UC Santa Cruz

The tropical Pacific has an outsized influence on global climate – it is the center of the El Niño-Southern Oscillation (ENSO), the dominant source of interannual climate variability, and the seat of the West Pacific Warm Pool (WPWP), a major source of energy to the atmosphere. The future evolution of the tropical Pacific is unclear; the WPWP’s response to increased pCO2 is not well constrained, and ENSO may either strengthen or weaken. To investigate ENSO's dependence on the mean climate, and on various positive and negative feedbacks, I collected paleo-ENSO data from time periods with different mean climate: the mid- and early Holocene, and the Pliocene. My data are based on Mg/Ca measurements of individual foraminifera from marine sediment cores from the central and eastern equatorial Pacific. By measuring many individuals in a sample, I reconstruct the distribution of temperatures. Differences in the warm “tail” of the distribution are attributable to changes in El Niño amplitude.

I found that El Niño amplitude was dampened (relative to the late Holocene) during the mid- and early Holocene and throughout the early Pliocene, whereas during the mid-Pliocene, El Niño amplitude varied on centennial and/or orbital timescales, in agreement with modeling studies. Though modeling studies agree on changes in past ENSO, they disagree on the mechanisms of change, and here the proxy data (on both ENSO and mean climate) provide key constraints for model validation. The dampening mechanism best supported by proxy data, and which provides a unified explanation of our findings from all time periods, is that a deeper thermocline in the mid- and early Holocene and in the early Pliocene weakened the upwelling and thermocline feedbacks, thus weakening ENSO. This work highlights the importance of the thermocline, which should help predict ENSO’s response to anthropogenic change.

To investigate the WPWP’s response to varying pCO2, I focus on the Pliocene, the most recent epoch in which pCO2 was higher than preindustrial. Only two Pliocene temperature records exist from the heart of the WPWP, and they show different trends. The foraminiferal Mg/Ca-based SST record shows Pliocene WPWP temperatures similar to today, but the TEX86 temperature proxy indicates a WPWP cooling trend since the Pliocene. The TEX86 studies, which claim that Pliocene WPWP temperatures were warmer than today, echo the claims of modeling studies, which produce a warmer WPWP whenever pCO2 is higher than preindustrial. Though much of the debate over Pliocene WPWP SSTs has focused on changes in seawater Mg/Ca, spatial variations in proxy agreement point to dissolution as a key factor. Dissolution, which imparts a cool bias to Mg/Ca temperatures, varies across ocean basins depending on Δ[CO32-], the difference from the carbonate ion concentration needed for calcite saturation. By necessity, dissolution corrections use the modern value of Δ[CO32-] for the entire record, so it is possible that Pliocene proxy discrepancies could stem from varying Δ[CO32-] over time. To constrain the effect of changing dissolution on the Mg/Ca SST record, I collected benthic foraminiferal B/Ca data (a proxy for Δ[CO32-]) from the WPWP spanning the past 5.5 Myr.

I found no long-term trend in Δ[CO32-] over the past 5.5 Ma, implying no dissolution bias in the trend of the Mg/Ca record. After accounting for changes in seawater Mg/Ca, I estimate the temperature of the WPWP during the Pliocene to be ~1°C warmer than today. As such, the 2-2.5°C trend shown in TEX86 records is not supported by the Mg/Ca data, and likely stems from a bias in the TEX86 data toward subsurface temperatures.