{
  "id": "ice-sheet",
  "title": "Ice Sheet",
  "lang": "en",
  "source_lang": "en",
  "summary": "An ice sheet is a continent-scale mass of land ice that flows under its own weight and strongly influences climate, landscapes, and sea level.",
  "category": "science",
  "subcategory": "earth-science",
  "type": "concept",
  "aliases": [
    "Continental ice sheet",
    "Continental glacier"
  ],
  "generated_at": "2026-11-03",
  "model": "pro/gpt-6.1-sol",
  "url": "https://aiwiki.page/en/ice-sheet/",
  "translations": {
    "zh-Hans": {
      "url": "https://aiwiki.page/zh/ice-sheet/",
      "json": "https://aiwiki.page/zh/ice-sheet.json"
    }
  },
  "sections": [
    {
      "heading": "",
      "text": "An ice sheet is a large, persistent body of glacial ice covering more than approximately 50,000 square kilometres of land. It accumulates from snowfall and spreads outward under its own weight, often burying most of the underlying landscape. Only two ice sheets exist today: those of Greenland and Antarctica. Together they contain roughly 99 percent of the world’s glacier ice and enough water to raise global mean sea level by about 65 metres if completely lost, although such complete loss is not a near-term projection.\n\nIce sheets are major components of the cryosphere, interacting with the atmosphere, ocean, solid Earth, and global water cycle. They preserve records of past environmental conditions, modify continental landscapes, and regulate freshwater storage over centuries to millions of years. Their present imbalance is an important contributor to sea-level rise. Understanding them requires distinguishing slow changes in their enormous interiors from potentially rapid changes at their margins, where melting, ocean circulation, and ice flow interact."
    },
    {
      "heading": "Definition, extent, and internal organization",
      "text": "An ice sheet is a type of glacier, distinguished principally by its continental scale rather than by a different material or a unique mode of motion. The conventional area threshold of 50,000 square kilometres separates ice sheets from smaller ice caps. Both typically spread across underlying topography, whereas valley glaciers are more strongly confined by mountain valleys. An ice field differs because its surface and flow remain substantially controlled by the terrain beneath it. These categories are useful descriptive conventions, not sharp physical boundaries.\n\nThe Greenland Ice Sheet covers approximately 1.7 million square kilometres and contains about 2.9 million cubic kilometres of ice. Its thickness exceeds 3 kilometres in the interior. The Antarctic Ice Sheet covers roughly 14 million square kilometres when its floating margins are included, and its ice volume is approximately 26–27 million cubic kilometres. Antarctic ice averages around 2 kilometres thick and reaches more than 4 kilometres in some deep interior basins. Estimates differ somewhat with mapping methods and with the inclusion of peripheral ice and floating shelves.\n\nAntarctica is commonly divided into the East Antarctic Ice Sheet, the West Antarctic Ice Sheet, and the ice of the Antarctic Peninsula. The Transantarctic Mountains form a major boundary between East and West Antarctica. Much of East Antarctica rests on a large continental landmass, although extensive subglacial basins lie below sea level. Much of West Antarctica occupies a bed below sea level and is therefore especially exposed to interactions between grounded ice and the ocean. Greenland has a more compact arrangement, with a high interior feeding outlet glaciers through coastal mountains and fjords.\n\nIce-sheet surfaces form broad domes and ridges separated into drainage basins. An ice divide marks a boundary across which ice generally flows in different directions, broadly analogous to a watershed boundary. Flow converges downstream into outlet glaciers and fast-moving ice streams, which transport a disproportionate share of the ice reaching the coast. Some outlet glaciers terminate directly in seawater; others feed floating ice shelves.\n\nThe distinction between grounded and floating ice is fundamental. Grounded ice rests on bedrock or sediment and transmits some of its weight to that bed. Floating ice is supported by seawater. Their boundary, called the grounding line, is not necessarily fixed: it can migrate with thinning, thickening, tides, and changes in ocean conditions. Ice-sheet area alone consequently gives an incomplete measure of condition. Thickness, volume, grounding-line position, and discharge are all necessary to describe how an ice sheet is changing."
    },
    {
      "heading": "Snow accumulation and the development of glacial ice",
      "text": "Ice sheets form where snowfall survives successive summers and accumulates over long periods. Newly fallen snow contains abundant interconnected air spaces and has a much lower density than solid ice. Burial compresses the snow, while grain rearrangement, sublimation, and recrystallization alter its structure. Surviving snow eventually becomes firn, an intermediate material between fresh snow and glacier ice. Continued densification closes the connected pore spaces and traps bubbles of ancient air within the ice.\n\nThe time required for this transformation varies greatly. In relatively warm, snowy areas, pore closure may occur within decades; in cold, dry Antarctic interiors, it may take centuries or longer. Bubble close-off commonly occurs at densities around 830 kilograms per cubic metre, although the process extends over a depth interval rather than occurring at one perfectly defined surface. Pure glacier ice has a density close to 917 kilograms per cubic metre near its melting point at atmospheric pressure.\n\nBecause air can move through porous firn before the pores close, the gas trapped in an ice layer is younger than the ice surrounding it. This ice-age–gas-age difference is essential when comparing past atmospheric carbon dioxide concentrations with temperatures reconstructed from ice. At low-accumulation Antarctic sites, the offset can amount to thousands of years. Firn physics therefore links the formation of an ice sheet directly to the interpretation of its climate archive.\n\nAccumulation is geographically uneven. Coastal mountains can receive heavy snowfall as moist air rises and cools, whereas the high Antarctic interior is a polar desert with very low annual precipitation. Wind redistributes snow, creates dunes and crusts, and can expose older ice in blue-ice areas. Surface mass input is therefore not simply equal to snowfall measured at a nearby weather station. Wind erosion, deposition, sublimation, and the freezing of liquid water all affect the amount retained.\n\nIce-sheet formation also requires a favourable balance between accumulation, summer melting, and ice export. High latitude alone is insufficient: snowfall supply, elevation, seasonal sunlight, and atmospheric circulation all matter. As an ice sheet thickens, its surface rises into colder air, helping snow survive. Its bright surface increases albedo, reflecting more incoming sunlight than exposed rock or vegetation. These elevation and reflectivity effects can reinforce growth, although increasing ice thickness also increases outward flow.\n\nIce sheets are consequently dynamic constructions rather than static piles of frozen precipitation. Snow deposited near an interior divide may eventually travel hundreds or thousands of kilometres before being lost at a margin. Interior ice can be hundreds of thousands of years old, while much younger ice is continually added above it and transported toward the coast."
    },
    {
      "heading": "Ice deformation and movement over the bed",
      "text": "Although ice is brittle over short timescales, it deforms as a slowly flowing crystalline material under sustained stress. Gravity drives an ice sheet from its elevated interior toward lower margins. A useful approximation for the gravitational driving stress is\n\n\\[\\tau_d \\approx \\rho_i g H \\sin\\alpha,\\]\n\nwhere \\(\\rho_i\\) is ice density, \\(g\\) is gravitational acceleration, \\(H\\) is ice thickness, and \\(\\alpha\\) is surface slope. Even a very gentle slope can generate substantial stress when the ice is several kilometres thick.\n\nMuch large-scale ice-sheet modelling represents deformation using Glen’s flow law, developed from laboratory and field studies associated with John Glen in the 1950s. In a simplified scalar form,\n\n\\[\\dot{\\varepsilon}=A(T)\\tau^n,\\]\n\nwhere \\(\\dot{\\varepsilon}\\) is strain rate, \\(\\tau\\) is an effective stress, and \\(A(T)\\) describes the temperature-dependent softness of the ice. A stress exponent \\(n\\) near 3 is widely used. The relationship is nonlinear: doubling stress can increase deformation by much more than a factor of two. Real ice also responds to crystal orientation, impurities, grain size, and previous deformation.\n\nIce sheets move through a combination of internal deformation and motion at the base. Where ice is frozen firmly to its substrate, deformation within the ice may dominate. Where the bed reaches the local melting temperature, ice can slide over rock or move over deformable sediment. Subglacial water reduces effective pressure—the ice overburden pressure minus water pressure—and can reduce resistance to motion. The relationship is not universal because bed roughness, cavities, sediment properties, and drainage configuration all matter.\n\nTemperature strongly controls these processes. Cold ice is generally harder to deform than warm ice. Heat enters from the bed through geothermal flux, is generated by deformation and basal friction, and is exchanged with the atmosphere at the surface. Thick ice insulates its base, so basal melting can occur beneath a surface that remains far below freezing. Conversely, cold ice carried downward or inward by flow can help maintain a frozen bed.\n\nFlow speeds vary by several orders of magnitude. Ice near a major divide may move only metres per year or less, while major outlet glaciers can move kilometres per year. Jakobshavn Isbræ in Greenland has at times exceeded 10 kilometres per year. Rapid motion is often concentrated within narrow corridors bounded by strong lateral shear, even where the surrounding ice moves much more slowly.\n\nThis spatial organization explains why small coastal regions can control large inland volumes. Changes in basal resistance or the forces acting on an outlet glacier can alter upstream velocity and thickness. The resulting thinning can propagate inland through a drainage basin, linking coastal disturbances to changes far from the ocean."
    },
    {
      "heading": "Surface energy, meltwater, and subglacial drainage",
      "text": "Surface melting depends on the balance of incoming and outgoing energy, not on air temperature alone. Solar radiation, thermal radiation, turbulent heat exchange, rain, and conduction into the snow all contribute. Melting begins when the surface reaches the melting point and additional available energy supplies the latent heat of fusion. Fresh snow reflects much of the incoming sunlight, while bare ice, dust, soot, and some biological growth can lower reflectivity and increase absorbed energy.\n\nGreenland experiences widespread seasonal melting, particularly around its lower margins. Exceptional warm summers can produce melt even across much of the high interior, as occurred in July 2012. Most of the Antarctic interior is too cold for substantial surface melting under present conditions, but melt occurs on parts of the Antarctic Peninsula and on some coastal ice shelves. Surface melt must be distinguished from net mass loss: water that refreezes within the snow remains part of the ice sheet.\n\nFirn acts as a temporary water reservoir. Meltwater can percolate downward, refreeze, and warm the firn by releasing latent heat. In parts of southeastern Greenland, liquid water persists within perennial firn aquifers. Elsewhere, repeated refreezing forms relatively impermeable ice layers that limit further infiltration and increase surface runoff. The storage capacity of the firn therefore affects how a given amount of melting translates into water reaching the ocean.\n\nWater also travels through crevasses and near-vertical shafts called moulins. If water fills a fracture, its pressure can help drive the crack deeper, a process known as hydrofracture. Water reaching the bed enters a drainage system consisting of cavities, linked passages, channels, sediments, and sometimes large subglacial lakes. The hydraulic gradient depends on both bed elevation and pressure beneath the overlying ice, so subglacial water can move uphill relative to the bed surface while still flowing down hydraulic potential.\n\nDrainage organization influences ice motion. Early in a melt season, water entering a poorly connected system can raise basal water pressure and accelerate sliding. Later, efficient channels may develop and carry more water at lower pressure, reducing the acceleration. Consequently, more surface melt does not necessarily produce a proportionate annual increase in ice velocity. Some rapid changes reflect short-lived water-pressure fluctuations rather than a sustained dynamical response.\n\nAntarctica contains extensive subglacial water systems despite its cold surface. Lake Vostok lies beneath roughly 4 kilometres of ice, and satellite observations have identified numerous lakes that fill and drain. Their activity can change surface elevation and locally modify basal motion. Subglacial hydrology also transports dissolved substances and fine sediment, connecting ice-sheet dynamics with chemical weathering and downstream marine environments."
    },
    {
      "heading": "Mass balance and contemporary ice loss",
      "text": "An ice sheet’s mass balance is the difference between all mass entering and leaving it. Snow accumulation is the principal input. Losses include meltwater runoff, sublimation, wind-driven removal, basal melting, and the export of ice across its margins. For a grounded ice sheet, a useful accounting relationship is\n\n\\[\\frac{dM}{dt}=B_s+B_b-D,\\]\n\nwhere \\(M\\) is grounded ice mass, \\(B_s\\) is surface mass balance, \\(B_b\\) is basal mass balance expressed as positive for net basal freezing, and \\(D\\) is discharge across the boundary. The precise boundary and sign conventions must be specified when comparing estimates.\n\nSurface mass balance includes snowfall, refreezing, condensation, runoff, and sublimation. Melting is an internal transformation until the meltwater leaves the system: snow that melts and refreezes contributes no immediate mass loss. Likewise, the breakup of an already floating iceberg does not represent a new transfer of grounded ice to the ocean. A glacier can retreat while temporarily thickening elsewhere, so terminus position alone cannot determine whole-basin mass balance.\n\nGreenland’s recent losses arise from both increased surface runoff and changes in glacier discharge. Their relative importance varies by period and region. Warm summers can sharply increase annual loss, while unusually snowy or cool years can reduce it. Antarctic grounded ice loss is more strongly associated with discharge changes in particular coastal sectors, especially where ocean-driven thinning of floating ice reduces resistance to inland flow. Surface snowfall variability remains important and can partly offset dynamic losses during some intervals.\n\nA large international assessment covering 1992–2020 estimated that Greenland and Antarctica together lost approximately 7,600 billion tonnes of ice, adding about 21 millimetres to global mean sea level. Greenland accounted for most of that combined loss. The estimate integrates different observational techniques and carries uncertainty; it does not imply uniform loss through time or across either continent. West Antarctica and the Antarctic Peninsula have shown pronounced losses, while East Antarctica exhibits substantial regional contrasts and greater uncertainty in its overall balance.\n\nFor scale, about 360 billion tonnes of water corresponds to approximately 1 millimetre of global mean sea-level rise. This conversion is a global average rather than a prediction of the rise at any particular coast. It also requires appropriate treatment of marine-grounded ice, whose bed lies below sea level.\n\nAnnual mass balance is noisy because atmospheric circulation and snowfall vary strongly. Identifying long-term change therefore requires multi-year records and careful separation of persistent trends from short-lived anomalies. Mass loss can accelerate through either increased surface losses or faster ice discharge, but these mechanisms have different causes and different implications for future behaviour."
    },
    {
      "heading": "Ocean contact, ice shelves, and grounding-line instability",
      "text": "At marine margins, the ocean can affect an ice sheet without extensively melting its upper surface. Water circulating beneath an ice shelf transfers heat to its underside, causing basal melting. The relevant temperature is the water temperature relative to the local seawater freezing point, which decreases with increasing pressure. Water only modestly above that freezing point can produce substantial melting when it is continually supplied to an ice-shelf cavity.\n\nFloating shelves transmit stresses between grounded ice, their lateral boundaries, and elevated seabed features. A shelf confined by embayment walls or grounded on pinning points can exert buttressing, resisting the seaward flow of ice upstream. Removing or thinning such a shelf may accelerate its tributary glaciers. The collapse of the Larsen B Ice Shelf in 2002 was followed by major acceleration and thinning of several formerly restrained Antarctic Peninsula glaciers, demonstrating that floating ice can influence sea level indirectly.\n\nFloating ice already displaces approximately its own mass of seawater, so its melting has little direct effect on sea level compared with the loss of grounded ice. Nevertheless, shelf removal can change the discharge of a much larger grounded reservoir. Calving—the breaking off of icebergs—is therefore not automatically evidence of ice-sheet imbalance. Calving is a normal component of many stable margins; its significance depends on whether ice supply, shelf geometry, and grounding-line position are changing.\n\nA particular concern is marine ice-sheet instability. In a simplified setting, grounding-line retreat into a basin that deepens inland places thicker ice at flotation. Thicker ice can support greater discharge, causing further thinning and retreat. This positive feedback can allow retreat to continue after an initiating disturbance. A bed that deepens inland is an important condition, but not a complete prediction: lateral confinement, bed bumps, shelf buttressing, basal resistance, and changing ocean access can stabilize or redirect retreat.\n\nThwaites and Pine Island glaciers in West Antarctica illustrate the interaction of ocean heat, shelf thinning, and retreating grounding lines. Warm deep water reaching the continental shelf can enter pathways toward their ice cavities. The pathways depend on seafloor topography, winds, and ocean circulation, making local conditions more complex than a direct relationship with global average ocean temperature.\n\nAnother proposed mechanism, marine ice-cliff instability, concerns the possible structural failure of very tall, exposed ice cliffs after protective shelves disappear. Its importance at continental scale remains uncertain, particularly because fracture, mélange, and compensating changes in glacier geometry are difficult to represent. It is distinct from grounding-line instability and is treated as a major structural uncertainty in some high-end sea-level projections, not as an established universal response of marine ice margins."
    },
    {
      "heading": "Ice sheets in geological history and landscape development",
      "text": "Ice sheets have repeatedly expanded and contracted during Earth history. Geological evidence records major glaciations long before the modern polar ice sheets, including Neoproterozoic episodes more than 600 million years ago. The present Antarctic ice sheet developed through a complex history, with a major expansion around 34 million years ago near the Eocene–Oligocene boundary. Declining atmospheric carbon dioxide was an important influence, alongside changes in ocean circulation, geography, and orbital conditions.\n\nLarge Northern Hemisphere ice sheets became especially prominent during the late Cenozoic. During the Pleistocene, their growth and retreat responded to changes in Earth’s orbit and axial orientation, amplified by greenhouse-gas, albedo, and ocean feedbacks. Orbital forcing changes the seasonal and geographic distribution of sunlight rather than supplying a large change in globally averaged annual solar energy. Summer conditions at high northern latitudes are particularly important for whether winter snow survives.\n\nDuring the Last Glacial Maximum, approximately 26,500–19,000 years ago, extensive ice sheets covered much of northern North America and northern Europe. The Laurentide Ice Sheet occupied a large part of present-day Canada and extended into the northern United States; the Cordilleran Ice Sheet covered mountainous western North America. The Eurasian ice-sheet complex included the Fennoscandian and British–Irish ice sheets. Global mean sea level was approximately 125–135 metres below its modern level because more water was stored on land.\n\nIce sheets reshape their beds through abrasion, quarrying, and sediment deformation. Embedded rock fragments scrape bedrock, while basal stresses and pressurized water help detach blocks. These processes can deepen valleys, overexcavate basins, and leave polished or striated surfaces. Deposited sediments form landforms including moraines, drumlins, and till plains. Meltwater constructs additional features such as eskers and outwash plains. Their geometry allows past ice-flow directions and margin positions to be reconstructed.\n\nThe weight of an ice sheet also depresses the solid Earth. After ice loss, the crust and mantle respond through glacial isostatic adjustment, producing continuing uplift in formerly glaciated regions and more complex vertical motion elsewhere. Parts of Scandinavia and Canada are still rising in response to the disappearance of Pleistocene ice. This delayed response affects relative sea level, drainage patterns, and interpretations of modern satellite gravity measurements.\n\nPast warm intervals provide evidence that ice sheets can be smaller than today. During the Last Interglacial, roughly 129,000–116,000 years ago, global mean sea level stood several metres above the present level. Determining how much came from Greenland and from different Antarctic sectors remains an active research problem. Such reconstructions constrain ice-sheet sensitivity, but past conditions are not exact analogues for modern warming because orbital forcing, warming rates, and the duration of exposure differ."
    },
    {
      "heading": "Observation, climate archives, and numerical modelling",
      "text": "Modern ice-sheet science combines field measurements, satellite observations, and mathematical models. Stakes and automatic weather stations measure snowfall, ablation, and surface energy exchange. Satellite positioning instruments track motion, while boreholes provide information on temperature, deformation, and basal conditions. Airborne and ground-based radar detect internal layers and the bed, allowing ice thickness and buried landscapes to be mapped. Seismic surveys help distinguish bedrock, sediment, and water beneath the ice.\n\nSatellite radar and laser altimetry measure changes in surface elevation. These observations must be corrected for processes that change height without an equivalent change in ice mass, especially firn compaction and variations in snow density. Radar measurements also require attention to changes in signal penetration into snow. Laser altimetry has different sampling and surface-reflection characteristics, so agreement between methods is valuable rather than automatic.\n\nThe GRACE satellite mission, launched in 2002, and GRACE Follow-On, launched in 2018, detect changes in Earth’s gravity field that reveal large-scale mass redistribution. These measurements provide estimates of ice-sheet mass change independently of surface-height conversion, but they have limited spatial resolution and require correction for solid-Earth motion. Satellite radar interferometry and feature tracking map ice velocities. Combining velocity with ice thickness estimates the amount of ice discharged through selected cross-sections.\n\nDeep ice cores provide a different kind of observation: a layered archive of past conditions. Water-isotope ratios help reconstruct climatic changes, while bubbles preserve samples of ancient atmosphere. Dust, volcanic material, sea salts, and other constituents record changes in circulation, eruptions, and environmental conditions. The EPICA Dome C Antarctic core yielded a continuous climate and greenhouse-gas record extending about 800,000 years. Dating combines layer counting where possible, identifiable events, ice-flow modelling, and comparisons among records.\n\nNumerical models apply conservation of mass, momentum, and energy to evolving ice geometry. They represent ice deformation, basal sliding, temperature, accumulation, melting, and interaction with the ocean. Computationally efficient approximations are appropriate for some slowly varying interiors, whereas grounding zones and fast outlet glaciers often require more detailed stress calculations. Coupling to atmospheric and ocean models supplies changing boundary conditions but also introduces uncertainty from those systems.\n\nObservation and modelling are closely linked through calibration and data assimilation. Present-day velocities may help infer basal resistance, while historical retreat and thickness changes test model behaviour. Matching a modern snapshot, however, does not prove that a model will respond correctly to a new climate. Bed topography, subglacial drainage, fracture, and ocean melting remain incompletely observed, particularly beneath thick ice and floating shelves. Ensemble calculations explore these uncertainties rather than eliminating them."
    },
    {
      "heading": "Climate feedbacks and future sea-level commitment",
      "text": "Ice sheets influence climate through reflectivity, elevation, freshwater release, and their effects on atmospheric and ocean circulation. Bright snow-covered surfaces absorb less sunlight than darker ground or water. Loss of snow and ice can therefore amplify local warming through an albedo feedback. Large ice sheets also create high plateaus and strong temperature contrasts that influence winds. Cold, dense air flowing downslope generates katabatic winds, especially around Antarctica.\n\nFreshwater from melting ice and exported icebergs changes ocean salinity and density. Around Antarctica, these changes interact with sea-ice formation and the production of dense water. In the North Atlantic, Greenland freshwater is one influence on the conditions governing deep convection and large-scale overturning circulation. The strength and geographic pattern of these effects depend on where and when freshwater enters the ocean, rather than on total meltwater volume alone.\n\nFuture ice-sheet change depends on the magnitude, duration, and regional pattern of warming. A warmer atmosphere can carry more moisture and increase snowfall in some cold regions, partly offsetting losses. It can also increase surface melting, shift precipitation from snow toward rain, and alter the winds and ocean conditions affecting marine margins. Additional snowfall is therefore not equivalent to protection against warming, particularly where faster discharge dominates the mass budget.\n\nThe Greenland Ice Sheet contains about 7.4 metres of global mean sea-level equivalent, and Antarctica about 58 metres. These values describe their approximate total potential contribution, not expected losses over the twenty-first century. Greenland responds strongly to atmospheric warming and elevation feedback: thinning lowers the surface into warmer air, promoting further melt. Antarctica’s nearer-term response is particularly sensitive to ocean forcing and the stability of marine-grounded sectors.\n\nIce-sheet response times extend far beyond the period over which atmospheric temperatures might stabilize. Sustained warming can commit an ice sheet to losses that unfold over centuries or millennia. A threshold for long-term retreat is therefore different from the time at which that retreat is completed. Threshold estimates depend on the model, warming duration, spatial climate pattern, and assumptions about poorly constrained processes. Retreat may also be hysteretic: restoring previous temperatures need not immediately restore the former ice extent.\n\nRegional sea-level change differs substantially from the global mean. Ice sheets gravitationally attract nearby ocean water; as they lose mass, that attraction weakens, the solid Earth deforms, and Earth’s rotation changes slightly. These effects produce characteristic sea-level fingerprints, with reduced rise or even a fall near a shrinking ice sheet and greater-than-average rise in some distant regions. Local projections must combine those fingerprints with ocean circulation, thermal expansion, land subsidence or uplift, and other water-storage changes. The geographic source of future ice loss is consequently important alongside its total volume."
    }
  ],
  "links": [
    "cryosphere",
    "water-cycle",
    "sea-level-rise",
    "glacier",
    "drainage-basin",
    "ice-shelf",
    "density",
    "firn",
    "carbon-dioxide",
    "albedo",
    "friction",
    "energy",
    "latent-heat",
    "surface-runoff",
    "earth",
    "last-glacial-maximum",
    "glacial-isostatic-adjustment",
    "interferometry",
    "ice-core",
    "climate"
  ],
  "sources": [],
  "license": "https://creativecommons.org/licenses/by/4.0/"
}