The sea level anomaly (SLA) gridded time series obtained from TOPEX/Poseidon and Jason-1 satellite altimetry are available from 10.01.1993 to 14.07.2003 with sampling interval of approximately 10 days and spatial resolution of 1° ×1°. It is widely known that the SLA data limited to the equatorial Pacific comprise the El Niño/Southern Oscillation (ENSO) signal. The potential departures from the normal distribution of the SLA data may be due to the ENSO impact and can significantly increase the errors of sea level change predictions in this region.
The objective of this paper is to check if the SLA gridded data for the global ocean being monitored by TOPEX/Poseidon and Jason-1 satellite altimetry follow the normal distribution and to link possible departures from this distribution for the equatorial Pacific with the ENSO events. In order to evaluate this we produced the maps demonstrating the spatial variability of the SLA statistical properties, i.e. trend, standard deviation, skew and kurtosis. In addition, the Shapiro-Wilk test for normality was applied. The assessments were made using the de-trended gridded SLA data, after subtracting or not the model of 365-, 182 and 62- day oscillations. The trend and seasonal oscillations were determined using the least-squares fit. The SLA data in many parts of the global ocean were found to depart from the normal distribution. In particular, the strong departures were distinguished in two zones in the eastern and central-western equatorial Pacific Ocean. The considerable skew in the eastern equatorial Pacific is probably linked to the predominance of El Niño’s signal amplitude over La Niña’s signal amplitude in the SLA data. The low values of skew for the central-western equatorial Pacific may suggest the opposite relation. The high values of kurtosis are interpreted as evidence of high intermittency of the ENSO signal in the SLA time series in the equatorial Pacific. High values of skew and kurtosis explain possible causes of the SLA prediction errors obtained elsewhere using linear autoregressive technique. The non-linear models are recommended for forecasting sea level change in the equatorial Pacific.