The El Nino/Southern Oscillation (ENSO) phenomenon is the prominent mechanism of year-to-year climate variability in the tropical Pacific with far-reaching impacts on weather, ecosystems, and the economy worldwide.

ENSO activity varies at multidecadal-to-centennial time scales as evident in the instrumental record, paleoclimate proxy records, and long simulations with global climate models. Some periods are characterized by strong El Nino and La Nina events, while others by muted ENSO activity. Such activity fluctuations are associated with changes in the mean climate state of the tropical Pacific, arising either from external climate forcing (e.g., from greenhouse-gas forcing or volcanic eruptions) or from internally-generated climate variability. Most research on determining the causal links between mean state changes and ENSO activity has focused on the impact of the former on that latter; however, ENSO activity itself can modulate the mean state: In particular, periods of increased ENSO activity may decrease the zonal sea surface temperature gradient in the tropical Pacific due to residual heating from large El Nino events (referred to as “ENSO rectification on the mean state”). This project studied the causal links between ENSO activity and the mean state using a combination of climate model simulations, paleoclimate proxy records and theoretical ENSO models.

The study of climate model simulations and paleoclimate proxy records of the last millennium revealed an alternating sign of the correlation between ENSO variance and the zonal temperature gradient in the equatorial Pacific -an indicator of its mean state- at multidecadal-to-centennial time scales. This work showed that the assumption of a consistent multidecadal ENSO-mean relationship, either in direction or strength, is not valid. However, the greater frequency of a negative ENSO-zonal gradient correlation during the last millennium in the proxy data and in models with better skill in simulating ENSO indicate that the hypothesized mechanisms that produce such a negative relationship may be at work more frequently (e.g., the aforementioned ENSO rectification mechanism).

This project developed a novel metric for characterizing ENSO nonlinearities and diversity in observations and in climate models; this metric was named “ENSO coefficient or parameter α” and has been since used in numerous high-profile studies: Sub-selecting climate models based on their performance in this metric has led to new discoveries, including that models with better simulation of α project relative less warming of the east Pacific in response to greenhouse-gas forcing. This metric is now becoming increasingly used in new studies of model skill in simulating ENSO and its diversity.

The multidecadal ENSO mean-state relationship in climate model simulations of the last millennium was found to be different than in simulations -with the same models- of preindustrial climate without external forcing. Given that volcanic forcing is a primary external climate forcing in the last millennium, this project included an investigation of the modulation of ENSO activity by volcanic eruptions. Targeted model simulations were conducted with a state-of-the-art earth system model, and showed that latitudinal shifts in the Intertropical Convergence Zone and changes in extratropical climate are the dominant mechanisms for volcanic impacts on ENSO. Depending on the initial state of the ocean at the time of the eruption, Northern Hemisphere eruptions can lead to the development of a strong El Nino event or the suppression of a developing La Nina event, and vice versa for Southern Hemisphere eruptions. These findings help explain both the predominance of posteruption El Nin?o events and the occasional posteruption La Nin?a and neutral events in observations and paleoclimate reconstructions.

Persistent multidecadal shifts in ENSO activity may be related to so-called “regime-behavior” in ENSO. Whether the different ENSO flavors and their relative frequency are a result of the existence of different ENSO regimes or are simply different noise-driven manifestations of a single regime remains an open question. This project approached this question via a theoretical ENSO model, which provided a parsimonious explanation for the existence of two ENSO regimes (strong nonlinear, and moderate ENSO events). The model showed that a simple threshold nonlinearity in the evolution of sea surface temperature anomalies is sufficient to produce a bimodal distribution of peak ocean temperature anomalies associated with two El Nin?o regimes, lending support to the idea of a multi-regime ENSO.

This project also created new educational material for training a new generation of interdisciplinary paleoclimate scientists, versed in both data development and climate modeling. A new course was developed and taught jointly via teleconferencing at the two collaborating institutions, the University of Hawaii at M?noa and the University of Illinois at Urbana-Champaign. The course's online teaching and active-learning configuration provided a basis for continuing remote learning during COVID-19 pandemic conditions.

This project was interdisciplinary in nature, trained two tenure-track professors and three students from underrepresented groups in STEM, and was conducted in a minority-serving institution. The PIs were actively engaged in numerous outreach activities with local and international media and K-12 institutions.

Last Modified: 09/28/2020
Modified by: Christina Karamperidou