The recent paper by Ernst-Georg Beck, ‘180 Years of Atmospheric C02 Gas Analysis by Chemical Methods’
and his supporting data has been discussed here. Briefly, Beck looks at historical records of atmospheric CO2 measurements since 1812, and finds that many scientists recorded measurements much greater than the 290 parts per million (ppm) which has been accepted by the Intergovernmental Panel on Climate Change as being the pre-industrial level of atmospheric CO2, before the increasing usage of fossil fuels began to raise atmospheric CO2.
This raises a number of questions. How accurate are the old measurements? Were they contaminated by nearby sources of CO2 emissions? How did the IPCC come to accept 290 ppm as the pre-industrial atmospheric CO2 level?
On the question of accuracy, Beck mentions three methods of measuring CO2, and explains that the apparatus used was calibrated against gas with a known CO2 content, and scientists also calibrated their apparatus against the equipment of other scientists. The accuracy of the most common technique, the Pettenkofer process, is +/- 3%. Obviously this is much less accurate than modern methods, but still enough to be confident that the results are not wildly inaccurate by tens or hundreds of parts per million.
Could the samples have been contaminated, such as by war activity, industrial processes, or other local sources of CO2? It is certainly possible. For example, a series of 25,000 measurements taken at Giessen, Germany, between 1939 and 1941, averaged 438.5 ppm. The influence of a city is estimated to be between 10 and 70 ppm increase in CO2 levels. Even allowing a 70 ppm increase for the proximity of the city gives a background level of about 370ppm, comparable to present day levels, but much higher than is generally thought to have occurred at that time.
Other sites are unlikely to have been contaminated. Lockhart and Court found CO2 levels in Antarctica between 200 and 1700 ppm, in 1940 and 1941. Hock, et al found CO2 levels averaging 400 ppm between 1947 and 1949 at Point Barrow in northern Alaska. Once again, these are much higher than the generally accepted values of that time.
So how did the pre-industrial figure come to be accepted as 290 ppm? As mentioned in Beck’s paper, Guy Callendar, a British engineer and scientist, was influential. He examined 19th and 20th century CO2 measurements and rejected those he considered inaccurate, the ones he selected leading him to conclude that the pre-industrial CO2 level was about 290 ppm (G. S. Callendar, “The Composition of the Atmosphere through the Ages,” The Meteorological Magazine,vol. 74, No. 878, March 1939, pp. 33-39.). Callendar was a proponent of the theory that CO2 emissions from industrial activity would raise global temperatures, and had written a paper to that effect in 1938, at a time when Europe had just experienced five warm years.
Among the criteria that Callendar used to reject measurements, were any that deviated by 10% or more from the average of the region, and any taken for special purposes such as such as “biological, soil air, atmospheric pollution”. The first criteria is a rather circular argument, while the second seems to ignore the accuracy of the results. Whatever the validity of these exclusions, it turned out that the mean of 19th century samples he included was 292 ppm. The mean of the samples he had available to include was 335 ppm.
Not everyone agreed with Callendar. Giles Slocum pointed out in 1955 that Callendar’s exclusions from the 19th century data were mostly higher than the ones he included, while those from the 20th century that he excluded were lower than the ones he included, in line with his theory that CO2 levels had risen and were causing increased temperatures. As Slocum diplomatically put it ” Much seems to depend on the objectivity of Callendar’s decisions as to which data to keep.”
The other official source of pre-industrial CO2 levels is, of course, ice core readings. Not everyone is happy with those either, as I will show in a later post.
Beck shows, in figure 14, that CO2 levels and temperature are correlated, if the historical CO2 measurements are used instead of the IPCC approved figures. This figure also shows that the chemical measurement of CO2, which ended about 1957, matches well with the Mauna Loa measurements, which began in 1958, with readings of about 315 ppm.
So were pre-industrial CO2 levels stable until Industrial Man disturbed the balance, or has there always been an ebb and flow? Beck’s paper certainly raises some interesting questions.
Adelaide, South Australia