Palaeo-ice stream landsystem

Ice streams are corridors of fast-flowing ice within ice sheets that are flanked on either side by slowly moving ice1. Palaeo-ice streams are ice streams that existed in former ice sheets2,3, such as the continental ice sheets that grew during the last Ice Age. Glaciologists know that these palaeo-ice streams existed as they left a clear imprint on the landscape over large parts of North America4, Scandinavia5, and Britain6.

Landsat 7 ETM+ satellite image of Byrd Glacier, an ice stream in West Antarctica. Ice flow is towards the top of the image. Note how flow converges into the main ice stream trunk. Also, note the sharp boundary between fast- and slow-flowing ice. Image: NASA.

Why are ice streams important?

Ice streams in Greenland and Antarctica are the main control on ice sheet mass balance and discharge to the world’s oceans7. Understanding how ice streams behave and change over time, therefore, is important for predicting and managing the impacts of future climate change.

But this is easier said than done…

Firstly, records of modern-day ice stream activity only cover the most recent ~50 years (the length of the satellite record), which is not enough to confidently predict how they may change in the future.

Secondly, it is almost impossible for glaciologists to study the processes that occur at ice stream beds – which control fast ice flow and, ultimately, ice stream discharge to the oceans1 – owing to the great thickness (up to ~3 kilometres) of ice sheets.

Fast-flowing ice streams (blue and white areas) drain the interior of the East and West Antarctic ice sheets, controlling ice sheet mass balance and discharge to the oceans. Image: Jonathan Bamber

Why study palaeo-ice streams?

Therefore, the landforms and sediments left behind by palaeo–ice streams in areas like North America, Scandinavia, and Britain, are very important.

Firstly, they allow glaciologists to study how ice streams have evolved over thousands to tens of thousands of years, through important stages, such as ice sheet build-up, at a glacial maximum, and during deglaciation2,3,8,9.

Secondly, the landform record offers a window into the processes that occurred at former ice stream beds, allowing researchers study how they flowed, shifted, turned ‘on’ and ‘off’, and interacted with the landscape2.

The palaeo–ice stream record, therefore, can be used to better understand how ice streams change over long timescales and under different climate conditions, in order to improve predictions of future ice sheet change.

The palaeo–ice stream landsystem

Ice streams have three important characteristics that are reflected in the landforms they create10,11,12. First, they flow very rapidly – orders of magnitude faster than a typical valley glacier13 – by a combination of internal deformation, sliding, and subglacial deformation1,10. Second, they have convergent onset zones1,10 (onset zones are areas where ice flow changes from slow- [sheet flow] to fast-moving [stream flow] at the head of an ice stream). Third, their lateral margins are very sharp1,10.

Characteristics of an ice stream (fast-flowing ice, a convergent onset, and sharp lateral margins) displayed at Byrd Glacier, West Antarctica. Image: NASA.

Fast ice flow

Mega-scale glacial lineations are the most striking landforms created by fast ice flow in palaeo–ice streams14,15. They are streamlined sediment ridges formed at the bed in the main ice stream trunk zone16. You can think of these landforms as ‘stretched’ out flutes or drumlins, as they are similar in shape, but much larger and more elongate14,15.

In size, mega-scale glacial lineations are between 10–100 kilometres long and 200–1300 metres wide11, making it difficult to identify them on the ground. Instead, they are most easily mapped from satellite images (see below). When viewed from space, it is also obvious that mega-scale glacial lineations are not isolated features, but occur together in large groups. Within these groups, they run parallel to one another over great distances11,14,15.

Mega-scale glacial lineations formed at the bed of the Duawnt Lake palaeo-ice stream in Canada (see ref. 23). Note how individual lineations are highly elongate and closely parallel each other. In this example, the palaeo-ice stream flowed from right to left. Image: Google Earth.

Convergent onset zones

Shorter subglacial bedforms, such as flutes and drumlins, form in palaeo–ice stream onset zones, where ice velocity is lower than in the ice stream trunk zone11,17. These landforms are arranged in a fan-like pattern that flows in toward (or converges on) a narrower corridor of fast-flow landforms that include mega-scale glacial lineations.

Convergent flow in the onset zone of the Transition Bay palaeo-ice stream, Arctic Canada (see ref. 17). Also, note how this ice stream flow path (white) crosscuts an older ice stream flow path (black). Image: Google Earth.

Sharp ice stream margins

In modern ice streams, shear zones – areas of intense deformation several kilometres wide, marked by crevassing at the ice-surface18 – develop at the margins of ice streams, where fast- and slow-moving ice meet19.

Surface crevasses in a shear zone at Recovery Glacier ice stream in East Antarctica Image: NASA.

Ice stream shear margin moraines are sediment ridges deposited subglacially in the shear zone20. At first glance, they look similar to mega-scale glacial lineations, but they are generally wider and longer20. Shear margin moraines can be used to identify the edges (and thus lateral extent) of palaeo–ice streams11,12.

Shear margin moraine (arrowed) with a fast-flow assemblage (e.g. drumlins, mega-scale glacial lineations) ‘inside’ the palaeo-ice stream flow path (right of shear moraine) and ice-stagnation landforms ‘outside’ the ice stream flow path (left of shear moraine). Example from the M’Clintock Channel palaeo-ice stream in Arctic Canada (see ref. 20). Image: Google Earth.

Flow-direction changes

Ice streams do not always follow the same flow pathway; they are capable of switching flow-direction over time owing to glaciological (e.g. ice thickness) or topographic (e.g. basin infilling) changes9,21.

In the palaeo–ice stream landsystem, flow-direction changes can be mapped where one group of flow assemblages (e.g. drumlins) crosscuts another11,12,14. It is usually possible to work out the relative order of flow changes by studying the pattern of crosscutting (see the Transition Bay palaeo-ice stream diagram above).

Ice stream shutdown

While the palaeo–ice stream landsystem is dominated by features relating to fast ice-flow (e.g. mega-scale glacial lineations), these may be overprinted by other landform assemblages. For example, during deglaciation, moraine ridges and ice-stagnation landforms may be deposited over the top of fast-flow landforms as the active ice-front moves back2,11,12.

Similarly, ribbed moraines (transverse sediment ridges) may form over the top of glacial lineations22. Ribbed moraines are thought to form where ice-flow changes from an extensional (ice streaming) to a compressional regime. Where they lie on top of glacial lineations, therefore, they may record the slowing or shutdown of palaeo-ice streams during ice sheet deglaciation22.

Ribbed moraines lying on top of glacial lineations at the bed of the former Dubawnt Lake palaeo-ice stream. This ordering of landform assemblages records ice stream shutdown during deglaciation (see ref. 22).


Ice streams shape the land surface they flow over, leaving behind a distinctive landsystem11 that includes mega-scale glacial lineations, which record the passage of fast-moving ice14, convergent bedforms in onset zones, and shear margin moraines that mark their sharp lateral margins20. In addition, the palaeo–ice stream landsystem often displays evidence of dynamic ice sheet changes5,6, such as switches in flow-direction9,21 (crosscutting landforms) and velocity.

Related content

Professor Chris Clark’s Sheffield University webpages also host a wealth of information on mega-scale glacial lineations, drumlins, and ribbed moraines!


1. Bennett, M.R., 2003. Ice streams as the arteries of an ice sheet: their mechanics, stability and significance. Earth-Science Reviews61, 309-339.

2. Stokes, C.R. and Clark, C.D., 2001. Palaeo-ice streams. Quaternary Science Reviews20, 1437-1457.

3. Livingstone, S.J., Cofaigh, C.Ó., Stokes, C.R., Hillenbrand, C.D., Vieli, A. and Jamieson, S.S., 2012. Antarctic palaeo-ice streams. Earth-Science Reviews111, 90-128.

4. Margold, M., Stokes, C.R., Clark, C.D. and Kleman, J., 2015. Ice streams in the Laurentide Ice Sheet: a new mapping inventory. Journal of Maps11, 380-395.

5. Kleman, J., Hättestrand, C., Borgström, I. and Stroeven, A., 1997. Fennoscandian palaeoglaciology reconstructed using a glacial geological inversion model. Journal of glaciology43, 283-299.

6. Hughes, A.L., Clark, C.D. and Jordan, C.J., 2014. Flow-pattern evolution of the last British Ice Sheet. Quaternary Science Reviews89, 148-168.

7. Rignot, E., Velicogna, I., van den Broeke, M.R., Monaghan, A. and Lenaerts, J.T., 2011. Acceleration of the contribution of the Greenland and Antarctic ice sheets to sea level rise. Geophysical Research Letters38 (5).

8. Stokes, C.R., Margold, M., Clark, C.D. and Tarasov, L., 2016. Ice stream activity scaled to ice sheet volume during Laurentide Ice Sheet deglaciation. Nature530, 322-326.

9. Ó Cofaigh, C., Evans, D.J. and Smith, I.R., 2010. Large-scale reorganization and sedimentation of terrestrial ice streams during late Wisconsinan Laurentide Ice Sheet deglaciation. GSA Bulletin122, 743-756.

10. Clark, C.D., 1999. Glaciodynamic context of subglacial bedform generation and preservation. Annals of Glaciology28, 23-32.

11. Clark, C.D and Stokes, C.R. 2003. Palaeo-ice stream landsystem. In Evans, D.J.A. (Ed.) Glacial Landsystems. Hodder–Arnold, UK.

12. Stokes, C.R. and Clark, C.D., 1999. Geomorphological criteria for identifying Pleistocene ice streams. Annals of Glaciology28, 67-74.

13. Rignot, E., Mouginot, J. and Scheuchl, B., 2011. Ice flow of the Antarctic ice sheet. Science333, 1427-1430.

14. Clark, C.D., 1993. Mega‐scale glacial lineations and cross‐cutting ice‐flow landforms. Earth Surface Processes and Landforms18, 1-29.

15. Stokes, C.R. and Clark, C.D., 2002. Are long subglacial bedforms indicative of fast ice flow? Boreas31, 239-249.

16. King, E.C., Hindmarsh, R.C. and Stokes, C.R., 2009. Formation of mega-scale glacial lineations observed beneath a West Antarctic ice stream. Nature Geoscience2, 585-588.

17. Angelis, H.D. and Kleman, J., 2008. Palaeo‐ice‐stream onsets: examples from the north‐eastern Laurentide Ice Sheet. Earth Surface Processes and Landforms, 33, 560-572.

18. Raymond, C., 1996. Shear margins in glaciers and ice sheets. Journal of Glaciology42, 90-102.

19. Schoof, C. 2004. On the mechanics of ice-stream shear margins. Journal of Glaciology50, 208-218.

20. Stokes, C.R. and Clark, C.D., 2002. Ice stream shear margin moraines. Earth Surface Processes and Landforms27, 547-558.

21. Winsborrow, M.C., Stokes, C.R. and Andreassen, K., 2012. Ice-stream flow switching during deglaciation of the southwestern Barents Sea. GSA Bulletin124, 275-290.

22. Stokes, C.R., Lian, O.B., Tulaczyk, S. and Clark, C.D., 2008. Superimposition of ribbed moraines on a palaeo‐ice‐stream bed: implications for ice stream dynamics and shutdown. Earth Surface Processes and Landforms33, 593-609.

23. Stokes, C.R. and Clark, C.D., 2003. The Dubawnt Lake palaeo‐ice stream: evidence for dynamic ice sheet behaviour on the Canadian Shield and insights regarding the controls on ice‐stream location and vigour. Boreas32, 263-279.

Introduction to glacial landsystems

What are glacial landsystems | The landsystems approach | Studying landform–sediment assemblages | Why are glacial landsystems useful? | Summary | Key terms | References

What are glacial landsystems?

Research in glacial geology has increasingly concentrated on glacial landsystems1,2. In broad terms, the landsystems concept attempts to understand how a landscape was created through investigation of the complete collection of landforms and sediments within it.

Formally, a landsystem can be defined as1:

An area with a common set of features that is different to that of neighbouring areas. This includes the surface topography (of which landforms are a part), but also its underlying sediments and soils, and overlying vegetation.

There are two key principals to the landsystem approach:

The first is that landforms (such as an esker, moraine, or roche moutonnée) and sediments are considered not in isolation, but in combination as landform–sediment assemblages that make up a landscape (see “Key terms” at bottom of page).

The second is that, in landsystems research, the emphasis is on linking landform–sediment assemblages to the processes that create them – to produce process–form models – which, when applied to a particular landscape, can even be linked to local environmental (e.g. climate) and geological (e.g. topography) controls.

The glacial tongue and foreland of Skaftafellsjökull glacier in Iceland, with assemblages of closely-spaced sawtooth push moraines. Iceland has been an important testing ground for the study of process–form models at active glacier margins (see Evans, 2003; ref. 1) Photo: Chensiyuan

The landsystems approach

For the landsystems approach to be effective, glacial geologists must carefully study the glacial landform–sediment assemblages of two settings:

One – they investigate the landforms and sediments being actively created in currently glaciated areas (such as Iceland, the European Alps, or the Southern Alps of New Zealand) to establish clear links with the glaciological processes that produce them, and;

The partially debris-covered terminus of the Fox Glacier, South Island, New Zealand, in 2013. Here, a glacial geologist could study the process–form relationships of a temperate valley glacier with a high supraglacial debris load. Photo: M. Basler

Two – they study the landforms and sediments of areas where glaciers are no longer present (such as the British Isles) to reconstruct past glacial systems and the processes that operated within them.

The cirque floor of Cwm Idwal, North Wales, with a chain of moraine ridges flanking Llyn (lake) Idwal and a staircase sequence of lateral moraines descending the opposite valley side (right of image). This is a good example of landform–sediment assemblages produced by a small cirque glacier during the Loch Lomond Stadial in upland Britain. Photo: J. Bendle.

From this, you should see that the accuracy of the landsystem approach – and the reconstruction of former glacier systems – relies on a clear understanding of the processes that create specific landform–sediment assemblages at active glaciers – i.e. on process–form relationships.

Studying landform–sediment assemblages

In practice, studying process–form relationships at active glaciers, and the application of the landsystem approach to a formerly glaciated landscape, involve broadly the same method: a detailed physical inspection of the landscape to identify landform–sediment assemblages and patterns in their spatial distribution.

This is typically achieved by mapping the type, size, and shape of landforms from aerial and satellite imagery, or through fieldwork (this can even be achieved using Google Earth imagery, making it an interesting and accessible topic for A-level and undergraduate research projects).

Google Earth image of the Skálafellsjökull glacier margin and immediate foreland (Iceland). Glacial geologists use satellite images such as this to map the type and distribution of landform–sediment assemblages.

In addition, the sedimentological characteristics of landform-assemblages – for example, the size, shape, and roundness of particles in a moraine – are recorded in the field or analysed in a laboratory. This extra information tells us much about how a feature was formed, such as whether the sediment was laid down directly by ice or by glacial meltwater2,3.

Exposure in the ‘Little Ice Age’ moraine of the Exploradores Glacier, Patagonia (South America). Glacial geologists carefully record the constituents (e.g. the size, roundness and shape of particles) of landforms such as moraines to better understand how they were created. Photo: J. Bendle

Why are glacial landsystems useful?

One of the main advantages of the landsystems approach is the ability to reconstruct – not just the size and shape of former glaciers – but the distinct characteristics of these glaciers and the processes that once operated there, more accurately and in much greater detail than is possible by studying individual landforms in isolation1.

One example would be the ability to determine the thermal regime of a former glacier based on the suite of landform-sediment assemblages it left behind. Active temperate (warm-based) glaciers, for example, are often associated with low-amplitude push, dump and squeeze moraines, flutes, drumlins, and glaciofluvial features, such as eskers and outwash plains4,5. These assemblages develop under wet-based conditions that encourage basal sliding and subglacial deformation. Cold-based glaciers, by contrast, which are frozen to their beds, produce different landform–sediment assemblages typified by ice-contact fans, thrust-block moraines, and periglacial screes, with no subglacial features like flutes and drumlins6,7.

Google Earth image of the Svínafellsjökull glacier (Iceland) showing the landform–sediment assemblages typical of an active temperate glacier margin, such as flutes and debris stripes, sawtooth push moraines, and outwash deposits (see ref. 8)

Another advantage of the landsystem approach occurs where distinctive landform-sediment assemblages overprint or overlap one another, as they contain evidence of temporal changes in glacier processes or characteristics. One example relates to assemblages of cross-cutting drumlins (and/or other subglacially moulded landforms) or bedrock striations, which indicate changes in the direction of ice flow over time.


The landsystems approach is a holistic method of studying glacier and landscape history that (i) makes inferences using the full suite of landform–sediment assemblages that constitute a landscape, and (ii) is supported by process–form models established at active glaciers.

Other pages in this section of the website give examples of the main glacial landsystems to be identified in both actively and formerly glaciated areas.

Key terms

Landform-sediment assemblage: a distinctive group of landforms and sediments that together reflect a common process or age. A glaciated landscape is typically made up of lots of distinctive landform–sediment assemblages related to different (e.g. subglacial, supraglacial, ice-marginal) processes.

Process–form model: a theoretical model (based on detailed observation) that links physical processes to the landforms and sediments they create. In glaciated systems, processes of glacier erosion, ice and debris transfer, and deposition create landform–sediment assemblages.


[1] Evans, D.J.A., 2003. Glacial landsystems. Hodder–Arnold.

[2] Benn, D.I., and Evans, D.J.A., 2014. Glaciers and glaciation. Routledge.

[3] Evans, D.J.A., and Benn, D.I., 2004. The practical guide to the study of glacial sediments. Hodder–Arnold

[4] Evans, D.J.A., and Twigg, D.R., 2002. The active temperate glacial landsystem: a model based on Breiðamerkurjökull and Fjallsjökull, Iceland. Quaternary Science Reviews21 (20-22), 2143-2177.

[5] Evans, D.J.A., 2003. Ice-marginal terrestrial landsystems: active temperate glacier margins (ed.) in Evans, D.J.A., Glacial Landsystems. Hodder–Arnold, pp. 89-110.

[6] Fitzsimons, S.J. 2003. Ice-marginal terrestrial landsystems: polar continental glacier margins (ed.) in Evans, D.J.A., Glacial Landsystems. Hodder–Arnold, pp. 89-110.

[7] Hambrey, M.J., and Fitzsimons, S.J., 2010. Development of sediment–landform associations at cold glacier margins, Dry Valleys, Antarctica. Sedimentology57 (3), 857-882.

[8] Evans, D.J.A., Ewertowski, M.W., and Orton, C., 2019. The glacial landsystem of Hoffellsjökull, SE Iceland: contrasting geomorphological signatures of active temperate glacier recession driven by ice lobe and bed morphology. Geografiska Annaler: Series A, Physical Geography101 (3), 249-276.

Types of glaciers

Earth’s glaciers are incredibly varied in their size and shape, ranging from small ice masses that cling precariously to steep mountain sides, to vast ice sheets that submerge entire continents below kilometres thick ice1,2.

The form, shape and structure – known as the morphology – of these two extreme examples, as well as all glacier types in between, is a function of two key variables: climate and topography.


Climate controls the annual temperature cycle of a region as well as the amount of precipitation that falls as snow. Because of this, climate governs the annual mass balance of glaciers and hence their size (a key part of glacier morphology).

Where climatic conditions lead to mass inputs (e.g. snowfall) that are larger than mass outputs (e.g. melting) a glacier will grow. Conversely, where mass outputs exceed mass inputs a glacier will shrink.

Regular and heavy snowfall over Monte San Valentín (4058 m) and the glacier accumulation zone(s) of the North Patagonian Icefield contribute to regional mass balance. Photo: M. Foubister

All other factors being equal, therefore, it follows that the coldest places on Earth’s surface, the polar regions, will contain the largest and most extensive glaciers. However, climate is only one part of the story.


Topography is also a major control on glacier morphology. Topography not only provides the land surface (e.g. high altitude mountains) on which glacial ice can develop, but it also controls the physical dimensions of glaciers and how they flow.

The Alps mountains above Chamonix, France, not only rise to a high enough altitude that glaciers can exist there due to the cold conditions, but the very steep slopes running away from Mont Blanc (top left) dictate the form and flow of Glacier des Bossons (front) and Glacier de Tacconaz (behind). Photo: S. Räsänen

Consider this example. A deep valley that is several kilometres in length will contain a thicker and longer ice mass than a small mountain cirque. Because of its greater thickness, this hypothetical valley glacier will, in turn, flow more rapidly because thicker ice increases driving stresses at the glacier bed and raises basal temperatures3, which increase the rate of ice deformation and basal sliding.

Types of glacier

Bearing in mind the combined influences of climate and topography in shaping glacier morphology, the broad range of glacier types at Earth’s surface fit into two main groups, known as unconstrained glaciers and constrained glaciers1,2, which are defined as follows:

  • Unconstrained glaciers have a morphology and flow pattern that is in the most part independent of underlying topography, whereas;
  • Constrained glaciers have a morphology and flow pattern that is strongly dependent on underlying topography.

Unconstrained glaciers

Ice sheets and ice caps

Ice sheets and ice caps take the same basic form, having a broad, upstanding, and slowly moving ice dome at their centre, with channels of faster moving ice that transfer mass to their margins.

Surface elevation maps of the Greenland and Antarctic ice sheets, showing their dome-like structure (IPCC, AR5)

However, they differ in terms of scale2. Ice sheets are larger, being more than 50,000 km2 in size, with ice domes that can be more than 3000 m thick. In contrast, ice caps only reach thicknesses of several hundred metres.

GoogleEarth image of the Vatnajökull ice cap, Iceland, with a central ice dome drained by valley glaciers along its southern margin. The snowline is marked by the boundary between bare ice (grey and black) and snow (white).

There are two more key features of ice sheet and ice cap morphology. Firstly, they almost completely submerge the landscape, with only the tips of mountain peaks (known as nunataks) piercing the ice surface.

Starr nunatak rising above the ice surface in the Victoria Land region of the Antarctic Ice Sheet. Photo: S. Bannister

Secondly, their flow patterns are (at least in the most part) unaffected by underlying topography. The exception to this general rule are the fast-flowing ice streams and outlet glaciers that often reside within glacial troughs closer to the periphery of ice sheets and ice caps4,5.

Ice streams

Ice streams are corridors of rapidly moving ice in an ice sheet4. A feature unique to ice streams is that they are bordered on either side not by bedrock, but by slowly moving ice.

The crevassed surface of the Recovery ice stream that drains part of the East Antarctic Ice Sheet. Photo: NASA

Ice streams are extremely large (>20 km wide and >150 km long) and when viewed from space, we observe that they are fed by numerous tributaries that are connected to a central ice dome6,7. Ice streams are critically important to the overall dynamics and mass balance of ice sheets as they control the vast majority (~90% in Antarctica) of ice and sediment discharge to the oceans4,6,7.

Ice streams of Antarctica draining the ice sheet interior. From: Rignot et al. (2011)

Constrained glaciers

Ice fields

Unlike ice caps, ice fields do not have a simple dome-like structure. Instead, their morphology and flow are controlled by topography. Ice fields (such as the Patagonian ice fields) develop in mountainous terrain where the land surface reaches an altitude that enables snow and ice to accumulate. They are drained by large valley glaciers.

The North Patagonian Icefield of southern South America with its numerous radiating valley glaciers. Image: NASA

Valley glaciers

Valley glaciers (as their name suggests) exist within bedrock valleys and are overlooked by ice-free slopes. They are found in many alpine and high mountain environments, including the European Alps, Southern Alps of New Zealand, the Andes, and the Himalayas (to name just a few).

The Aletsch Glacier (or ‘Aletschgletscher’ in German) in the Swiss Alps is a classic example of a valley glacier. Note in the upper reaches that the glacier has several cirque basin tributaries that feed the main glacier trunk. Photo: D. Beyer

Valley glaciers are fed in their upper parts by ice and snow discharged from surrounding ice fields or cirques (see the Aletsch Glacier above) in addition to snow and ice avalanches from overlooking slopes. In terms of morphology, valley glaciers can be single features or made up of a branching network of tributaries (see image below), and range in length from several kilometres to over 100 kilometres.

Large valley glaciers in Alaska (USA) seen from space by the Sentinel-2 satellite. Notice the
tributary glaciers feeding the main trunk of Columbia Glacier (centre).

Transection glaciers

Transection glaciers are, in essence, a system of interconnected valley glaciers that flow in several different directions, often in a radiating (or web-like) pattern. Transection glacier networks develop where bedrock valleys are deeply dissected, allowing ice to overflow the cols between adjacent valleys.

GoogleEarth image of the Spitsbergen island of the Svalbard archipelago, showing transection glaciers.

Examples of active transection glaciers can be found in Greenland, Svalbard (see above), and Alaska. Such systems also developed during the last glacial period in the European Alps8, and parts of the Loch Lomond Stadial ice cap in Scotland are also thought to have formed transection glacier networks9.

Piedmont glaciers

Piedmont glaciers have a distinctive form characterised by large terminal ice lobes that splay outwards onto lowland terrain after exiting a confining bedrock valley. Topography, therefore, exerts varying degrees of control on piedmont glacier morphology and flow at different points along the glacier length.

The Agassiz (left) and Malaspina (right) piedmont glacier lobes spilling out from the St. Elias mountains, Alaska, on to flat coastal plains. Image: NASA

Another common feature is that large areas of a piedmont glacier are situated below the equilibrium line altitude in the ablation zone. The Malaspina Glacier in Alaska (see image above) is the most famous example of a piedmont glacier. This glacier, which drains the Mt. St. Elias ice field, has a terminal lobe that is around 40 km long and almost 65 km across at its widest.

Cirque glaciers

Cirque glaciers are among the most common types of glacier on Earth, being found in nearly all alpine landscapes that support ice accumulation. Cirque glaciers are either localised to armchair-shaped bedrock hollows on a mountain side (see image below), or to the uppermost parts of a glacial trough, where they flow into larger valley glaciers.

Small cirque glacier (Styggebrean) in Jotunheimen National Park, Norway. Photo: J. Bendle.

The morphology of a cirque glacier largely depends on the topography in which it sits. The cirque basin itself dictates the size and shape of the cirque glacier and directs its flow, while the terrain surrounding a cirque basin is an important source of wind-blown snow and therefore glacier mass balance10.

Niche glaciers

Smaller in size than cirque glaciers, niche glaciers form where ice accumulates in a mountain side recess (or niche), such as a rock bench, couloir, or depression. Niche glaciers represent the early stages of glacier development and are commonly found in climatically favourable settings, such as in shaded north-facing slopes of mountains in the Northern Hemisphere11. Similar to niche glaciers, but adhering to steep mountain sides, are ice aprons.

Niche glacier occupying a bedrock recess at the summit of Blick vom Gatschkopf (2945 m) in the Austrian Alps. Photo: Kogo


Using GoogleEarth (or similar) explore Earth’s mountain regions and, using the definitions and images in this article, try to identify examples of unconstrained and constrained glacier types. While doing this, think about the possible climatic and topographic factors that control the size, shape, and flow of glaciers.

You may also like to compare the size of different glacier types, as well as other physical metrics such as ice surface gradient. You can do this by experimenting with the “Measure Tool” in GoogleEarth, which enables you to measure distance and area.


[1] Sugden, D.E., John, B.S., 1976. Glaciers and Landscape: Arnold.

[2] Benn, D.I., Evans, D.J.A., 2010. Glaciers and Glaciation, 2nd edition: Routledge.

[3] Cuffey, K.M. and W.S.B. Paterson, 2010. The Physics of Glaciers, 4th edition: Academic Press.

[4] Bennett, M.R., 2003. Ice streams as the arteries of an ice sheet: their mechanics, stability and significance. Earth-Science Reviews, 61, 309-339.

[5] Winsborrow, M.C.M., Clark, C.D., Stokes., C.R. 2010. What controls the location of ice streams? Earth-Science Reviews, 103, 45-59.

[6] Joughin, I., Smith, B.E., Howat, I.M., Scambos, T., Moon, T., 2010. Greenland flow variability from ice-sheet-wide velocity mapping. Journal of Glaciology, 56, 415-430.

[7] Rignot, E., Mouginot, J., Scheuchl, B., 2011. Ice flow of the Antarctic ice sheet. Science, 333, 1427-1430.

[8] Wirsig, C., Zasadni, J., Christl, M., Akçar, N., Ivy-Ochs, S., 2016. Dating the onset of LGM ice surface lowering in the High Alps. Quaternary Science Reviews, 143, 37-50.

[9] Golledge, N.R., Hubbard, A., 2005. Evaluating Younger Dryas glacier reconstructions in part of the western Scottish Highlands: a combined empirical and theoretical approach. Boreas, 34, 274-286.

[10] Lie, Ø., Dahl, S.O., Nesje, A., 2003. A theoretical approach to glacier equilibrium-line altitudes using meteorological data and glacier mass-balance records from southern Norway. The Holocene, 13, 365-372.

[11] Harrison, S., Knight, J., Rowan, A.V., 2015. The southernmost Quaternary niche glacier system in Great Britain. Journal of Quaternary Science, 30, 325-334.

Moraine formation

Ridges, mounds and hummocks formed at the margin of glaciers are generally termed moraines. The study of moraines is particularly useful as it can shed light on the physical processes occurring at both active and former ice margins1,2 and because moraines are markers of former glacier extent, so can be used to track glacier change (e.g. size) over time3.

Moraine ridge forming at the terminus of Easton Glacier, Washington, USA. Photo: W. Siegmund

How do moraines form?

Moraines form through several main processes, which may vary from glacier to glacier, on a temporal (e.g. seasonal basis), and with changes in climate. The key moraine-forming processes are shown in the diagram below and explained through this page.

Summary of the three main moraine-forming processes. Push moraines (top) form during periods of ice-front stillstand or advance that bulldoze proglacial sediments. Dump moraines (middle) consist of rock and sediment that fall, flow and slump from the ice margin by gravity. Ablation moraines (bottom) form due to the varying rates of ice melt across the snout. Where debris cover is sparse (i.e. where the ice is ‘clean’) melting is relatively rapid and the glacier surface lowers quickly. Where surface debris cover is thick, the ice is insulated from melting and ice-cored moraines can exist. Created by J. Bendle.

Push moraines

Push moraines form at the snout of active glaciers. Rock and sediment debris at the ice margin is moulded into ridges by the bulldozing of material (ice pushing) by an advancing glacier4,5.

Due to the nature of their formation, push moraines tend to take on the shape of the ice margin during the time at which they formed4,5 (see image below). They are often found at the margin of active temperate glaciers (such as those found in southern Norway and Iceland) that experience brief periods ice-front stability or advance despite a general pattern of recession4,5. In some cases, a series of annual push moraines may form, where low-relief ridges are formed during winter advances of the glacier snout, leaving behind a detailed record of glacier extent over time6-10.

Push moraine ridges formed at the retreating terminus of Skálafellsjökull, Iceland. The moraines display a ‘sawtooth’ planform that closely mimics the ice margin geometry. These push moraines have been shown to form annually, driven by local climate conditions (Chandler et al., 2016; ref. 10). Image from GoogleEarth.

Debris squeezing

As well as the bulldozing of debris, sediment may also be squeezed out from beneath the glacier margin, either as a glacier advances in winter, or in the ablation season when till becomes water-soaked and easily displaced by the weight of overlying ice4,11. This process also contributes to the formation and growth of push moraines.

Dump moraines

Dump moraines form where debris flows or falls from a glacier surface due to gravity and accumulates at the ice front or side as a ridge. They form where the ice front is stationary and there is a regular supply of debris to the snout, normally due to the melt-out of rock debris stored in the ice4.

Dump moraine size is related to the amount of debris accumulating at the snout and the length of time the glacier margin is stationary. The volume of debris on the glacier surface is high where (i) the debris content within the ice is high; and (ii) where ice velocity is high, as faster flowing ice can transfer more debris to the margin11.

Loch Lomond Stadial moraines

Debris dumped from the ice front may be bulldozed into push moraines by advance(s) of the glacier margin2,12. Moraines formed by a combination of both the dumping and pushing of debris include those constructed by certain Scottish cirque and valley glaciers during the Loch Lomond Stadial2,13 (see image below). These moraines are similar in their genesis and morphology to those created by Icelandic glaciers today, which suggests that Loch Lomond Stadial glaciers in Britain were likely temperate and active during deglaciation2,13.

Loch Lomond Stadial moraine ridges formed by a combination of the dumping and bulldozing of rock and sediment debris, Coire Ardair. Photo: B. Davies.

Ablation moraines

Ablation moraines form where rock and sediment debris accumulate on the glacier surface near the margin and subsequently undergo melt-out4,11. The accumulation of dark-coloured material on the glacier surface lowers the ice albedo (i.e. its reflectiveness) and increases the amount of solar radiation absorbed at the glacier surface, which causes ice melt to speed up. However, where the debris layer is more than a few centimetres thick it insulates the ice surface from heating, slowing the rate of ice melt. Where the debris cover is extensive across a large part of the snout, the ice margin may detach completely from the main body of the glacier and become stagnant (see image below).

The debris-covered and stagnant ice margin of Exploradores Glacier in central Patagonia, Chile. Photo: J. Bendle.

When a debris-covered snout melts over time material is gradually let down from the ice surface to produce an area of ‘hummocky moraine’. This melt-out process can produce a variety of moraine types, from a chaotic assortment of sediment mounds and hollows (see image below)1 to more regular transverse ridges (often termed controlled moraines) that reflect the former pattern of debris in a parent glacier14.

Example of chaotic mounds and hollows in southern Patagonia, South America, which are interpreted to have formed by ice stagnation (see Darvill et al., 2017; ref. 15).


[1] Kjær, K.H. and Krüger, J., 2001. The final phase of dead‐ice moraine development: processes and sediment architecture, Kötlujökull, Iceland. Sedimentology48, 935-952.

[2] Lukas, S., 2005. A test of the englacial thrusting hypothesis of ‘hummocky’ moraine formation: case studies from the northwest Highlands, Scotland. Boreas34, 287-307.

[3] Schomacker, A. 2011. Moraine (Eds.) Singh, V.P., Singh, P. and Haritashya, U.K. Encyclopedia of Snow, Ice and Glaciers. Springer.

[4] Benn, D.I. and Evans, D.J.A., 2010. Glaciers and Glaciation. Hodder Education. 

[5] Boulton, G.S., 1986. Push‐moraines and glacier‐contact fans in marine and terrestrial environments. Sedimentology33, 677-698.

[6] Sharp, M., 1984. Annual moraine ridges at Skálafellsjökull, south-east Iceland. Journal of Glaciology30, 82-93.

[7] Bradwell, T., 2004. Annual moraines and summer temperatures at Lambatungnajökull, Iceland. Arctic, Antarctic, and Alpine Research36, 502-508.

[8] Beedle, M.J., Menounos, B., Luckman, B.H. and Wheate, R., 2009. Annual push moraines as climate proxy. Geophysical Research Letters36.

[9] Lukas, S., 2012. Processes of annual moraine formation at a temperate alpine valley glacier: insights into glacier dynamics and climatic controls. Boreas, 41, 463-480.

[10] Chandler, B.M., Evans, D.J. and Roberts, D.H., 2016. Characteristics of recessional moraines at a temperate glacier in SE Iceland: Insights into patterns, rates and drivers of glacier retreat. Quaternary Science Reviews135, 171-205.

[11] Bennett, M.M. and Glasser, N.F. 2009. Glacial Geology: Ice Sheets and Landforms. John Wiley & Sons.

[12] Boulton, G.S. and Eyles, N., 1979. Sedimentation by valley glaciers: a model and genetic classification. Moraines and varves33, pp. 11-23.

[13] Jones, R.S., Lowe, J.J., Palmer, A.P., Eaves, S.R. and Golledge, N.R., 2017. Dynamics and palaeoclimatic significance of a Loch Lomond Stadial glacier: Coire Ardair, Creag Meagaidh, Western Highlands, Scotland. Proceedings of the Geologists’ Association128, 54-66.

[14] Evans, D.J.A., 2009. Controlled moraines: origins, characteristics and palaeoglaciological implications. Quaternary Science Reviews28, 183-208.

[15] Darvill, C.M., Stokes, C.R., Bentley, M.J., Evans, D.J. and Lovell, H., 2017. Dynamics of former ice lobes of the southernmost Patagonian Ice Sheet based on a glacial landsystems approach. Journal of Quaternary Science32, 857-876

Moraine types

Moraines are distinct ridges or mounds of debris that are laid down directly by a glacier or pushed up by it1. The term moraine is used to describe a wide variety of landforms created by the dumping, pushing, and squeezing of loose rock material, as well as the melting of glacial ice.

Moraine ridges on the forefield of the Matanuska Glacier, Alaska. Photo: Frank K.

In terms of size and shape, moraines are extremely varied. They range from low-relief ridges of ~1 m high and ~1 m wide formed at the snout of actively retreating valley glaciers2, to vast ‘till plains’ left behind by former continental ice sheets3.

Low-relief moraine ridges on the forefield of the actively retreating Skaftafellsjökull Glacier in Iceland. The moraines mark former ice extent and mirror the shape of the glacier terminus at the time of formation. Photo: TommyBee

Moraines consist of loose sediment and rock debris deposited by glacier ice, known as till. They may also contain slope, fluvial, lake and marine sediments if such material is present at the glacier margin, where it may be incorporated into glacial ice during a glacier advance, or deformed by glacier movement4,5.

Moraine composed of loose rock and sediment forming at the lateral margin of the Boulder Glacier, Washington, USA. Photo: W. Siegmund.

Moraines are important features for understanding past environments. Terminal moraines, for example, mark the maximum extent of a glacier advance (see diagram below) and are used by glaciologists to reconstruct the former size of glaciers and ice sheets that have now shrunk or disappeared entirely6.

Summary of the main moraine types and their spatial patterns. The top diagram is a cross-section through a cirque glacier. The bottom diagram is drawn in plan view, looking down on the surface of a valley glacier made up of several tributaries. Image created by J. Bendle.


The most common moraine types are defined below:

A terminal moraine is a moraine ridge that marks the maximum limit of a glacier advance. They form at the glacier terminus and mirror the shape of the ice margin at the time of deposition. The largest terminal moraines are formed by major continental ice sheets and can be over 100 m in height and 10s of kilometres long7,8.

Terminal moraine marking the limit of the former Patagonian Ice Sheet at the Last Glacial Maximum (~25 to 18 thousand years ago). Photo: J. Bendle.

Recessional moraines are found behind a terminal moraine limit and form during short-lived phases of glacier advance or stillstand that interrupt a general pattern of glacier retreat. In some cases, recessional moraines form on a yearly basis (normally as a result of winter glacier advances) and are known as annual moraines9,10,11.

Recessional moraines (arrowed) marking the shrinkage of a South American valley glacier. The glacier (not shown) retreated towards the south-west, leaving behind a moraine-dammed glacial lake. Imagery from GoogleEarth, diagram created by J. Bendle.

Lateral moraines form along the glacier side and consist of debris that falls or slumps from the valley wall or flows directly from the glacier surface12 (see image below). Where the rate of debris supply is high, lateral moraines can reach heights of more than 100 metres12–15.

Lateral moraine of the Callequeo Glacier of the San Lorenzo Icefield in central Patagonia, South America. Photo: J. Martin.

The term latero-frontal moraine is used where debris builds up around the entire glacier tongue14. These moraine types are common in mountain settings such as the European Alps, the Southern Alps of New Zeland (see the Mueller Glacier moraines below) and the Himalayas, where the high supply of rock debris from unstable valley sides, rapidly build up at the glacier margins.

Latero-frontal moraine complex of the Mueller Glacier, South Island, New Zealand. The debris-covered and downwasting Mueller Glacier is flanked by lateral moraines of ~100 m in height, which continue down valley and merge into terminal moraines. Imagery from GoogleEarth, diagram created by J. Bendle.

Medial moraines are debris ridges at the glacier surface running parallel to the direction of ice flow4,5. They are the surface (or supraglacial) expression of debris contained within the ice. Medial moraines form where lateral moraines meet at the confluence of two valley glaciers, or where debris contained in the ice is exposed at the surface due to melting in the ablation zone16.

Medial moraines on the surface of an Alaskan valley glacier. In this example, surface debris is concentrated at the point where two glaciers merge. Imagery from GoogleEarth, diagram created by J. Bendle.

Ground moraine is a term used to describe the uneven blanket of till deposited in the low-relief areas between more prominent moraine ridges6. This type of moraine, which is also commonly referred to as a till plain, form at the glacier sole as due to the deformation and eventual deposition of the substratum.


1. Hambrey, M. J. 1994. Glacial Environments. UCL Press.

2. Krüger, J., Schomacker, A. and Benediktsson, Í.Ö., 2010. 6 Ice-Marginal Environments: Geomorphic and Structural Genesis of Marginal Moraines at Mýrdalsjökull. Developments in Quaternary Sciences13, 79-104.

3. Dyke, A.S. and Prest, V.K. 1987. Late Wisconsinan and Holocene history of the Laurentide Ice Sheet. Geographie Physique et Quaternaire XLI, 237–63.

4. Benn, D.I. and Evans, D.J.A., 2010. Glaciers and Glaciation. Hodder Education. 

5. Bennett, M.M. and Glasser, N.F. 2011. Glacial Geology: Ice Sheets and Landforms. John Wiley & Sons.

6. Schomacker, A. 2011. Moraine (Eds.) Singh, V.P., Singh, P. and Haritashya, U.K. Encyclopedia of Snow, Ice and Glaciers. Springer.

7. Dyke, A.S., Andrews, J.T., Clark, P.U., England, J.H., Miller, G.H., Shaw, J. and Veillette, J.J., 2002. The Laurentide and Innuitian ice sheets during the last glacial maximum. Quaternary Science Reviews21, 9-31.

8. Glasser, N.F., Jansson, K.N., Harrison, S. and Kleman, J., 2008. The glacial geomorphology and Pleistocene history of South America between 38°S and 56°S. Quaternary Science Reviews27, 365-390.

9. Sharp, M., 1984. Annual moraine ridges at Skálafellsjökull, south-east Iceland. Journal of Glaciology30, 82-93.

10. Bradwell, T., 2004. Annual moraines and summer temperatures at Lambatungnajökull, Iceland. Arctic, Antarctic, and Alpine Research36, 502-508.

11. Beedle, M.J., Menounos, B., Luckman, B.H. and Wheate, R., 2009. Annual push moraines as climate proxy. Geophysical Research Letters36.

12. Lukas, S., Graf, A., Coray, S. and Schlüchter, C., 2012. Genesis, stability and preservation potential of large lateral moraines of Alpine valley glaciers–towards a unifying theory based on Findelengletscher, Switzerland. Quaternary Science Reviews38, 27-48.

13. Benn, D.I. and Owen, L.A., 2002. Himalayan glacial sedimentary environments: a framework for reconstructing and dating the former extent of glaciers in high mountains. Quaternary International97, 3-25.

14. Benn, D.I., Kirkbride, M.P., Owen L.A. and Brazier, V. 2003. Glaciated Valley Landsystems (Ed.) Glacial Landsystems, Arnold, London.

15. Evans, D.J., Shulmeister, J. and Hyatt, O., 2010. Sedimentology of latero-frontal moraines and fans on the west coast of South Island, New Zealand. Quaternary Science Reviews29, 3790-3811.

16. Eyles, N. and Rogerson, R.J., 1978. A framework for the investigation of medial moraine formation: Austerdalsbreen, Norway, and Berendon Glacier, British Columbia, Canada. Journal of Glaciology20, 99-113.


Glacial cirques, known locally as corries or coires (Scotland) and cwms (Wales), are large-scale erosional features common to many mountainous regions1,2. Classic cirques take the form of armchair-shaped hollows (see image below), with a steep headwall (which often culminates in a sharp ridge, or arête) and a gently-sloping or overdeepened valley floor (see diagram below).

Classic glacial cirque basin. Cwm Clyd in the Glyderau mountains of Snowdonia. Image from GoogleEarth.
Cross-section of a classic glacial cirque with an overdeepened (and lake filled) valley floor and a steep headwall mantled with slope deposits, such as scree. Image created by J. Bendle based on Barr & Spagnolo (2015; ref. 2)

In actively glacierized terrain, cirques are important basins for the accumulation of snow. They may host small cirque glaciers (see image below) that are confined to their bedrock hollows, or act as the source area for larger valley glaciers.

Cirque glacier (Styggebrean) in Jotunheimen National Park, Norway. Photo: J. Bendle.

In other mountainous areas, such as the British uplands, the occurrence of ice-free cirques (see image below) serve as a reminder of past glacier activity by recording former sites of glacier build-up3,4,5.

Cwm Cau, a formerly glacierized cirque basin in Snowdonia, Wales. Photo: J.Bendle.

Types of cirques

Far from being the same in all mountain areas, a wide range of cirque types occur. The most common are1,6:

  • Simple cirques, which are distinct and independent features
  • Compound cirques, where the upper part of a cirque basin contains two similarly sized simple cirques
  • Cirque complexes, where the upper part of a cirque basins contains more than two similarly sized simple cirques
  • Staircase cirques, where one cirque occurs above another
  • Cirque troughs, where a cirque basin occurs at the upper end of a glacial trough
Different types of glacial cirques. The top three examples are drawn in plan view, whereas the bottom two are drawn in cross-section. Image created by J. Bendle based on Barr & Spagnolo (2015; ref. 2)

The formation and growth of cirques

Cirques form through the gradual expansion of mountainside hollows associated with earlier fluvial, volcanic, or mass movement (e.g. landsliding) activity7. When these hollows become filled with snow8 they start to enlarge by nivation (a group of processes that includes freeze-thaw activity, chemical weathering, and seasonal snow melt)9.

True cirque growth only occurs once the thickness of snow patches increases to a point at which glacier ice can form by compaction. Once formed, glaciers widen and deepen cirques by subglacial abrasion and quarrying of the hollow floor and lower headwall3 (see diagram below). Cirques can also grow by backwards headwall erosion (wear back) due to frost-action, free-thaw, and mass movement3,10.

Cirque glaciers erode their hollows by subglacial plucking and abrasion, which are most effective under a warm-based, sliding glacier. Meltwater that drains to the bed through the randkluft (the gap between the glacier and headwall), bergshrund (a large crevasse near, but not touching, the headwall) or other crevasses, promotes subglacial erosion. Periglacial erosion (e.g. freeze-thaw) occurs on the headwall and in the randkluft. Image created by J. Bendle.

Case study: glacial cirques of Snowdonia

The glacial cirques of Snowdonia formed over several glaciations, and have a long history of investigation, first being visited by Charles Darwin over 150 years ago11. The most recent period of glacier activity in Snowdonia was during the mountain glaciation of upland Britain in the Loch Lomond Stadial (between ~12 and 10 thousand years ago)5,12,13.

Loch Lomond Stadial (~12 to 10 thousand years ago) cirque glaciers in Snowdonia, North Wales. Image from Bendle & Glasser (2012; ref. 5)

Why are cirques important?

Because cirques are areas of snow accumulation, the direction in which they point (their aspect) can tell us something about the links between climate and glacier growth in the past2,14.

If looking from above (see image above), an interesting observation is that most cirques in Snowdonia face to the north or east14 and these also held most (as well as the largest) Loch Lomond Stadial glaciers5,12.

Controls on cirque aspect

This is due to two factors. Firstly, north-facing cirques receive less solar radiation than south-facing cirques (in the Northern Hemisphere), resulting in lower air temperatures and less ice-melt across the year15.

Secondly, where prevailing winds blow mainly from the west, the snow on high ground will be blown down into east-facing cirques, adding to glacier mass5,15.


Using GoogleMaps or GoogleEarth, enter “Snowdon” in the navigation search bar and explore the cirques of Snowdonia.

Try to identify different cirque types (e.g. ‘simple’, ‘compound’, ‘complex’), and compare their sizes, shapes, and aspects.


[1] Benn, D.I. and Evans, D.J.A., 2010. Glaciers and Glaciation. Hodder Arnold.

[2] Barr, I.D. and Spagnolo, M., 2015. Glacial cirques as palaeoenvironmental indicators: their potential and limitations. Earth-Science Reviews151, 48-78.

[3] Evans, I.S., 2006. Allometric development of glacial cirque form: geological, relief and regional effects on the cirques of Wales. Geomorphology80, 245-266.

[4] Ballantyne, C.K., 2007. Loch Lomond Stadial glaciers in North Harris, Outer Hebrides, North-West Scotland: glacier reconstruction and palaeoclimatic implications. Quaternary Science Reviews26, 3134-3149.

[5] Bendle, J.M. and Glasser, N.F., 2012. Palaeoclimatic reconstruction from Lateglacial (Younger Dryas Chronozone) cirque glaciers in Snowdonia, North Wales. Proceedings of the Geologists’ Association123, 130-145.

[6] Gordon, J.E., 1977. Morphometry of cirques in the Kintail-Affric-Cannich area of northwest Scotland. Geografiska Annaler: Series A, Physical Geography59, 177-194.

[7] Turnbull, J.M. and Davies, T.R., 2006. A mass movement origin for cirques. Earth Surface Processes and Landforms 31, 1129-1148.

[8] Sanders, J.W., Cuffey, K.M., MacGregor, K.R. and Collins, B.D., 2013. The sediment budget of an alpine cirque. Geological Society of America Bulletin125, 229-248.

[9] Thorn, C.E., 1976. Quantitative evaluation of nivation in the Colorado Front Range. Geological Society of America Bulletin87, 1169-1178.

[10] Sanders, J.W., Cuffey, K.M., Moore, J.R., MacGregor, K.R. and Kavanaugh, J.L., 2012. Periglacial weathering and headwall erosion in cirque glacier bergschrunds. Geology40, 779-782.

[11] Darwin, C.R., 1842. Notes on the effects produced by the ancient glaciers of Caernarvonshire, and on the boulders transported by floating ice Lond. Edinb. Dublin Philos. Mag. J. Sci. 21, 180-188.

[12] Gray, J.M., 1982. The last glaciers (Loch Lomond Advance) in Snowdonia, N. Wales. Geological Journal17, 111-133.

[13] Hughes, P.D., 2009. Loch Lomond Stadial (Younger Dryas) glaciers and climate in Wales. Geological Journal44, 375-391.

[14] Evans, I.S., 2006. Local aspect asymmetry of mountain glaciation: a global survey of consistency of favoured directions for glacier numbers and altitudes. Geomorphology73, 166-184.

[15] Evans, I.S., 1977. World-wide variations in the direction and concentration of cirque and glacier aspects. Geografiska Annaler: Series A, Physical Geography59, 151-175.

Glaciers as a water resource

Mountains as Water Towers of the World

In many mountainous parts of the world with a seasonal rainfall, glaciers are a reliable water resource in the dry season. Mountains could be called the “Water Towers of the World”1, providing water from glacier melt and orographic rainfall to lowland regions. 

Glacierised drainage basins cover 26% of the global land surface outside of Greenland and Antarctica, and are populated by almost one-third of the World’s population2. Upland areas (above 2000 m above sea level) in southeast Asia supply the five basins of the Indus, Ganges, Yellow, Brahmaputra and Yangtze rivers, providing water to 1.4 billion people (over 20 % of the global population).

The Himalayan river basins and the number of people living in each one.

(Source: Redrawing the map of the world’s international river basins)

High Mountain Asia river basins


Glacier meltwater and runoff

Glacier meltwater and runoff contribute to and module downstream water flow, affecting freshwater availability for irrigation, hydropower, and ecosystems3.

Glacier runoff is typically seasonal, with a minimum in the snow-accumulation season, and a maximum in the melt season. This meltwater can compensate for seasons or years with low streamflow or droughts in downstream regions4.

Mountain glacier and lake in Peru

Global glacier recession

Mountain glaciers around the World are currently shrinking5-7, and this is expected to continue throughout the next century. Globally, glaciers are shrinking by 227 ± 32 gigatonnes per year8, enough to raise global sea levels by 0.63 ± 0.08 mm per year.

The areas shrinking fastest are in north America (-50 gigatonnes per year), northern Arctic Canada (60 gigatonnes per year), the Himalaya region (26 gigatonnes per year), and South America (29 gigatonnes per year)8.

World glaciers and ice sheets mass balance. Glaciers are shown in black. Green circles show glacier area, red circles are how much ice is lost annually.

Glacier “Peak meltwater”

As glaciers shrink, meltwater is released from storage within the glacier. Annual meltwater therefore increases, until a maximum is reached3,9. This maximum has been called ‘Peak Meltwater’.

After Peak Meltwater, runoff decreases as smaller glacier volumes can no longer support rising meltwater volumes. As the glacier retreats and disappears, annual runoff from direct precipitation may return to something like the original value, as water is no longer stored as snow. However melt-season runoff may decline substantially, as the glacier no longer acts as a reservoir. Seasonality of water availability may therefore increase, leading to droughts in dry years or dry seasons.

Essentially, as the glaciers shrink, they provide less and less melt water from long-term storage, which impacts seasonal freshwater availability3.  

Peak Meltwater and glacier recession under a warming climate.

Adapted from Huss and Hock (2018) and Rowan et al. (2018).

The degree to which glacier runoff contributes to downstream meltwater varies according to the basin, with glacier contributions being as much as 25% of the annual water budget. In many of these basins, peak meltwater is expected to have passed (e.g., Ref. 10), or will be passed in the next 20-30 years (e.g. Ref. 11). Ultimately, some projections suggest that up to half of the world’s population could be living in water scarcity by 2100 AD12.

Meltwater stream on Mendenhall Glacier, Alaska. From: Gillfoto, Wikimedia Commons

Global scale peak meltwater?

A recent study by Huss and Hock (2018, Nature Climate Change) computed glacier runoff changes for the Earth’s 56 large-scale (>5000 km2) glacierised drainage basins with at least 30 km2 of ice to 2100 AD, and analysed the effect of glacial recession on streamflow.

In half of the basins, peak meltwater has already been reached. In the remaining basins, the modelled annual glacier runoff continues to rise until the maximum is reached, and then runoff declines. Peak water tends to occur later in basins with larger glaciers and higher ice-cover fractions3.

The researchers used a glacier model and climate model outputs forced by three different emissions scenarios, with peak emissions occurring at 2020 AD (RCP 2.6), 2050 AD (RCP 4.5) and after 2100 AD (RCP 8.5)3.  RCP 2.6 is the closest scenario to the targets of the Paris 2015 climate agreement. Projected temperature increases between 1990-2010 and 2080-2100, range from 1.6 ± 1.1°C (RCP 2.6) to 5.4 ± 2.2°C.

Between 2010 and 2100 AD, glacier volume in the 56 investigated basins was projected to decrease by 43±14% (RCP 2.6), 58±13% (RCP 4.5) and 74±11% (RCP 8.5). For the mid-range RCP 4.5, glacier volume reductions in the individual basins ranged from 37 to 99%1.

Reaching Peak meltwater

Peak meltwater has already been reached in 45% of the basins (year 2017 AD), but annual runoff is expected to continue to rise beyond 2050 AD in 22% of the basins. Basins with larger glaciers and high glacier cover (e.g. Susitna, Jökulsá) tend to reach peak meltwater towards the end of the twenty-first century.

In basins dominated by small glaciers (e.g. western Canada, central Europe, South America), peak meltwater has already passed and meltwater will decline over coming decades.

In most basins fed by High Mountain Asia (Aral Sea, Indus, Tarim, Brahmaputra), annual glacier runoff is projected to rise until the middle of the century, followed by steadily declining glacier meltwater runoff thereafter3.

By the end of the twenty-first century, the seasonal glacier runoff maximum is reduced in 93% of the basins compared with the 1990-2010 average, and runoff is less concentrated during the melt season.

Colours show the modelled year of peak water computed from 11-year moving averages of annual glacier runoff from all the glaciers located in the 56 investigated drainage basins, aggregated in 0.5 × 0.5° grid cells. Peak water is also shown with grey scales for all the macroscale basins, classified in 30-year intervals. The results refer to runoff from the initially glacierized area, and are based on the multimodel mean of 14 GCMs and the RCP4.5 emission scenario. The numbers in brackets below the basin names refer to basin glacierization in per cent. The insets show the modelled annual glacier runoff normalized with the average runoff in 1990–2010 for three selected basins. Triangles depict peak water (± standard deviation), thin lines show results for individual GCMs and G denotes the percentage ice cover. From Huss and Hock, 2018

In 19 of the 56 basins, the glacier runoff change between 2000 and 2090 AD accounts for at least one melt-season month with a reduction in runoff of at least 10% (i.e. glacier runoff reduction exceeds 10% of the basin runoff). This is sufficient to cause water scarcity in these basins.

The most significantly affected basins are in High Mountain Asia (Aral Sea, Indus, Tarim, Balkhash), Peru (Santa), South America (Colorado, Baker, Santa Cruz), and North America (Fraser, Skeena, Taku, Nass)3.

The ratio of glacier runoff change to basin runoff is evaluated for the period July to October (January to April for the southern hemisphere, and throughout the year in the tropics). For basins with substantial glacier runoff decreases in at least one month, the ratio refers to the month (given in brackets below the basin names) with the largest reduction in glacier runoff. Basins with negligible glacier impact (|ΔQ′g/Qbasin|< 5%) are shown in grey, and the remaining basins, which show glacier runoff increases that exceed 5% in at least one month, in dark blue. The results refer to multi-GCM means and RCP4.5. Small dots refer to population density > 100 km−2 on a 0.5 × 0.5° grid as an indicator for potential downstream socio-environmental impacts.

Case study: Glaciers and water resources in the Himalaya

In the Himalaya, Karakorum and Hindu Kush mountains, millions of people rely on the 90,000 glaciers as a water resource9. These glaciers form the headwaters of the Indus, Ganges and Brahmaputra rivers. Glaciers here are highly sensitive to climate change, and are rapidly shrinking7,13. The developing countries in these catchments use this water for agriculture and hydropower, and are vulnerable to changes in their water supply14.

The contribution of glaciers to runoff varies in each basin, ranging from 18.8% in the Dudh Koshi catchment (a major tributary to the Ganges), to 80.6% in the Hunza catchment, which drains into the Indus basin9.

In High Mountain Asia, the glacial ice acts to protect against extreme water shortages on seasonal and longer timescales, because the glacial melt is sustained through droughts while all other stores of water in the basin decline14. Hydrological modelling predicts a decline in glacial meltwater contribution to the overall catchment hydrology by 2065 AD of -8% in the Indus, -18% in the Ganges and -20% in the Brahmaputra1.

In southern China, just north of the border with Nepal, one unnamed Himalayan glacier flows from southwest to northeast, creeping down a valley to terminate in a glacial lake. At the end of the glacier’s deeply crevassed snout sits a glacial lake, coated with ice in this wintertime picture. Just as nearby mountains cast shadows, the crevassed glacier casts small shadows onto the lake’s icy surface. This glacial lake is bound by the glacier snout on one end, and a moraine—a mound formed by the accumulation of sediments and rocks moved by the glacier—on the other. Source:

Glacier water resources in the Indus catchment

In the westerly Indus catchment, meltwater dominates water inputs during drought summers, and predicted glacier loss will add considerably to drought-related water stress14. The Indus and Aral basins are dominated by wet winters, dry summers, and have extensive glaciation14. The summer monsoon in these more westerly basins is also less dominant than that further east.

Map of the Indus River basin with tributaries labeled. Yellow regions are non-contributing parts of the watershed (e.g. the Thar Desert). From Wikimedia Commons (Keenan Pepper,

In these basins, the highest proportion of glacial melt to overall basin hydrology occurs in the upper basins, closer to the glaciers. In the Indus basin, two thirds of the population (>120 million people) lives in the middle altitudes, where glacial meltwater is more significant. The use of water for hydropower and irrigation is concentrated at dams and barrages with average altitudes of 936-1484 m above sea level, in these middle altitudes14. In the Indus, 121 of 143 existing or planned dams are glacier-fed. In the upper Indus, without glaciers, summer monthly water flows would be reduced by 38%, and up to 58% in drought years. Water stress is likely to peak in the relatively dry summers in drought years as the glacier melt declines.

Further reading


1              Immerzeel WW, van Beek L P H & P, B. M. F. Climate change will affect the Asian water towers. Science 328, 1382–1385 (2010).

2              Beniston, M. Climatic Change in Mountain Regions: A Review of Possible Impacts. Climatic Change 59, 5-31, doi:10.1023/a:1024458411589 (2003).

3              Huss, M. & Hock, R. Global-scale hydrological response to future glacier mass loss. Nature Climate Change 8, 135-140, doi:10.1038/s41558-017-0049-x (2018).

4              Barnett, T. P., Adam, J. C. & Lettenmaier, D. P. Potential impacts of a warming climate on water availability in snow-dominated regions. Nature 438, 303, doi:10.1038/nature04141 (2005).

5              Bamber, J. L., Westaway, R. M., Marzeion, B. & Wouters, B. The land ice contribution to sea level during the satellite era. Environmental Research Letters (2018).

6              Zemp, M. et al. Historically unprecedented global glacier decline in the early 21st century. Journal of Glaciology 61, 745-762, doi:10.3189/2015JoG15J017 (2015).

7              Zemp, M. et al. Global glacier mass changes and their contributions to sea-level rise from 1961 to 2016. Nature, 1 (2019).

8              Gardner, A. S. et al. A Reconciled Estimate of Glacier Contributions to Sea Level Rise: 2003 to 2009. Science 340, 852-857, doi:10.1126/science.1234532 (2013).

9              Rowan, A. V. et al. The sustainability of water resources in High Mountain Asia in the context of recent and future glacier change. Geological Society, London, Special Publications 462, 189-204 (2018).

10           Frans, C. et al. Implications of decadal to century scale glacio‐hydrological change for water resources of the Hood River basin, OR, USA. Hydrological processes 30, 4314-4329 (2016).

11           Immerzeel, W., Pellicciotti, F. & Bierkens, M. Rising river flows throughout the twenty-first century in two Himalayan glacierized watersheds. Nature geoscience 6, 742 (2013).

12           Hejazi, M. I. et al. Integrated assessment of global water scarcity over the 21st century under multiple climate change mitigation policies. Hydrology and Earth System Sciences 18, 2859-2883 (2014).

13           Bolch, T. et al. The state and fate of Himalayan Glaciers. Science 336, 310-314 (2012).

14           Pritchard, H. D. Asia’s glaciers are a regionally important buffer against drought. Nature 545, 169-174, doi:10.1038/nature22062 (2017).

Roches moutonnées

Roches moutonnées are asymmetric bedrock bumps or hills with a gently sloping and abraded upglacier (stoss) face and a quarried (or plucked) downglacier (lee) face that is typically blunter1,2. A good example of a roche moutonnée is shown in the image below.

Roche moutonnée from near Castle Loch, southwest Scotland, with a gently sloping (abraded) stoss face and a blunt (quarried) lee face. Ice flow was from left to right. Photo: David Baird

Roches moutonnées range in size from several metres to several hundreds of metres across, and often occur in clusters1 (see image below). They may be found emerging from beneath actively deglaciating ice masses (see image below), or on the sides and bottom of deglaciated valleys where they were once overridden by glacial ice3,4. Their distinctive form, which is partly linked with the orientation of glacier flow, make roches moutonnées useful to glaciologists aiming to reconstruct the flow direction of former glaciers.

Cluster of roches moutonnées (white arrows) in Porsangerfjorden, northern Norway. Ice flow was from right to left. Photo: Arnstein Rønning

Roches moutonnées emerging from beneath Goldbergkees glacier (Austria) as the ice thins and retreats. Photo: Ewald Gabardi

How do roches mountonnées form?

Roches mountonnées develop their distinctive morphology due to the pattern of stress on a bedrock surface beneath a sliding glacier, as shown in the diagram below. On the stoss side of bedrock bumps, normal stresses are relatively high and particles embedded in the ice are moved across the underlying surface where they carry out abrasion5,6. The evidence of such abrasion is the common occurrence of striations (i.e. scores and scratches on bedrock) on the sloping upper surface and flanks of roches moutonnées (see image below).

Formation of a roche moutonnée as a result of stress differences over the bedrock surface. High normal stress (pressure) on the stoss face results in bedrock abrasion, whereas lower normal stresses (pressure) on the lee face often allow a cavity to form, which promotes quarrying of bedrock along lines of existing weakness (e.g. bedrock joints). Diagram: Jacob M. Bendle

Striations on the flank of a roche moutonnée in Mount Rainier National Park, USA, giving evidence of glacial abrasion. Photo: Walter Siegmund

On the lee side of bedrock bumps, normal stresses are lower, which allows a cavity to form between the ice and bed (see diagram above) and prevents abrasion. In its place, bed cavities increase stress build up in the bedrock immediately upstream of the cavity, causing rock fracture and erosion by quarrying (or plucking). This process is particularly efficient where water pressure at the bed regularly changes3,7,8,9 (see diagram below).

The importance of bed cavities in roche moutonnée formation. In T1, the water pressure (pw) present in the bed cavity in the lee of a bedrock bump offsets the downward directed ice overburden pressure (pi), preventing bedrock fracture. However, in T2, the water has drained, and a high stress zone (red) develops in the bedrock around the cavity edge, which causes rock fracture and quarrying (plucking) to occur. Diagram: Jacob M. Bendle

The quarrying of rock at the lee end of roches mountonnées is also strongly influenced by the joint distribution in the parent rock3, and determines the size and shape of quarried rock fragments (see diagram below).

The importance of bedrock joint structure in the evolution of quarrying of a roche moutonnée lee face. The orange lines depict the progressive upglacier migration of the lee face as bedrock fragments are progressively plucked along lines of weakness (joints). Diagram: Jacob M. Bendle

What do roches mountonnées tell us about former glaciers?

Through an understanding of how roches mountonnées are formed, glaciologists are able to make inferences about the nature of past glacier systems where such landforms are found.

As roches mountonnées are most likely to form where cavities exist at the glacier bed, it is common for them to develop where the ice overburden pressure is low (i.e. where ice is relatively thin). Such conditions occur beneath thin cirque or valley glaciers, or near the margins of ice sheets3,4,10. This also means that roches moutonnées may be more likely to develop during deglaciation, when a glacier or ice sheet thins, ice overburden pressure decreases, and gaps between the ice and bed open up11 (see diagram below).

During full glacial conditions, when ice is at its thickest, ice overburden pressure (pi) is high and the glacier presses down into bumps in the bed. As the ice thins during deglaciation, the ice overburden pressure (pi) decreases and cavities open up at the bed, promoting favourable conditions for roche moutonnée formation. Diagram: Jacob M. Bendle (based on Roberts and Long, 2005)

Because roche mountonnée formation is also aided by fluctuations in basal water pressure, they are most likely to occur beneath warm-based (temperate) glaciers with hydrological systems that direct meltwater the bed10. The fact that they contain abraded (i.e. polished and striated) surfaces (see image above) also informs glaciologists that the ice responsible for their formation was (at least at times) warm based and moving by basal sliding, as well as carrying a basal debris load.


[1] Bennett, M.R., and Glasser, N.F. (2009) Glacial Geology: Ice Sheets and Landforms. Wiley-Blackwell.

[2] Benn, D.I., and Evans, D.J.A. (2010) Glaciers and Glaciation. Routledge.

[3] Sugden, D.E., Glasser, N.F., and Clapperton, C.M. (1992) Evolution of large roches moutonnées. Geografiska Annaler, 74A, 253-264.

[4] Glasser, N.F. (2002) Scottish Landform Example 28: The large roches moutonnées of upper Deeside. Scottish Geographical Journal, 118, 129-38.

[5] Boulton, G.S. (1974) Processes and patterns of glacial erosion. In Glacial Geomorphology (ed. D.R. Coates) Springer, Dordrecht, pp. 41-87.

[6] Hallet, B. (1979) A theoretical model of glacial abrasion. Journal of Glaciology, 23, 39-50.

[7] Iverson, N.R. (1991) Potential effects of subglacial water pressure fluctuations on quarrying. Journal of Glaciology, 37, 27-36.

[8] Hallet, B. (1996) Glacial quarrying: a simple theoretical model. Annals of Glaciology, 22, 1-8.

[9] Cohen, D., Hooyer, T.S., Iverson, N.R., Thomason, J.F. and Jackson, M. (2006). Role of transient water pressure in quarrying: A subglacial experiment using acoustic emissions. Journal of Geophysical Research: Earth Surface, 111(F3).

[10] Hall, A.M., and Glasser, N.F. (2003) Reconstructing the basal thermal regime of an ice stream in a landscape of selective linear erosion: Glen Avon, Cairngorm Mountains, Scotland. Boreas, 32, 191-208.

[11] Roberts, D.H., and Long, A.J. (2005) Streamlined bedrock terrain and fast flow, Jakobshavns Isbrae, West Greenland: implications for ice stream and ice sheet dynamics. Boreas, 34, 25-42.


Subglacial erosion

What is subglacial erosion?

Subglacial erosion refers to processes that act at a glacier or ice sheet bed that cause the Earth’s surface to be worn down, broken up, and transported by ice. These processes leave behind some of the classic signs of glacial activity, in the form of erosional landforms and landscapes.

Subglacial erosion is one of the key components of the glacial system, yet it remains poorly understood despite decades of research. This is largely due to the inaccessible nature of glacier beds, which limit the opportunity for direct observations or measurements1,2.

Because of this, processes of subglacial erosion have been based on theoretical models3-6 or inferred through investigations of landforms left behind in deglaciated areas. Nonetheless, direct access to glacier beds has been achieved in rare instances7,8, by accessing natural or artificial tunnels, which allow the installation of monitoring equipment and direct measurements to be made.

These approaches have identified two main mechanisms of subglacial erosion:

  • Glacial abrasion, the wearing down of bedrock surfaces
  • Glacial plucking or quarrying, the removal of rock fragments and blocks from the bed

Glacial abrasion

Glacial abrasion is the wear of a bedrock surface by rock fragments transported at the glacier base. This can happen by (i) the scoring (striation) of bedrock by rock particles (usually > 1 cm) embedded in the glacier sole, due to ice flow across a rock surface (see image below); and (ii) the polishing of bedrock surfaces by smaller, silt-sized particles that are dragged across the bedrock1.2.

Fine-grained debris frozen to the basal ice of Nigardsbreen glacier, west Norway, with debris coming into contact with underlying bedrock. Photo: Jacob M. Bendle

Scoring results in the formation of thin, linear grooves across a bedrock surface (see image below). These are known as striations (or striae). While striations may appear smooth, close inspection of striae beds show they form by a series of small rock fractures due to the build-up of stress below a mobile rock particle9.

Crossing-cutting glacial striations in bedrock, Maine, USA. Photo: Neil P. Thompson

Polishing, on the other hand, results in the overall smoothing down of rough areas of the bed (see image below). This process can be likened to the effect of sandpaper on wood.

Glacially smoothed bedrock recently uncovered due to retreat of Nigardsbreen glacier, west Norway. Photo Jacob M. Bendle

Controls on glacial abrasion

Rate of basal sliding

As glacial abrasion is caused by the movement of rock particles across bedrock, it is closely associated with basal sliding1,2. In warm-based (temperate) glaciers, where ice exists at the pressure melting point throughout, basal sliding occurs and a high flux of debris is dragged across a bedrock surface. By contrast, cold-based glaciers are frozen to their beds, so sliding rates are very low and the ability to abrade the bed limited1,2.

Debris concentration

Along with the rate of basal sliding, the amount of debris embedded in the glacier base also influences the rate of abrasion5. However, it is not as simple as a higher debris load resulting in faster rates of abrasion. In fact, a glacier with a high basal debris concentration results in friction between the ice and its bed, slowing the rate of basal sliding (see diagram below). Instead, glacial abrasion is most effective where basal debris is relatively sparse, as the reduced friction promotes faster sliding1.

The effect of basal debris concentration and glacier sliding on abrasion rate. Graph redrawn after Bennett and Glasser (2009)

Glacial plucking or quarrying

Plucking or quarrying is the fracture and removal of larger rock fragments (>1 cm) from the bed. In much the same way as the striation process, plucking occurs where stress build-up beneath an overriding rock particle results in the expansion of pre-existing cracks in the bedrock and the detachment of rock fragments1,2 (see image below). The fractured bedrock can then entrained by the overriding glacier and transported downglacier.

Vertical  joints and fractures observed in the bedrock of the formerly glaciated Ibañez valley, central Patagonia. Ice flow was from left to right. In the foreground, a former zone of plucking is illustrated by the ‘missing’ blocks of bedrock at the downglacier end of a roches mountonnée. Photo: Jacob M. Bendle.

There are several main ways by which plucked rock fragments can be entrained into the base of a glacier:

  • Debris can be frozen-on to the glacier sole as meltwater refreezes in the low-pressure zone in the lee (i.e. the downglacier end) of bedrock obstacles (see diagram below); this process is therefore strongly associated with the regelation mechanism of basal sliding
  • Debris can be frozen-on to the glacier sole at the boundary between warm-based (temperate) and cold-based ice, for example, approaching the glacier snout where ice thickness decreases and pressure melting of basal ice is inhibited (see diagram below)
  • Debris can be simply dragged from the bedrock and enveloped into basal ice, particularly where it is very loose

Freeze-on of plucked (quarried) debris in a low pressure zone at the downglacier end of a bedrock bump, due to refreezing of meltwater associated with regelation. Source: Jacob M. Bendle.

Meltwater and debris frozen in to basal ice layers at the transition from warm based to cold based ice (in this instance, at the glacier margin). Source: Jacob M. Bendle

Controls on plucking

Bedrock lithology

As described above, plucking tends to be focused along pre-existing cracks in bedrock8. The lithology (or rock type) of the bed will therefore influence its resistance to erosion. For example, in well-jointed rocks with deep, near-continuous cracks (e.g. shale), plucking rates will be higher than in rocks with fewer or more widely-spaced joints and cracks (e.g. granite).


Both theoretical models10 and direct observations beneath modern glaciers8 show that the presence of cavities at the bed is an important control on plucking. When a cavity is filled with water, water pressure offsets the overburden pressure resulting from the weight of overlying ice (‘T1’ in diagram below). In this situation, stresses in the bed are highest adjacent to the cavity. If water leaves a cavity and the water pressure drops, stresses in the bedrock increase considerably and lead to plucking of rock fragments (‘T2’ in diagram below).

In T1, the water pressure (pw) associated with a water-filled cavity in the lee of a bedrock bump offsets the downward directed ice overburden pressure (pi), preventing bedrock fracture. In T2, the water has drained, and a high stress zone (red) develops in the bedrock around the cavity edge, which may result in bed fracture and plucking. Source: Jacob M. Bendle

Therefore, plucking rates will be highest where the bed surface is undulating (i.e. there are abundant sites for cavities to form) and when the supply of meltwater causes fluctuation in water pressure, such as during the ablation season, where diurnal (day-night) melting patterns often develop1,2.


Subglacial erosion occurs at all ice masses, from small cirque glaciers to large continental ice sheets. It is also fundamentally linked to ice motion (e.g. sliding) and, in turn, mass balance regime and glacier thermal regime. Subglacial erosion processes therefore offer an excellent example of the connections between various components of the glacial system.

Other pages in this section of the site explore the effect of glacial erosion on Earth’s surface morphology.


[1] Bennett, M.R., and Glasser, N.F. (2009) Glacial Geology: Ice Sheets and Landforms. Wiley-Blackwell.

[2] Benn, D.I., and Evans, D.J.A. (2010) Glaciers and Glaciation. Routledge.

[3] Boulton, G.S. (1974) Processes and patterns of glacial erosion. In Glacial Geomorphology (ed. D.R. Coates) Springer, Dordrecht, pp. 41-87.

[4] Hallet, B. (1979) A theoretical model of glacial abrasion. Journal of Glaciology23, 39-50.

[5] Hallet, B. (1981) Glacial abrasion and sliding: their dependence on the debris concentration in basal ice. Annals of Glaciology2, 23-28.

[6] Iverson, N.R. (2012) A theory of glacial quarrying for landscape evolution models. Geology40, 679-682.

[7] Cohen, D., Iverson, N.R., Hooyer, T.S., Fischer, U.H., Jackson, M. and Moore, P.L. (2005). Debris‐bed friction of hard‐bedded glaciers. Journal of Geophysical Research: Earth Surface110(F2).

[8] Cohen, D., Hooyer, T.S., Iverson, N.R., Thomason, J.F. and Jackson, M. (2006). Role of transient water pressure in quarrying: A subglacial experiment using acoustic emissions. Journal of Geophysical Research: Earth Surface111(F3).

[9] Drewry, D.J. (1986) Glacial Geologic Processes. Edward Arnold, London.

[10] Morland, L.W., and Morris, E.M. (1977) Stress in an elastic bedrock hump due to glacier flow. Journal of Glaciology, 18, 67-75.