Considerable progress has been made since the TAR in understanding the response of the climate system to external forcings. Periods like the mid-Holocene and the LGM are now used as benchmarks for climate models that are used to simulate future climate (Chapter 6). While considerable uncertainties remain in the climate reconstructions for these periods, and in the boundary conditions used to force climate models, comparisons between simulated and reconstructed conditions in the LGM and mid-Holocene demonstrate that models capture the broad features of changes in the temperature and precipitation patterns. These studies have also increased understanding of the roles of ocean and vegetation feedbacks in determining the response to solar and greenhouse gas forcing. Moreover, although proxy data on palaeoclimatic interannual to multi-decadal variability during these periods remain very uncertain, there is an increased appreciation that external forcing may, in the past, have affected climatic variability such as that associated with ENSO.
The understanding of climate variability and change, and its causes during the past 1 kyr, has also improved since the TAR (IPCC, 2001). There is consensus across all millennial reconstructions on the timing of major climatic events, although their magnitude remains somewhat uncertain. Nonetheless, the collection of reconstructions from palaeodata, which is larger and more closely scrutinised than that available for the TAR, indicates that it is likely that NH average temperatures during the second half of the 20th century were warmer than any other 50-year period during the past 1.3 kyr (Chapter 6). While uncertainties remain in temperature and forcing reconstructions, and in the models used to estimate the responses to external forcings, the available detection studies, modelling and other evidence support the conclusion that volcanic and possibly solar forcings have very likely affected NH mean temperature over the past millennium and that external influences explain a substantial fraction of inter-decadal temperature variability in the past. The available evidence also indicates that natural forcing may have influenced the climatic conditions of individual periods, such as the cooler conditions around 1700. The climate response to greenhouse gas increases can be detected in a range of proxy reconstructions by the end of the records.
When driven with estimates of external forcing for the last millennium, AOGCMs simulate changes in hemispheric mean temperature that are in broad agreement with proxy reconstructions (given their uncertainties), increasing confidence in the forcing reconstructions, proxy climate reconstructions and models. In addition, the residual variability in the proxy climate reconstructions that is not explained by forcing is broadly consistent with AOGCM-simulated internal variability. Overall, the information on temperature change over the last millennium is broadly consistent with the understanding of climate change in the instrumental era.