonizing their accounting boundaries. Revisions to United Nations energy data and methodological refinements can change estimates, and emissions from non-energy uses remain particularly uncertain because product flows and the share of carbon retained in long-lived products are difficult to establish. The authors conclude that the CDIAC-FF update summarizes emissions through 2022 and illustrates the value of national and international data, while emphasizing differences in emissions profiles as countries address global emissions reductions.
Corresponding Author: Michael Erb
Original Source: https://doi.org/10.1007/s44246-026-00304-1
Contributions: Michael Erb and Gregg Marland contributed to the concept of this study and the writing of the paper. Michael Erb updated the code and data used to produce the new CDIAC-FF dataset and led analysis of the results, including figure creation. Gregg Marland provided guidance and insight about data and methods.
Journal
Carbon Research
Method of Research
Data/statistical analysis
Subject of Research
Not applicable
Article Title
Global carbon dioxide emissions from 1750 to 2022: the CDIAC-FF annual product
Article Publication Date
21-Oct-2026
COI Statement
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.