It has been acknowledged, that providing improved uncertainty estimates is an urgent need for models and policy applications, but also that it is hard to achieve on the short term other than through collective expert judgement. Within the IPCC Inventory Programme a special activity has recently been started aiming at providing default values and a better framework for estimating and reporting of uncertainties. A similar activity has started within GEIA, aiming at the same results at country level as well as on grid level, but for a more extended group of compounds and in a more scientific setting. Comparisons of different datasets, e.g. as done with the EDGAR data, may be an important input to this process.
 Therefore, it was decided by the EDGAR team that it was at present not feasible to go beyond the uncertainty tables compiled for Version 2.0 (see Table 1 and 2). However, in addition some apparent conclusions on uncertainty can be drawn from the comparison of V2.0 with other datasets as part of the validation and check for urgent adaptations. For more detailed info on uncertainty in specific inventories we refer to the papers on N2O, NH3 and CO (Bouwman et al. 1995, 1997; Olivier et al, 1999) and conclusions on validation in Olivier et al. (2001b).
 To illustrate the usefulness of the uncertainty indications presented here, we translated for methane the qualifications into the percentage ranges mentioned in the notes of these tables. Next, we compared the results with the uncertainty estimates used by IPCC Working Group I for the preparation of the Third Assessment Report (Fig. 1). The overall picture is rather similar, which shows that this simple approach and interpretation of uncertainty ranges is still quite useful for application at global levels.

Table 1. Indication of uncertainty estimate for greenhouse gases. Source: Olivier et al., 1999a.

Notes:  Expert judgement of uncertainty ranges, which were assigned with the following classification in terms of order of magnitude of the uncertainty in mind: S = small (10%); M = medium (50%); L = large (100%); V = very large (>100%).
"-" Indicates that the compound is not applicable for this source or that emissions are negligible.

Table 2. Indication of uncertainty estimate for ozone and aerosol precursors. Source: Olivier et al., 1996b

Notes:  Expert judgement of uncertainty ranges, which were assigned with the following classification in terms of order of magnitude of the uncertainty in mind: S = small (10%); M = medium (50%); L = large (100%); V = very large (>100%).
"-" Indicates that the compound is not applicable for this source or that emissions are negligible.

Figure 1. Comparison of uncertainty estimates for major global methane sources using the uncertainty estimates by the EDGAR team.

Figure 2. Comparison of uncertainty estimates for major global methane sources using the compilation for the Third Assessment Report of IPCC Working Group I.