Reconstructing Tropical Pacific Climate Variability from Coral Archives

Abstract

The tropical Pacific Ocean is a critical component of the climate system, and its mean state and variability are linked to effects across the globe. Our understanding of the area is limited by the scarcity of direct observations of important climatic variables prior to the mid-twentieth century, without which it is difficult to characterize longer timescale variability and trends. Various paleoclimate proxies have been used to extend the observational record into the past; coral records are especially useful as they can offer high-resolution snapshots of sea surface temperature (SST) variability, sometimes as long as hundreds of years. Taken together, these records can be a powerful tool with which to reconstruct climate across the tropical Pacific basin. In the present study, I use two new coral paleoclimate records to examine trends and variability in the eastern equatorial Pacific. I then apply reduced dimension reconstruction techniques to a network of tropical Pacific coral records to evaluate the possibility of reconstructing the Pacific SST field into the past. First, my coauthors and I generate a 1940-2010 Sr/Ca-SST reconstruction from two Wolf Island corals, in the northern Galápagos archipelago. We apply trend analysis to this and several other twentieth century eastern tropical Pacific coral and instrumental datasets, showing that on multidecadal timescales, the entire area has warmed in response to radiative forcing. In recent decades, though, increases in strength in upwelling and the Equatorial Undercurrent have led to spatially complex trends, including cooling in the eastern tropical Pacific during boreal fall and winter. In the second chapter, my colleagues and I report on a second coral SST reconstruction from the northern Galápagos, this one from Darwin Island. Based on coral δ18O, the reconstruction runs from 1866-2015 and is one of few seasonally resolved SST reconstructions from the eastern Pacific to span the nineteenth to twenty-first centuries. The Darwin reconstruction shows unique features relative to other tropical Pacific coral records, including antiphasing with central and western Pacific coral records that suggests it may be monitoring decadal-scale variability related to Central Pacific El Niño activity. Finally, we use the PAGES2k network of tropical coral records to reconstruct annual SSTs across the Indo-Pacific basin over the last several centuries. We test the ability of two related climate field reconstruction techniques, both reduced dimension approaches based in optimal interpolation, to estimate the leading climate features of the field from this sparse proxy network. Evaluation of reconstruction skill and uncertainty suggests that both methods are negatively affected by proxy system error, as well data availability in space and time.