Treatment of Uncertainties for National Estimates of Greenhouse Gas Emissions



  1. Introduction
Global Atmosphere Division within the Department of the Environment, Transport and the Regions (DETR) has the UK policy lead on climate change. In order to develop policies to limit the emissions of greenhouse gases from the UK, it is necessary to understand the sources of these emissions in detail and how they will change with time. Annual updates of the UK Greenhouse Gas Inventory are reported to the UN Framework Convention on Climate Change (UNFCCC). The guidelines encourages parties to assess uncertainties in the reported estimates but does not give any definitive guidelines to be followed.

It is important to DETR that it understands uncertainties and the impact they might have on the way in which UK's reported emissions are interpreted, and the likelihood that the UK will achieve its targets. This particular study aims to examine uncertainties in the UK emission inventory in more detail than has been achieved to date.

This study has two main objectives:

  1. To review of methods of assessing uncertainty - in the context of their applicability to the UK emissions inventory
  2. To estimate as far as is practical the uncertainty in estimates - both of emissions in 1990 and in projected emissions for the year 2010, and in the projected percentage change from 1990 to 2010

Work addressing the first objective above is summarised in Appendix A of this report. The main section of this document focuses on the second objective, that of addressing current uncertainties in the UK inventory together with the corresponding uncertainties in predictions of greenhouse gas emissions and sinks. From the analysis for projected inventories and uncertainties the role of uncertainty in assessing compliance with targets is then explored.

This study examines the UK emissions and sinks of the six greenhouse gases that are limited by the Kyoto agreement:


In order to provide a firm basis for the uncertainty analysis, information relating to UK inventories of these greenhouse gases was first reviewed. Much of this information was previously presented and summarised in the recent compilation [], although other sources were used to supplement this where practicable. In parallel, the features, advantages and disadvantages of the main methods for assessing uncertainties were studied. This review is presented as Appendix A.

On the topic of the methods for addressing and then assessing uncertainty, the study team felt that significant benefit would obtain if one method could be used for the entire analysis. In this way the overall presentation of the level of uncertainty (or data quality, if the degree of quantitativeness were insufficient) could be given with respect to a common baseline. As is made clear in Section 2, in the context of the UK National Greenhouse Gas Inventory [1], the quantitative methods for uncertainty analysis outlined were found to be appropriate.

The way in which the uncertainties in the individual components of the overall emissions were derived depended crucially on the nature of the information available in each case. Where possible, directly relevant measurements (such as variations in the carbon content of fuels) were used in the formulation of uncertainty estimates, whereas in other cases particular estimates were derived using information from similar sources or associated sectors. In the absence of sufficient numerical data, but with some appreciation of the technical issues involved in modelling the magnitude of emissions, distribution functions for parameters were formulated using expert judgement. In these last-mentioned situations, owing to the lack of directly applicable data emission estimates were assigned comparatively large uncertainties.

Using this general approach, estimates of uncertainties in the emissions inventory of the gases listed above were generated by sector, and nationally, for 1990 and 2010. Using the uncertainties in the two selected years, the changes in emissions during this time interval were also presented in the form of distributions. These reflect the range of increases or decreases in emissions which the data infer would take place during this time.

No assessment was made in this study of the uncertainty of any projected fuel consumption or economic data. This study is limited to the uncertainties that arise from the emission inventory calculation methodology itself.

The estimates of uncertainty in the overall contributions to global warming from the six Kyoto gases (weighted by their respective 100 year global warming potentials (GWPs)) were also evaluated for 1990 and 2010, together with the change between these times.

The final stage of the analysis undertaken involved identifying and ranking the most important contributors to the uncertainties in emissions of individual gases, and hence to total current and projected emissions. This was accomplished using an established statistical method.

Inevitably, an assessment such as this involves the capture and processing of large quantities of data. In order to keep this report to a manageable size the general sources of information are referred to, as well as the strategy adopted in the analysis, but it was not felt useful to reproduce the entire input dataset, or to record every individual decision or judgement of the study team. The large spreadsheets that provide such a comprehensive record of the assessment can be made available separately to DETR if required.