Wavelet Transform analysis of the length of the day and the El Niño/Southern Oscillation variations at intraseasonal and interannual time scales.

Over the last decade, much progress has been made concerning the determination of the variations of the Earth Rotation Parameters and their physical origins. In particular, variations of Universal Time and of the derived length of day (LOD) are attributed to principally zonal wind motions for periods ranging from a few days up to about two years. More recently, effects due to the atmospheric-oceanic EI Nino-Southern Oscillation (ENSO) phenomena and to the Quasi-Biennal Oscillation have been detected in LOD on the interannual time-scale. The studies performed on this subject use classical statistical methods, assuming linear relationships between the variations of the atmosphere/ocean system and stationarity of the various series involved in the analyses; this is not realistic and may be misleading in the interpretation of the results. In the present analysis, a new approach based on the recent concept of the Wavelet Transform is used. This method gives a time-scale decomposition of the signal into localized contributions. It has been applied to parallel studies of LOD and the Southern Oscillation Index (SOI) over the internal 1962-1991.7 in order to explore the relationships between these series on interannual and intraseasonal time scales. The results show significant similar features between both series at low frequencdy, suggesting in particular the possible occurrence of a new El-Nino event for 1990. For the SOI, the low-frequency oscillations seem triggered by high-frequency signals. This is not the case for LOD series, for which no relationship appears between large and small scales