Some statistical characteristics of El Niño/southern oscillation and North Atlantic oscillation indices

Abstract

The interannual variations of the El Niño/Southern Oscillation phenomenon (ENSO) and the North Atlantic Oscillation (NAO) are statistically analysed in the time and frequency domain using time series of the following monthly indices: Two Pacific sea surface temperature indices reflecting the El Niño events (1870-1983) and two sea level pressure indices monitoring the Southern Oscillation (1882-1987) and the NAO (1881-1984). The variance spectrum analysis of these time series (with a high spectral resolution) in an "integrated" (conventional) and a "dynamic" (moving with time) way is based on both the autocovariance and the maximum entropy method. The "integrated" variance spectra reveal significant peaks at periods of 2.3, 2.9, 3.5 and 6 years for the ENSO indices and of 1.7, 2.2 and 7.5 years for the NAO index. Some of these spectral features appear to be quasi-stationary in the "dynamic" view. The statistical results are helpful in identifying and interpreting physical mechanisms associated with these large-scale oceanic/ atmospheric circulation patterns, particularly with respect to the important role of oceanic wave dynamics in the ENSO cycle.