Major Sectors or Activities


Current Uncertainty for Annual Emissionsa

Possible Improvements in 3-5 Years

Uncertainty of Improved Methods

CFCs, PFCs, HFCs, and SF6

Industrial processes

Atmospheric measurements and models


Develop gridded inventories, improved tracer-transport inversion, and measurement of correlated variations of gases


NOTES: 1 = <10% uncertainty; 2 = 10-25%; 3 = 25-50%; 4 = 50-100%; 5 = >100% (i.e., cannot be certain if it is a source or sink). Ranges represent emission uncertainties in different countries (e.g., 1-3 means that uncertainties are <10% in some countries, 10-25% in some, and 25-50% in others). Uncertainty levels correspond to two standard deviations. Shaded rows are the self-reported values; unshaded rows are the independent checks on the self-reported values from independent methods.

aUncertainties for the magnitudes of decadal changes in national emissions can be computed from the numbers in the table using standard statistical methods. Decadal changes (the cumulative change in emissions over 10 years) are reported in the rows requiring OCO measurements because early estimation biases will be reduced in calculation of a decadal change. The uncertainty of a trend is reported as a percentage of the emissions at the beginning of the decade.

bBased on 2006 data reported by five developed countries (Denmark, Greece, Portugal, the United States, and Poland) with a range of institutional capabilities for compiling inventories. In countries where AFOLU sources dominate energy and industrial sources, the uncertainties for total anthropogenic emissions would be much higher.

reported national inventories depend on the data and methods used to create them, which in turn depend on each country’s institutional and technical capabilities. In many developed countries, uncertainties are reported to be less than 5 percent for national CO2 emissions from fossil-fuel use (Table S.1), which is the dominant source. With the exception of a few minor sources in the industrial sector, uncertainties are much higher for other greenhouse gases and sources and vary greatly by country. Uncertainties for the net CO2 emissions from agriculture, forestry, and other land uses and for emissions of CH4, N2O, PFCs, HFCs, CFCs, and SF6 from all sectors can be less than 25 percent in some countries and greater than 100 percent in others.

The second method for estimating greenhouse gas emissions, called tracer-transport inversion, is based on atmospheric and/or oceanic measurements of the gases and mathematical models of air and water flow. Tracer-transport inversion estimates the net sum of anthropogenic and natural sources and sinks. Uncertainties inferred from tracer-transport inversions are less than 10 percent for the net global CO2 flux to the atmosphere but greater than 100 percent for anthropogenic CO2 fluxes at national scales (Table S.1). These large uncertainties arise because of the small size of the anthropogenic signal relative to the large and uncertain natural cycles of emissions and uptake, errors in the reconstruction of atmospheric transport, and the paucity and limited distribution of observations. Tracer-transport estimates of emissions of N2O, CH4, and the synthetic fluorinated gases are currently too uncertain to verify national emissions.

The third method estimates emissions of CO2, CH4, and N2O using methods that are conceptually similar to those used for UNFCCC inventories, but can be made using independent information on land cover. It can be used to estimate emissions from both natural sources and land-use activities, such as agriculture and forestry. Satellite imagery provides the remote information on land surface characteristics and change. This information is converted into estimates of emissions using biogeochemical models constrained by measurements of greenhouse gas exchange between the land and the atmosphere. Satellite imagery is particularly useful for constraining forestry activities and can be used to determine the area of deforestation and forest degradation. The total annual change in forest area has an uncertainty of 10-25 percent in northern forests and up to 100 percent in tropical forests (Table S.1). Uncertainties in emissions from deforestation, reforestation, and forest degradation are high for both annual values and trends, ranging from 25 to 100 percent, because of uncertainties in parameters used to translate deforested area into CO2 emissions. Land remote sensing can also be used to estimate agricultural emissions by identifying the areas using certain agricultural practices, such as paddy rice. Annual uncertainties in CH4 emissions from rice production are 25-50 percent, and uncertainties in N2O emissions from synthetic fertilizer use and manure production are 50-100 percent.

The National Academies of Sciences, Engineering, and Medicine
500 Fifth St. N.W. | Washington, D.C. 20001

Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement