Can we get rid of uncertainties in GHG inventories?

摘要

Uncertainties in typical GHG emission inventories are dominated by the contribution from the non-CO_2 gases. The sources of these gases can be complex and emissions variable in space and time and the estimates can contain large systematic errors due to lack of knowledge. The available data indicate that the uncertainty in overall emission level can be highly influenced by the uncertainty in nitrous oxide from agricultural soils. The trend uncertainty can be influenced by gases where the emission level is changing rapidly, even if their contribution to total emissions is small, e.g. HFCs and N_2O from road transport. The uncertainty in emission inventories can be estimated by methodologies at different levels of sophistication. While simple error propagation can be used to roughly assess the uncertainty in emission level, the trend uncertainty needs to be modelled, taking into account correlations between parameters in the base and end year. Both these methodologies require assessments of the uncertainties in individual parameters. These are often based on expert judgments and may be rather subjective. This means that uncertainty estimates of inventories not necessarily are comparable. The typical uncertainty in the total emission level of high quality inventories is higher than desired for scientific and policy use of inventories, more than 10 % (excluding LUCF). The few available studies indicate that the trend uncertainty can be higher than desired for assessing compliance with the Kyoto protocol, about 4-5 percentage points. It is expected that the uncertainty of emission estimates from individual sources can be reduced in the future by better scientific knowledge and improvements in inventory methodologies. In most inventories the gain is highest by reducing the uncertainty of the non-CO_2 gases. This will reduce the overall uncertainty in emission level, but only to some extent the uncertainty in the trend.