Environment Statistics

Collecting, Analyzing and Disseminating data, one Country at a time!

Forest CO2 emissions and removals, 1990 – 2020

Rome, July 21 2020. FAO releases today new estimates of forest emissions and removals of CO2 to the atmosphere, based on newly released data from the Forest Resources Assessment 2020 (FRA). Emission and removal statistics are disseminated by country, with regional and global coverage, over the period 1990-2020. Data are aggregated at national level for 205 countries and territories, These FAOSTAT statistics represent the only available global dataset in the world today showing country, regional and global time series on forest emissions and removals based on the FRA 2020, including emissions from deforestation and emissions/removals on remaining forest land.

Statistics of forest land area and biomass stock are collected from countries, analysed and disseminated by FAO in five-year cycles through the Forest Resources Assessment. They relevant for monitoring trends in sustainable land management, forest dynamics and forest health. FAOSTAT uses the FRA data as input for estimating CO2 emissions and removals from forest (Federici et al., 2015), an integral component of total greenhouse gas emissions from land, disseminating the information via the FAOSTAT Land Emissions-Forest Land domain. The new FRA 2020 data allowed for updating the FAOSTAT domain with emissions and removals from forests for the period 1990-2020. The Forest Land domain contains information for 214 countries and territories, with complete coverage over the 1990-2020 time series of statistics on: forest land area, area of planted forests and naturally regenerating forests, carbon stock and emissions from deforestation and emissions/removals from remaining forest land, by country, with regional and global coverage. The forest land domain features a significant update to ‘’t minus zero,’’ only possible in FRA years, through which the most recent information on forest emission trends is made available. This analysis focuses on new trends in forest emissions in terms of deforestation, carbon storage on remaining forest land, and overall net impact of forests on climate. In this respect, it is a notable contribution to ongoing efforts to update the global carbon cycle, to which it regularly contributes through analyses published in high-impact peer-reviewed journals (e.g, LeQuere et al., 2019) and international assessments (e.g., IPCC, 2019). It should be noted that all emissions/removals in this report are all treated as anthropogenic. While there is no doubt that emissions stemming from deforestation activities are indeed attributable to human action, not all of the other dynamics on forest land are necessarily due to human action. It is nonetheless a fact that the vast majority of forest land on earth is actively or passively managed, protected or covered by administrative regulations, and in this respect it can be considered to be under human influence.

Figure 1.  Global emissions and removals from forests, 1990-2020 (Gt CO2 yr-1)




Overall emissions/removals. Forests’ global net CO2 flux to the atmosphere was small in 2020 (0.5 Gt CO2 yr-1) and had steadily decreased over the period 1990-2020. Remarkably, the new estimates based on FRA 2020 data suggest that forests acted globally as a net carbon sink (-0.6 Gt CO2 yr-1) in the period 2011-2015. This is the first time that an overall sink has been calculated from the underlying forest area and biomass statistics, and it remains to be verified independently, including through careful observation of the atmospheric CO2 record during that period. Compared to FAOSTAT estimates made with the FRA 2015, the FRA 2020 data led to significantly reduced estimates of total forest emissions—by about half on average over the full period 1990-2020.

Deforestation. Globally, deforestation carbon fluxes decreased by about one-third from 1990 to 2020, reaching about 3 Gt CO2 yr-1 in 2020.

Forest Land. Remaining forest land, i.e., the area not being deforested, acted globally as a significant net carbon sink over the entire period 1990-2020. The 2020 sink, at -2.6 Gt CO2 yr-1, was nonetheless roughly one-quarter smaller than in 2000, and thirty-five percent less than in 2015. The FRA 2020-based estimates predicted a stronger global forest sink compared to previous estimates, particularly during the period 2011-2015 (-4 Gt CO2 yr-1)


Figure 2.  Regional distribution of deforestation emissions, 1990-2020.


Overall emissions/removals. The global forest land flux was determined in 2020 by large net fluxes to the atmosphere from Africa (>1 GtCO2 yr-1) and the Americas (0.5 Gt CO2 yr-1). Counterbalanced by net sinks from Europe (1 Gt CO2 yr-1) and Asia (0.5 Gt CO2 yr-1).

Deforestation. America remained in 2020 the region with the highest deforestation emissions (1.3 Gt CO2 yr-1), with Africa reaching these levels in the same period. While in the Americas emissions had declined since 2010 by about one-half, deforestation emissions in Africa increased by nearly one-fourth from 2010 to 2020 (Figure 2).

Forest Land. The strongest sinks were in Europe, Asia and the Americas (-1 Gt CO2 yr-1). However they declined strongly in 2020 compared to 2015, by nearly 50 per cent in Europe, and by nearly one-third in Asia and the Americas (Figure 3)


Figure 3.  Regional distribution of forest land emissions and removals, 1990-2020


Deforestation. Brazil and DRC were the countries with the highest estimated emissions from deforestation during the recent 2016-2020 period, around 600 Mt CO2 yr-1. Indonesia was a distant third with about 200 Mt CO2 yr-1. Forest Land. During the recent period 2016-2020, countries with the largest forest sinks were    China, Russia (~ 750 Mt CO2 yr-1), followed by USA and Brazil (~ 350 Mt CO2 yr-1) (Figure 4).


Figure 4.  Country level forest land emissions and removals, 2015-2020

In terms of comparing the FAOSTAT estimates to emissions, data reported by countries to the Climate Convention (UNFCCC), FRA 2020-based estimates for Forest Land were in good agreement with Annex I country aggregates (within 20% in 2000 and 5% in 2015), with significant improvements compared to FRA 2015 estimates in most countries Estimates for the countries with the three largest forest land sink countries were in very good agreement with UNFCCC data (within , 7% for Russian Federation; 15% for China; 30% for USA) Despite less available data for non-Annex I countries, comparisons of deforestation emissions for major contributing countries Brazil and Indonesia indicated good agreement with reported data


Figure 5. Comparisons of FAOSTAT Estimates (FRA 2020 and 2015) and UNFCCC country data, 1990-2020




The Forest Land domain disseminates information on emissions and area change due to gain and losses of carbon stocks in living (aboveground and belowground) tree biomass. Estimates are computed following land Approach 1 and Tier 3 ( stock difference method) of the 2006 IPCC Guidelines for National GHG Inventories (IPCC, 2006). Data on carbon stocks and forest area used as input are those of the Forest Resources Assessment 2020 (FAO, 2020). Data are available by country, with global coverage and relative to the period 1990-2020.

It contains data on net CO2 emissions/removals, associated implied emission factors and underlying activity data. The FAOSTAT Emissions data are estimated by FAO and do not coincide with GHG data reported by Parties to UNFCCC. FAOSTAT data and estimates support member countries to assess and report their emissions and removals. FAOSTAT emissions data are a global knowledge product.


FAO, 2020. Global Forest Resources Assessment 2020, Desk Reference. Food and Agriculture of the United Nations, Rome. Available at

IPCC, 2006. 2006 IPCC Guidelines for National Greenhouse Gas Inventories, Prepared by the National Greenhouse Gas Inventories Programme, Eggleston H.S., Buendia L., Miwa K., Ngara T. and Tanabe K. (Eds), IGES, Hayama, Japan. Available at

Federici S., F.N. Tubiello, M. Salvatore, H. Jacobs, J. Schmidhuber, 2015. New estimates of CO2 forest emissions and removals: 1990-2015. Forest Ecology and Management 352 (2015) 89-98. Available at:


This analytical brief was prepared by Francesco Nicola Tubiello, with support from Nathan Warren, Giulia Conchedda and Griffiths Obli-Layrea (Statistics Division) and inputs from Anssi Pekkarinen, Lars Marklund and Mette Wilkie (Forestry Division)

Cover photo: Francesco N Tubiello