Patterns of decadal climate variability and their impact on global rainfall

Abstract

An analysis is made of global sea surface temperature (SST) data sets over the past 110 years to determine the principal patterns of climate variability on time scales longer than ENSO, and to relate these to likely dynamical processes. Taking 5-year running means, the most recent versions of the interpolated global data from the UK Hadley Centre and the US NOAA are analysed using singular value decomposition, and described as coherent global patterns that have a physical/dynamical basis. These patterns are: Global Warming, the Pacific Decadal Oscillation, the Atlantic Multi-decadal Oscillation and the Pacific Gyre Oscillation. Each of these patterns, as represented by SST, is described in both data sets with real and complex empirical orthogonal functions, and current understanding of their dynamical basis is described. An example of the use of such patterns is given through an application to the recent trend pattern in global rainfall. Complete agreement between the two SST data sets is lacking, but both agree that major contributors to this rainfall pattern are global warming and the Pacific Decadal Oscillation, where the latter may be regarded as the low-frequency signal of ENSO.