the climate system, such as the one involving the increase in water vapor with increasing temperature. Climate models are often used to estimate the strength of the various feedbacks in the climate system and the overall sensitivity of the Earth’s global mean surface temperature to a prescribed forcing, such as a doubling of atmospheric carbon dioxide concentration.
Climate sensitivity can also be estimated by forcing climate models with the observed or reconstructed external forcings of the climate system over a certain time period and comparing the model response to the observed or reconstructed surface temperature during the same period. This strategy can be applied to past climatic variations on timescales ranging from a few years (in the case of a single volcanic eruption) to tens of thousands of years (as in the simulation of the Ice Ages). Modeling climate variations on the timescale of the last 2,000 years is particularly challenging because the external forcings that operate on this timescale are relatively small and are not as well known as the forcings in the above examples.
What is our current understanding of how the hemispheric mean or global mean surface temperature has varied over the last 2,000 years?
To understand the current state of the science surrounding large-scale surface temperature reconstructions, it is helpful to first review how these efforts have evolved over the last few decades. In a chapter titled “Observed Climate Variability and Change,” IPCC (1990) presented a schematic depiction, reproduced in Figure O-3, of global temperature variations extending from 1975 back to A.D. 900. The Medieval Warm Period and Little Ice Age labels that appear in the graphic refer to features in European and other regional time series that were assumed to be indicative of global mean conditions. The peak-to-peak amplitude of the temperature fluctuations was