Reconstructing Pacific Trade-wind Variability: Assessing and Extending the Novel Coral Mn/Ca Proxy

Abstract

Tropical Pacific trade winds are a critical component of the Pacific Ocean’s climate state; by driving atmosphere and ocean circulation, they facilitate heat storage in the sub-surface of the ocean and thus modulate the rate of climate change. Trade winds weaken and strengthen on interannual and decadal timescales in accordance with the warm and cold phases, respectively, of the El Niño-Southern Oscillation (ENSO) and Pacific Decadal Variability (PDV), which have global climate and societal impacts. Despite their importance, trade-wind observations in this region are sparse; they span only the past 30-40 years and are thus insufficient for analyzing trade-wind variability on longer timescales. Previous work has demonstrated that reef-building corals in the equatorial Pacific may extend this limited record of historical wind observations. The ratio of manganese-to-calcium (Mn/Ca) in the calcium carbonate skeleton of corals from Tarawa, Kiritimati, and Butaritari atolls in the west-central tropical Pacific (Republic of Kiribati) reflects local trade-wind behavior, with spikes in Mn/Ca linked to westerly-wind activity, which is more frequent during warm phases of ENSO and PDV. Long coral Mn/Ca records therefore hold great promise in improving our understanding of past trade-wind behavior. However, coral Mn/Ca measurements to date have revealed variable magnitudes of spikes among corals (both between and within sites) as well as lags between the timing of westerly winds and Mn/Ca spikes in certain corals. The studies included in this dissertation can be seen as performing “due diligence” in testing and addressing uncertainties for the development of a robust, novel climate proxy. My co-authors and I build upon previous work to systematically investigate potential uncertainties in the coral Mn/Ca-wind relationship with the goal of applying this proxy broadly across spatiotemporal scales to reconstruct trade-wind behavior across the equatorial Pacific.In the first study, we test the mechanism of the coral Mn/Ca-wind relationship that was originally proposed at Tarawa atoll. After examining the various reservoirs of Mn that play a role in the mechanism (i.e., sediment, porewater, water) at Kiritimati atoll, an island over 3,000 km to the east of Tarawa that also has a west-facing lagoon (though with a different morphology than Tarawa’s), and confirm the validity of this mechanism. However, a comparison of sediment porewater Mn concentration in the lagoons of Tarawa and Kiritimati suggests that the depth and morphology of the lagoon may influence the concentration of Mn that is eventually incorporated into the coral, thus raising additional questions. Nonetheless, this work confirms the overall mechanism linking winds to Mn/Ca at a different site. further strengthening the foundation upon which this coral Mn/Ca-wind relationship is based. In the second study, we address the question of whether the coral’s mineralogical properties, i.e., the structure of aragonite crystals grown by the coral, influence the incorporation of Mn into the skeleton from which it is measured (thus impacting the resulting interpretation of wind behavior). We find that changes in the dimensions of the aragonite building blocks (i.e., unit cell parameters: a-, b-, c-axis lengths and volume) of the coral skeleton through time do not have a significant impact on its Mn content. However, we do observe significant relationships between unit cell parameters and the concentrations of other trace elements that are commonly used to reconstruct paleoclimate conditions, such as Li, Mg, and Ba. Furthermore, we discover that the nature of these relationships change for sections of the coral skeleton that have been precipitated during periods of thermal stress. While not necessarily relevant to the coral Mn/Ca proxy, these findings are the focus of this manuscript. The third study extends the novel Mn/Ca proxy to new sites to create a robust, multi-site reconstruction of westerly wind variability, and by doing so tackles the remaining uncertainties surrounding the coral Mn/Ca proxy that were not addressed in previous work or the earlier studies in this dissertation. We find that lagoon size, depth, and morphology are correlated with the magnitude of Mn incorporated into the coral, and that the number of lagoon mouths and the distance of the coral from the lagoon mouth are linked to the lag time between westerly winds and coral Mn/Ca spikes. After accounting for these non-wind-related influences on coral Mn/Ca, we create two composite records that combine eleven Mn/Ca records from west and east of the Date Line, which track westerly-wind activity across the equatorial Pacific. We find that these coral-based composite wind records not only capture interannual variability related to ENSO, but also have the potential to differentiate between different flavors of El Niño events. In our final study, we take this final robust 120-year long coral Mn/Ca-based composite wind record and perform spectral analysis to extract information about significant timescales of variability contained in this record. We compare this record against wind observations, reanalysis data, and two climate model ensembles to see how well this coral-based record captures ENSO- and PDV-related variability and to determine the skill with which climate models represent wind variability over time. While the composite record successfully captures ENSO-related timescales of variability, it does not reproduce the decadal variability present in the observational wind record. Furthermore, a comparison of simulated wind records from climate models suggest that this decadal variability cannot be successfully reproduced without two-way feedbacks between the atmosphere and ocean and time-evolving external forcings. Taken together, these four studies systematically identify and address uncertainties in how coral Mn/Ca is linked to trade-wind behavior. The outcome is a solid foundational knowledge of the non-wind-related factors that impact coral Mn/Ca, which has enabled us to produce a composite record combining all existing coral Mn/Ca records—11 replicated records across 6 atolls east and west of the Date Line—to extend our knowledge of westerly-wind activity back to the late 19th century. Furthermore, this foundational knowledge of the mechanisms linking wind to coral Mn/Ca will continue to support the addition of new Mn/Ca records to the composite, thus extending the reach of this proxy and providing additional insight into how trade-wind behavior has changed in the past, and how it may change in the future.