Despite the obvious relationship between atmospheric CO2 and global temperature over the course of glacial-interglacial cycles (Figure 5.12), the precise role of CO2 in climate regulation remains uncertain. Specifically, it is not clear whether changes in CO2 were the effect of fluctuations in climate, which then acted as a feedback factor, or whether they were the primary cause of temperature variations, albeit driven by Milankovitch-type insolation variations. Clearly, changes in ocean circulation, productivity and chemistry all played a significant role in the events, particularly during glacial-interglacial transitions. Nevertheless, existing data from ice core does not allow the resolution of the phase relationship between CO2 changes and temperature variations (Siegenthaler, 1988). The cause-and-effect relationships operating within the internal dynamics of the climate system over such time scales of millennia and tens of millennia must therefore remain unsettled.
Figure 5.16 schematises a possible scenario of events associated with the last glacial (Würm) episode. Changes in insolation at high latitudes of the Northern Hemisphere, steered by Milankovitch-type variations, acted as a global pacemaker, driving or regulating the complex non-linear internal variations within the climate system. Such internal variations may include the growth and decay of ice sheets, changes in vertical ocean circulation and productivity, and CO2 feedback.
According to this view, CO2 changed in response to oceanographic events in the North Atlantic, induced by the Northern Hemisphere cooling (due to external insolation changes) and ice growth (due to ice-albedo feedback) at the beginning of the last glacial episode, and then acted as a climatic feedback. The time delay between climate and CO2 may have been rather small; the chemistry of the ocean has been shown to adjust itself within a few centuries to changes in circulation or productivity (Siegenthaler & Wenk, 1984). Clearly, however, more higher resolution proxy data and improved modelling studies are required if the full history of the Würm glacial, and indeed the glacial-interglacial cycles of the Pleistocene, is to be unravelled.