Science Wiki

Sea ice is frozen seawater that floats on the surface of the ocean. It forms, grows, and melts in the polar regions, and it covers about 15 million square kilometers at its seasonal maximum in the Arctic and about 18 to 19 million square kilometers at the maximum in the Antarctic. It differs from the ice of ice sheets, glaciers, and ice shelves, which form on land from compacted snow and are made of fresh water. Because sea ice is already floating, its melting does not change sea level in a significant way, although melting of land ice does. It is a component of the cryosphere, and it strongly influences the surface energy balance, ocean circulation, weather, and ecosystems at high latitudes. It is studied in oceanography, climatology, glaciology, and Earth system science.

Arctic sea ice has declined in extent and thickness since satellite observations began in 1979. The decline is one of the clearest and fastest signals of climate change. The sea ice cover is also variable from year to year, because it responds to winds, ocean currents, and weather, and the Arctic and Antarctic differ in their trends and in their physical settings.

Freezing of seawater

The freezing point of seawater is lower than that of fresh water, because of the dissolved salt. For a salinity of 35 g/kg, it is about −1.8 °C. The empirical relation, with a small correction for pressure, is approximately:

where T_f is in °C and S is the salinity (practical salinity units), which gives −1.9 °C for S = 35. Fresh water has a maximum density at 4 °C, but the temperature of maximum density of seawater with a salinity above about 24.7 is below its freezing point, so that surface seawater continues to become denser as it cools until it freezes. This allows convective mixing of the upper ocean to continue to the freezing point, and makes deep convection possible.

Ice forms in a sequence. Small needle-like crystals (frazil ice) form in turbulent supercooled water and collect at the surface as a slurry (grease ice). In calm conditions, a thin, continuous sheet (nilas) forms. In waves, the crystals consolidate into rounded pancake ice, which freezes together into a continuous cover. Once a cover has formed, further growth is by the freezing of water at the bottom of the ice (columnar ice). In the Antarctic, flooding of the surface of the ice by seawater, when the snow load pushes the surface below sea level, and its freezing contributes much of the ice mass (snow ice).

Salinity and brine

The crystal lattice of ice does not accommodate salt ions easily. As seawater freezes, most of the salt is excluded from the ice crystals, but some is retained in pockets and channels of concentrated brine within the ice. Newly formed ice has a salinity of about 5 to 10, and, over time, brine drains by gravity and by flushing, so that multi-year ice has a salinity of about 1 to 3. The brine that is rejected into the water below increases its salinity and density (brine rejection). The relative volume of brine ν_b within the ice depends on the salinity of the ice S_i and its temperature T (in °C below 0), and a commonly used approximation, valid for roughly −22 °C < T < −2 °C, is:

For example, ice of salinity 5 at −5 °C has a brine volume fraction of about 5 × (9.84 + 0.53) / 1,000 ≈ 0.052, which is about 5 percent. At about 5 percent brine volume, ice becomes permeable (the "rule of fives"), which allows the movement of brine and nutrients, and the drainage of meltwater.

Thickness and growth

The growth of ice is controlled by the conduction of heat through it. A simple model, due to Stefan, treats the ice as a slab with a surface temperature T_s below the freezing point T_f at the bottom. The latent heat released at the base is conducted through the ice, so the thickness h grows as:

where k is the thermal conductivity of ice (about 2.1 W·m⁻¹·K⁻¹), ρ_i is the density of ice (about 910 kg·m⁻³), L_f is the latent heat of fusion (3.34 × 10⁵ J·kg⁻¹), and t is time. For a temperature difference of 20 K over 90 days (7.8 × 10⁶ s), h ≈ √[2 × 2.1 × 20 × 7.8 × 10⁶ / (910 × 3.34 × 10⁵)] ≈ √(2.16) ≈ 1.5 m. Snow on the ice is a good insulator, and it slows growth, so that real growth is slower, and the thickness grows with the square root of time, so that thin ice grows faster than thick ice. This negative feedback is a reason why thin ice can regrow in winter.

Types of ice

Sea ice is classified as first-year ice, which has not survived a summer melt season, and multi-year ice, which has survived at least one. Multi-year ice is thicker (about 3 to 4 m, compared with about 1 to 2 m for first-year ice), is less saline, and is stronger. A large part of the multi-year ice in the Arctic has been lost since the 1980s. The ice is also divided into fast ice, which is attached to the coast or to the seabed, and pack ice, which drifts with winds and currents. Dynamic processes deform the pack, and cause ridges, in which ice blocks pile up (and the keels below, which can reach tens of meters), and leads, which are cracks of open water. Polynyas are persistent areas of open water within the ice that are produced by winds or by the upwelling of warm water. They are places of intense ice formation, because the open water is exposed to cold air.

Albedo and the ice-albedo feedback

The albedo of a surface is the fraction of incident shortwave radiation that it reflects. Fresh snow on sea ice has an albedo of about 0.8 to 0.9. Bare, cold sea ice has about 0.5 to 0.7. Melting ice, covered by melt ponds, has about 0.2 to 0.5, and open ocean has about 0.06 or less when the Sun is high. The reflection of sunlight by ice is not "up to 90 percent" for all ice, but the highest values are for fresh, dry snow. A loss of ice therefore increases the absorption of solar energy. In the Arctic summer, the Sun is up for most of the day, and the change of albedo from 0.6 to 0.07 for an incoming flux of 250 W·m⁻² increases the absorbed flux from 100 to 232 W·m⁻². This is the ice-albedo feedback, which amplifies the warming and is one of several contributors to Arctic amplification, in addition to changes in lapse rate, atmospheric and oceanic heat transport, and cloud and water vapor feedbacks. The ice-albedo feedback is a positive feedback and it is partly offset in the long term by other effects. Melt ponds on the surface of ice and the loss of snow also change the albedo, and the transmission of sunlight through thin ice supports algae under the ice.

Insulation and air-sea exchange

Sea ice is a barrier between the ocean and the atmosphere. In winter, the ocean is much warmer than the air (the air can be below −30 °C), so that, where open water exists, there is a large flux of heat, of the order of several hundred W·m⁻² in leads and polynyas, whereas through thick ice it is a few W·m⁻². This is why the loss of sea ice increases the transfer of heat and moisture from the ocean to the atmosphere in autumn and winter, which warms the lower atmosphere, increases cloud cover and snowfall, and affects regional weather. Whether the loss of Arctic sea ice influences the weather at mid-latitudes, through changes in the jet stream and the polar vortex, is a subject of active research and debate, and the evidence from models and observations is mixed. Sea ice also reduces the exchange of gases, including carbon dioxide and methane, and it reduces the waves and the transfer of momentum from the wind to the ocean.

As sea ice forms, the rejected brine makes the water below it denser. In regions with intense ice formation, such as the Antarctic continental shelves (the Ross Sea, the Weddell Sea, and the coastal polynyas) and, to a lesser extent, the Arctic shelves, this dense water sinks and, together with water from ice shelf cavities, forms Antarctic Bottom Water, which spreads in the deep ocean and is a major part of the deep limb of the global overturning circulation. Dense shelf water is formed by brine rejection and by the cooling of the ocean, with a contribution from the interaction with the floating ice shelves. The statement that the brine "violently plunges" and is the sole driver of the thermohaline circulation is not accurate. The overturning circulation is also driven by the sinking of cold, salty water in the North Atlantic (which is not produced by sea ice formation alone), by winds over the Southern Ocean, and by mixing. In the Arctic, the loss of sea ice, and the freshening of the upper ocean from ice melt and river runoff, can increase stratification, but it is uncertain how this affects the Atlantic overturning. The overturning circulation transports heat, carbon, and nutrients, and ventilates the deep ocean, and a change in the production of dense water has potential consequences for them. Observations indicate that the formation of Antarctic Bottom Water has declined in recent decades, with the freshening of shelf waters and the melting of ice shelves as possible causes.

Sea ice is a habitat. Brine channels and the underside of the ice host a community of sea ice algae (diatoms are common), bacteria, archaea, protists, and small animals, which is called the sympagic community. Algae under the ice start to grow in early spring, when light penetrates the ice and snow. They are an important source of food for zooplankton, such as Arctic copepods and krill (in the Southern Ocean, juvenile Antarctic krill feed on the algae on the underside of the ice in winter), and the carbon that they produce is exported to the seafloor or is eaten in the pelagic food web. In the Arctic, the ice algae account for a small to moderate part of the primary production (about 4 to 25 percent in many estimates, and more in the central Arctic), but they are present early in the season. At the spring melt, ice algae are released into the water, and the melting also increases stratification and the availability of light, which can lead to a bloom of phytoplankton, and under the ice, blooms have also been observed where melt ponds and thin ice allow the passage of light. The ice is also a platform for animals, such as polar bears, ringed seals and other ice-breeding seals, walruses, and emperor penguins, which rely on sea ice for hunting, breeding, or resting. Changes in the ice extent, the timing of the melt, and the thickness of the snow affect them, and the loss of habitat is a threat to specialized species. The loss of ice also allows an increase of open-water production and the northward movement of species from lower latitudes, so that the effect on the food web is complex. It is also not the case that the loss of the ice "eradicates" all production, because the open water is productive, but the timing and the composition of the ecosystem change.

Arctic

Satellite passive microwave measurements since late 1978 show a decline of the Arctic sea ice extent in all months, with the largest in September, at the end of the melt season. The September extent has fallen by about 12 to 13 percent per decade relative to the 1981 to 2010 average, and the March extent at its maximum by about 2 to 3 percent per decade. The record low of the minimum in the satellite record was about 3.4 million square kilometers in September 2012, and the minimum in recent years has been about 4 to 4.7 million square kilometers. The area of multi-year ice has shrunk greatly, and the mean thickness of the ice in the central Arctic has fallen by more than half since the 1950s and 1960s according to submarine and satellite records, with uncertainty. Models of climate indicate an Arctic that is nearly ice-free in September (defined as less than 1 million square kilometers) at some point in this century under higher emissions. The timing is uncertain, because of natural variability, and it may be in the 2030s to 2050s in some scenarios, and later with strong reductions of emissions. A first ice-free September has not occurred. The summer ice loss is approximately linear in the cumulative CO₂ emissions, at about 3 square meters of September ice per metric tonne of CO₂, which is a relationship found in models and observations. The sea ice may recover quickly if the climate cools, because first-year ice grows rapidly, so that the loss is not an irreversible tipping point in the sense of a threshold, and most assessments consider the loss of summer sea ice to be reversible on decadal timescales if the global temperature falls. The statement that the loss of sea ice is a "tipping point" in the strict sense is not supported by the evidence.

Antarctic

Antarctic sea ice behaved differently. The extent increased slightly from 1979 to about 2014, with a maximum of about 20 million square kilometers in 2014, and then fell sharply in 2016 and again in 2023, when the winter maximum and the summer minimum were record lows in the satellite record. The causes of the earlier increase and of the recent decline are under investigation. They include changes in winds associated with the Southern Annular Mode, ocean stratification and the heat of the subsurface ocean, the freshening of surface waters, and natural variability, and whether the recent low extents indicate a long-term change of regime is not yet clear. Because the Antarctic ice is mostly first-year ice, which melts in summer, the extent responds more to the winds and the ocean than that of the Arctic.

The extent (the area with at least 15 percent ice concentration) and concentration are measured by passive microwave sensors (SMMR, SSM/I, SSMIS, AMSR2). The thickness is measured by submarines (upward-looking sonar), by laser and radar altimeters on satellites (ICESat, ICESat-2, CryoSat-2, and others), by airborne surveys (such as NASA's Operation IceBridge), by moored sonars, and by electromagnetic induction. Radar altimetry depends on the thickness and the density of the snow on the ice, which is a major source of uncertainty. The motion of the ice is tracked by buoys and by satellite imagery, and field campaigns, such as the MOSAiC expedition in 2019 to 2020, in which the ship Polarstern drifted with the ice for a year, provide detailed data. Sea ice is also represented in climate models and in sea ice models, and the models tend to underestimate the observed rate of the decline of the Arctic summer ice.

• Cryosphere
• Arctic sea ice decline
• Ice-albedo feedback
• Arctic amplification
• Antarctic Bottom Water
• Thermohaline circulation
• Brine rejection