Several extratropical Pacific mechanisms in both hemispheres have been proposed as precursors of the El Niño–Southern Oscillation (ENSO), but their evaluation has been largely statistical. Pegion et al. (2020, Climate Dynamics) used regularized linear regression on 1948–2015 observations to show that boreal-winter extratropical predictors relate to Central Pacific (CP) but not Eastern Pacific (EP) ENSO events at one-year lead, with a Western North Pacific (WNP) sea surface temperature (SST) dipole and the South Pacific Quadrupole (SPQ) as the dominant Northern and Southern Hemisphere predictors. The authors themselves left open whether these conclusions survive beyond their linear, observational, a-priori-predictor framework: whether nonlinear precursor–ENSO relationships carry additional information, whether predictors of Atlantic or other origin alter the picture, what the dynamical pathways (closed temperature budgets and sensitivity experiments in coupled models) connecting the precursors to CP ENSO development are, and how faithfully climate models and operational seasonal prediction systems represent the Northern and Southern Hemisphere extratropical mechanisms. Follow-up work between 2021 and 2026 has produced partial answers on each facet—mechanism-isolation experiments, machine-learning forecast studies, CMIP6 fidelity assessments, and greenhouse-warming response studies—while simultaneously reopening the statistical foundation through artefact critiques of precursor lead–lag relationships. A precise open core remains: the dynamical closure of the precursor-to-CP-ENSO pathway, the predictive and causal status of nonlinear and inter-basin predictors for ENSO diversity specifically, and the fidelity of operational prediction systems.

Background

The El Niño–Southern Oscillation (ENSO) is the dominant coupled ocean–atmosphere mode of year-to-year climate variability, originating in the tropical Pacific but affecting weather worldwide. ENSO events differ in where their maximum sea surface temperature (SST) anomalies occur, a property called ENSO diversity: Eastern Pacific (EP) events peak in the eastern equatorial Pacific cold tongue (the cold, upwelled-water region of the eastern basin), while Central Pacific (CP) events (also called El Niño Modoki or dateline events) peak near the equatorial date line and tend to have distinct teleconnections. A practical limit on ENSO forecasting is the spring predictability barrier: prediction skill drops sharply for forecasts initialized across boreal spring, roughly 6–9 months before events mature, so information available up to a year in advance is especially valuable.

Since the 2000s, a family of extratropical Pacific mechanisms has been proposed to stochastically force ENSO. In the Northern Hemisphere (NH) these include the North Pacific Oscillation (NPO; a sea-level-pressure seesaw between Hawaii and the Aleutian sector), the seasonal footprinting mechanism (SFM; winter extratropical atmospheric anomalies leave persistent SST footprints on the subtropical ocean that later couple to the tropics through the wind–evaporation–SST feedback, in which wind-speed anomalies alter evaporation, changing SST, which in turn alters the winds), the North Pacific Meridional Mode (NPMM; a subtropical coupled SST–wind variability pattern), the Victoria mode (the second-leading spatial pattern, or empirical orthogonal function, of North Pacific SST variability), and trade wind charging (TWC; anomalous subtropical trade winds that charge or discharge equatorial Pacific oceanic heat content). In the Southern Hemisphere (SH) the proposed precursors include the South Pacific Oscillation (SPO; an extratropical sea-level-pressure mode), the South Pacific Meridional Mode (SPMM), and the South Pacific Quadrupole (SPQ; a four-lobed SST anomaly pattern in the South Pacific). A further NH candidate is an east–west SST dipole in the Western North Pacific (WNP). These precursors overlap in physical space and are highly correlated in time, which makes it difficult to isolate their independent contributions; they are also correlated with ENSO itself, so contemporaneous ENSO influence must be removed before defining precursor indices.

Pegion, Selman, Larson, Furtado and Becker (2020) addressed this redundancy problem with linear regression methodologies drawn from information theory and machine learning—least absolute shrinkage and selection operator (LASSO) regression, a regularized linear method that selects sparse predictor sets, combined with Akaike information criterion model selection—applied to observational data for 1948–2015. Predictors were chosen a priori from the Pacific extratropical mechanisms listed above, with the ENSO signal linearly removed. ENSO diversity was measured with the standard CP and EP indices built from the first two principal components (time coefficients of the two leading spatial patterns of tropical Pacific SST variability), CP=(PC1+PC2)/2CP=(PC1+PC2)/\sqrt{2} and EP=(PC1−PC2)/2EP=(PC1-PC2)/\sqrt{2}. Their main findings: at one-year lead, boreal-winter (Northern Hemisphere winter, roughly December–February) extratropical precursors explain less than 10% of the variance of the EP index but about 35% of the variance of the CP index; SH winter precursors contribute to CP ENSO predictability at much longer leads than previously indicated; the dominant NH predictor of CP ENSO is the WNP SST dipole and the dominant SH predictor is the SPQ; combining NH and SH predictors explains more variance than either hemisphere alone; and CP events with strong extratropical forcing begin about one season earlier than events mainly forced from the tropics, with roughly one-third to one-half of CP events preceded a year in advance by at least twice as much extratropical as tropical forcing in a fitted-versus-residual decomposition.

The authors flagged several unresolved issues that jointly define the present problem. The analysis was restricted to linear regression, so nonlinear precursor–ENSO relationships were not captured, and low-frequency variability (decadal to multi-decadal changes, warming-related trends) was not addressed. The predictors were selected a priori from Pacific extratropical mechanisms, leaving open whether other predictors—including Atlantic-origin signals—would change the picture. The analysis was observational and statistical rather than dynamical, so the physical pathways connecting the precursors to ENSO remained only partially understood; the authors called for closed temperature-budget analyses in coupled ocean–atmosphere models, together with sensitivity experiments, for a rigorous dynamical explanation. Finally, it remained uncertain how well operational prediction systems and climate models represent the SH and NH extratropical mechanisms. These questions are pressed further by an adjacent line of work arguing that statistical precursor–ENSO lead–lag relationships can be artefacts of ENSO's own autocorrelation and of seasonality, which has already led to an active debate about whether some proposed inter-basin precursors are spurious.

Problem Statement

The overarching objective, as posed by the unresolved issues of Pegion et al. (2020), is to replace their linear, observational, a-priori-predictor account of extratropical forcing of ENSO diversity with a dynamically grounded and model-validated one. Specifically, for the Northern and Southern Hemisphere extratropical precursors identified in that study - most notably the Western North Pacific SST dipole and the South Pacific Quadrupole as the dominant predictors of Central Pacific (CP) ENSO at one-year lead - three linked questions remain open, each stated below with its two aspects (a) and (b):

(1) Physical pathways: what are the dynamical mechanisms by which these boreal-winter extratropical precursors influence the subsequent winter's CP ENSO development, and does the pathway reproduce the observed lead-lag relationships and the earlier onset of extratropically forced CP events? Its two aspects are (1a) the causal mechanism by which each boreal-winter precursor influences subsequent CP ENSO development, and (1b) whether that pathway reproduces the observed lead-lag relationships and the earlier onset of extratropically forced CP events.

(2) Predictor scope and nonlinearity: do nonlinear precursor-ENSO relationships, which linear regression cannot capture, and predictors outside the a-priori Pacific extratropical set - including Atlantic-origin signals, whose precursor status is itself contested - add predictive information or change the attribution of which mechanisms matter, specifically for CP versus Eastern Pacific (EP) ENSO diversity and across low-frequency (decadal and warming-related) background states? Its two aspects are (2a) nonlinear precursor-ENSO relationships and (2b) predictors outside the a-priori Pacific extratropical set, including the contested Atlantic-origin signals; for each of (2a) and (2b), whether it adds predictive information and whether it changes the attribution of which mechanisms matter, both for CP versus EP ENSO diversity and across the low-frequency background states.

(3) Model and operational representation: how faithfully do current climate models and operational seasonal prediction systems represent the SH and NH extratropical mechanisms and their influence on ENSO diversity, and what role do these mechanisms play in real-time prediction skill and in climate projections? Its two aspects are (3a) the representational fidelity of both climate models and operational seasonal prediction systems for the SH and NH extratropical mechanisms and their influence on ENSO diversity, and (3b) the role of these mechanisms in real-time prediction skill and in climate projections.

These three questions jointly support one objective - establishing whether extratropical forcing is a genuine, dynamically understood, and operationally represented source of long-lead predictability for ENSO diversity.

References

  1. Pegion, K., Selman, C. M., Larson, S., Furtado, J. C., & Becker, E. J. (2020). The impact of the extratropics on ENSO diversity and predictability. Climate Dynamics, 54(9-10), 4469-4484. DOI: 10.1007/s00382-020-05232-3. https://doi.org/10.1007/s00382-020-05232-3
  2. Chakravorty, S., Perez, R. C., Anderson, B. T., Larson, S. M., Giese, B. S., & Pivotti, V. (2021). Ocean dynamics are key to extratropical forcing of El Niño. Journal of Climate, 34(21), 8739-8753. DOI: 10.1175/JCLI-D-20-0933.1. https://doi.org/10.1175/jcli-d-20-0933.1
  3. Jia, F., Cai, W., Gan, B., Wu, L., & Di Lorenzo, E. (2021). Enhanced North Pacific impact on El Niño/Southern Oscillation under greenhouse warming. Nature Climate Change, 11(10), 840-847. DOI: 10.1038/s41558-021-01139-x. https://doi.org/10.1038/s41558-021-01139-x
  4. Zhang, W., Jiang, F., Stuecker, M. F., Jin, F.-F., & Timmermann, A. (2021). Spurious North Tropical Atlantic precursors to El Niño. Nature Communications, 12(1), 3096. DOI: 10.1038/s41467-021-23411-6. https://doi.org/10.1038/s41467-021-23411-6
  5. Wang, Z., Han, L., Ding, R., & Li, J. (2021). Evaluation of the performance of CMIP5 and CMIP6 models in simulating the South Pacific Quadrupole-ENSO relationship. Atmospheric and Oceanic Science Letters, 14(4), 100057. DOI: 10.1016/j.aosl.2021.100057. https://doi.org/10.1016/j.aosl.2021.100057
  6. Tseng, Y.-h., Huang, J.-h., & Chen, H.-c. (2022). Improving the predictability of two types of ENSO by the characteristics of extratropical precursors. Geophysical Research Letters, 49(3), e2021GL097190. DOI: 10.1029/2021GL097190. https://doi.org/10.1029/2021GL097190
  7. Zhao, Y., Jin, Y., Li, J., & Capotondi, A. (2022). The role of extratropical Pacific in crossing ENSO spring predictability barrier. Geophysical Research Letters, 49(15), e2022GL099488. DOI: 10.1029/2022GL099488. https://doi.org/10.1029/2022GL099488
  8. Song, W.-J., Yu, J.-Y., & Lian, T. (2023). The efficacy of tropical and extratropical predictors for long-lead El Niño-Southern Oscillation prediction: A study using a machine learning algorithm. International Journal of Climatology, 43(14), 6887-6899. DOI: 10.1002/joc.8241. https://doi.org/10.1002/joc.8241
  9. Jiang, F., Zhang, W., Jin, F.-F., Stuecker, M. F., Timmermann, A., McPhaden, M. J., Boucharel, J., & Wittenberg, A. T. (2023). Resolving the tropical Pacific/Atlantic interaction conundrum. Geophysical Research Letters, 50(13), e2023GL103777. DOI: 10.1029/2023GL103777. https://doi.org/10.1029/2023GL103777
  10. Osipov, A., & Gushchina, D. (2023). The heat budget of the tropical Pacific mixed layer during two types of El Niño based on reanalysis and global climate model data. Atmosphere, 15(1), 47. DOI: 10.3390/atmos15010047. https://doi.org/10.3390/atmos15010047
  11. Deng, L., Borhara, K., & Wang, S.-Y. (2024). Explainable AI in assessing multi-year ENSO predictability from the Western North Pacific precursor. Preprints (non-peer-reviewed). DOI: 10.20944/preprints202402.0302.v1. https://doi.org/10.20944/preprints202402.0302.v1
  12. Qin, B., Yang, Z., Mu, M., Wei, Y., Cui, Y., Fang, X., Dai, G., & Yuan, S. (2024). The first kind of predictability problem of El Niño predictions in a multivariate coupled data-driven model. Quarterly Journal of the Royal Meteorological Society, 150(765), 5452-5471. DOI: 10.1002/qj.4882. https://doi.org/10.1002/qj.4882
  13. Li, X., Ding, R., Tseng, Y.-h., Yu, J.-Y., & Xu, X. (2024). Mutual influences between ENSO and its two precursor modes in the extratropical South Pacific. Climate Dynamics, 62(5), 3109-3121. DOI: 10.1007/s00382-023-07054-5. https://doi.org/10.1007/s00382-023-07054-5
  14. Lu, Y., Feng, J., & Hu, D. (2024). CMIP6 models simulation of the connection between North/South Pacific Meridional Mode and ENSO. Journal of Oceanology and Limnology, 42(2), 439-453. DOI: 10.1007/s00343-023-3024-6. https://doi.org/10.1007/s00343-023-3024-6
  15. Tan, J., Feng, L., Jiang, H., Zhou, Q., & Huang, Y. (2024). A new ENSO statistical prediction model considering extratropical effects and its application to the prediction of the 2015/16 El Niño event. Weather and Forecasting, 39(11), 1635-1649. DOI: 10.1175/WAF-D-23-0173.1. https://doi.org/10.1175/WAF-D-23-0173.1
  16. Xu, J., Fan, H., Yang, S., Cai, Y., Collins, M., & Yu, W. (2025). Seasonality of the South Pacific Meridional Mode: Role of oceanic meridional advection feedback beyond thermodynamics dominance. Geophysical Research Letters, 52(23), e2025GL117920. DOI: 10.1029/2025GL117920. https://doi.org/10.1029/2025GL117920
  17. Jeffree, J., Maher, N., Amaya, D. J., & Dommenget, D. (2026). Adapting analogue forecasting to compare ENSO remote influences across different models and regions. Climate Dynamics, 64(7), 327. DOI: 10.1007/s00382-026-08258-1. https://doi.org/10.1007/s00382-026-08258-1
  18. Wang, Z., Zheng, Y., Song, L., Han, F., Wang, Z., Cao, X., Chen, W., & Chen, S. (2026). Diverse impact of the western North Pacific sea surface temperature anomalies on El Niño-Southern Oscillation among CMIP6 models: dependence on the background meridional sea surface temperature gradient. Climate Dynamics, 64(7), 301. DOI: 10.1007/s00382-026-08268-z. https://doi.org/10.1007/s00382-026-08268-z