HR: 1400h
AN: PP43C-16    [Abstracts]
TI: Reconstructing Tropical Sea Surface Temperatures Using Precipitation Proxy Records: Methods and Uncertainties
AU: * Furtado, J C
EM: jason.furtado@eas.gatech.edu
AF: Georgia Institute of Technology, Department of Earth and Atmospheric Sciences 311 Ferst Drive, Atlanta, GA 30332, United States
AU: Di Lorenzo, E
EM: edl@gatech.edu
AF: Georgia Institute of Technology, Department of Earth and Atmospheric Sciences 311 Ferst Drive, Atlanta, GA 30332, United States
AU: Cobb, K M
EM: kcobb@eas.gatech.edu
AF: Georgia Institute of Technology, Department of Earth and Atmospheric Sciences 311 Ferst Drive, Atlanta, GA 30332, United States
AB: Reconstruction of historical tropical sea surface temperatures (SSTs) over the last millennia and beyond is an active venture in current paleoclimate research. This study explores the potential of using tropical precipitation data from paleoclimate proxy locations to reconstruct tropical ocean SSTs. The goal is to quantify the range of uncertainties in the SST reconstruction and their dependence on both errors in the proxy data and distribution of the proxy network. Two reconstruction methods are presented and applied to observational and modeled datasets over the period 1950-2000. The first reconstruction method exploits the high correlation between the leading mode of variability of precipitation and SST, which corresponds to ENSO and accounts for more than 50% of the total covariance. The second method uses multiple modes of covariability between precipitation and SST. We find that in addition to the ENSO mode, the second mode of covariability (about 20% of the total covariance) captures variations in the spatial expression of ENSO. This second mode has a large-scale spatial signature and is robust in all the precipitation datasets used in this study. Inclusion of this mode in the linear model leads to an approximate 30% improvement in the overall reconstruction skill. Uncertainty estimates computed from the linear model are consistent with those obtained using Monte Carlo simulations. Given the few degrees of freedom in the covariability modes, we find that a relatively sparse network of individual precipitation time series captures the phase of the modes and leads to high SST reconstruction skills in the tropical Pacific and Indian Ocean.
DE: 1620 Climate dynamics (0429, 3309)
DE: 3260 Inverse theory
DE: 3354 Precipitation (1854)
DE: 4522 ENSO (4922)
DE: 4999 General or miscellaneous
SC: Paleoceanography and Paleoclimatology [PP]
MN: 2007 Joint Assembly