Ice Ages and Kelp: How Glacial Refugia Shaped Today's Forests

Walk through a kelp forest in British Columbia or Norway and you are visiting a young neighbourhood. At the height of the last ice age, roughly twenty thousand years ago, much of that coast lay under ice sheets kilometres thick or fronted a shoreline more than a hundred metres below today's. The kelp beds you see are recolonists — descendants of forests that survived the deep freeze in sheltered pockets scientists call glacial refugia. Those pockets, where they were and who escaped from them, still structure the genetics of kelp forests today. For anyone restoring kelp, that ancient history is surprisingly practical: it tells us where genetic variety hides, and how carefully we must move it.

The ice-age coastline was a different map

During the Last Glacial Maximum — broad brush, about 26 to 19 thousand years ago — a few transformative things happened to the world's kelp coastlines:

• Sea level fell dramatically, by around 120 metres give or take, as water stacked up in ice sheets. Continental shelves became coastal plains; today's shallow reefs were clifftops. Species that wanted rock in well-lit water had to move downslope and seaward, tracking a shoreline that migrated kilometres or hundreds of kilometres.
• Ice sheets scraped whole coasts clean. Scandinavia, most of Britain, Canada, Alaska's mainland rim — the kelp habitat was not merely chilled, it was ground away and buried under till.
• Southern kelp zones mostly shifted rather than vanished. South of the ice, the subantarctic and temperate coasts were cold but largely unglaciated. Kelp there faced range compression and harsher storms, not erasure.

This asymmetry — scraped north, squeezed south — is the master key to reading kelp biogeography today. Northern forests are largely post-glacial seedlings; many southern forests are ancient.

What exactly is a refugium?

A glacial refugium is any pocket where a species' basic requirements — rocky bottom, light, tolerable temperature, freedom from scouring ice — persisted through a time when its broader range became unliveable. Refugia for kelp have been proposed or genetically inferred in places like these (the strength of evidence varies by species and study, and several remain debated):

• North Pacific pockets around parts of Haida Gwaii, the Aleutian arc and Kamchatka — coasts that stood partly free of ice even at the glacial maximum, fed by a Pacific that never fully froze. The kelp forests of Haida Gwaii today may be heirs to some of the oldest continuously kelped shoreline in the North Pacific.
• Southwest European edges — Brittany, southwest Ireland, Iberia — south of the British and Fennoscandian ice sheets, where Atlantic kelps likely waited out the cold.
• Southern hemisphere strongholds: the Chilean channels, the subantarctic islands, stretches of southern Africa and Australasia, where kelps persisted with room to spare. Southern bull kelps in the Strait of Magellan and giant kelp across the subantarctic carry genetic signatures of deep, old continuity.

The genetic signature of survival: reading the map

Population geneticists read this history in DNA — one part of the toolkit we unpacked in our guide to kelp population genetics and provenance. Three signatures recur:

1. Diversity piles up at refugia and thins along the recolonisation route. Populations that endured in place accumulated mutations for tens of thousands of generations; forests founded by a handful of long-distance dispersers carry only a fraction of that variety. The classic pattern is bluntly named: "southern richness versus northern purity." In several northern kelps, postglacial expansions show exactly this gradient — a tight handful of genetic types across enormous northern ranges.

2. Contact zones mark where separate refugial lines met. When two recolonising fronts converge — say, an Atlantic lineage spreading north meeting a Pacific lineage coming around Alaska — their genes mingle in a narrow seam. Hybrid and transition zones in today's kelps are fossils of invasions that ended ten thousand years ago.

3. Deep splits reveal which refuge. When two modern populations differ at many genes across the genome, coalescent models can estimate when their lineages last shared ancestors. Kelps repeatedly show deep splits — Pacific versus Atlantic clades in sugar kelp, for instance — that imply survival in multiple separate refugia for multiple glacial cycles, not just one.

This is how we know, for example, that giant kelp's southern populations carry genetic depth the northern ones lack: the south was home through the ice ages; the north is a relatively recent frontier adventure.

Post-glacial recolonisation: the long swim back

As ice retreated and sea level rose — climaxing in a flooding reshuffle of every coast between roughly 15 and 7 thousand years ago — kelp spread poleward and shoreward at once. Recolonisation was powered by the same machinery that moves kelp today: spores' short hop, rafts of buoyant adults drifting long distances, and the lucky lottery of settlement we describe in how kelp travels.

But recolonisation was not a fair lottery. Each step of a range expansion is founded by few individuals, so every long hop costs genetic diversity — the so-called leading edge of an expanding range ends up genetically depauperate. Add ten thousand years of subsequent local evolution on top of that inherited skeleton, and you get the modern map: old southern diversity, young northern forests, and distinct "provinces" divided by ancient migration corridors rather than today's weather.

Why deep history matters for restoration

None of this is antiquarian. Refugial history is the operating manual behind two of restoration's hardest questions:

Where does genetic diversity live? Restoring a forest "with local stock" sounds simple until the local stock carries only the narrow gene set of a post-glacial frontier. A restorer wanting heat-tolerant or disease-resistant variety needs to know where it pooled — often in southern or refugial populations hundreds of kilometres away. Our overview of kelp genetics, populations and provenance lays out the practical framework.

How far can we move genes? The same history marks the boundaries. Recolonisation corridors were frequently blocked for millennia by vast ice barriers or unsuitable coasts, letting each refugium evolve its own solutions. Stirring the pot carelessly risks outbreeding depression — breaking up locally co-adapted gene combinations. The debate over assisted migration of warm-adapted kelp is really a debate about how much weight to give these ancient boundaries when the climate is shifting faster than kelp can migrate. At Tidewright, our (fictional) Holdfast hatchery program treats it as hard constraint: seed stock is matched by both thermal profile and lineage, favouring within-region sources whose shared post-glacial history keeps them genetically compatible.

Refugia old and new rhyme. The ice-age concepts generalise: pockets where a population endures its era's worst conditions are strategic assets. Today's analogues — thermal refugia, the cool pockets where forests survived recent marine heatwaves — are the same idea on a decade timescale, and protecting them borrows the same logic glacial refugia proved over millennia.

How scientists reconstruct the unrecorded past

No one watched kelp retreat from the Laurentide ice, so the entire reconstruction is inference — but converging, cross-checked inference:

• Genome scanning and phylogeography compare allele frequencies and genealogies across a species' range, fitting demographic models: when did lineages diverge, from how small a founding pool, expanding from where?
• Coalescent and simulation methods work backward from today's DNA to ask which ancient scenario — single refugium or several, large or bottlenecked — most likely produced the patterns we now see.
• Paleo-shoreline reconstruction supplies the stage: ice-margin maps, isostatic uplift models and drowned landscapes tell us where kelp could have lived, eliminating scenarios the geology rules out.
• Subfossil records — kelp holdfasts in old sediments, shells dated in raised beach deposits — anchor the models to at least a few direct physical traces.

Where these independent lines agree — as they do on the deep southern origins of several giant kelp lineages, or the deep Pacific-Atlantic split in sugar kelp — confidence is earned, not assumed.

Key takeaways

• At the last glacial maximum, sea level was ~120 metres lower and ice sheets erased northern kelp coasts; today's northern forests are largely post-glacial recolonists.
• Kelp survived in refugia — ice-free pockets in the Pacific northwest, southwest Europe and wide southern hemisphere strongholds — whose locations are reconstructed from genetic diversity patterns.
• Post-glacial expansions left legible signatures: low diversity at leading edges, contact zones where lineages met, and deep splits marking multiple ancient refuges.
• Refugial history locates where genetic variety pooled — essential knowledge when selecting restoration seed stock and when weighing assisted migration.
• The same logic applies today: thermal refugia protecting forests through heatwaves are the ice-age playbook running on a modern clock.

FAQ

Did kelp forests exist near the ice edge during the ice age?

Yes, wherever ice-free rocky coast persisted in tolerable temperatures. The ice-free fringes of the Aleutians and nearby North Pacific coasts, and Europe's unglaciated southwest shores, likely held working kelp ecosystems within sight of glacial landscapes.

How do scientists know where the refugia were?

Primarily from DNA: diversity is high and old at refugia, low and young along recolonisation routes. Coalescent models estimate divergence times between lineages, and reconstructions of ice margins and drowned shorelines confirm which locations were physically available.

Are northern kelp forests genetically "worse" because they're young?

Younger and less diverse, on average — which matters because genetic variety is the raw material for adapting to warming seas. Restoration programmes on northern coasts pay particular attention to sourcing diversity thoughtfully rather than assuming local stock holds every answer.

What's the difference between a glacial refugium and a modern thermal refugium?

Same concept, different stressor and timescale. A glacial refugium sheltered a population through tens of thousands of years of ice-age cold; a thermal refugium is a cool pocket — an upwelling shadow, a deep reef — where kelp endures today's marine heatwaves.

Does this history apply only to kelp?

Not at all. The refugia-and-recolonisation framework was built largely on land species — trees, mammals, insects — and applies across marine life. Kelp makes an unusually clean test case because it is strictly tied to rocky, sunlit shoreline, a habitat whose ice-age position we can map precisely.