Climate Change Cluedo: Anthropogenic CO2 - Notes

Original article.

1) Timing
2) Correlation
3) Mass Balance
4) C14
5) C13


1) Timing
Data from Boden, T.A., G. Marland, and R.J. Andres. 2011. Global, Regional, and National Fossil-Fuel CO₂ Emissions. Carbon Dioxide Information Analysis Center, Oak Ridge National Laboratory, U.S. Department of Energy, Oak Ridge, Tenn., U.S.A. doi 10.3334/CDIAC/00001_V2011
Data for Human emissions from land use change from 1850 - 2005 can be found from Houghton 2008:
Houghton, R.A. 2008. Carbon Flux to the Atmosphere from Land-Use Changes: 1850-2005. In TRENDS: A Compendium of Data on Global Change. Carbon Dioxide Information Analysis Center, Oak Ridge National Laboratory, U.S. Department of Energy, Oak Ridge, Tenn., U.S.A.

2) Correlation
To determine the correlation I took the emissions data for fossil fuel use and cement production from Boden et al (2011), and those for Land Use Change from Houghton et al (2008) (see above).  I combined these two to give a record of anthropogenic emissions from 1850 to 2005.  I also obtained the data for CO2 concentration from Mauna Loa, and the 20 year smoothed ice core data from Law Dome.   That gave me a record from 1832-1978.
I adjusted the Mauna Loa data by deleting the 1958 record, and by infilling missing months from 1964 by taking the average of the same month of the preceding and following years.  With the infilled data I obtained an annual average for 1964.  I then formed an extended record by taking the ice core data through to 1978, and then for each following year, adding the difference between the Mauna Loa record of that year and the Mauna Loa record of the preceding year.  Finally, I calculated the correlation over the record from 1850 to 2005,  and seperately with the Mauna Loa data itself from 1959-2008, the figures being reported above.

Also of interest are the correlations with the fossil fuel plus cement emissions only.  From 1833-2008, the correlation is 0.987 (r^: 0.975), while compared with the Mauna Loa record from 1959-2008, the correlation is a stunning 0.9996 (r^2: 0.9991).  I do not know why the fossil fuel data correlates better than does the combined data.  It is possibly because of inaccuracies in the LUC emissions, which have higher uncertainties; and possibly because there have been large changes in the amount and latitude of LUC emissions possibly resulting in changes in efficiency of uptake over time.  Of course, it is very possible the reason is something I have not thought of.

All calculations where performed using Open Office Calc, and the spread sheet is available on request.
Keeling, R.F., S.C. Piper, A.F. Bollenbacher and J.S. Walker. 2009. Atmospheric CO₂ records from sites in the SIO air sampling network. In Trends: A Compendium of Data on Global Change. Carbon Dioxide Information Analysis Center, Oak Ridge National Laboratory, U.S. Department of Energy, Oak Ridge, Tenn., U.S.A. doi: 10.3334/CDIAC/atg.035

Etheridge et al, 1998 data

3)  Mass Balance 

As a further extension of the spreadsheet, I took the growth in CO2 concentration in ppmv divided by (Fossil Fuel  + Land Use Change emissions data)/2113  to convert megatonnes to ppmv.  The most recent time in which CO2 concentration grew by more than emissions was 1884.  That is a fairly robust result.  I repeated the test requiring only that CO2 growth excede 80% of combined emissions and came up with a most recent year in which that occured as 1885.  That means that even allowing for uncertainties in the data, it is unlikely that CO2 growth has exceeded emissions growth in the 20th century.   The mean airbourne fraction from 1850 to 2005 was 56%, while the that for the Mauna Loa data alone (1959-2005) was 57.6%, compared to the 57% shown in the diagram for figure 2.

4) Declining C14 ratios 

One subtlety to watch for is that dissolved carbon has a much lower C14 content than atmospheric CO2 or land plants.  This means that a sufficiently large outgassing from the ocean would also depress the C14 concentration in the atmosphere, but not as rapidly as does the emissions of fossil fuels.

5) Declining C13/C12 ratio

C13/C12 ratios are measured as d¹³C, where:
d¹³C_sample = {((¹³C/¹²C sample) / (¹³C/¹²C standard)) - 1} x 1000
The standard is the C!3/C!2 ratio of Vienna Pee Dee Belemnite (VPDB), or 0.0112372
It follows that C13/C12_sample =  0.0112372 * ((d¹³C_sample/1000) + 1)
The following are the d¹³C (C13/C12) of various sources of CO2 emissions.
From the British Geological Survey article, Volcanic Contributions to the Global Carbon Cycle (PDF):
Volcanic emissions:
Divergent Plate Boundaries: -4 +/-2.5 parts per thousand C13 (0.0111923); Emissions: 66-97 Megatons Carbon per year
Intraplate Hotspots: -4 +/-2.5 parts per thousand C13 (0.0111923); Emissions: 80-132 Megatons Carbon per year
Convergent Plate Boundaries: +2.5 to -12 parts per thousand C13 (0.0112653 to 0.0111024); Emissions: 66-135 Megatons Carbon per year

From Quay et al (1992) and Fung et al (1997) (in brackets):
Fossil Fuels:  - 27 (-24 to -28) parts per thousand (0.0109338)
Biomass: -27 parts per thousand (0.0109338)
(C3 plants: -27.5 parts per thousand)
(C4 plants:  -3 to -6 parts per thousand)
Atmosphere: -8 parts per thousand (0.0111473)
Ocean: 0 to 2 parts per thousand (0.0112372 to 0.0112597)From Hudon (1975)


Cement: 0 parts per thousand (0.0112372)
I will assume (perhaps rashly) that the mean C13 concentration in emissions from convergent plate boundaries is the midpoint of the range of values given, ie, -4.75 parts per thousand.  In that event, the net volcanic emissions from convergent plate boundaries will increase the C13/C12 ratio in the atmosphere contrary to what has been occuring.  However, even if all convergent plate boundary emissions had a dC13 of -12, as they represent only a third of all volcanic emissions, the net impact would still be a dC13 for all volcanic emissions approximately equal to 7, still increasing the C13/C12 ratio.  Indeed, even if all convergent plate boundaries had a dC13 of -12, and they represented 50% of all volcanic emissions, the net effect of volcanism would be no change in atmospheric dC13.  The odds of that are, however, very remote on available information; and it must be considered a best case scenario for volcanic outgassing on available information.  Ergo, there is buckley's chance that volcanic emissions have caused any decline in the C13/C12 ratio.
Further, recalling from Miller et al that uptake by the ocean discriminates by 2 parts per thousand, and that consequently outgassing would discriminate in the reverse direction, we can see that the dC13 of CO2 released by oceanic outgassing would be at most a quarter of that of the atmosphere itself.  As with volcanic emissions, not only would oceanic outgassing not cause a fall in the C13/C12 ratio in the atmosphere, it would cause the reverse.