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.


Glacial depositional landforms

This section of the website includes many examples of landforms created underneath and around the margins of glaciers. These depositional landforms typically form in two domains: subglacial landforms and ice-marginal landforms.

Subglacial landforms include:

  • A continuum of lineated bedforms, ranging from small scale (flutes), through to intermediate scale (10s of metres; Drumlins), through to large scale (kilometres; Megascale glacial lineations).
  • Sediments and landforms associated with meltwater, such as eskers.

Ice-marginal landforms include:

  • Piles of debris formed at the ice margin, such as moraines;
  • Till plains formed underneath the ice sheet;
  • Fluvioglacial landforms such as kames, outwash plains, meltwater channels.

There are lots of examples of these types of landforms across the Patagonian Ice Sheet.

Improvements to

Here at we have been busy making many updates to the website. We are particularly keen to update the website to bring it in to line with the reformed A-Level syllabus, and also to update and rewrite some of the older content, and improve the website as a resource to promote public understanding of glaciers and climate change.

Since was founded 6.5 years ago, we have undergone substantial improvements and learned a lot over the years. This outreach endeavour, motivated by a desire to publicly communicate the risks that climate change and rising sea levels pose to our world’s glaciers and ice sheets, has evolved into one of the premier sites on this subject. This website aims to inspire both interested adults and also young people and school children with geology and geomorphology, and specifically targets teachers to supply them with engaging, original content that they can use in lesson planning.

Continue reading

Glacier accumulation and ablation

Glacier accumulation | Glacier ablation | Equilibrium line altitude | Glaciers as a system | Further reading | References | Comments |

Glacier accumulation

A glacier is a pile of snow and ice. In cold regions (either towards the poles or at high altitudes), more snow falls (accumulates) than melts (ablates) in the summer season. If the snowpack starts to remain over the summer months, it will gradually build up into a glacier over a period of years.

Unnamed Glacier, Ulu Peninsula, James Ross Island. Small valley glacier.

The key input to a glacier is precipitation. This can be “solid precipitation” (snow, hail, freezing rain) and rain1. Further sources of accumulation can include wind-blown snow, avalanching and hoar frost. These inputs together make up the surface accumulation on a glacier.

The Glacier as a System. Inputs are largely from precipitation, and also from wind-blown snow and avalanches. The glacier loses mass (ablates) mainly by the processes of calving and surface and subaqueous melt. After Cogley et al., 2011.

In general, glaciers receive more mass in their upper reaches and lose more mass in their lower reaches. The part of the glacier that receives more mass by accumulation than it loses by ablation is the accumulation zone.

Heavy snowfall over Monte San Valentín (4058 m asl) and in the accumulation zone of the North Patagonian Icefield. Photo: Murray Foubister Wikimedia Commons.

Formation of glacial ice

Over time, the snowfall (by far the most important input to a glacier) is gradually compressed and compacted by the weight of further snowfall on top it. The beautiful pointy edges of the snowflake gradually lose their tips and shape, becoming first granular ice, then firn, and finally glacial ice.

Layers of ice on Davies Dome Glacier, James Ross Island, Antarctic Peninsula.

The processes of transformation from snow to ice include partial melting, refreezing and fusing. The rate of transformation varies according to climate (temperature and precipitation regimes). The image below is from an ice core. Note the summer and winter layers in the ice. You can also no longer see the individual crystals that make up the glacier ice at this depth.

This 19 cm long of GISP2 ice core from 1855 m depth shows annual layers in the ice. This section contains 11 annual layers with summer layers (arrowed) sandwiched between darker winter layers. From the US National Oceanic and Atmospheric Administration, Wikimedia Commons.

Glacier ice is a crystalline material, and the crystal size and depth varies with the history of the ice.

Glacier ablation

As ice flows downhill, it either reaches warmer climates, or it reaches the ocean.  This causes various processes of melt, or ablation, to occur. In a land-terminating glacier (a glacier that ends on dry land), the main processes of ablation will be surface melt, because air temperatures generally increase as you lose altitude. This meltwater runs off the glacier and forms a number of rivers that typically drain the glacier.

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

This surface meltwater may runoff as surface runoff (as shown above; this is a supraglacial meltwater stream on the surface of the glacier), or it may make its way to the bed of the glacier through cracks in the ice (see the figure below). The water at the glacier bed eventually makes it way to the margin of the glacier, where it exits as a meltwater stream.

Meltwater propagates to the glacier bed through crevasses and moulins

Glaciers that reach the sea or terminate in a lake (Marine-terminating and lacustrine-terminating respectively) additionally will calve icebergs and melt underwater.   In large parts of Antarctica, melting underneath the base of floating ice shelves and calving from the margin of the glaciers dominate over surface melt.

Upsala Glacier, from the Southern Patagonian Ice Field, terminates in a large lake. Note the calved icebergs drifting out across the lake. Credit: NASA

The lower part of the glacier generally loses more mass from ablation than it receives from accumulation. This part of the glacier is the ablation zone.


Small tidewater (marine-terminating) glaciers calving into Croft Bay, Antarctic Peninsula

Equilibrium line altitude

Most glaciers receive more inputs and accumulation in their upper reaches, and lose more mass by ablation in their lower reaches. The Equilibrium Line Altitude (ELA) marks the area of the glacier separating the accumulation zone from the ablation zone, and were annual accumulation and ablation are equal2.

Equilibrium line altitudes in a hypothetical glacier

Glaciers as a system

Glacier ice is actually a viscous fluid, which flows and deforms under its own weight. Glaciers can therefore be thought of as systems, which receive snow and ice, flow downslope, and melt. Snow and ice are stored in the glacier until they melt as the glacier reaches lower elevations. This concept is explored in more detail in the Introduction to Glacier Mass Balance page and the pages on Glacier Flow.

In the European Alps and North America, most glaciers receive snowfall throughout the winter, and the main glacier ablation occurs in the summer. The Mass Balance, the balance of accumulation and ablation, is usually therefore positive in the winter and negative in the summer3. These glaciers, which receive more snow in winter and less in summer, are known as Winter Accumulation Type Glaciers. These glaciers form the majority of the world’s glaciers4.

In contrast, in places like the Himalaya, the monsoon brings more precipitation in the summer and less in the relatively cold, dry winter. These glaciers therefore receive more accumulation in the summer, and are known as Summer Accumulation Type Glaciers.

Further reading


1              Cogley, J. G. et al. Glossary of Glacier Mass Balance and related terms.  (IHP-VII Technical Documents in Hydrology No. 86, IACS Contribution No. 2, UNESCO-IHP, 2011).

2              Bakke, J. & Nesje, A. in Encyclopedia of Snow, Ice and Glaciers   (eds Vijay P. Singh, Pratap Singh, & Umesh K. Haritashya)  268-277 (Springer Netherlands, 2011).

3              Naito, N. in Encyclopedia of Snow, Ice and Glaciers   (eds Vijay P. Singh, Pratap Singh, & Umesh K. Haritashya)  1107-1108 (Springer Netherlands, 2011).

4              Kumar, A. in Encyclopedia of Snow, Ice and Glaciers   (eds Vijay P. Singh, Pratap Singh, & Umesh K. Haritashya)  1227-1227 (Springer Netherlands, 2011).


Glacial geomorphology of the Patagonian Ice Sheet

This page is based largely on Bendle et al. (2017) and summarises the glacial geomorphology of the North Patagonian Icefield region (46–48°S).

Glaciers and the Patagonian landscape

The Patagonian Ice Sheet has expanded and contracted at least five times during the last million years1. During glacial periods, large outlet glaciers discharged along major valleys (see map below), moving mass from the ice-sheet interior to its margins2. These glaciers also eroded, entrained, transported, and deposited rock and sediment, moulding the landscape and creating glacial landforms.

The Patagonian Ice Sheet at the Last Glacial Maximum (LGM; 21,000–19,000 years ago) with extent of contemporary glaciers shown in blue. Copyright: J. Bendle.

Why study glacial landforms?

The sediments and landforms left behind by glaciers are window into the evolution of the Patagonian Ice Sheet through time. For example, ice-marginal landforms, such as moraines, tell us about the extent or thickness of former glaciers. Other landforms provide us with information about the conditions at glacier beds or margins. Glacial erosional landforms, for instance, give us clues about the bed temperature, hydrology, and flow behaviours of former glaciers, and hence glacier thermal regime. When combined with methods of dating glacial deposits, glacial geomorphology allows us to reconstruct former glacier behaviour.

How are glacial landforms mapped?

From ‘remotely sensed’ data

When at its maximum extent, the Patagonian Ice Sheet was very large2, with outlet glaciers of similar size to those draining the Greenland Ice Sheet today. Therefore, it would be almost impossible to map its glacial geomorphology solely on foot. To get around this problem, we use ‘remotely sensed’ imagery (i.e. information about Earth’s surface obtained from aircraft or satellite).

Glacial geomorphology of the Lago Buenos Aires basin, based on satellite image interpretation (Bendle et al., 2017). Copyright: J. Bendle.

Satellite images are useful because they cover large areas, making it possible to assess regional landform distribution quickly and efficiently (see image above)2,3,4. Digital Terrain Models are also valuable in areas of steep or undulating relief, as they provide 3D representations of Earth’s surface that may give clues about the origin of landforms. For example, while a moraine typically has a crest with two sloping sides, a lake shoreline has a flat upper surface with a single sloping side (see image below).

Hillshade model with draped Digital Elevation Model (DEM), showing moraine complexes and glacial lake shorelines (note shading on one side of feature only).

‘Virtual globe’ software, such as GoogleEarth, is also useful in landform mapping as it allows the mapper to view the landscape in 3D, and customise the angle of viewing (see image below)4. With GoogleEarth it is also possible to measure the size (e.g. length, width, area) and elevation of features very quickly and simply, which may help an accurate landform interpretation be made.

Oblique view generated using GoogleEarth software, and enabling the mapping of moraine ridges. Image: GoogleEarth. Compiled by J. Bendle.

On foot, in the field

Field mapping is used to check whether features mapped from satellite images have been correctly identified, and to identify additional landforms that are not easily recognised in satellite datasets, such as small-scale or low-relief features, or landforms that may be hidden beneath vegetation or clouds in satellite imagery3,4.

Field mapping of glacial landforms in the Lago General Carrera–Buenos Aires valley. Photo credit: V. Thorndycraft.

Ice-marginal landforms

Ice-marginal landforms were produced in abundance at the snout and sides of former Patagonian glaciers2,3,4,5, due to the dumping of debris from the ice surface, the pushing and squeezing of debris at the snout, or by the flow of meltwater.

Latero-frontal moraines

On the forelands of former glaciers, moraine ridges mark out former ice limits. In central Patagonia, these ridges reach 50 m relief and can run continuously for tens of kilometres in latero-frontal arcs (see image below)2,3,4. The overall distribution of moraines can be used to reconstruct the pattern of ice retreat (see image below)6.

Latero-frontal moraine arc marking the former limit of the Lago General Carrera–Buenos Aires outlet glacier during the Last Glacial Maximum (~21,000-19,000 years ago). Photo credit: J. Bendle.


Sharp lines on valley sides, known as trimlines, are common closer to the contemporary North Patagonian Icefield7. Trimlines mark the boundary between terrain that has been recently covered by ice, and terrain that has been ice free for a longer period of time. Therefore, trimlines provide information about glacier thickness change.

Valley side trimlines (labelled with white arrows) marking the former thickness of the Callequeo Glacier, Monte San Lorenzo. Note moraine ridges in the foreground. Photo credit: J. Martin.

Meltwater channels

Meltwater channels come in many varieties. Most commonly they follow the lateral margins of former glaciers3,4,8, and can be straight, sinuous, or meandering (see image below). They start and end abruptly, and rarely contain rivers in the present day. Like moraines, meltwater channels are useful for reconstructing former glacier extent.

Meltwater channels cut into the surface of outwash deposits, and dissecting moraine ridges (right). Image from: GoogleEarth.

Outwash plains

Outwash plains are surfaces of glaciofluvial sand and gravel that build up in front of a stable ice margin. They may slope gently away from a former ice margin to form an expansive plain (see image below), or can accumulate in topographic lows between moraine ridges (see image above)3,4. The vast outwash plains of Patagonia2 tell us that meltwater streams carrying high sediment loads were common around former ice margins3,4.

Outwash plain grading away from moraine ridges in the Lago Cochrane–Pueyrredón valley. Photo credit: J. Bendle.

Subglacial landforms

Subglacial landforms are produced at the bed of former glaciers and commonly relate to patterns of former ice or meltwater flow.


Eskers are straight to sinuous ridges of glaciofluvial sand and gravel, which are formed in subglacial, englacial, or supraglacial channels. They can form single ridges (as shown in the image below) or consist of a braided network. Eskers give an indication of meltwater flow patterns within or beneath former glaciers9, which is important for ice dynamics (e.g. motion).

A sinuous esker ridge, and several smaller eskers (all shown in green), mapped from satellite imagery in the Lago Cochrane-Pueyrredón valley. Copyright: J. Bendle.

Ice-moulded bedrock

In the major valleys, the bedrock has been widely scoured or smoothed by subglacial erosion processes (see image below)4. This suggests that the outlet glaciers that once overrode this terrain were (at least sometimes) warm-based, fast-flowing, and carrying a significant basal debris load10.

Glacially-smoothed bedrock outcrops in the Ibañez valley, Chile. Photo credit: J. Bendle.

In some areas, the bedrock has been moulded into very long, linear ridges called glacial lineations (see image below). When mapped at a regional-scale, these landforms reveal the flow pathways of former outlet glaciers (see map below)4.

Glacial lineations formed in bedrock south of the Lago Cochrane-Pueyrredón. Image: GoogleEarth. Compiled by J. Bendle.

Map of glacial lineations (red lines) formed at the bed of the Lago General Carrera-Buenos Aires and Lago Cochrane-Pueyrredón outlet glaciers. Interpretation of glacier flow direction is shown by dashed lines. Copyright: J. Bendle.


Drumlins are streamlined hills formed under moving glacier ice11. They are elongated in the direction of ice flow. Like glacial lineations, therefore, they can be used to reconstruct glacier dynamics. Their length may also be related to ice velocity12, with more elongated drumlins formed under faster flowing ice. While not that common around the Northern Patagonian Icefield (see image below), drumlin swarms occur more widely around the Southern Patagonian Icefield, and across Tierra del Fuego3,13.

Drumlin swarm in the Lago Cochrane-Pueyrredón valley. Ice flow from left to right. Image from: GoogleEarth.

Glaciolacustrine landforms

These landform types are produced in association with glacial lakes, which were common around the Patagonian Ice Sheet margin14,15.

Glacial lake shorelines

Shorelines are long, continuous terraces with flat surfaces that mark the level of a former glacial lake. Around Lago General Carrera–Buenos Aires, extensive ‘flights’ of shorelines are seen (see image below) and reflect changing lake level through time4,14,15.

Top: Relict glacial lake shorelines above the present-day Lago General Carrera-Buenos Aires, marking the level of a former glacial lake. Photo credit: J. Bendle. Bottom: GoogleEarth image showing a flight of former lake shorelines.

Raised deltas

Raised deltas are flat-topped accumulations of sand, gravel, and cobbles (see image below). They mark the point at which rivers entered former glacial lakes and deposited their sediment load. Like shorelines, therefore, they provide an indication of glacial lake level14.

Raised lacustrine delta approximately 100 m above the modern Lago General Carrera-Buenos Aires, marking the level of a former glacial lake. Photo credit: J. Bendle.

What does the pattern of landforms tell us?

The geomorphology of the North Patagonian Icefield region indicates that (1) when outlet glaciers reached their maximum extent they remained stable on flat plains, forming large moraines and outwash plains. Subglacial landforms (e.g. glacial lineations) suggest that (2) the glaciers were (at least sometimes) fast-flowing and warm-based. The presence of shorelines and raised deltas show that (3) large proglacial lakes developed around the retreating ice margins during deglaciation.


[1] Coronato, A. & Rabassa, J. 2011. Pleistocene glaciations in Southern Patagonia and Tierra del Fuego. In Ehlers, L., Gibbard, P.L., Hughes, P.D. (Eds.) Developments in Quaternary Sciences, 15, Elsevier. pp. 715–727.

[2] Glasser, N.F., Jansson, K.N., Harrison, S. & Kleman, J. 2008. The glacial geomorphology and Pleistocene history of South America between 38°S and 56°S. Quaternary Science Reviews, 27, 365–390.

[3] Darvill, C.M., Stokes, C.R., Bentley, M.J. & Lovell, H. 2014. A glacial geomorphological map of the southernmost ice lobes of Patagonia: the Bahía Inútil–San Sebastián, Magellan, Otway, Skyring and Río Gallegos lobes. Journal of Maps, 10, 500–520.

[4] Bendle, J.M., Thorndycraft, V.T. & Palmer, A.P., 2017. The glacial geomorphology of the Lago Buenos Aires and Lago Pueyrredón ice lobes of central Patagonia. Journal of Maps, 13, 654–673.

[5] García, J.L., Hall, B.L., Kaplan, M.R., Vega, R.M. & Strelin, J.A. 2014. Glacial geomorphology of the Torres del Paine region (southern Patagonia): Implications for glaciation, deglaciation and paleolake history. Geomorphology, 204, 599–616.

[6] Darvill, C.M., Stokes, C.R., Bentley, M.J., Evans, D.J. & Lovell, H., 2017. Dynamics of former ice lobes of the southernmost Patagonian Ice Sheet based on a glacial landsystems approach. Journal of Quaternary Science, 32, 857–876.

[7] Glasser, N.F., Harrison, S. & Jansson, K.N., 2009. Topographic controls on glacier sediment–landform associations around the temperate North Patagonian Icefield. Quaternary Science Reviews, 28, 2817–2832.

[8] Bentley, M.J., Sugden, D.E., Hulton, N.R. and McCulloch, R.D., 2005. The landforms and pattern of deglaciation in the Strait of Magellan and Bahía Inútil, southernmost South America. Geografiska Annaler: Series A, Physical Geography, 87, 313–333.

[9] Storrar, R.D., Evans, D.J., Stokes, C.R. & Ewertowski, M., 2015. Controls on the location, morphology and evolution of complex esker systems at decadal timescales, Breiðamerkurjökull, southeast Iceland. Earth Surface Processes and Landforms, 40, 1421–1438.

[10] Glasser, N.F. & Jansson, K.N. 2005. Fast-flowing outlet glaciers of the last glacial maximum Patagonian Icefield. Quaternary Research, 63, 206–211.

[11] Stokes, C.R., Spagnolo, M. and Clark, C.D., 2011. The composition and internal structure of drumlins: complexity, commonality, and implications for a unifying theory of their formation. Earth-Science Reviews, 107, 398-422.

[12] Stokes, C.R. & Clark, C.D., 2002. Are long subglacial bedforms indicative of fast ice flow?. Boreas, 31, 239-249.

[13] Lovell, H., Stokes, C.R., Bentley, M.J. & Benn, D.I. 2012. Evidence for rapid ice flow and proglacial lake evolution around the central Strait of Magellan region, southernmost Patagonia. Journal of Quaternary Science, 27, 625–638.

[14] Glasser, N.F., Jansson, K.N., Duller, G.A., Singarayer, J., Holloway, M. & Harrison, S. 2016. Glacial lake drainage in Patagonia (13-8 kyr) and response of the adjacent Pacific Ocean. Scientific Reports, 6. 21064.

[15] Bendle, J.M., Palmer, A.P., Thorndycraft, V.R. and Matthews, I.P., 2017. High-resolution chronology for deglaciation of the Patagonian Ice Sheet at Lago Buenos Aires (46.5°S) revealed through varve chronology and Bayesian age modelling. Quaternary Science Reviews, 177, 314–339.

The Patagonian Icefields

Geographic setting

Patagonia, between ~40°S to 56°S, is the most southerly part of the South American continent. The landscape of this region is one of contrasts. Dense temperate rainforests cover the western coast, whereas the eastern plains are flat, vast, and arid. Perhaps most striking, however, are the high, Patagonian Andes, which rise steeply (up to around 4000 m asl) above deeply carved glacial fjords and valleys, and are home to the Patagonian Icefields.

The rugged, and densely vegetated west coast of Patagonia (Pacific fjords in distance). The high, ice-covered Andes are shown in the foreground. Photo: Miguel Vieira Wikimedia Commons.

Climate of Patagonia

Patagonia is one of the windiest and wettest places on Earth. The region has a temperate maritime climate, with a strong west-east precipitation gradient as a result of the year-round passage of westerly winds over the Patagonian Andes(1,2). On the west coast, annual precipitation reaches up to 7,500 mm year, whereas less than 1,500 mm a year falls east of the icefields(3,4). The snowfall brought by the westerly winds are the main source of accumulation for the glaciers of this region.

Windswept Nothofagus Antarctica tree, Ushuaia, Tierra del Fuego, Patagonia, Argentina. Photo: Leonardo Pallotta, Wikimedia Commons.

The Patagonian Icefields

While the Patagonian Andes are home to hundreds of small caps and valley glaciers(5,6,7,8), most ice is locked up in three large icefields, called the North, South, and Cordillera Darwin Icefields (see map below). Together, these icefields contain around 5,500 gigatons of ice, enough to raise global sea level by around 15 mm if completely melted(9).

The three major Patagonian Icefields, which occur south of ~46°S, are the largest expanse of ice in the Southern Hemisphere outside Antarctica(10). While large, today’s icefields are the remnants of a much larger Patagonian Ice Sheet, which formed during the global LGM around 21,000 years ago.

The icefields and glaciers of Patagonia. Major outlet glaciers are labelled. Glacier outlines from Pfeffer et al. (2014). Copyright J. Bendle

Dynamic outlet glaciers

Outlet glaciers of the Patagonian Icefields are fed by heavy snowfall (up to 10,000 mm per year) in the accumulation area, and have high melt rates at lower elevations. As a result, they have high mass balance gradients and are very dynamic(10), especially on the western coast snowfall is highest(11).

Heavy cloud cover and snowfall over Monte San Valentín (4058 m asl) in the accumulation zone of the North Patagonian Icefield. Photo: Murray Foubister Wikimedia Commons.

North Patagonian Icefield

Size, elevation, and structure

The North Patagonian Icefield (NPI) stretches over 100 km from 46°30’S to 47°30’S (see map above). Its current size is estimated at ~3,700 km2(8,12).

The NPI is drained by 29 main outlet glaciers (>10 km2), with glacial tongues that terminate in different settings(7). 11 (38%) outlet glaciers terminate on land, 17 (59%) terminate in lakes (e.g. Glacier Exploradores), and only one (3%), Glacier San Rafael, terminates in the sea.

The average elevation of NPI glaciers is around 400 m lower on the western side of the icefield (1240 m asl) than on the east (1640 m asl), reflecting the higher rates of snowfall in the west(8).

The North Patagonian Icefield seen from space. Image: NASA.

Glacier Flow speeds

The fastest-flowing glacier of the NPI is Glacier San Rafael (NASA Fig). This glacier is flowing at 7.6 km per year (around 20 metres per day11). In total, three NPI glaciers flow faster than 1 km per year (San Rafael, San Quintín, and Colonia11).

Outlet glacier flow speeds calculated from satellite imagery. The fastest-moving glaciers (shown by green and yellow colours) terminate in the Pacific Ocean fjords and embayments, or in large glacial lakes (e.g. Glacier Upsala). Image: NASA (based on data from Mouginot and Rignot, 2015)

South Patagonian Icefield

Size, elevation, and structure

The South Patagonian Icefield (SPI) is the largest of the South American icefields. It stretches over 350 km from 48°20’S to 51°30’S, and its current size is estimated at ~12,200 km2 (around three times larger than the NPI8).

The SPI is drained by 53 main outlet glaciers (>10 km2) that are generally larger than those of the NPI. The SPI has a similar amount of lake-terminating glaciers (59%), but more (17 or 32%) marine-terminating glaciers(7).

The South Patagonian Icefield seen from space. Image: NASA

Glacier Flow speeds

The SPI also has some of the fastest flowing glaciers in the world (see glacier flow speed diagram above). Of the ten fastest SPI outlet glaciers, which all move faster than 2.5 km per year, eight flow out into Pacific Ocean fjords (e.g. the Jorge Montt glacier shown below) and two into large glacial lakes (e.g. Glacier Upsala). The fastest, Glacier Penguin, flows at 10.3 km per year (around 28 metres per day11). These data suggest that ocean heat plays an important role in melting at glacier fronts, and rapidly drawing down ice from the SPI interior.

The Jorge Montt outlet glacier of the Southern Patagonian Icefield, flowing out into a Pacific Ocean fjord that is choked with icebergs. Image: NASA.

The Upsala outlet glacier of the Southern Patagonian Icefield, flowing into a iceberg filled proglacial lake (Lago Argentino). Image: NASA

Cordillera Darwin Icefield

Size, elevation, and structure

The Cordillera Darwin Icefield (CDI) is the smallest (~2600 km2), and southernmost of Patagonia’s icefields(13), existing between 54°40’S to 55°00’S. It is also the least studied. Most CDI glaciers descend to sea level, and many flow into the Pacific Ocean(8).

Glacier Flow speeds

Most CDI outlet glaciers have flow speeds of between 1 and 3 metres per day(13). However, the marine-terminating Marinelli Glacier (the largest and fastest moving of the CDI glaciers) and Darwin Glacier, flow much faster, at around 8-10 metres per day(13).

The Cordillera Darwin Icefield seen from space. Image: NASA

Glacier change

Most glaciers of the Patagonian Icefields are experiencing negative mass balance, as a result of glacier thinning and the widespread retreat of ice-fronts(5,11,14). Only Glacier Pío XI (the largest glacier in South America) from the South Patagonian Icefield has grown in recent years(15,16). The overall pattern of ice loss make the Patagonian Icefields one of the largest current sources of global sea level rise (they contribute ~10% of that from all glaciers and ice caps worldwide17,18).

Since the Little Ice Age

Since the end of the Little Ice Age at around 1870 AD, over 90% of Patagonian outlet glaciers have shrunk(5,8,19) (see GIF below). Reconstructions of Little Ice Age glacier extents(19) show that the North Patagonian Icefield has lost around 103 ± 20.7 km3 of ice, and the South Patagonian Icefield around 503 ± 101.1 km3. This gradual melting of the icefields has contributed around 0.0018 mm to global seal level rise per year (or 0.27 mm in total) since 1870 AD(19).

Retreat of the North Patagonian Icefield between 1870 AD (the end of the Little Ice Age) to 2011.

Over the 21st century

During the last 40-50 years, however, the rate of Patagonian ice loss has sped up substantially. Between 1975 and 2000, the North and South Patagonian Icefields have lost a combined 15.0 ± 0.7 gigatons of ice per year(17). Between 2000 and 2011, the rate of ice loss increased further to 24.4 ± 1.4 gigatons per year(20,21), and has since remained similar (between 2011 and 2017 the icefields lost 21.29 ± 1.98 gigatons per year16).

Retreat of the HPS-12 outlet glacier of the Southern Patagonian Icefield between 1985 and 2017. The glacier terminus has retreated approximately 13 km in the last 30 years. Images from NASA. Compiled by: J. Bendle

Contribution to rising sea level

The current contribution of the Patagonian Icefields to global sea level rise is around 0.067 ± 0.004 mm per year(21), over one order of magnitude greater than the long-term contribution since the Little Ice Age(19). The rate of glacier melting is expected to continue, and maybe even increase, in coming decades(22).

Causes of icefield retreat: a warming atmosphere

The rapid melting of the Patagonian Icefields during the 21st century has several possible causes. For example, a warming atmosphere has resulted in a greater number of warm ‘summer’ days each year, and more melting at the glacier surface(23). At 50°S, for example, a warming of 0.5°C between 1960 to 1999 has resulted in a 0.5 m water equivalent increase in annual glacier melt in the ablation area(24).

Less snowfall, more rainfall

The amount of snowfall falling over the Patagonian Icefields decreased by around 5% between 1960 and 1999, whereas the amount of rainfall increased, both as a result of warmer air temperatures(24). Alongside faster surface melting, therefore, the amount of annual accumulation is getting smaller, preventing glacier growth.

Continued warming in Patagonia, therefore, will have a major impact on glacier accumulation and ablation trends(25), and will ultimately dictate the fate of the Patagonian Icefields.


[1] Garreaud, R.D., Vuille, M., Compagnucci, R. and Marengo, J., 2009. Present-day South American climate. Palaeogeography, Palaeoclimatology, Palaeoecology281, 180-195.

[2] Garreaud, R., Lopez, P., Minvielle, M. and Rojas, M., 2013. Large-scale control on the Patagonian climate. Journal of Climate26, 215-230.

[3] Carrasco, J.F., Casassa, G. and Rivera, A., 2002. Meteorological and climatological aspects of the Southern Patagonia Icefield. In The Patagonian Icefields (pp. 29-41). Springer, Boston, MA.

[4] Schneider, C., Glaser, M., Kilian, R., Santana, A., Butorovic, N. and Casassa, G., 2003. Weather observations across the southern Andes at 53°S. Physical Geography2, 97-119.

[5] Davies, B.J. and Glasser, N.F., 2012. Accelerating shrinkage of Patagonian glaciers from the Little Ice Age (~AD 1870) to 2011. Journal of Glaciology58, 1063-1084.

[6] Falaschi, D., Bravo, C., Masiokas, M., Villalba, R. and Rivera, A., 2013. First glacier inventory and recent changes in glacier area in the Monte San Lorenzo Region (47 S), Southern Patagonian Andes, South America. Arctic, Antarctic, and Alpine Research45, 19-28.

[7] Pfeffer, W.T., Arendt, A.A., Bliss, A., Bolch, T., Cogley, J.G., Gardner, A.S., Hagen, J.O., Hock, R., Kaser, G., Kienholz, C. and Miles, E.S., 2014. The Randolph Glacier Inventory: a globally complete inventory of glaciers. Journal of Glaciology60, 537-552.

[8] Meier, W.J-H., Grießinger, J., Hochreuther, P. and Braun, M.H., 2018. An updated multi-temporal glacier inventory for the Patagonian Andes with changes between the Little Ice Age and 2016. Frontiers in Earth Science, 6, 62.

[9] Carrivick, J.L., Davies, B.J., James, W.H., Quincey, D.J. and Glasser, N.F., 2016. Distributed ice thickness and glacier volume in southern South America. Global and Planetary Change146, 122-132.

[10] Warren, C.R. and Sugden, D.E., 1993. The Patagonian icefields: a glaciological review. Arctic and Alpine Research25, 316-331.

[11] Mouginot, J. and Rignot, E., 2015. Ice motion of the Patagonian icefields of South America: 1984–2014. Geophysical Research Letters42, 441-1449.

[12] Dussaillant, I., Berthier, E. and Brun, F., 2018. Geodetic Mass Balance of the Northern Patagonian Icefield from 2000 to 2012 using two independent methods. Frontiers in Earth Science6, 8.

[13] Melkonian, A.K., Willis, M.J., Pritchard, M.E., Rivera, A., Bown, F. and Bernstein, S.A., 2013. Satellite-derived volume loss rates and glacier speeds for the Cordillera Darwin Icefield, Chile. The Cryosphere7, 823-839.

[14] Rivera, A., Benham, T., Casassa, G., Bamber, J. and Dowdeswell, J.A., 2007. Ice elevation and areal changes of glaciers from the Northern Patagonia Icefield, Chile. Global and Planetary Change59, 126-137.

[15] Schaefer, M., Machguth, H., Falvey, M., Casassa, G. and Rignot, E., 2015. Quantifying mass balance processes on the Southern Patagonia Icefield. The Cryosphere, 9, 25-35.

[16] Foresta, L., Gourmelen, N., Weissgerber, F., Nienow, P., Williams, J.J., Shepherd, A., Drinkwater, M.R. and Plummer, S., 2018. Heterogeneous and rapid ice loss over the Patagonian Ice Fields revealed by CryoSat-2 swath radar altimetry. Remote Sensing of Environment211, 441-455.

[17] Rignot, E., Rivera, A. and Casassa, G., 2003. Contribution of the Patagonia Icefields of South America to sea level rise. Science302, 434-437.

[18] Gardner, A.S., Moholdt, G., Cogley, J.G., Wouters, B., Arendt, A.A., Wahr, J., Berthier, E., Hock, R., Pfeffer, W.T., Kaser, G. and Ligtenberg, S.R., 2013. A reconciled estimate of glacier contributions to sea level rise: 2003 to 2009. Science340, 852-857.

[19] Glasser, N.F., Harrison, S., Jansson, K.N., Anderson, K. and Cowley, A., 2011. Global sea-level contribution from the Patagonian Icefields since the Little Ice Age maximum. Nature Geoscience4, 303-307.

[20] Willis, M.J., Melkonian, A.K., Pritchard, M.E. and Ramage, J.M., 2012. Ice loss rates at the Northern Patagonian Icefield derived using a decade of satellite remote sensing. Remote Sensing of Environment117, 184-198.

[21] Willis, M.J., Melkonian, A.K., Pritchard, M.E. and Rivera, A., 2012. Ice loss from the Southern Patagonian ice field, South America, between 2000 and 2012. Geophysical Research Letters39. L17501.

[22] IPCC, 2013. The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, UK.

[23] Monahan, P.A. and Ramage, J., 2010. AMSR-E melt patterns on the Southern Patagonia Icefield. Journal of Glaciology56, 699-708.

[24] Rasmussen, L.A., Conway, H. and Raymond, C.F., 2007. Influence of upper air conditions on the Patagonia icefields. Global and Planetary Change59, 203-216.

[25] Schaefer, M., Machguth, H., Falvey, M. and Casassa, G., 2013. Modelling past and future surface mass balance of the Northern Patagonia Icefield. Journal of Geophysical Research: Earth Surface118, 571-588.