Analysis of Forcing Response, and Feedbacks in a Paleoclimate Modeling Experiment

摘要

It is often argued that paleoclimate studies are necessary to determine whether climate models and their predictions of future climate change can be trusted. An overall measure of the sensitivity of global mean surface temperature to a given radiative perturbation is provided by the global climate sensitivity parameter. In climate model experiments, this parameter appears to be moderately independent of the cause of the perturbation, but it may differ from one model to the next by as much as a factor of three. Moreover, there are some scientists who claim that all models are much more sensitive than the climate system itself. Thus it would be valuable to determine which models (if any) are consistent with the paleoclimate record and what factors are responsible for model differences in sensitivity. In an analysis of the Paleoclimate Modeling Intercomparison Project (PMIP) simulations of the Last Glacial Maximum (LGM) of 21,000 years ago, we have calculated how the 'forcing' and feedbacks determine the climatic response. In the PMIP context, the ice sheet distribution is prescribed and the resulting increase in planetary albedo is the most important 'forcing' factor. Also important are radiation perturbations induced by changes in atmospheric CO2 concentration. Here we describe a new, approximate method for estimating the strength of forcing and feedback factors from commonly archived model output. We also summarize preliminary results from the PMIP experiment, which show that differences in forcing and to a lesser extent differences in feedbacks can explain differences in surface temperature response.