2017

The Carbon Dioxide Information Analysis Center (CDIAC), located at the U.S. Department of Energy’s (DOE) Oak Ridge National Laboratory, has published annual Carbon Emissions from Fossil Fuels and cement production by country since 1959 (‘Global Carbon Project‘), but now this continuous time series has come to an end and 2015 will be the last data-year (as it seems).

Since carbon emissions data from CDIAC (Global Carbon Project) is the core ‘Indicator‘ in ClimatePositions’ calculation of Climate Debt, carbon emissions data will be replaced with nearly similar data from ‘EDGAR‘ (‘European Commission‘ / ‘Climate Action‘), retroactively since 1990, in connection with the coming updates [done 16-08-2017].

The following describes the differences between CO2 Emissions data from Global Carbon Project (CDIAC) and EDGAR (sourced: European Commission, Joint Research Centre (JRC)/Netherlands Environmental Assessment Agency (PBL). Emission Database for Global Atmospheric Research (EDGAR)), and the consequences in terms of Climate Debt in ClimatePositions – illustrated with a range of country examples. Note that other sources, such as ‘IEA‘, ‘EIA‘ and ‘BP‘, provides CO2 Emissions data-sets different from the ones of CDIAC and EDGAR.

Before proceeding, the table below shows the ten largest CO2 emitters in terms of 1) Per capita Climate Debt calculated with CDIAC’s CO2 Emissions data (the current) with the per capita CO2 Emissions 2015 in brackets, and 2) Per capita Climate Debt calculated with EDGAR’s CO2 Emissions data (the coming) with the per capita CO2 Emissions 2015 in brackets. Increased Climate Debt with EDGAR is red and decreased is green. Note that the Climate Debts are calculated on the bases of emissions-data from 1990s (baseline), in comparison with emissions-data since 2000.

The differences in national CO2 Emissions at CDIAC and EDGAR are caused by numerous complex explanations. The list below is neither adequate nor scientifically profound, but it gives a sense of the comprehensive scientific disputes and uncertainties behind the final figures.

• Both CDIAC and EDGAR includes CO2 Emissions from oil, natural gas, coal, and cement production (per ton of cement produced), however the latter are taking into account the decreasing share of clinker in cement. As this fraction has be decreased significantly in most countries in the last decades, thereby proportionally decreasing the emission factor expressed in per tons of cement produced, the EDGAR emissions are about 20% lower than the unadjusted values in the CDIAC dataset
• In addition to cement production, EDGAR also includes other non-combustion industrial processes, such as the production of lime and soda ash and carbon used in metal production. All sources of CO2 related to nonenergy/feedstock uses of Fossil Fuels are estimated according to IPCCs 2006-guidelines. Collectively, the other carbonate sources added about 30% to CO2 Emissions from global cement production in 2010.
• Different data release dates and conversion factors used. EDGAR, for example, has incorporated the recent major revisions of China’s Fossil Fuel statistics since 2000 (in particular coal) in the data set.
• Country-specific emission factors for key sources, sometimes differ from international default factors recommended by the IPCC. Also, countries may use default values when country-specific data are missing or for relatively small sources.
• The IPCC guidelines recommend emission factors expressed as “per unit of energy” for fuel. However, these recommendations are sometimes disregarded and replaced with tons or m3 (EDGAR respect the guidelines).
• Different definitions might be used for specific emission sources. For Fossil Fuel combustion, for example, CO2 Emissions related to coal and coke inputs in blast furnaces and coke ovens (and carbon losses in these processes) may be partly or fully reported under industrial processes or fugitive sources or under fuel combustion. Also, the level of detail of the methodology used can be different: detailed fossil-fuel types or only aggregated ones (e.g. coal, oil products and natural gas), corrections made for non-energy uses of fuels (e.g. natural gas for ammonia production), and fuels used for international transport.
• CO2 Emissions from the use of oil and gas for non-energy use (for example as chemical feedstock) may be calculated with different methods and may be included under Fossil Fuel combustion instead of under industrial processes.
• Uncertainty concerning fuel data and variations in carbon content per fuel type causes an overall margin of error. Differences between official national CO2 Emissions (in different data-sets) are generally within 5% for OECD countries and around 10% for countries with less well-developed statistical systems. The uncertainty in EDGAR’s total national CO2 Emissions from Fossil Fuel use and other, non-combustion sources is estimated at about 5% for OECD countries and around 10% for countries of the former Soviet Union, such as Russia and Ukraine. For other countries which are not annually reporting national emissions inventories to UNFCCC, the EDGAR uncertainty estimates of national CO2 Emissions vary between 5% for countries with well-developed statistical systems (including India) and around 10% or more for countries with less-developed statistical systems.
• EDGAR (as IEA) has 42 different Fossil Fuel types in the ‘Database‘ distinguished for CO2 calculation, while CDIAC has 4 (hard coal, brown coal, gas and liquid)¹.
• Moreover, energy statistics for fast changing economies (such as China) are less accurate than those for the other countries within the OECD.
• CO2 emission trends over recent years, estimated using energy data published annually by BP, appear to be reasonably accurate for estimating global annual CO2 trends.
• For example, based on older BP energy data, the increase in 2005 in global CO2 emissions from fuel combustion compared to 2004 was estimated at 3.3%, globally. With more detailed statistics by the International Energy Agency (IEA) for 2005, which became available two years later, the increase is estimated at 3.2%.
• The uncertainty in CO2 emissions from fossil-fuel combustion using international statistics is discussed in detail in Marland et al. (1999) and Andres et al. (2012), and general uncertainty characteristics in global and national emission inventories are discussed in Olivier and Peters (2002). Andres et al. (2012) evaluate several studies on the uncertainty of CO2 emissions from fossil-fuel use and cement production and conclude that they range from between about 3% and 5% for the United States, to between 15% and 20% for China, based on a comparison of CO2 estimates based on national coal statistics and on the sum of provincial coal statistics (Gregg et al., 2008), to estimates of 50% or more for countries with poorly maintained statistical infrastructure (Marland et al., 1999). In spite of the large national efforts to provide accurate estimates, emission inventories in non-OECD countries are generally less accurate than those in OECD1990 countries.
• Although the level of detail for the national fuel use calculations differs substantially, the differences in the global level are relatively small (the difference in global level between CDIAC and EDGAR Fossil Fuel related CO2 Emissions are very small).
• For flaring EDGAR reports values about twice as high as CDIAC (and EIA), which is remarkable since the CDIAC and EIA data also include venting. The difference can be explained by the different estimation method for the activity data, which is mainly based on reported energy statistics for CDIAC and EIA, but mainly on satellite data for EDGAR.
• Neither CDIAC nor EDGAR include emissions from international shipping or bunker fuels (some international data sets, notably those of the EIA and BP, do not separate the use of bunker oil for international shipping and international aviation from a country’s oil consumption and report those emissions as part of total national emissions).

The next table shows the ten largest per capita CO2 emitters (again) in terms of 1) Per capita Climate Debt calculated with CDIAC’s emissions data (per capita CO2 Emissions 2015 in brackets), and 2) Per capita Climate Debt calculated with EDGAR’s emissions data (per capita CO2 Emissions 2015 in brackets).

One country-example explained: The reason why the accumulated per capita Climate Debt of United Arab Emirates decreases after shifting to EDGAR although CO2 Emissions is higher, is that the average emissions between 2000 and 2015 has decreased more in comparison with 1990s (baseline), than it has with CDIAC’s data-set.

The table below shows the ten countries with the largest per capita Climate Debt. Among these extremely climate-destroying countries only Bahrain benefit significantly from the shift to EDGAR.

The last table shows ten countries with significant change in per capita Climate Debt after shifting the data-source. Singapore goes from “good” to “very bad” (although lower 2015-emissions with EDGAR) because emissions with EDGAR was 2.2 tons higher (than with CDIAC) in 1990s (baseline) and therefore must reduce emissions at an even steeper rate.

¹Source: ‘http://edgar.jrc.ec.europa.eu/news_docs/jrc-2014-trends-in-global-co2-emissions-2014-report-93171.pdf‘ (Table 2.3).

.

CDIACs data on carbon emissions (CO2 Emissions) from Fossil Fuels (without bunkers) and cement production are from ‘Global Carbon Project (CDIAC)‘ (National Emissions 2016). National Population data is from ‘Worldbank‘.

EDGARSs data on per capita CO2 Emissions from Fossil Fuels (without bunkers) and industrial processes (cement production, carbonate use of limestone and dolomite, non-energy use of fuels and other combustion, chemical and metal processes, solvents, agricultural liming and urea, waste and fossil fuel fires) are from ‘EDGAR‘ (Download of timeseries 1970-2015 for all countries, xls) . Source: European Commission, Joint Research Centre (JRC)/Netherlands Environmental Assessment Agency (PBL). Emission Database for Global Atmospheric Research (EDGAR). Population data used in EDGAR are from ‘UNDP‘ (World Population Prospects (WPP), Population Division).

Main source of the differences between CO2 Emissions data from CDIAC, EDGAR, etc.: ‘Trends in global CO2 Emissions, 2016 report, pdf‘ (Netherlands Environmental Assessment Agency).

Drawing by Claus Andersen, 2017.