What is actual definition of total emission

[jheeffer]

Probably correct definition, at least for 1751-2013 version, is

Total emission = Fossil fuels + cement production + gas flaring

In 1751-2013 nation csv's header is

Total CO2 emissions from fossil-fuels and cement production (thousand metric tons of C)

In EMS file it says

Total fossil fuels emission

But when adding up the values to total it seems total value is total of everything except bunker fuels.
In other words: fossil fuels, cement production and gas flaring but without bunker fuels.

In global ems file there is no bunker fuel and the formula at the start seems correct.

In global csv file it says Total carbon emissions from fossil fuel consumption and cement production (million metric tons of C) but values again match with above formula (so including gas flaring.

Please clean up indicator names of totals and make sure they match the data well (correcting apparent errors in cdiac naming). Then write what we did in readme.md
Make sure to also check other versions if definition changed between versions. If they do, we might have to change our structure or drop version alltogether (and use latest version).

See below comment

[jheeffer]

We'll take the fast way:

Just call the concept total_carbon_emission. Remove version entity domain from dataset and only add data for latest version (1751-2013).

Instead we can make release branches for old releases. That job will be in the backlog, no need to do it now.

[jheeffer]

See #3 (comment) for proposal. Discuss or comment there.

[semio]

I compared the aggregated sum of nation data and global data (all indicator except per capita emission, because per capita data can't be compared like this), they mostly have small differences. For example

year	carbon_emissions_cement_production (nation)	carbon_emissions_cement_production (global)
2009	415154.0	415000.0
2010	447500.0	448000.0
2011	496228.0	496000.0
2012	520063.0	520000.0
2013	554122.0	554000.0

But there is one indicator have bigger diffs, carbon_emissions_liquid_fuel_consumption. The diffs in this indicator are not constant, there are years data are very close in nation and global. not sure why

year	carbon_emissions_liquid_fuel_consumption (nation)	carbon_emissions_liquid_fuel_consumption (global)
2009	2685213.0	3065000.0
2010	2721127.0	3129000.0
2011	2717870.0	3158000.0
2012	2777257.0	3214000.0
2013	2774646.0	3216000.0

@jheeffer do you think it safe to assume the liquid fuel consumption in nation file is same concept as in global file?

[jheeffer]

Good question. I think it is but I'll send an email to cdiac. Wonder if they'll reply :D. Let's assume it's the same for now

[jheeffer]

Just found this on http://cdiac.ornl.gov/pns/faq.html
Q. Why is the sum of all national and regional CO2 emission estimates less than the global totals?

A. The difference between the sum of the individual countries (or regions) and the global estimates is generally less than 5%. There are four primary reasons for this.

• global totals include emissions from bunker fuels whereas these are not included in national (or regional) totals. Bunker fuels are fuels used by ships and aircraft in international transportation,
• global totals include estimates for the oxidation of non-fuel hydrocarbon products (e.g., asphalt, lubricants, petroleum waxes, etc.) whereas national totals do not,
• national totals include annual changes in fuel stocks whereas the global total does not, and
• due to statistical differences in the international statistics, the sum of exports from all exporters is not identical to the sum of all imports by all importers. [TAB]

[jheeffer]

But I'm sending the email anyways, because differences are 12-14%.

[jheeffer]

They misunderstood the question about liquid fuels, but nonetheless their answer explains some things.

Reply:

For the cement data, this is simply a rounding issue. In presenting global estimates we round to million metric tons to better reflect the precision of the data (i.e., we doubt we know cement production at the global scale to the nearest thousand metric ton). For national estimates, particularly for small countries, we present in thousand metric tons so as not to lose national year-to-year features. For example, if Chad had 3478 and 3003 thousand metric tons of emissions in successive years one gets a better appreciation of the year-to-year change than if we reported both as 3 million metric tons. In summary, this is a balance between reporting an appropriate number of significant digits while maintaining the trends and features of the data.

Global estimates are based on energy production data which are more abundant and accurate than detailed consumption data. National estimates are based on production data and trade data (imports, exports, stock changes) or what we call “net apparent consumption”. Consumption data, especially for developing countries, are simply too sparse to produce a bottom-up estimate.

Negative emissions are typically either artifacts of underlying data errors or our methodology. For example, if the UN reports Saudi Arabia exported more crude oil than they produced this may or may not be an error. If the Saudis are pulling crude out of stock reserves this may be possible but otherwise would indicate a data reporting error. To demonstrate how negative emissions are possible in our methodology, let assume a country imports lots of crude oil and converts the crude to diesel for export. Our emissions estimates may well show negative emissions for liquid fuels because the exports for the country would exceed their within country production of diesel and the resulting emissions for the consumption of the diesel fuel would appear under the country that imported the diesel from the country.

[jheeffer]

In another email I asked:

1. About the specific differences in liquid fuel emissions (and them being so large compared to other indicators)
2. About Western Sahara not being in CDIAC 2013 release.

Reply:

1. The reasons for the differences between global liquid emissions estimates and national liquid emissions estimates are the differences I mentioned between using production data versus “net apparent consumption” data, the circumstances that result in negative emissions, and simple data errors (e.g., the sum of all crude oil imports does not equal the sum of all crude oil exports). This is not a simple rounding issue like cement although we do present global liquid emissions estimates in million metric tons and national estimates in thousand metric tons for the same reasons I mentioned for cement. Apologies I did not make this clearer.
2. The 2013 release of the UN energy database did not contain any data for “Western Sahara”. According to the UN, “We have no direct data for the Western Sahara. A large part is controlled by Morocco and national publications suggest that energy data for that part are included in Morocco's data. There is no explicit declaration in their data submissions to us though.”