Date of Award

May 2020

Degree Type


Degree Name

Doctor of Philosophy (PhD)


Earth Sciences


Linda C. Ivany


Eocene Epoch, oxygen isotope paleothermometry, paleoceanography, paleoclimate, sclerochronology, sea surface temperature

Subject Categories

Physical Sciences and Mathematics


Climate change is arguably the most important issue facing modern society. One of the best tools we have for constraining future climate conditions comes from looking at warm and transitional intervals in Earth’s geologic past, such as the Eocene Epoch (~56-34 Ma). The Eocene Epoch was a time of large-scale global climate change, bookended by both the warmest temperatures of the Cenozoic (i.e., the Paleocene-Eocene Thermal Maximum) and the onset of southern hemisphere glaciation (i.e., the Eocene-Oligocene Transition). While mean global climatic conditions across the Eocene, inferred from a compilation of oxygen isotopes of benthic foraminifera, are well constrained and document a clear cooling trend, the few and geographically disparate records of local sea surface temperature (SST) from this interval are often conflicting and difficult to reproduce with climate models. Likewise, multi-proxy studies from the same location frequently yield diverging SST estimates.

These inconsistencies within the climate record inhibit our ability to identify the mechanisms responsible for late Eocene cooling, and call into question our understanding of fundamental aspects of climate dynamics and the underlying assumptions guiding our interpretation of proxy data. Further, they highlight one of the dominant shortcomings of paleoclimate studies; namely the propensity to express climate variability in terms of global or latitudinal averages, while overlooking local and regional scale climate heterogeneity. Distilling global climate to single numbers (e.g., the 2oC global warming threshold) or metrics (e.g., meridional temperature gradients) is appealing, as it allows for direct comparison of different climate states, however oversimplifying conditions by ignoring natural spatial heterogeneity may lead to erroneous paleoclimate interpretations and contribute to the frequent need to set unrealistic boundary conditions in climate modelling studies. In fact, inspection of modern SST data reveal abundant variability along individual latitudinal bands. This contradicts the simplifying assumption of homogenous zonal paleotemperatures and suggests that improving our understanding on the controls on modern SSTs may hold the key to better understanding ancient climate systems.

The ultimate goal of my dissertation is to provide the tools to facilitate a more robust evaluation of ancient climate dynamics, and thereby improve the fidelity of proxy-based paleoclimate reconstructions and future climate predictions. In Chapter 2, I use analyses of modern SST data to identify sampling biases in the paleo record and propose a new framework within which to more meaningfully interpret annually- and seasonally-resolved SST proxy data. In Chapter 3, I develop a bivalve growth rate model, which accounts for variable intra-annual growth rates and facilitates the temporal alignment of serially-sampled geochemical proxy data, increasing the reliability and applicability of paleo-seasonality interpretations. In Chapters 4 and 5, I apply these approaches to reconstruct seasonal changes in nearshore waters off the eastern margin of the Antarctic Peninsula between the middle and late Eocene. Proxy data are evaluated using climate models and modern analog analyses, supplemented with seasonal precipitation data, and contextualized with existing SST data from the Eocene Southern Ocean, resulting in a holistic assessment of climatic conditions during this critical time interval.

The findings of these studies: 1) demonstrate the utility of seasonal data in distinguishing between the mechanisms responsible for large-scale climate change and identifying seasonal biases in other SST proxy data, 2) suggest that initial late Eocene Antarctic cooling was driven by changes in ocean circulation, rather than pCO2, 3) reveal how sampling location biases can generate spurious climate interpretations, and 4) illustrate that recognition of and correction for these biases can allow for a more comprehensive and accurate understanding of ancient climate dynamics conditions.


Open Access

Available for download on Friday, July 02, 2021