This study addresses the changes in the probability density function (PDF) of the 500-hPa geopotential heights during El Niño and La Niña events using the dynamical seasonal hindcasts made at the National Center for Environmental Prediction. First, two-way layout analysis of variance (ANOVA) is applied to the hindcasts to assess the effects of lead time on the simulated climatological mean and variance. Results demonstrate that there is no statistically significant effect of lead time in DJF. However, there is a significant effect of lead time on the climatological mean for JJA over northeastern China, due to the initial soil wetness. This effect on anomalies can be eliminated by using the climatological mean with different lead times.
Based on this analysis, we combined four 10-member ensembles with different lead times to increase the ensemble members to 40, which allows us to examine the statistical significance of the change in the PDF. We applied non-parametric Kolmogorov-Smirnov tests to the changes for the El Niño and La Niña events. Results indicate that the changes in the PDF are significant during these events over most of the globe and have unique features for each event, but the major factor is the change in the mean.
View full abstract