Nonlinear complex principal component analysis of the tropical Pacific interannual wind variability

Abstract

Complex principal component analysis (CPCA) is a linear multivariate technique commonly applied to complex variables or 2‐dimensional vector fields such as winds or currents. A new nonlinear CPCA (NLCPCA) method has been developed via complex‐valued neural networks. NLCPCA is applied to the tropical Pacific wind field to study the interannual variability. Compared to the CPCA mode 1, the NLCPCA mode 1 is found to explain more variance and reveal the asymmetry in the wind anomalies between El Niño and La Niña states.

Artificial Intelligence Methods in the Environmental Sciences

“…The tropical Pacific wind anomalies (expressed as w = u + iv) have been analyzed by NLCPCA in Rattan and Hsieh (2004), where a comparison between the first mode of CPCA and that of NLCPCA revealed a large difference in the spatial anomaly patterns during strong El Niño episodes but a much smaller difference during strong La Niña episodes, indicating stronger nonlinearity was manifested in the El Niño side than the La Niña side of the oscillation.…”

Nonlinear principal component analysis of the tidal dynamics in a shallow sea

“…NLPCA has an important advantage over the linear techniques mentioned above, which produce linearly independent, but statistically correlated modes. For the theory of NLPCA and its applications in climate studies, the reader is referred to Hsieh and Tang [1998], Monahan [2000], Monahan et al [2001], Rattan and Hsieh [2004], Hsieh [2001, 2004b], and Rattan et al [2005]. Suitability of NLPCA to studies concerning the medium‐term variability of hydrodynamic processes in the coastal zone is demonstrated below.…”