No document available.
Abstract :
[en] The atmospheric CO2 mixing ratio has fluctuated widely over the Phanerozoic, according to the estimates from available proxy records. Because atmospheric CO2 is a major greenhouse gas, these fluctuations should have led to significant climatic variations. The "classical" view is indeed that atmospheric CO2 has been the main driver of the Earth's climate history. On long-term time scales, the atmospheric CO2 level is the result of the balance between CO2 inputs from volcanoes or oxidation of old organic carbon (kerogen) in exposed rocks and outputs through silicate weathering or organic carbon deposition. Existing model reconstructions of the Phanerozoic history of atmospheric CO2 are based on such budgets. Recent data and model experiments currently challenge these models. First, the carbon cycle may be more complex than represented in the earliest models. In particular, silicate weathering depends on numerous factors, which are not obvious to model or are poorly known over the Phanerozoic. Mountain uplift is one such factor, which has been much debated in the last decade. Lithology is another example: basalts weather much more rapidly than other silicate rocks and the emplacement of large basaltic areas on the continents may trigger glaciations. Continental configuration is also more important than previously thought, as indicated by recent model experiments on super-continent fragmentation coupling geochemical and climate models. Problems of "classical" Phanerozoic CO2 models are also well illustrated by the fact that the most recent estimates of CO2 degassing show very little variation between the Cretaceous and the present, a period when large changes in CO2 have occurred, whereas degassing is the most important forcing of CO2 evolution in long-term carbon cycle models. Second, CO2 is not the only driver of climate evolution. This obvious fact has largely been forgotten in Phanerozoic studies. What the proxies tell us on paleo-atmospheric CO2 is not always in line with what we know about paleoclimatic records. For instance, the proxies suggest relatively high CO2 levels during the Late Ordovician glaciations. Similarly, the Late Jurassic now appears to be colder than earlier thought, while again proxies suggest high atmospheric CO2 at that time. The mid-Miocene climate warming, which occurs simultaneously with a drop in CO2, provides another example. This latter change in CO2 is unanimously reflected in all proxies and, so, this decoupling between CO2 and climate cannot arise from uncertainties on the reconstructed CO2 levels or from dating problems, as might be the case of the former two examples. Other climatic drivers than CO2 clearly need to be considered. In this respect, vegetation- climate feedbacks have been completely disregarded in long-term climatic studies. Cenozoic cooling is, however, accompanied by a progressive transition from closed forests to more widespread grasslands and deserts on the continental areas, a change which must have had major impacts on the surface albedo and the water cycle.