Earth’s poles are uniquely sensitive to climate system perturbations; in recent decades, Arctic temperatures have warmed at twice the global average. Antarctic warming has been slower to emerge, but climate models project long-term changes in both polar regions, posing severe consequences for societies, ecosystems, and global weather patterns. Managing these consequences necessitates a detailed physical understanding of sea ice and its central role in governing the high-latitude energy budget. To address this need, my thesis investigates ice-ocean-atmosphere interactions in the polar regions. I quantify causal relationships among the physical processes shaping high-latitude climate, characterizing how sea ice both drives and responds to climate variability and change in each hemisphere. The results of my research provide physical insights towards more accurate climate models and guide future observations in these remote, data-sparse regions.

In Chapter Two (Kaufman et al. 2020) I study the relationship between Southern Ocean polynyas and high-latitude climate variability. These anomalous ice-free ocean regions, enclosed by the winter sea-ice pack, have been observed to occasionally release heat from the deep ocean to the overlying atmosphere. Yet, most standard resolution climate models represent these features poorly. I analyzed output from a fully coupled model that effectively simulates polynyas due to its uniquely high-resolution seafloor bathymetry. I found that the reduction of surface heat fluxes during periods of full ice cover is not fully compensated by poleward heat transport. This imbalance increases ocean heat content, supplies polynya heat loss, and drives higher atmospheric temperatures. The results disentangle the complex processes that both enable polynyas’ existence and respond to them, providing a robust physical description of these rare, but impactful, events. This research was conducted in collaboration with the Climate, Ocean, and Sea Ice Modeling (COSIM) group at Los Alamos National Laboratory, where the sea ice model is developed.

In Chapter Three (Kaufman and Feldl 2022), my research addresses polar climate dynamics in the Arctic, where I analyze the atmosphere’s response to increased carbon dioxide concentrations. During polar winter, the Arctic’s lower atmosphere warms more than the air aloft; this bottom-heavy vertical warming structure reduces radiative cooling to space and increases the regional climate sensitivity. However, the physical processes that set the warming structure have remained unclear. Using output from a 21st century warming simulation, I quantified the relative influence of various Arctic heating sources, identifying sea-ice loss as the dominant mechanism driving bottom-heavy winter warming. This interseasonal effect is mediated by changes in Arctic clouds and air-sea energy fluxes. By uncovering the causes of vertically non-uniform warming, my results provide key constraints on future Arctic climate change projections.

The final project of this thesis, Chapter Four, seeks to clarify what climate conditions enable the Warm Arctic-Cold Eurasia pattern, a notable feature of historical temperature observations. Observed Eurasian cooling is often attributed to sea-ice loss, implying an anthropogenic cause. However, comprehensive climate changesimulations do not produce Eurasian cooling, and the observational record is too short to rule out a role for unforced atmospheric variability. I reconcile this discrepancy by identifying atmospheric blocking over the Ural mountains as a robust driver of the Warm Arctic-Cold Eurasia pattern across a range of background climate states. By accounting for distinct transient responses to Ural blocking variability in observations and models, I am able to bring their seemingly divergent midlatitude temperature trends into agreement.

A common challenge in each of these projects is defining the routes of causality between coupled physical processes. I address this challenge with causal discovery techniques, which provide an alternative to computationally expensive model perturbation experiments. These statistical techniques identify causal mechanisms in any time series, observational or simulated, and were an essential component of my research approach. By integrating statistical reasoning with climate dynamics theory, I am able to advance understanding of the distinct complex systems comprising Earth's two polar regions.