The lost lakes of Britain (Part 3): Landscapes that vanished, rivers that remain
Jon Trevelyan (UK)
Simplified summary of Britain’s geological history from the Precambrian to the present day. The coloured timeline shows the major geological periods discussed throughout this series, while the illustrations above represent some of the dominant landscapes and environments that characterised each interval. The red line(s) indicates the geological period(s) covered by this article.
This final part of the series brings together the two very different lake worlds explored so far. In Part 1, we reconstructed the long-lived Devonian lakes of the Orcadian Basin; in Part 2, we examined the short-lived but dramatic ice-dammed lakes of Glen Roy. Here, we broaden the view to consider the wider pattern of lost lakes across Britain – how they formed, how they drained, and how their influence persists in modern landscapes, particularly in the behaviour of rivers and the shape of valleys.
Seen together, the Devonian lakes of the Orcadian Basin and the Ice Age lakes of Glen Roy represent two ends of the same spectrum: long-lived systems shaped by sediment and climate, and short-lived systems driven by ice and drainage.
At the end of the last Ice Age, large parts of Britain were occupied by lakes formed when advancing ice blocked natural drainage routes. These short-lived but extensive bodies of water reshaped river systems and left subtle traces across the landscape. By combining geological evidence with modern observations, it is possible not only to reconstruct these lost environments, but to recognise their imprint directly in the field.
A landscape that shouldn’t exist
Britain’s rivers behave themselves. They flow downhill, follow valleys, and ultimately find the sea. Valleys drain, slopes shed water, and – outside of flood conditions – water rarely lingers for long.
Yet scattered across the landscape are clues that this tidy picture is incomplete. Broad, flat valleys with rivers that seem too small for them. Gorges cut through resistant rock that appear wildly oversized. Subtle benches along valley sides that hint at former water levels. Extensive lowlands that seem too flat, too uniform, too quietly featureless.
These are the fingerprints of a very different Britain.
During the last Ice Age, large parts of the country were occupied not by flowing rivers, but by standing water. Lakes – sometimes local, sometimes vast – formed wherever advancing ice blocked natural drainage routes. Rivers that once flowed freely were trapped, and water spread across valleys and lowlands.
For a time – sometimes thousands of years – these lakes dominated the landscape. Then, as the ice retreated, they drained, often rapidly. The water vanished. But the changes it made remained.
How to make a lake out of ice
The basic mechanism is simple. A glacier advances across a landscape and blocks a valley. Water continues to flow into the basin, but now it has nowhere to go. Gradually, it accumulates, forming a lake ponded against the ice (Figs. 1 and 2).
Simple though it is, the consequences are profound. As the lake grows, its level rises until it finds a new equilibrium. That equilibrium may be temporary. Water may spill gently over a low point, establishing a new outlet. Alternatively, it may rise until a barrier fails, releasing water more dramatically.
These different modes of drainage leave different signatures. Slow overtopping tends to produce stepped shorelines and terraces. More rapid drainage can carve deep spillways and gorges.
In reality, many lake systems probably passed through both phases: gradual rise, episodic overflow, and eventual lowering as new drainage routes became established.
A Midlands lake that rewrote the map
The clearest and most instructive example is Lake Lapworth, which once occupied large parts of the English Midlands (Fig. 3).
The significance of Lake Lapworth lies not just in its existence, but in how it was recognised.
In the late nineteenth century, geologists were faced with a problem. The Ironbridge Gorge, through which the River Severn flows, appeared far too large for the river that occupied it (Fig. 4). The surrounding landscape showed signs of unusual sedimentation. Drainage patterns were difficult to reconcile with gradual river erosion.
Individually, these observations were puzzling. Together, they demanded explanation.
It was Charles Lapworth (Fig. 5) who provided it. Better known for his work at Dob’s Linn and for defining the Ordovician System, Lapworth approached the problem by stepping back from individual features and asking a broader question: what kind of landscape would produce all of these observations at once?
His answer was bold but elegant: a large lake, held in place by ice that had eventually drained southwards. This was not a matter of spotting a single feature. It was a synthesis. The oversized gorge, the unusual sediments, the disrupted drainage – all became coherent once the presence of a former lake was accepted.
Standing above the gorge today, Lapworth’s reasoning becomes immediately intuitive. The valley is simply too large. The river too small. The mismatch demands explanation. Lapworth’s work illustrates a fundamental point: geological understanding often depends not on new observations, but on new ways of connecting existing ones. By recognising that disparate features formed part of a single system, he transformed a series of puzzles into a coherent story.
But the gorge is only part of the story. Across the Midlands, fine-grained sediments—silts and clays – indicate prolonged standing water. Their distribution suggests a basin rather than a river system. Variations in thickness hint at changing water levels. Subtle terraces mark former shorelines.
Each piece of evidence is modest. Together, they are compelling.
From local to vast: the scale of glacial lakes
Lake Lapworth was only part of a much larger pattern. In eastern England, Glacial Lake Humber occupied an even broader region (Fig. 7)
Here, the scale shifts dramatically. Instead of a confined basin in the Midlands, we are dealing with a wide, low-lying landscape extending across the Vale of York (Fig. 8) and beyond. Ice advancing from the north and east blocked drainage to the North Sea, and water spread across this gently undulating terrain, forming a lake whose margins were controlled less by steep topography than by subtle variations in elevation.
Unlike Lake Lapworth, which is anchored by a dramatic spillway at Ironbridge, Lake Humber does not reveal itself through a single defining feature. Instead, its presence must be inferred from the overall character of the landscape. The Vale of York, as seen in Fig. 8, is remarkably flat and laterally extensive. This subdued relief reflects prolonged occupation by standing water, which smoothed and simplified the terrain while allowing fine sediments to accumulate across a broad area.
The margins of such a lake were not sharply defined. Rather than forming a clear shoreline, water would have extended irregularly across the basin, advancing and retreating in response to small changes in level. Low ridges, slight rises in ground, and subtle changes in slope would have influenced the position of the lake edge, producing a shifting, complex boundary rather than a single fixed margin.
This has important implications for how the lake is recognised today. Instead of dramatic features, we see widespread but understated evidence: extensive spreads of fine-grained sediments, low-relief topography, and drainage patterns that do not quite match the underlying geology. The scale of the system is best appreciated not from any one location, but by considering the region as a whole.
At times, the Humber system may not even have been a single continuous body of water. Fluctuations in water level, combined with the gentle relief of the landscape, could have produced a series of interconnected basins that merged and separated over time. What we describe as “Lake Humber” is therefore best understood as a dynamic system, responding to the advance and retreat of ice and the availability of drainage routes.
Lake Pickering: a lake between the hills
The story of Glacial Lake Pickering provides an ideal bridge between the vast scale of Lake Humber and the more general principles of glacial lake evolution. Occupying the Vale of Pickering between the North York Moors and the Yorkshire Wolds, it was smaller than Humber but preserved many of the same characteristics. Ice advancing southwards blocked drainage towards the North Sea, allowing water to accumulate within the low-lying basin.
Unlike Lake Humber, whose margins are difficult to define precisely across its broad lowlands, Lake Pickering was more strongly constrained by the surrounding topography. The hills bordering the vale acted as natural basin margins, while the ice front formed a temporary dam to the east. As water levels rose, the lake expanded westwards and southwards, until it eventually found an escape route through what is now Kirkham Gorge, establishing a new drainage pathway to the River Derwent.
Lake Pickering also illustrates an important point about glacial lakes: they were rarely static. As the ice margin advanced and retreated, the position of the dam changed, altering both the size of the lake and the elevation of its outlet. The lake probably expanded and contracted several times before finally disappearing altogether, as the ice withdrew from eastern Yorkshire.
Today, evidence of the lake survives in several forms. Extensive deposits of fine silts and clays record prolonged periods of standing water across the floor of the Vale of Pickering. Elsewhere, subtle shoreline features occur along the valley margins, although these are far less conspicuous than the celebrated “parallel roads” of Glen Roy. Together with the overflow channel at Kirkham, these deposits allow geologists to reconstruct not simply the existence of the lake but the sequence of events through which it formed, evolved and finally drained.
In many respects, Lake Pickering occupies a middle ground within Britain’s Ice Age lakes. It is large enough to demonstrate regional-scale drainage reorganisation, yet sufficiently confined that the relationship between basin, ice dam and overflow can still be readily understood from the modern landscape.
These lakes were not static bodies of water. Their margins shifted, their levels fluctuated, and their internal structure likely varied through time. In some cases, separate basins may have coalesced into larger systems before fragmenting again as drainage evolved.
Reading the vanished shoreline
One of the most tangible pieces of evidence comes in the form of terraces.
One of the most direct ways to recognise former lake levels is through the presence of terraces – flat or gently sloping benches cut into valley sides when water remains at a stable level for a prolonged period.
In principle, these features are straightforward. A lake surface defines a horizontal plane. Where that surface intersects a slope, wave action and sediment processes can cut a narrow bench. If the water level remains stable, the feature becomes more pronounced; if levels fluctuate, a series of benches may develop at different elevations.
In practice, however, the expression of such features varies enormously depending on the landscape. In Glen Roy, as discussed in Part 2 of this series, the ‘parallel roads’ provide a striking example: sharply defined, laterally continuous shorelines that can be traced for kilometres along the valley sides. There, the combination of steep topography and relatively stable water levels has preserved the features with exceptional clarity.
By contrast, in much of England, the situation is very different. Around basins such as the Vale of Pickering (where the glacial Lake Pickering formed when Late Devensian glaciation blocked meltwaters) or the margins of the Humber system, relief is low and slopes are gentle. Under these conditions, former shorelines are rarely expressed as crisp, continuous terraces. Instead, they appear as subtle breaks in slope, slight benches, or barely perceptible changes in gradient that extend discontinuously across the landscape.
Recognising these features requires a shift in perspective. Rather than looking for dramatic landforms, the observer must look for consistency: faint, near-horizontal alignments that persist across fields or along valley margins, often only becoming apparent when viewed from a distance or under favourable lighting conditions. Even then, interpretation is rarely straightforward. Agricultural modification, erosion and later sedimentation can all obscure or fragment the original features.
For this reason, terraces in English glacial lake basins are rarely identified on the basis of a single exposure. Instead, they are recognised through the accumulation of evidence – topographic patterns, sediment distribution, and their relationship to reconstructed water levels. Individually, these clues may be ambiguous. Taken together, they provide a coherent picture of former lake surfaces and their changing levels through time.
Where to see the evidence
The traces of Britain’s Ice Age lakes can still be recognised.
- Ironbridge Gorge – a classic glacial spillway.
- River Severn valley – reorganised drainage.
- Vale of York – former lake basin.
- York/Ouse floodplain – low-gradient, water-spread landscape.
- Upland terraces – subtle shoreline remnants.
Spillways: the escape routes of vanished lakes
If terraces record still water, spillways record movement.
When lake levels rose high enough, water spilled over the lowest point in the basin. Initially, this may have been a shallow overflow. But once flow began, erosion intensified, deepening the channel and lowering the lake level. And this process can be self-reinforcing. As the channel deepens, more water is routed through it, increasing its erosive power. Over time, a substantial valley or gorge may be carved.
The key point is that these features are not the product of long-term river erosion. They are the result of relatively short-lived, high-volume flow events associated with lake drainage. Although the mechanisms differ in detail, the fundamental controls – basin shape, water supply and available outlets – are the same as those that governed the Orcadian lakes hundreds of millions of years earlier.
When rivers forget their past
Before glaciation, rivers followed established routes. During glaciation, those routes were blocked, and lakes formed. When the lakes drained, new channels were cut. After the ice retreated, rivers adopted these new paths (Fig. 10).
This explains many anomalies in modern drainage. Rivers may cut across geological structures, flow through unexpectedly narrow gorges, or occupy valleys that seem disproportionate to their size.
They are, in effect, inheriting a landscape shaped under very different conditions.
Seasonal rhythms in stone
At a finer scale, lake conditions are recorded in the sediments themselves (Fig. 12).
In some deposits, sediment accumulates as paired light and dark layers known as varves. Each pair represents a single year. Lighter layers form during summer, when increased runoff delivers coarser sediment. Darker layers form in winter, when sediment input is reduced and finer particles settle slowly.
The preservation of these layers requires calm, undisturbed conditions. In particular, it implies that bottom waters were largely anoxic, preventing organisms from disturbing the sediment. The same principles of calm water and undisturbed sedimentation seen in Devonian deposits can also be recognised, in more limited form, in some Ice Age lake sediments.
These rocks therefore record not just the presence of a lake, but its seasonal pulse – year by year, layer by layer.
Lost lakes, lasting landscapes
The lakes themselves are gone. No trace of open water remains. Yet their influence is everywhere – etched into valley sides, carved into gorges, and written into the paths of modern rivers.
A terrace, a gorge, or an anomalous river may all be clues to a vanished lake. These landscapes only make sense once the water is put back.
The lakes have disappeared – but the rivers still remember them.
Taken together, these three studies show that Britain’s lost lakes were not a single phenomenon, but a recurring theme in its geological history. From the long-lived, stratified waters of the Devonian Orcadian Basin, to the rapidly changing ice-dammed lakes of the last Ice Age, each system reveals a different way in which water can shape the landscape. The lakes themselves have vanished, but their effects remain – etched into rock, preserved in sediment, and carried forward in the courses of modern rivers. To recognise them is to see familiar landscapes in a new way: not as static features, but as the latest stage in a long and dynamic history.