Author:
K. W. Smith ;
A. L. Aretxabaleta

Abstract:
Expectation maximization (EM) was used to estimate the parameters of a Gaussian mixture model spatial data series. The method is presented as an alternative and complement to Empirical Orthogonal Function (EOF) analysis. The weight, combining points with time component distributions are used to distinguish the physical regimes. The method is applied to data equatorial Pacific surface temperature of the sea from the TAO / TRITON mooring series. Indeed, the EM algorithm partitions of the time series in El Niño, La Niña and normal conditions. The EM method leads to a clear interpretation of the variability associated with each basic scheme of analysis EOF.

Journal: Nonlinear Processes in Geophysics

Technorati : El Niño, La Niña