『Abstract
　We investigated the Austrian national greenhouse gas emission inventory to review the reliability and usability of such inventories. The overall uncertainty of the inventory (95％ confidence interval) is just over 10％ of total emissions, with nitrous oxide (N2O) from soils clearly providing the largest impact. Trend uncertainty - the difference between 2 years - is only about five percentage points, as important sources like soil N2O are not expected to show different behavior between the years and thus exhibit a high covariance. The result is very typical for industrialized countries - subjective decisions by individuals during uncertainty assessment are responsible for most of the discrepancies among countries. Thus, uncertainty assessment cannot help to evaluate whether emission targets have been met. Instead, a more rigid emission accounting system that allows little individual flexibility is proposed to provide harmonized evaluation uninfluenced by the respective targets. Such an accounting system may increase uncertainty in terms of greenhouse gas fluxes to the atmosphere. More importantly, however, it will decrease uncertainty in intercountry comparisons and thus allow for fair burden sharing. Setting of post-Kyoto emission targets will require the independent evaluation of achievements. This can partly be achieved by the validation of emission inventories and thorough uncertainty assessment.

Keywords: model uncertainty; Monte Carlo simulation; greenhouse gases; inventory quality considerations; Kyoto Protocol』

1. Introduction
2. Methodology: How to assess the uncertainty of national emission inventories
3. Results
4. Discussion
　4.1. Excluding sources with high uncertainty
　4.2. Covariance and the definition of adequate system boundaries
　4.3. Significance of subjective interpretation of uncertainty
　4.4. Rigid accounting as a method of assessing emissions
5. Conclusions
References