Detecting the Nonstationary Response of ENSO to Greenhouse Warming

On the basis of the latest greenhouse warming experiment performed with the coupled general circulation model ECHAM4/OPYC it is shown that not only the climate mean but also the statistics of higher order statistical moments respond signi cantly to greenhouse warming. In particular the El Ni~ no-Southern Oscillation (ENSO) cycle obtains more energy and a tendency towards stronger cold events can be observed, relative to an overall warming trend. It is suggested that this information can be used in order to re ne climate change detection via the optimal ngerprinting strategy. An optimal spectral ngerprint is developed on the basis of linear perturbation theory of wavelet variances. In order to elucidate the potential of higher order statistical moments in the climate change detection context, the optimal spectral ngerprint technique is applied to the ECHAM4/OPYC greenhouse warming simulation. The results provide a rough estimate of the timescale over which anthropogenic changes in the ENSO statistics can be detected against the background noise. The results reveal in particular that recent observed changes of ENSO variability are consistent with the null hypothesis of natural climate variability. Furthermore, an information theoretical approach is adopted to investigate possible in uences of global warming on ENSO predictability.