COMPARISON OF MODELLED AND PROXY ESTIMATES OF SEA SURFACE TEMPERATURES DURING THE LAST GLACIAL MAXIMUM, DERIVED FROM UK37' AND COUPLED OCEAN-ATMOSPHERE GCM

One of the most stringent tests of the performance of a climate model is to use it to simulate the climate of the Last Glacial Maximum (LGM). Environmental conditions during that period were extreme, with large differences in the state of climate forcings in comparison to present. Hence, if a model is shown to be able to reproduce the climate of the LGM, its reliability to predict the future is strengthened. However, LGM conditions are not as well known as required to conduct unambiguous tests. This is not because of a shortage of studies, but because of the apparent conflicting views that arise from studies in a given area derived using different approaches. Therefore, concerted, international, efforts are required to summarise and reconcile results from multiple sources, which are also accompanied by an appraisal of the intrinsic uncertainties of the reconstructions. A pioneering effort in this respect was carried out by the CLIMAP project. But over the years some elements of the CLIMAP reconstruction have been questioned as new developments in modelling and proxy data have appeared, although the “bulk” of the reconstruction remains unchallenged.

In this communication, global data from one of such latter approaches, the UK37' index to estimate sea surface temperature, are discussed and compared to CLIMAP and general circulation model (GCM) data. The latter is obtained from PMIP [Joussaumme, 1995] and from some more recent modelling experiments, obtained from high resolution fully coupled atmosphere and ocean GCM that does not require flux adjustments (HadCM3) [Hewitt et al., 2002]. The comparison clearly shows that both models and palaeodata based reconstructions generally converge into a unified view of sea surface temperature conditions during the LGM. Hence, models seem to reflect accurately the primary drivers and forcings of glacial climates. Some notable differences are observed, however, that illustrate the lack of comprehensive understanding of the precise meaning of the proxy data, and the translation of global climate conditions into local/regional expressions. Comparison of model and proxy data, allows us to identify and improve our understanding of the shortcomings of the different approaches, and to develop strategies to address present unresolved uncertainties.