Early CO₂ emission estimates for 2016 based on Eurostat monthly energy data. Annual project report

The report verifies the results of the 2015 CO₂ estimates and shows the differences between the calculation of early CO₂ emission estimates and final GHG inventory CO₂ emission for the year 2015. It provides an analysis of the differences in the data sources used to calculate early CO₂ estimates and the final GHG inventory and finally presents the results of the early CO₂ estimates for the year 2016. Results of the comparison of the early CO₂ estimates and final inventory data for the year 2015 show a very good match on the level of the EU-28. Trend changes for total CO₂ emissions from energy consumption from the early CO₂ estimates show an increase between 2014 and 2015 of 0.6 % on the level of EU-28. Trend changes calculated with final GHG inventory data increased by 0.8 %. There are sixteen Member States with differences to final inventory data below 2 %, which contribute 77 % to total EU-28 emissions. Ten Member States show differences to final GHG inventory between 2 % and 5 % and contribute 22 % to final EU-28 emissions. Only two Member States show differences above 5 %, but contribute only 1 % to total EU-28 emissions. Thus the early CO₂ estimates give an early and good indication on the development of the CO₂ emissions on the level of EU-28 and for most MS also on individual country level. Most differences in the trend changes between early CO₂ estimates and final GHG inventory data can be explained by the quality of monthly Eurostat data. Improved data quality for the year 2015 affected the trend changes as well as deterioration of data quality in 2015. This might be caused by the fluctuation in monthly Eurostat data showing an over reporting in one year and an underreporting in the next year. Trend changes for the calculation of the early CO₂ estimates are only influenced by the data quality of monthly Eurostat data. The trend changes of the final GHG inventory data are further affected by changing shares in the amount of carbon stored in liquid, solid and gaseous fuels. This affects the results of the trend changes in Member States that show a high share of carbon sored in total fuel consumption. For carbon stored no data sources are available at the level of monthly energy statistics. Thus, this uncertainty will further remain in the methodology. In Chapter 3.2 the differences between monthly Eurostat data, annual Eurostat data and GHG inventory data are analysed. This is done for each Member State and for the aggregated fuel categories liquid, solid and natural gas. The analysis starts at the most aggregated level of trend changes for CO₂ emissions from liquid, solid and gaseous fuels, followed by the level of aggregated liquid (in kt), solid (in kt) and gaseous fuel (in TJ NCV) consumption and ends up with explanations on the level of single fuel categories for differences above ±3 % in total liquid, solid or gaseous fuel consumption. The analysis shows that on Member States level and on the level of aggregated fuel categories the differences are much bigger than the results of the early CO₂ estimates propose. On the level of liquid and solid fuel consumption the differences in the reporting between annual and monthly Eurostat data increased again in 2015. Only the reporting of natural gas improved in 2015. However, the deterioration of the data quality could not be identified with the outlier and gaps analysis which was carried out in spring 2016 for the monthly Eurostat data 2015. For the calculation of the 2016 early CO₂ estimates only a few outliers and gaps could be identified in the monthly Eurostat data for 2016. Most Member States that were asked to confirm data or send corrected values confirmed their data. Only in a few cases updated questionnaires have been sent. The early CO₂ estimates 2016 include corrections for hard coal consumption in the Netherlands (using market inland deliveries instead of gross inland consumption), for peat and hard coal consumption in Ireland and for liquid fuel consumption in Sweden. Further corrections are related to the calculation of trend changes in TJ instead of kt. This applies for liquid fuel consumption in Estonia and for solid fuel consumption in Spain, Hungary and Romania. Finally the application of the trend method to estimate early CO₂ emissions for all EU Member States represents a robust procedure that adjusts systematic errors of under- or over reporting in Eurostat monthly data. The advantage of the method is the simplicity that ensures a fast and straightforward calculation for each Member State. By applying the trend change method, inconsistencies in the reporting of Eurostat monthly energy data compared to Eurostat annual or GHG inventory data can be levelled out, if these inconsistencies persist throughout the entire time series of Eurostat monthly energy data. For many Member States that show differences above 3 % between monthly Eurostat data and the other data sources (annual Eurostat data and GHG inventory data), the use of the trend changes method levels out the differences and leads to good results for the early CO₂ estimates. 8