Concerns about impending, indeed actual, anthropogenic climate change brought on by rapidly rising levels of the greenhouse gas carbon dioxide have spurred efforts to quantify climates of the distant past. Beyond the CO2 record of the last 800 ka established from air bubbles trapped in glacial ice palaeoclimate researchers have had to depend on a range of proxies for the greenhouse effect. Those based on models linking plate tectonic and volcanic CO2 emissions with geological records of the burial of organic matter, weathering and limestone accumulation are imprecise in the extreme, although they hint at considerable variation during the Phanerozoic. Other proxies give a better idea of the past abundance of the main greenhouse gas, one using the curious openings or stomata in leaves that allow gases to pass to and fro between plant cells and the atmosphere. Well preserved fossil leaves show stomata nicely back to about 400 Ma ago when plants first colonised the land.
Stomata draw in CO2 so that it can be combined with water during photosynthesis to form carbohydrate. So the number of stomata per unit area of a leaf surface is expected to increase with lowering of atmospheric CO2 and vice versa. This has been observed in plants grown in different air compositions. By comparing stomatal density in fossilised leaves of modern plants back to 800 ka allows the change to be calibrated against the ice-core record. Extending this method through the Cenozoic, the Mesozoic and into the Upper Palaeozoic faces the problems of using fossils of long-extinct plant leaves. This is compounded by plants’ exhalation of gases to the atmosphere – some CO2 together with other products of photosynthesis, oxygen and water vapour. Increasing stomatal density when carbon dioxide is at low concentration risks dehydration. How extinct plant groups coped with this problem is, unsurprisingly, unknown. So past estimates of the composition of the air become increasingly reliant on informed guesswork rather than proper calibration. The outcome is that results from the distant past tend to show very large ranges of CO2 values at any particular time.
An improvement was suggested some years back by Peter Franks of the University of Sydney with Australian, US and British co-workers (Franks, P.J. et al. 2014. New constraints on atmospheric CO2 concentration for the Phanerozoic. Geophysical Research Letters, v. 41, p. 4685-4694; doi:10.1002/2014GL060457). Their method included a means of assessing the back and forth exchange of leaf gases with the atmosphere from measurements of the carbon isotopes in preserved organic carbon in the fossil leaves, and combined this with stomatal density and the actual shape of stomata. Not only did this narrow the range of variation in atmospheric CO2 results for times past, but the mean values were dramatically lessened. Rather than values ranging up to 2000 to 3000 parts per million (~ 10 times the pre-industrial value) in the Devonian and the late-Triassic and early-Jurassic, the gas-exchange method does not rise above 1000 ppm in the Phanerozoic.
The upshot of these findings strongly suggests that the Earth’s climate sensitivity to atmospheric CO2 (the amount of global climatic warming for a doubling of pre-industrial CO2 concentration) may be greater than previously thought; around 4° rather than the currently accepted 3°C. If this proves to be correct it forebodes a much higher global temperature than present estimates by the Intergovernmental Panel on Climate Change (IPCC) for various emission scenarios through the 21st century.
See also: Hand, E. 2017. Fossil leaves bear witness to ancient carbon dioxide levels. Science, v. 355, p. 14-15; DOI: 10.1126/science.355.6320.14.