07 Jan 2026 - Amy
Finally, a blog post about science! It is rather late, so apologies for that, but it was Christmas and I had parties to attend, cricket to be disappointed by, cheese to eat, and darts to watch. I’m here now—let’s see if it was worth the wait.
When I wrote the proposal for my current fellowship, my central argument was that sea ice is a fundamental component of the global climate system—one whose behaviour we can best understand by examining past periods of global climate change. My own expertise lies in the dating of sedimentary archives, which also represent some of our most valuable records of past sea ice variability. I proposed that by improving constraints on when sea ice advanced and retreated in the past, we might better understand how contemporary sea ice loss both responds to, and actively drives, changes in the cryosphere, oceans, and atmosphere.
Sea ice retreat is among the most visible consequences of anthropogenic climate change. This visibility does not imply simplicity—global sea ice is challenging to quantify—but the scale of change is undeniable. In their 2024 update, Roach and Meier report that global annual mean sea ice reached its second-lowest extent on record. My own focus is on the Arctic, which in 2024 experienced the sixth-lowest September sea ice extent, the lowest December extent, and record regional minima in the Canadian Archipelago and Hudson Bay.
So, why does sea ice matter? It has frequently been framed as a potential “tipping point” within the climate system. I will not dwell on that concept here—partly because it has been comprehensively reviewed elsewhere, notably in this excellent review by Wunderling and colleagues, and partly because I am not convinced that the tipping-point framework is always the most productive lens through which to view ongoing global change.
What is beyond doubt is that sea ice has played a central role in Earth’s climate system over at least the past million years. My own introduction to sea ice as a key climate component came during my PhD, when I was studying ecosystem change in Greece during the last glacial cycle, from about 50 to 35 thousand years ago. At first glance, the connection between Arctic sea ice and forest composition in the eastern Mediterranean may not be obvious. In practice, it is profound.
The last glacial cycle—sometimes called Marine Isotope Stage 4 to 2, and often simply “the ice age”—was a long-lived cold period spanning 71 to 14 thousand years ago, when the planet was considerably colder. The most obvious evidence of this icy past can be found in places like the British Isles and North America, where the present-day landscape still bears countless clues that these regions were once buried beneath vast ice sheets. Further south, in the Mediterranean, the climate was also markedly cooler and drier. In Greece, for example, extensive glaciers covered the Pindus Mountains, and ecosystems responded accordingly.
Pollen records from the foothills of the Pindus provide a detailed picture of these changes. During the last glacial cycle, species characteristic of the modern Mediterranean, such as olive and evergreen oak, were largely absent. Instead, vegetation was dominated by cold-tolerant taxa, including pine, spruce, and steppe grasses. Crucially, these ecosystems were not static. Pollen records show repeated alternations between grass-dominated landscapes and intervals during which deciduous trees expanded.
The leading explanation for these shifts is that they reflect global-scale climate oscillations occurring on millennial timescales, known as Dansgaard–Oeschger cycles. The idea that the last glacial climate was punctuated by abrupt changes predates modern ice-core science: early temperature records from the North Atlantic and vegetation change from northwest Europe had already hinted that glacial climates were unstable. Greenland ice cores, however, revealed the full magnitude and frequency of these abrupt warmings and coolings with in full colour HD.
The stories behind these earliest Greenland ice cores are fascinating and honestly deserve their own blog post—but Cold War history is outside my expertise, so that will have to wait.
The crux of the issue is this: our climate system naturally experienced repeated abrupt warmings and coolings throughout the last glacial cycle. Understanding the drivers and impacts of these natural climate changes provides a laboratory for testing climate models and identifying which components of the climate system respond most rapidly and which have the largest impacts.
For the abrupt Dansgaard–Oeschger warmings, it is becoming increasingly clear that the key to understanding their drivers lies in the Arctic Ocean and nearby seas. While records of past climate change show that sea ice, sea surface temperature, and ocean circulation all varied on thousand-year timescales during the last glacial cycle, the exact processes governing the onset of these warmings and coolings have long remained elusive.
Two major advances have helped to resolve this. The first is methodological: new geochemical tools now allow more robust reconstructions of past sea ice concentration. These records demonstrate that sea ice variability was a defining feature of millennial-scale climate change in the Nordic Seas, amplifying temperature swings across the North Atlantic region.
The second advance is in climate modelling. Earlier modelling efforts required strong external perturbations—such as large freshwater inputs—to reproduce glacial abrupt climate events. More recent Earth system models, however, are capable of generating self-sustained millennial-scale oscillations under glacial CO₂ concentrations alone. In these simulations, the timing of abrupt transitions is closely linked to internal variability in the coupled ocean–atmosphere–sea ice system.
Together, observational data and modelling studies increasingly point to Arctic and North Atlantic sea ice as a key regulator of the duration and pacing of Dansgaard–Oeschger cycles, largely through its influence on ocean circulation. The same high-latitude changes are closely connected to the ecosystem shifts I observed during my doctoral research in Greece, underscoring the global reach of Arctic climate processes.
That said, the last glacial cycle is not an ideal analogue for future climate change. Ice sheets were larger, atmospheric CO₂ concentrations were lower, and Earth’s orbit was different. To better understand how continued sea ice loss may shape future climate variability, we must look to past intervals characterised by warmer polar conditions.
The most recent of these is the Last Interglacial (approximately 129 to 116 thousand years ago, Marine Isotope Stage 5e), a period marked by a warm Arctic and global mean sea levels at least five metres higher than today. Despite its relevance, there are remarkably few reconstructions of sea ice and ocean change from this interval in the Arctic Ocean and Nordic Seas. The primary reason is technical: marine sediments in these regions are notoriously difficult to date with sufficient precision.
We have, however, very few reconstructions of Last Interglacial sea ice and ocean change in the Arctic Ocean and Nordic Seas, largely because of the huge challenges associated with dating marine sediments in the region. Which is where I come in! Where this blog post ends, my project begins. I’ll be focusing on determining when sea ice retreated in the Arctic Ocean by identifying volcanic ash layers within the sediments used to reconstruct past sea ice change.
Where this post ends, the research begins. More on that soon—but for now, I will leave it there.
I’ll see you next month.