Thayer Watkins
Silicon Valley
& Tornado Alley

The Early History of the Role
of Carbon Dioxide in Climate Analysis

The history of carbon dioxide and climate usually starts with the Swedish scientist, Svante Arrhenius, who in 1896 published a paper arguing that increased levels of carbon dioxide could raise global temperatures. He was primarily concerned with this as a factor that might prevent the Earth falling into another ice age. Other scientists considered Arrhenius' assertion and rejected it on the basis their belief that carbon dioxide did not absorb any wavelength of radiation that was not also absorbed by water vapor. The level of carbon dioxide in the atmosphere is so small compared to the level of water vapor that if it only duplicates the effect of water vapor its effect would be insignificant. Given the presumption that carbon dioxide did not absorb any wavelengths that were not absorbed by water vapor the conclusion that increased carbon dioxide levels could not affect the atmospheric temperatures of the Earth was inescapable. Therefore for forty to fifty years there was a scientific consensus, subscribed to by thousands of scientists, that carbon dioxide could not affect Earth's temperatures. So much for the consensus of scientists!

It was not until the late 1930's that this consensus was challenged. The British researcher, Guy Callendar, working at the Ministry of Supply in London began publishing papers in science journals on the issue. Callendar first compiled temperature statistics for different regions and concluded that there had been increases in temperature. His statistics (Callendar, 1961), shown below, are almost pathetic by modern standards, but he is thought to have made a plausible case that temperature had increased over the 1885-1950 interval of his data by something on the order of ½°C.

The data plotted in the graph are the deviations from the average for five-year moving averages. Of the five regions shown above only the British Isles, New Zealand and Turkestan showed a temperature increase.

Here is the graph of the temperatures in Central England over the same period. The temperatures at the end of the interval were declining rather than increasing.

It is also notable that the temperatures prior to the interval used by Calendar, say 1830, were significantly higher than those of the starting year of 1850 that he used. The temperature in 1960 was not any higher than the temperature in 1830.

It is notable that the temperature records for the U.S. or any region there of were not included by Callendar in his analysis.

Callendar then presented his temperature data by broad zonal regions.

Of the six regions one, the western Pacific, did not have a significant change. For two, the Ocean Islands and Africa, the change was about a quarter degree Celsius. Again it is notable that North America did not seem to have quality temperature records! (Or,more likely, they did not show what Callendar wanted them to show.)

Amazingly the temperature trends all over the world were, according to Callendar's results, remarkably the same.

There was however a greater temperature change for the sub-Arctic than for the Tropics region.

The final results are remarkably uniform and smooth.

The proponents of the carbon dioxide theory of global warming seem to be like the Canadian Royal Mounted Police who, not matter how tough the territory, always get their man. The CO2 proponents, not matter how rough the data, always seem to get the result they are looking for.

After having established the global temperature trend Callendar then examined the possible explanations for the global warming he found. The prime candidate at that time was an increase in the intensity of radiation from the Sun. This intensity had been increasing and the past correlation looked impressive, as is seen from below.

Callendar rejected this explanation. He considered the possibility of changed levels of dust in the atmosphere, such as from volcanoes, and rejected this as well. He then concluded it had to be the level of carbon dioxide. At the time Callendar was writing it was not well established that the level of carbon dioxide had increased. The concentration of carbon dioxide is so small that it was very difficult to find historical measurements of sufficient precision and accuracy to establish that there had been an increase. The graph below, for Fonselius et al(1956) shows the measurement record at the time Callendar was writing.

Callendar had to reject most of the past measurements. He was undoubtably correct but it is a dangerous proposition for a researcher to be able to select data that fits his hypothesis and reject all data that does not fit. Such a procedure can be used to prove anything.

The picture that Callendar presented (Callendar, 1958) was much more orderly than the full record:

Again like a Royal Mounty, Callendar got the result he was looking for.

Callendar did not make a rigorous case for the role of carbon dioxide in increasing global temperature. His results for the British Isles seem to have come from the careful selection of the data interval. He used data of dubious accuracy from such places as Turkestan. However about the same time that he was writing physicists were getting more accurate measurements of the absorption spectrum of carbon dioxide. They found that carbon dioxide absorbed infrared radiation in a band of wavelengths around 15 microns (micrometers) which is not absorbed by water vapor.

A Johns Hopkins University researcher, Gilbert N. Plass, took this information and made calculations to show that increased levels of carbon dioxide could raise atmospheric temperature. It was this research that convinced climatologists that carbon dioxide could affect climate. Water vapor and carbon dioxide are both greenhouse gases but the crucial thing is that they are not exactly equivalent. Carbon dioxide does not just increase the saturation of the absorption by water vapor in the wavelength bands where their spectra overlap; it blocks a window in the atmosphere which is not blocked by water vapor. There needs to be a more refined terminology. Labeling water vapor and carbon dioxide as merely greenhouse gases obscures the issue. Carbon dioxide is a greenhouse gas with a difference.

There are places on Earth where an increase in CO2 would have a large impact; i.e., those places with very little water vapor in the atmosphere. These are roughly speaking the desert areas but this is not precise enough. Deserts are defined in terms of a lack of precipitation but there are some deserts that have a degree of humidity but it not get precipitated out. The polar areas are deserts from a lack of precipitation but the air there is in contact with ice and snow and sublimation keeps a degree of humidity in the air. The deserts where an increase in CO2 would have a more significant impact are those where summer temperatures prevent any preservation of snow and ice. Such areas exists in central Siberia and northwest Canada and it is these two regions which account for an inordinate amount of the increase in the average global temperature.


Guy S. Callendar, "Temperature fluctuations and trends over the Earth," Quarterly Journal of the Royal Meteorological Society, vol. 87, no. 371 (January 1961), pp. 1-12.

Guy S. Callendar, "On the Amount of Carbon Dioxide in the Atmosphere," Tellus, vol. 10, no. 2 (1958), pp. 243-248.

Gilbert N. Plass, "The Carbon Dioxide Theory of Climate Change," Tellus, vol. 8, no. 2 (1956), pp. 140-154.

Stig Fonselius, Folke Koroleff and Karl-Erik Wärme, "Carbon Dioxide Variations in the Atmosphere," Tellus, vol. 8, no. 2 (1956), 176-183.

HOME PAGE OF applet-magic
HOME PAGE OF Thayer Watkins