Space-time decomposition of global Sea Surface Temperature variability using Multichannel Empirical Orthogonal Teleconnection analysis
Abstract
With earth observation data, one of the primary concerns is the discovery of recurrent patterns over time. For example, the ENSO phenomenon is a major climatological pattern of global significance. As a spatial/two-dimensional extension of Singular Spectrum Analysis (SSA), Multichannel Singular Spectrum Analysis (MSSA) seeks to uncover the temporal evolution of recurrent space-time patterns within a specified time frame (known as the embedding dimension) by a method of spectral decomposition equivalent to Extended Principal Components Analysis. However, it suffers from the same limitations as PCA with regard to the propensity to develop components that are mixtures of multiple dominant patterns. In this paper we introduce a novel procedure we call Multichannel Empirical Orthogonal Teleconnection (MEOT) analysis as a simple extension of the logic of Empirical Orthogonal Teleconnections (EOT). A global sea surface temperature dataset spanning the 1982-2007 time period is utilized to explore the similarities and differences between MSSA and MEOT. The techniques are applied with a 13 month embedding dimension to extract spatio-temporal patterns that exhibit clear basis vectors in quadrature. Findings indicate that MEOT is capable of detecting more patterns in quadrature than MSSA. MEOT identifies three climate events as quadratures corresponding to the El Niño Southern Oscillation (ENSO), the Atlantic Meridional Mode (AMM) and the Atlantic Niño/ Tropical Southern Atlantic (TSA) mode. All of these climate events have phase change within a year. MSSA in contrast, only identified the ENSO event. Moreover, since MEOT does not suffer from a bi-orthogonality constraint, it is capable of extracting fewer mixed modes of variability than MSSA. Thus, results suggest a better identification and representation of individual climate events by the MEOT method.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2011
- Bibcode:
- 2011AGUFMGC23B0953P
- Keywords:
-
- 1616 GLOBAL CHANGE / Climate variability;
- 3305 ATMOSPHERIC PROCESSES / Climate change and variability;
- 4215 OCEANOGRAPHY: GENERAL / Climate and interannual variability;
- 4513 OCEANOGRAPHY: PHYSICAL / Decadal ocean variability