HR: 09:00h
AN: PP51G-05    [Abstracts]
TI: Long simulations with General Circulation models as a testbed for empirical climate reconstructions methods.
AU: * Zorita, E
EM: zorita @gkss.de
AF: GKSS Research Centre, Max-Planck-Strasse 1, Geesthacht, 21502 Germany
AU: Gonzalez-Rouco, F
EM: fidelgr@fis.ucm.es
AF: Universidad Complutense, Ciudad Universitaria, Madrid, 28040
AU: von Storch, H
EM: storch@gkss.de
AF: GKSS Research Centre, Max-Planck-Strasse 1, Geesthacht, 21502 Germany
AB: The climate of the past centuries simulated by three-dimensional coupled atmosphere-ocean models may deviate from the true evolution of the past climate, due to uncertainties in the external forcing, model deficiencies and internal variability. However, they may be considered as plausible, and useful, realization of the climate of the last centuries, where the skill of proxy-based empirical reconstruction methods may be tested. In this approach, sometimes called the pseudo-proxy approach, the artificial proxy-indicators are constructed by adding stochastic noise to the simulated grid-point temperature or precipitation. The output of the reconstruction method, applied to the pseudo-proxies, may be compared to the simulated target variables, allowing for an estimation of the bias and uncertainty bounds at centennial timescales. Several empirical reconstructions methods have been recently put forward in the literature. They may differ in several important aspects: their statistical methods (e.g. regression-based versus simple proxy-averaging and rescaling), in their filtering prior to calibration (interannual or decadal), in their consideration of the 20th trend (detrended or undetrended calibration), etc. In this contribution several of the recent empirical reconstruction methods are tested in the pseudo-proxy approach, analyzing the influence that white or red noise in the pseudo-proxies may have on the final result. The robustness of the results to the particular model chosen for the analysis can be estimated by using the output of different models (though to date few long simulations of the past centuries have been performed). The role of the particular forcing history chosen to drive the model can also be estimated by aplying the pseudo-proxy approach to long control simulations with the same models.
DE: 1616 Climate variability (1635, 3305, 3309, 4215, 4513)
DE: 3305 Climate change and variability (1616, 1635, 3309, 4215, 4513)
DE: 4215 Climate and interannual variability (1616, 1635, 3305, 3309, 4513)
DE: 4902 Anthropogenic effects (1803, 4802)
DE: 4928 Global climate models (1626, 3337)
SC: Paleoceanography and Paleoclimatology [PP]
MN: Fall Meeting 2005