An ice core is a sample of ice taken from an ice sheet or a glacier on a high mountain. Ice forms layer by layer over many years, with older layers found deeper in the ice. Ice cores can be drilled using hand tools for shallow holes or powered machines for deeper ones. These cores can reach depths over two miles (3.2 km) and contain ice that is up to 800,000 years old.
Scientists study the physical properties of the ice and materials trapped inside to learn about past climates. The amounts of different types of oxygen and hydrogen atoms in the ice help determine ancient temperatures. Tiny bubbles of air trapped in the ice can be tested to find the levels of gases like carbon dioxide in the atmosphere long ago. Because heat moves very slowly through thick ice sheets, the temperature inside the drill hole can also show past temperatures. This information helps scientists compare data to find the best climate model.
Impurities in ice cores vary depending on where they are taken. Ice from coastal areas often includes materials from the ocean, such as salt. Ice cores from Greenland contain layers of dust blown by the wind, which are linked to times in the past when the climate was cold and dry. Radioactive materials, either naturally occurring or from nuclear testing, help scientists date the layers of ice. Some powerful volcanic eruptions left traces in ice cores from many places, which scientists use to match the timing of events across different cores.
Scientists have studied ice cores since the early 1900s. Many projects were started during the International Geophysical Year (1957–1958), when cores were drilled to depths over 400 meters. This record was later improved to 2,164 meters at Byrd Station in Antarctica in the 1960s. The Soviet Union conducted long-term projects at Vostok Station in Antarctica, with the deepest core reaching 3,769 meters. Other deep ice cores have been drilled in Antarctica through projects like the West Antarctic Ice Sheet project, the British Antarctic Survey, and the International Trans-Antarctic Scientific Expedition. In Greenland, team efforts began in the 1970s with the Greenland Ice Sheet Project. Follow-up projects have continued, with the most recent, the East Greenland Ice-Core Project, originally planned to complete a deep core in east Greenland by 2020 but delayed since.
Structure of ice sheets and cores
An ice core is a long tube of ice taken from a glacier. It shows the layers of snow and ice that formed over many years. Each year, new snow falls and covers older layers. As more snow piles up, the weight of the snow above presses down, making the layers below denser. Over time, this compacted snow becomes firn, which is not dense enough to trap air completely. When the density reaches about 830 kg/m³, the firn turns into ice, and the air inside becomes trapped in bubbles. These bubbles preserve the air from the time the ice formed. The depth at which this happens varies by location. In Greenland and Antarctica, it occurs between 64 meters and 115 meters deep. The age of the ice when it forms depends on how much snow falls each year. For example, at Summit Camp in Greenland, ice 230 years old is found at 77 meters deep. At Dome C in Antarctica, ice 2,500 years old is found at 95 meters deep. As more layers build up, pressure increases. At about 1,500 meters deep, the ice’s crystal structure changes from hexagonal to cubic, allowing air molecules to move into the cubic crystals and form a special type of ice structure called a clathrate. The bubbles disappear, and the ice becomes clearer.
About two to three feet of snow can become less than one foot of ice. The weight of the ice above pushes deeper layers outward, making them thinner. Ice is lost at the edges of glaciers through melting or by breaking off as icebergs. Because of this, the overall shape of a glacier does not change much over time. However, the movement of ice can twist the layers, making it harder to study them. To avoid this, scientists drill deep ice cores in areas where ice flows very slowly. These locations can be found using maps that show how ice moves.
Impurities in the ice, such as soot, ash, and particles from forest fires or volcanoes, provide clues about the environment when they were added. Other impurities include special types of atoms like beryllium-10 created by cosmic rays, tiny particles from space called micrometeorites, and pollen. The bottom layer of a glacier, called basal ice, is often made of water from under the glacier that refroze. This layer can be up to 20 meters thick and may contain life forms that live in the ice. However, it usually does not keep clear records of how the layers were arranged.
Ice cores are often drilled in places like Antarctica and central Greenland, where temperatures are cold enough to prevent melting. However, summer sunlight can still affect the snow. During the summer, the Sun is visible for long periods, causing some snow to sublimate, which means it turns directly from solid to gas. This leaves the top inch of snow less dense. When the Sun is lower in the sky, temperatures drop, and hoar frost forms on the surface. Over time, this frost is buried by new snow and compressed into lighter layers than the winter snow. This process creates alternating bands of lighter and darker ice that can be seen in an ice core.
Coring
Ice cores are collected by cutting around a cylinder of ice so it can be brought to the surface. Early ice cores were often collected with hand augers, which are still used for short holes. A design for ice core augers was patented in 1932, and they have changed little since. An auger is a cylinder with metal ribs wrapped around the outside, and cutting blades at the bottom. Hand augers can be turned using a T handle or brace handle, and some can be attached to electric drills. With the help of a tripod, cores up to 50 meters deep can be retrieved, but the practical limit is about 30 meters for engine-powered augers and less for hand augers. Below this depth, electromechanical or thermal drills are used.
The cutting part of a drill is at the bottom of a drill barrel, a tube that surrounds the core as the drill moves downward. Cuttings, or pieces of ice removed by the drill, must be removed from the hole or they will slow down the drill. They can be removed by pressing them into the hole walls or core, using air (dry drilling), or using a drilling fluid (wet drilling). Dry drilling is limited to about 400 meters deep because the hole might close up due to ice pressure.
Drilling fluids are chosen to keep the hole stable. The fluid must be thin enough to reduce the time needed to move the drill equipment up and down the hole. Since retrieving each core segment requires moving the equipment, slower movement through the fluid can add significant time to a project. The fluid must not damage the ice, be safe, be affordable, and be easy to transport. Historically, three types of fluids were used: kerosene-like mixtures with fluorocarbons, alcohol solutions, and esters. Newer options include ester-based fluids, silicone oils, and kerosene mixed with foam agents.
Rotary drilling is a common method for drilling minerals and has been used for ice. It involves rotating a drill pipe from the top and pumping fluid down through the pipe and back up. Cuttings are removed at the top, and the fluid is reused. This method requires long trip times because the entire drill string must be lifted and reconnected. This is difficult in remote areas, making traditional rotary drills less practical. Wireline drills allow the core barrel to be removed from the drill assembly at the bottom of the hole, retrieved, and then reconnected. Another option is flexible drill-stem rigs, where the drill string can be coiled at the surface, avoiding the need to disconnect and reconnect pipes.
Drillpipes can be avoided by suspending the drill assembly on a strong cable that powers the drill motor. These cable-suspended drills work for shallow and deep holes and use anti-torque devices to prevent rotation. Drilling fluid is usually circulated around the drill and back up between the core and barrel, with cuttings stored in a chamber. Some drills can retrieve a second core outside the main one, using the space between them for fluid circulation. Cable-suspended drills are the most reliable for deep ice drilling.
Thermal drills use heat to melt ice but have disadvantages. They use a lot of power, and heat can damage the ice core. Early versions could not go deep without fluid, and later versions slowed trip times. They are also bulky and hard to use in remote areas. Newer designs use antifreeze to reduce power needs. Hot-water drills use hot water jets to melt ice, but they are hard to control and may damage the core.
In temperate ice, thermal drills work better than electromechanical (EM) drills because EM drills can have ice refreeze on their bits, slowing them down. EM drills may also break ice cores under stress.
Deep holes require drilling fluid, so the hole must be lined with a casing to prevent the fluid from being absorbed by snow and firn. Casing reaches impermeable ice layers. A shallow auger can create a pilot hole, which is then expanded to fit the casing. Alternatively, a large auger can be used without reaming. Water can also be used to fill porous snow and firn, which eventually freezes.
Scientists do not need ice cores from all depths equally, leading to shortages at certain depths. To solve this, technology has been developed to drill extra cores from the sides of the borehole. This was successfully done at WAIS Divide in 2012–2013 at four depths.
Drilling projects are complex because they often occur in remote, high-altitude areas. Large projects take years to plan and execute and are usually run by international teams. For example, the EastGRIP project in Greenland is led by a Danish institute and includes scientists from 12 countries. During drilling seasons, many people work at the camp, and logistics include air transport using planes provided by the US Air National Guard.
Core processing
After drilling, several steps are required before ice cores are stored permanently. The drill removes a ring-shaped piece of ice around the core but does not cut beneath it. A spring-loaded lever, called a core dog, can break the core free and hold it in place as it is brought to the surface. Once removed from the drill barrel, the core is usually placed flat to slide out onto a prepared surface. Drilling fluid must be removed from the core as it slides out; for the WAIS Divide project, a vacuum system was used for this purpose. The surface where the core lands should match the drill barrel’s position to reduce stress on the core, which can break easily. Temperatures are kept well below freezing to prevent sudden temperature changes from damaging the core.
A record is kept for each core, noting its length and depth. The core may be marked to show its orientation and is often cut into shorter sections, typically 1 meter long in the United States. Cores are stored on site, usually in spaces below snow level to maintain stable temperatures, though refrigeration may also be used. If more drilling fluid remains, air may be blown over the cores. Samples for initial analysis are collected before the core is sealed in plastic bags, often polythene, and packed with protective materials for transport. During transport, aircraft flight decks are not heated to keep temperatures low, and refrigeration units are used for shipping by boat.
Ice cores are stored in specialized facilities worldwide, such as the National Ice Core Laboratory in the United States. These facilities provide samples for scientific testing, and a large portion of each core is saved for future study.
In a layer of ice called the brittle ice zone, air bubbles are trapped under high pressure. When brought to the surface, these bubbles can cause cracks in the ice because the pressure exceeds the ice’s strength. At deeper levels, the air dissolves into ice structures, and the ice becomes stable again. At the WAIS Divide site, the brittle ice zone was located between 520 meters and 1,340 meters deep.
Ice from the brittle zone is more likely to break and produce lower-quality samples. To reduce damage, liners can be placed inside the drill barrel to protect the core, though this makes cleaning harder. In other drilling methods, special equipment can bring cores to the surface without causing stress, but this is too costly for remote locations. Keeping processing areas very cold reduces thermal stress. Breaking cores into 1-meter sections in the hole before retrieval helps prevent damage. Placing cores in a net as they are removed from the drill barrel can keep them intact if they break. Brittle cores are sometimes stored for up to a year at the drilling site to allow the ice to gradually stabilize.
Ice core data
Scientists use many different methods to study ice cores. These methods include counting visible layers, testing for electrical conductivity and physical properties, and checking for gases, particles, radioactive elements, and other molecules. To understand past climates, scientists must know how deep layers of ice relate to their age. One simple way is to count annual snow layers, but this is not always possible. Another method is to model how snow accumulates and moves through ice to predict how long it takes for snow to reach a certain depth. A third method is to match radioactive elements or gases in the ice with other time markers, like changes in Earth's orbit.
A challenge in dating ice cores is that gases can move through firn, the porous snow layer. This means the ice at a certain depth may be older than the gases trapped in it. As a result, ice cores have two different age records: one for the ice itself and one for the trapped gases. Scientists use models to predict when gases become trapped at different locations, but these models are not always accurate. In places with very little snowfall, like Vostok, the difference between ice and gas ages can be more than 1,000 years.
The size and density of bubbles in ice can show the size of ice crystals when the bubbles formed. Crystal size depends on temperature, so scientists can use bubble properties, along with snowfall rates and firn density, to estimate past temperatures.
Radiocarbon dating can be used to study carbon in trapped CO₂. In polar ice, there is about 15–20 micrograms of carbon in each kilogram of ice, and some carbon may come from dust. Scientists isolate CO₂ by subliming ice in a vacuum, keeping the temperature low to prevent carbon from dust from mixing with the CO₂. Results must be adjusted for carbon created by cosmic rays, which depends on the ice core's location. Carbon from nuclear testing has less impact. Carbon in dust can also be dated by testing organic parts of the dust. Because only small amounts of carbon are found, this method requires at least 300 grams of ice, making it less precise for dating deep layers.
Ice cores from the same hemisphere can often be matched using layers from volcanic events. Matching cores from different hemispheres is harder. The Laschamp event, a change in Earth's magnetic field 40,000 years ago, can be identified in cores. Elsewhere, gases like methane can link Greenland and Antarctic cores. Volcanic layers can be dated using argon/argon dating, providing fixed points for ice dating. Uranium decay has also been used. Scientists use Bayesian probability to combine multiple records for accurate dating. This method, developed in 2010, became a software tool called DatIce.
The boundary between the Pleistocene and Holocene, about 11,700 years ago, is now defined using Greenland ice core data. Formal definitions of climate boundaries help scientists compare findings globally. While ice cores lack fossils, they provide precise climate data that can be linked to other climate records.
Dating ice sheets is essential for understanding past climates. As Richard Alley said, "Ice cores are like 'rosetta stones' that help scientists create a global network of accurately dated climate records."
Ice cores show visible layers that match annual snowfall. If two pits are dug in snow with a thin wall between them and one is covered, the layers can be seen through the light. A six-foot pit may show snow from less than a year to several years, depending on the location. Poles left in the snow show yearly snow accumulation, confirming that visible layers match single years.
In central Greenland, a typical year may add two to three feet of winter snow and a few inches of summer snow. When this becomes ice, the layers form less than a foot of ice. Summer layers have larger bubbles, making alternating layers visible. As ice becomes denser, bubbles disappear, and dust layers may appear. Greenland ice contains dust carried by wind, which appears as cloudy grey layers. These are more visible during cold, dry, and windy periods.
Layer counting becomes harder as ice flows and layers thin with depth. This is worse in areas with high snowfall. In places like central Antarctica, other methods are needed. For example, at Vostok, layer counting works only to about 55,000 years ago.
Summer melting causes snow to refreeze lower in the snow, forming ice with few bubbles. These layers can be identified visually and by measuring core density. This helps calculate the melt-feature percentage (MF), which shows how often melting occurred. MF data is averaged over time to show climate trends. Recent data shows melting rates have increased since the late 20th century.
In addition to manual checks, cores can be scanned digitally by cutting them lengthwise to create a flat surface for imaging.
The oxygen in ice cores can help scientists study past temperatures. Oxygen has three stable forms, with O-18 being heavier than O-16. Water with O-16 is more likely to turn into vapor, while water with O-18 is more likely to condense into snow. At lower temperatures, this difference is more noticeable. Scientists measure the ratio of O-18 to O-16 in ice to estimate past temperatures. This ratio is compared to a standard called standard mean ocean water (SMOW) using the formula:
δ¹⁸O = [(¹⁸O/¹⁶O)sample / (¹⁸O/¹⁶O)SMOW − 1] × 1000‰.
History
In 1841 and 1842, Louis Agassiz drilled holes in the Unteraargletscher in the Alps. He used iron rods to drill, but the holes did not produce cores. The deepest hole reached 60 meters. During Erich von Drygalski’s Antarctic expedition in 1902 and 1903, 30-meter holes were drilled in an iceberg near the Kerguelen Islands. Scientists measured the temperature inside the holes. James E. Church was the first scientist to create a tool for sampling snow. Pavel Talalay called Church “the father of modern snow surveying.” In the winter of 1908–1909, Church built steel tubes with slots and cutting heads to collect snow cores up to 3 meters long. Similar tools are still used today, modified to reach about 9 meters deep. These tools are pushed into the snow and turned by hand.
Ernst Sorge conducted the first organized study of snow and firn layers during the Alfred Wegener Expedition to central Greenland in 1930–1931. Sorge dug a 15-meter pit to examine snow layers. His findings later became known as Sorge’s Law of Densification, named by Henri Bader. Bader later collected ice cores in northwest Greenland in 1933. In the early 1950s, the SIPRE expedition collected pit samples across the Greenland ice sheet, obtaining early data about oxygen isotope ratios. Three other expeditions in the 1950s began ice coring: the Norwegian-British-Swedish Antarctic Expedition (NBSAE) in Queen Maud Land, the Juneau Ice Field Research Project (JIRP) in Alaska, and the Expéditions Polaires Françaises in central Greenland. The quality of the ice cores was poor, but scientists still studied the samples.
The International Geophysical Year (1957–1958) increased glaciology research worldwide, with a focus on drilling deep ice cores in polar regions. SIPRE tested drilling at Site 2 in Greenland in 1956 (305 meters) and 1957 (411 meters). The 1957 core was better preserved. In Antarctica, a 307-meter core was drilled at Byrd Station in 1957–1958, and a 264-meter core was drilled at Little America V on the Ross Ice Shelf the next year. The success of the IGY drilling led to improved ice coring techniques. A CRREL project at Camp Century in the early 1960s drilled three holes, with the deepest reaching 1387 meters in 1966. The drill later reached 2164 meters at Byrd Station before being frozen in the borehole by meltwater.
In the 1960s and 1970s, French, Australian, and Canadian projects collected ice cores. A 905-meter core was drilled at Dome C in Antarctica by CNRS. ANARE drilled a 382-meter core at Law Dome starting in 1969. A Canadian team recovered cores from the Devon Ice Cap in the 1970s. Soviet projects began in the 1950s in Franz Josef Land, the Urals, Novaya Zemlya, and Antarctica at Mirny and Vostok. Not all early holes retrieved cores. Work continued in Asia, with drilling focused on Mirny and Vostok in Antarctica. At Vostok, the first deep hole reached 506.9 meters in 1970, and by 1973, drilling reached 952 meters. Vostok 2 reached 450 meters (1971–1976), and Vostok 3 reached 2202 meters in 1985. Vostok 3 retrieved ice from 150,000 years ago. Drilling paused after a fire in 1982 but resumed in 1984, reaching 2546 meters in 1989. A fifth Vostok core began in 1990, reaching 3661 meters in 2007 and 3769 meters later. Ice at 3310 meters is estimated to be 420,000 years old, but data below that is unclear due to mixing.
EPICA, a European ice coring project formed in the 1990s, drilled two holes in East Antarctica. One at Dome C reached 2871 meters in two seasons and 3260 meters to bedrock after four more years. A hole at Kohnen Station reached bedrock at 2760 meters in 2006. The Dome C core had low snow accumulation, allowing climate data to extend 800,000 years back. Other Antarctic cores included a Japanese project at Dome F, reaching 2503 meters in 1996 (330,000 years old) and 3035 meters in 2006 (720,000 years old). US teams drilled at McMurdo Station, Taylor Dome (554 meters in 1994), and Siple Dome (1004 meters in 1999), with cores from the last glacial period. The West Antarctic Ice Sheet Project (WAIS Divide) reached 3488 meters in 2011. The WAIS Divide core provided climate data back to 115,000 years ago.
Ice cores have also been drilled in non-polar regions, such as the Himalayas and the Andes. Some of these cores provide information about climate changes in those areas.
Future plans
IPICS (International Partnerships in Ice Core Sciences) has created a series of white papers that describe future challenges and scientific goals for the ice core science community. These include plans to:
A warming climate causes glacial meltwater to remove layers of trapped aerosols that show the order of past environmental events. The Ice Memory Foundation plans to store more ice cores in Antarctica before this loss of data happens.