ALBERT

Description: The gradual cooling of the climate during the Cenozoic has generally been attributed to a decrease in CO2 concentration in the atmosphere. The lack of transient climate models and in particular the lack of high-resolution proxy records of CO2, beyond the ice-core record prohibit however a full understanding of for example the inception of the Northern Hemisphere glaciation and mid-Pleistocene transition. Here we elaborate on an inverse modelling technique to reconstruct a continuous CO2 series over the past 20 million year (Myr), by decomposing the global deep-sea benthic d18O record into a mutually consistent temperature and sea level record, using a set of 1-D models of the major Northern and Southern Hemisphere ice sheets. We subsequently compared the modelled temperature record with ice core and proxy-derived CO2 data to create a continuous CO2 reconstruction over the past 20 Myr. Results show a gradual decline from 450 ppmv around 15 Myr ago to 225 ppmv for mean conditions of the glacial-interglacial cycles of the last 1 Myr, coinciding with a gradual cooling of the global surface temperature of 10 K. Between 13 to 3 Myr ago there is no long-term sea level variation caused by ice-volume changes. We find no evidence for a change in the long-term relation between temperature change and CO2, other than the effect following from the saturation of the absorption bands for CO2. The reconstructed CO2 record shows that the Northern Hemisphere glaciation starts once the long-term average CO2 concentration drops below 265 ppmv after a period of strong decrease in CO2. Finally, only a small long-term decline of 23 ppmv is found during the mid-Pleistocene transition, constraining theories on this major transition in the climate system. The approach is not accurate enough to revise current ideas about climate sensitivity.

Description: On a million-year time scale the global carbon cycle and atmospheric CO2 are assumed to be largely determined by the so-called solid Earth processes weathering, sedimentation, and volcanic outgassing. However, it is not clear how much of the observed dynamics in the proxy data constraining the carbon cycle over the Cenozoic might be determined by internal processes of the atmosphere-ocean-biosphere subsystem. Here, we apply for the first time a process-based model of the global carbon cycle in transient simulations over the last 20 Myr to identify the contributions of terrestrial carbon storage, solubility pump and ocean gateways on changes in atmospheric CO2 and marine δ13C. We apply the isotopic carbon cycle box model BICYCLE, which consists of atmosphere, terrestrial biosphere and ocean reservoirs, the latter containing the full marine carbonate system. Our simulation results show that the long-term cooling since the Mid Miocene Climatic Optimum (about 15 Myr BP) leads to an intensification of the solubility pump, and a drop in atmospheric CO2 of up to 100 ppmv. This oceanic carbon uptake is largely counterbalanced by carbon loss from the terrestrial biosphere. The reduction in terrestrial C storage over time including the expansion of C4 grasses during the last 8 Myr might explain half of the long-term decline in deep ocean δ13C and would support high CO2 (400 to 450 ppmv) around 15 Myr BP. The closure of the Tethys and the Central America ocean gateways explains the developing gradient in deep ocean δ13C between the Atlantic and Pacific basin. We furthermore calculate the residuals, which are unexplained by our results and are therefore caused by solid Earth processes. From the residuals ocean alkalinity rising over time is detected as the main reason for declining atmospheric CO2 which led to Earth’s long-term cooling observed since the Mid Miocene Climate Optimum. A combination of two processes — a reduction in volcanic out-gassing of CO2 together with increasing continental weathering rates — might explain the rising alkalinity pattern. The reduced volcanic activity probably caused by shrinking seafloor spreading rates started around 16 Myr BP is connected with a prominent regime shift in the carbon cycle-climate system. The existence of such a regime shift is confirmed if we extend our analysis to deep ocean records of δ18O and δ13C over the whole Cenozoic.