The Indo-Pacific region is the arena of the planet's most powerful climate oscillations: El Niño, the Indian Ocean Dipole, and monsoons. However, extracting multi-decadal signals from short observational records is difficult, and classical methods such as principal component analysis require pre-filtering that can distort the physical picture. This work stands alongside studies of oceanic heat pumps, lava oceans, and dust storms on rogue planets, but it differs by focusing on the internal coupling of atmosphere and ocean in Earth's climate. The authors use a nonlinear method that can simultaneously extract modes at different timescales without pre-processing the data, reducing the risk of signal mixing.
The authors applied nonlinear Laplacian spectral analysis (NLSA) to sea surface temperature data from a 1300-year control run of the CCSM4 model and satellite observations from HadISST for 1973–2009. Unlike classical principal component analysis, NLSA works with raw data and does not require band-pass filtering, allowing separation of modes with close frequencies. For each identified mode, they constructed spatiotemporal reconstructions not only of temperature but also of near-surface winds, precipitation, and thermocline depth, using data from the MERRA and C-GLORS reanalyses. Validation included testing the stability of the modes to changes in method parameters and to data noise.
The identified El Niño modes and their combination harmonics with the annual cycle exhibit a seasonally synchronized southward shift of equatorial zonal winds followed by thermocline recovery, consistent with the mechanism that terminates El Niño and La Niña events. These same modes create teleconnections between the Pacific and Indian Oceans, forming a dipole pattern of temperature anomalies characteristic of the Indian Ocean Dipole. The tropospheric biennial oscillation modes reproduce the biennial variability of the Asian–Australian monsoon and the Walker circulation. On multi-decadal timescales, the western Pacific multidecadal mode (WPMM) significantly modulates the activity of El Niño and the biennial oscillation: in the cold phase of WPMM (negative temperature anomalies in the western tropical Pacific), the amplitude of El Niño and monsoons increases. The correlation between WPMM and decadal rainfall over Australia reaches −0.62, whereas for the classical Pacific Decadal Oscillation it is only 0.44.
The work demonstrates that nonlinear data analysis methods can uncover internal modes of multi-decadal climate variability that escape traditional linear approaches. Unlike neighboring studies devoted to individual phenomena (for example, Enceladus's heat pumps or lava oceans), here a holistic framework is proposed for describing atmosphere–ocean interactions across a wide range of scales. This paves the way for more accurate indices of climate oscillations and improved decadal forecasts.
The topic could develop toward using the identified modes to build statistical and dynamical forecast models. Promising directions include extending the analysis to other climate archives, including paleodata, and incorporating additional variables such as ocean salinity or heat content. It is also important to test the reproducibility of WPMM in other climate models and in real data over longer periods.
The results will affect research on the predictability of El Niño and monsoons, as well as understanding of decadal rainfall variability in Australia and possibly other regions influenced by Indo-Pacific climate.
Next steps include searching for similar modes in longer observational records and paleoreconstructions, and assessing their predictive value in operational systems.
The work directly addresses the unresolved problem of the origin of decadal climate variability: whether it is an internal dynamical process or a residual effect of interannual fluctuations. The results show that, at least in the CCSM4 model, there is a physically significant multi-decadal mode that modulates El Niño, supporting the hypothesis of an internal source.
In the 1300-year CCSM4 model run, the cold phase of the western Pacific mode can cause anomalously wet periods over Australia lasting nearly a century.