This is an update of the dataset documented in: Andrew, R.M., 2019. Global CO2 emissions from cement production, 1928–2018. Earth System Science Data 11, 1675–1710. https://doi.org/10.5194/essd-11-1675-2019. Data in this release cover the period 1880–2021. Note that emissions from use of fossil fuels in cement production are not included in this dataset since they are usually included elsewhere in global datasets of fossil CO2 emissions. The process emissions in this dataset, which result from the decomposition of carbonates in the production of cement clinker, amounted to ~1.7 Gt CO2 in 2021, while emissions from combustion of fossil fuels to produce the heat required amounted to an additional ~1.0 Gt CO2 in 2021. September 2022 release (220915): Major changes Brazil: Use of reported clinker ratio 1970-2000 South Africa: Revised calculation for 1996-2020 Jamaica: Updated with 2021 annual report of Caribbean Cement Corrected missing emissions in CZE, SVK, BIH, HRV, MKD, SVN, SRB Mexico: Used monthly cement production data Added source information for every data point in cement production dataset The Cement Production dataset Cement production data by country are primarily derived from USGS statistics. The construction of this dataset begins with production back-calculated from CDIAC's 2019 edition cement emissions data, which are a direct function of cement production (from the 2020 edition CDIAC has changed its methodology). Then using available data for some former Soviet states before the dissolution of the Soviet Union, Soviet states are disaggregated for all years before dissolution. Data obtained directly from USGS are used to overwrite from 1990 onwards, with a small number of additional corrections. Countries for which cement production is not available in the most recent years are extrapolated simply. Finally, country-specific cement production data are overwritten where better and/or more recent sources are available. Note that many zeros in the cement production dataset are propagated from CDIAC and should probably be NODATA. The approach used for each country is summarised in the file "6. cement_production_method.csv" with full, element-wise sources provided in the file "4. cement_production_5_sources.csv". Emissions calculation Emissions for all UNFCCC Annex I countries ("developed" countries) are derived from their official submissions to the UNFCCC in Common Reporting Format (structured Excel files), for which data are available from 1990 (slightly earlier for some Economies in Transition). For non-Annex I countries clinker ratios derived from the Getting the Numbers Right (GNR) cement sustainability initiative are applied to the cement production dataset to derive approximate clinker production by country, from which emissions are calculated using IPCC default factors. Country-specific methods are used for China, India, Japan, Thailand, Turkey, USA. The combined_cement_data.xlsx file is used to overwrite emissions with superior data, in most cases as reported in official reporting to the UNFCCC, e.g. Biennial Update Reports, National Communications, and National Inventory Reports. Some countries do not report time-series of emissions, but do supply some isolated estimates in their official reporting to the UNFCCC, and these are used in some cases to constrain estimates. A number of countries state in their official reporting to the UNFCCC that they have never produced clinker, so emissions are set to zero for all years for these countries. In other cases, statements are made that no clinker was produced before a certain year, and this information is also incorporated. The information available usually covers a number of years, up to 3 decades. These are then extrapolated by combining available data and assumptions about historical developments in clinker ratios to produce longer time series of emissions based on the longer cement production dataset. More detail on this method are given in the accompanying journal paper.