Revisions In United Nations Energy Statistics Data: Can Changes To The Past Improve Our Understanding Of The Present?

Before data can be analyzed effectively, it needs to be evaluated to determine its quality. With the rise in interest in combating global climate change by reducing carbon dioxide (CO2) emissions, understanding data quality and certainty is important for producing accurate estimates of emissions from fossil fuel combustion and other industrial processes. These estimates are essential for monitoring the prospects and progress towards national targets for reduction. Accurate fossil fuel CO2 (FFCO2) estimates, along with climate modeling and current atmospheric conditions, are also essential to forecast future global temperatures. Most fossil CO2 emissions inventories rely on energy statistics to generate country and year specific estimates of CO2. Every year, the United Nations publishes an updated Energy Statistics Database that provides annual statistics on the production, processing, trade, and use of fuels for over 230 countries, going back as far as 1950. Each subsequent publication of the database provides additional entries and updated values to the previous years’ data. This opens up questions about these data revisions: what changes are taking place, how are countries revising their data, and are there recognizable patterns that provide information to anticipate changes in future database releases? The purpose of this thesis is to learn from the revisions to address questions and concerns surrounding this Energy Statistics Database. This begins with an investigation into the changes occurring in the datasets year-to-year and ends with an inference into patterns within these changes. The additional insight this presents allows a deeper level of understanding for where in the database revisions are occurring, to better characterize the uncertainty of present CO2 value estimates.