A groundbreaking study has unveiled a remarkable evolution in the behavior of the El Niño–Southern Oscillation (ENSO) over the past 7,000 years, revealing an increased frequency of multi-year ENSO events throughout the Holocene epoch. By synthesizing detailed proxy data derived from fossil coral oxygen isotope records in the equatorial central Pacific, researchers have reconstructed a high-resolution picture of ENSO dynamics that challenges previously held assumptions about the stability and variability of this pivotal climate phenomenon.
At the heart of this research lies the comprehensive analysis of coral proxy data from Kiritimati and Fanning Atolls, strategically located within or near the Niño 3.4 region—an ENSO-active area spanning from 5° S to 5° N and 170° W to 120° W. These coral records collectively span over 1,200 years and chronicle oceanographic conditions over nearly the entire Holocene, approximately 7,000 years. Notably, only coral slices longer than 20 years were considered, ensuring robust and reliable datasets with an average length of close to six decades. The oxygen isotopic composition (δ¹⁸O) of coral skeleton carbonate serves as a nuanced proxy, simultaneously reflecting sea surface temperature (SST) and seawater salinity, the latter closely linked to oxygen isotopic variations.
Crucially, coral δ¹⁸O anomalies align well with ENSO phases: El Niño episodes correspond to warmer, wetter conditions yielding negative δ¹⁸O anomalies, while La Niña events coincide with cooler, drier conditions leading to positive δ¹⁸O anomalies. This relationship, well established in modern coral records, grants researchers unprecedented insight into ENSO variability over millennial timescales. However, other fossil proxy records like bivalve δ¹⁸O from the equatorial eastern Pacific, despite offering monthly resolution, were not included due to their shorter duration.
Complementing the paleo-proxy reconstructions, the study leveraged advanced transient climate simulations contributed by eight international modeling groups. These simulations extend from early to mid-Holocene boundary conditions (spanning approximately 6,000 to 10,000 years ago) and incorporate continuous orbital forcing changes primarily driven by insolation variations. By regridding the monthly climate outputs onto consistent spatial scales, the researchers obtained monthly SST and surface air temperature anomaly data crucial for identifying ENSO characteristics in model environments.
A key performance metric employed was the ratio of multi-year to single-year ENSO events (rENSO), calculated within the Niño 3.4 region’s SST anomalies. Comparing model outputs to the proxy-informed coral data and extended instrumental sea surface temperature datasets revealed that only four models accurately captured the observed multi-year event ratios during the last millennium. These four—EC-Earth3-LR, IPSL, MPI-ESM, and LOVECLIM1.3—were selected for further ensemble analyses, as other models failed to reproduce realistic ENSO periodicity, instead exhibiting overly regular or quasi-biannual cycles inconsistent with observations.
Despite these advances, the models—like previous generation climate simulations—tend to underestimate the prevalence of multi-year ENSO occurrences, suggesting persistent shortcomings rooted in factors such as background climatology biases, inaccurate feedback process representations, and limited spatial resolution. These challenges underscore the need for sustained model improvements to deepen understanding of ENSO’s intricate dynamics.
Methodologically, defining multi-year ENSO events across heterogeneous datasets presented a challenge due to variable ENSO amplitudes simulated by models and recorded in proxy archives. The study adopted a relative threshold approach, using 0.5 standard deviations of the Niño 3.4 SST anomalies within moving 100-year windows to identify events. This dynamic threshold accommodates changes in ENSO variability intensity over time. ENSO events were then classified based on whether the anomalies persisted above or below this threshold across two consecutive years during the ENSO season, centered on the peak month, typically December. This approach simplifies longer events into overlapping two-year intervals, capturing sustained ENSO phases effectively.
Spectral analysis was applied to both proxy and model time series to characterize ENSO periodicity and amplitude robustly. By filtering out variability outside the scientifically justified ENSO band of 1.5 to 7 years, the study focused on dominant interannual fluctuations. Rather than merely identifying the period of maximum spectral power, it computed weighted average periods (centroids) within this range, acknowledging the irregular and multimodal nature of ENSO oscillations.
To evaluate the significance of long-term trends in multi-year ENSO frequency and period shifts, the non-parametric Mann–Kendall statistical test was employed. This rank-based method is particularly suitable for climate time series, which often deviate from normal distributions. The test rejected the null hypothesis of trend absence at conventional significance levels, underscoring meaningful changes in ENSO behavior over the Holocene.
Beyond empirical reconstructions and climate model analyses, the investigators leveraged the conceptual Recharge Oscillator (RO) model to isolate the role of thermocline depth dynamics—specifically, the depth of the zonal mean equatorial thermocline—in modulating ENSO periodicity. The RO model simplifies coupled ocean-atmosphere interactions to a set of nonlinear differential equations capturing the evolution of SST anomalies and thermocline depth. Crucially, in its classical form, the model does not accommodate multi-year ENSO events but is adept at reflecting changes in oscillation periods driven by oceanic adjustment processes.
The researchers enhanced the RO model by introducing a parameter (λ) representing perturbations in mean thermocline depth, effectively decoupling or modulating the SST dynamics from subsurface ocean processes. Sensitivity experiments demonstrated that varying λ significantly influences ENSO periodicity: when λ approaches 1, SST dynamics become disengaged from thermocline depth anomalies, leading to distinct oscillation characteristics. The robustness of this relationship persists even when accounting for seasonal cycle effects on λ, shedding light on how thermocline depth variations likely governed past ENSO frequency changes.
This multimethod analysis revealed a consistent Holocene trend toward increasing multi-year ENSO events, linked mechanistically to a gradual deepening and modulation of the tropical Pacific thermocline. Such changes could be attributed to evolving orbital forcing patterns affecting solar insolation distribution and, consequently, ocean-atmosphere coupling strength. These findings not only elucidate past climate variability but also hold implications for understanding ENSO behavior under future climate change scenarios.
This study represents a significant leap forward in paleoclimatology and climate modeling integration, merging nuanced proxy reconstructions with sophisticated simulations and theoretical frameworks. It underscores the importance of multi-year ENSO events as a fundamental aspect of climate variability that has been underappreciated in long-term assessments. The improved understanding of ENSO frequency modulation may enhance predictive models and deepen our grasp of tropical climate dynamics under shifting boundary conditions.
The intricate approach combining multiple coral records and selective climate models illustrates the necessity of multi-disciplinary collaboration to unravel complex climate phenomena. By pinpointing both the empirical trends and underpinning physical mechanisms, this research offers a comprehensive narrative of Holocene ENSO evolution that enriches the climate science discourse.
As climate models continue to evolve and proxy records accumulate, future studies will likely refine our comprehension of ENSO’s multi-decadal to millennial variability further. This work sets a methodological precedent and demonstrates how integrating diverse data types can yield insights into climate system behaviors that single approaches might miss.
Ultimately, the enhanced knowledge of multi-year ENSO processes opens avenues for better anticipating long-lasting teleconnection impacts on global weather patterns, ecosystems, and human societies. It emphasizes the dynamic nature of ENSO and the importance of capturing its full temporal complexity to predict its future in a warming world correctly.
This research not only chronicles an ancient climate signal but also invigorates ongoing scientific efforts to decode one of Earth’s most influential and enigmatic climate oscillations.
Subject of Research: Holocene evolution of multi-year El Niño–Southern Oscillation (ENSO) events using coral proxy reconstructions, climate model simulations, and theoretical modeling.
Article Title: Increased frequency of multi-year El Niño–Southern Oscillation events across the Holocene.
Article References:
Lu, Z., Schultze, A., Carré, M. et al. Increased frequency of multi-year El Niño–Southern Oscillation events across the Holocene. Nat. Geosci. 18, 337–343 (2025). https://doi.org/10.1038/s41561-025-01670-y
Image Credits: AI Generated
