Uncertainty in Emissions Estimates 

The emissions numbers presented in this report are, for the most part, estimates based on estimated activity data and estimated emission factors. As such, they have an element of uncertainty, given that the activity data and emission factors on which the emission estimates are based also have a range of possible values. The activity data and emission factors can themselves be characterized by systematic biases and/or random errors. In 2000, EIA employed a Monte Carlo analysis to estimate the range of uncertainty, at a 95-percent confidence level, around estimated emissions of carbon dioxide, methane, and nitrous oxide (HFCs, PFCs, and SF₆ were not part of the analysis).⁸ 

The Monte Carlo simulations revealed that uncertainty varies by type of gas. There is less uncertainty around the simulated mean for carbon dioxide (-1.4 percent to +1.3 percent) than for methane (-15.6 percent to 16.0 percent) or nitrous oxide (-53.5 percent to +54.2 percent). The simulations also showed that the uncertainty around the simulated mean of total greenhouse gas emissions (excluding HFCs, PFCs, and SF₆) is -4.4 percent to +4.6 percent. 

The certainty of emissions data varies by category and by source. For example, methane emissions from existing underground coal mines are relatively certain. In general, however, the estimates for carbon dioxide emissions are more certain than the estimates for other gases. It is likely that the estimate of total U.S. carbon dioxide emissions is accurate to within 5 percent. For methane emissions, most of the estimates are much more uncertain, with a level of uncertainty that may exceed 30 percent. Estimates of methane emissions may also understate actual emissions as a result of the exclusion of sources that are unknown or difficult to quantify. For example, EIA does not include sources such as abandoned coal mines and industrial wastewater. Nitrous oxide emissions estimates are much less certain than those for carbon dioxide or methane emissions, in part because nitrous oxide emissions have been studied far less than emissions of the other greenhouse gases and in part because the largest apparent sources of nitrous oxide emissions are area sources that result from biological activity, which makes for emissions that are highly variable and hard to measure or characterize. The uncertainty for nitrous oxide emissions may exceed 100 percent.

Notes