HR: 10:35h
AN: PP52A-02 INVITED     [Abstracts]
TI: The Role of Global Palaeodata Syntheses in Benchmarking Climate Model Simulations
AU: * HARRISON, S P
EM: sandy.harrison@bristol.ac.uk
AF: University of Bristol, School of Geographical Sciences, University Road, Bristol, BS8 1SS United Kingdom
AB: Palaeo-records from the ice, marine and terrestrial domains are providing an ever more detailed record of past changes in regional climates and environments during the Quaternary. International initiatives over the last decade have led to the creation of a number of global syntheses of palaeodata. The ability to reproduce the spatial patterns of regional changes shown in these syntheses is a crucial measure of the confidence we can have in the ability of climate models to simulate climates radically different from the present day climate. Comparisons of climate model simulations with palaeoenvironmental observations have been an integral part of the Palaeoclimate Modelling Intercomparison Project (PMIP). Data-model comparisons carried out in the first phase of PMIP enabled the community to pinpoint areas in which atmospheric general circulation models (AGCMs) were unable to reproduce the full magnitude of past climate changes and led to the recognition of the importance of ocean and land-surface feedbacks in modulating past climate changes. PMIP is currently running palaeoclimate simulations with fully coupled ocean-atmosphere and ocean-atmosphere-vegetation models. The available global datasets will be used used in a systematic fashion to evaluate these state-of-the-art models. A new focus will be on the simulation of interannual to interdecadal variability, and the stability of teleconnections associated with modes of variability. Thus, the exercise of coupled models in the second phase of PMIP poses a new challenge to the palaeodata community requiring the synthesis and analysis of records of short-term climate variability. Palaeoenvironmental benchmarking could also play a significant role in other initiatives to quantify the uncertainties in future climate prediction. Several groups are now investigating the range of uncertainties inherent in model parameterisations of key physical processes by running large ensembles of experiments in which parameter values are changed within plausible ranges. Initial results suggest that the suite of versions of the model that are capable of reproducing the modern climate nevertheless differ considerably in their projections of future climate. There is an urgent need to explore whether constraining these experiments to reproduce key states of the climate in the past could help to reduce this large uncertainty.
DE: 3339 Ocean/atmosphere interactions (0312, 4504)
DE: 3344 Paleoclimatology
DE: 1600 GLOBAL CHANGE (New category)
DE: 1620 Climate dynamics (3309)
DE: 0315 Biosphere/atmosphere interactions
SC: Paleoceanography and Paleoclimatology [PP]
MN: 2004 AGU Fall Meeting