Text and pictures © 2005-2021 Guillaume Dargaud
Last updated on 2020/03/19
"You can't dig a different hole by digging the same one deeper." — Edward de Bono (1933-).
Left: The drill coming up.
Epica is an acronym for 'European Project for Ice Coring in Antarctica'. Launched by various european laboratories, it drills at various sites: Dome C from 1996 to 2004, Talos Dome since 2005, Dronning Maud land, etc... after similar projects were completed on the Greenland ice cap. The large ice sheets of Greenland and Antarctica are the single most important source of information about past climates, in particular past atmospheric concentrations in carbon dioxide, methane and nitrous oxide. Past temperatures can also be reconstructed by analyzing proportions of isotopic HDO (deuterium water) and H218O. Sodium, calcium and sulfates provide information on past sea salt concentration, soil dust, marine life as well as volcanic activity.
Right: The ice core is visible, still inside the drill.
In this series of images I show the cycle of bringing a drill up, recovering the ice core from it before sending it back down. This cycle is repeated every meter of ice, sometimes less, while it takes more than half an hour for the drill to drop down. So it can be two hours between the recovery of each core, the deeper the slower.
Left: Disassembling the drill. Several meters long, the drill contains many different parts: a cutting head, an internal space where the ice core slides, anti-return blades to maintain the core in place, side blades to maintain the outer part fixed in the ice while the head and internal screw rotate and bite the ice, a top part with engine and control systems, etc...
Right: Taking the drill apart. The fluid bath is open and ready to receive the inner drill chamber containing the core.
Previous to the Dome C core, there had been only two deep drillings in Antarctica at Vostok and Dome Fuji, neither of which reached bedrock. And contrary to Vostok, the site of Dome C needs only minor corrections for ice flow. Drilling in the north and south also helps determine whether the sudden changes found in the records are limited to an hemisphere (or less) or are indeed global. Another important question is whether the last 10000 years of climate stability are unusual or not in respect with the usual variability of ice-age periods.
Left: Extracting the screw containing the ice core from the drill casing. Chips soaked in drill fluid are falling off.
It has taken 10 years for the Epica project to reach rock bottom. Rock bottom meaning complete success in that case: 3270 meters of ice extracted from underneath our feet. When the project started drilling in 1996, there wasn't much: a basic drill, not tent over it, no workshop and just a few containers at the camp. That first year, they reached only 130m deep, but it was most of the toughest work: setting up the equipment and casing the hole, as the first hundred meters or so is made of soft snow which breaks and can plug the hole too easily. Below the hole can also close due to the pressure on the ice, so a fluid of the same density as the ice (923kg/m³ at -53°C), made mostly of special kerosene fuel, is poured into the hole to balance the pressure.
Right: The drill inside a kerosene bath at a fixed temperature, slightly below freezing, so as to unlock the drill.
Humidity at Dome C originates mostly from the southern Indian and Pacific ocean. Another Epica drilling site is at Dronning Maud land, where the influence is mostly from the south Atlantic. Before drilling took place, a detailed survey of the surface and bedrock topography was performed using radars.
Left: Removing the drill from the kerosene bath after it's warmed up enough. The ice core, visible inside, should now be removable.
So they started digging, and cycle after cycle brought more ice cores to the light of day. Ice cores containing precious information about past climates while at the same time the debate on global warming was picking up in the rest of the world. In early 1998, the hole was 364m deep, in 1999 it got stuck at 781m and had to be restarted; but by 2001 it was already at 1459m, in 2002 2864m and in 2003 3201m. The target being 'bedrock' when you actually touch the ground under the ice and even retrieve pieces of it, a feat performed in Greenland where mud and refrozen water cores where brought up.
Analysis of an ice core consists of various methods: analysis of gas trapped in air bubbles, analysis of isotopes present in the ice or the air bubbles, optical analysis of the ice (polarity of thin slices, transmission properties), size of the crystals, electrical properties of the ice (electrical resistance depending on dissolved salts), temperature where it was extracted, thickness of each layer, time reconstruction, search for inclusive extraneous materials (normally near bedrock)... Instruments used for ice-core analysis can be as diverse as mass spectrometers, polarizing microscopes, band saws, microtomes (an ultra-thin slicer), ohmmeter...
Right: Scrubbing the ice core to remove some of the contaminating drill fluid.
Left: Measuring the diameter of the ice core, to check for potential problems or wear-out of the mechanisms.
Right: Breaking the crust of ice formed over the drill engine mechanisms, so the parts are accessible for cleaning.
Left: All the snow and ice removed from the drill and not part of the core itself is recovered and put into a centrifugal machine to recover and recycle the drill fluid and avoid contamination of surface snow.
Right: Taking the drill apart for complete cleanup.
The drilling rig is a complex electromechanical system. A winch unrolls a cable on a pulley, which takes down the drill itself. The outer part of the drill is smooth metal, except for 4 anti-torque blades (we want the head to rotate and cut the ice, not the whole drill). Inside, from top to bottom, there's an electronic control system, an engine and associate gearbox, a tank to store ice chips with a driveshaft going through, a core chamber surrounded by a spiral screw, and finally a bladed cutting head. The drill can be brought to a horizontal position for easier work once it is out of the hole. Other drill designs exist, such as using an electric heater to melt the ice instead of cutting it with a blade.
Left: The 'chip evacuation screw' drill ready for cleaning.
The drill creates a hole 129.6mm in diameter, while the internal diameter for storing the core is 98mm. The head has 3 blades in hardened steel cutting at a 45° angle. There are also 3 locking blades to break the core and hold it in position until the drill reaches the surface. Three spirals above the head carry the cut chips to a compartment above for storage: the hole must stay clean of ice.
Right: Cleaning the drill parts after removal of the core, a spray of drill fluid is used as all those operations are performed well below freezing temperatures.
Left: Reassembling the drill.
Right: Preparing a bag of drill fluid to insert into the drill before sending it down.
The drilling tower is 13 meters high, with part of it underground, and it can stand 8 tons of weight on the cable. A load sensor avoids undue strain in case the drill gets stuck, something that sometimes happen near the surface, where soft snow can change shape or break easily, or near bedrock where warmer ice turns to slush and mud. The 4km of cable run up to 1.4m/s on a winch with 100Nm of torque. The 200kg of cable have a double use: hold the weight of the drill (elastic limit around 3 tons) and transmit electricity used to power and control the drill. Yes, it is a coaxial cable.
Left: Adding a plastic bag of drill fluid of a lesser density, so that it breaks apart when the drill reaches the ice. It helps avoiding lockups of the mechanism.
The drill fluid poured into the hole is a mix of white spirit and forane with ratios adjusted for the proper ice density. 18 liters of fluid are used for every meter drilled, bringing the total to 53 thousand liters poured down the hole.
Right: Reassembling the drill parts.
Left: Inserting the drill into its casing. This space holds the ice chips.
Right: Reassembling the head of the drill. The sharp blades are clearly visible.
Left: Reassembly of the drill.
The drilling tent, nicknamed the cathedral, contains a control cabin, the winch tower and drill, a table to extract the core, a kerosene bath and all the equipment necessary to perform the drill opening and reassembly. No analysis is performed there, but the core is carried to other Epica buildings at the summer camp.
Left: Getting ready to shift from horizontal to vertical position.
From the control cabin, one has access to various parameters: motor current, winch rotation speed, head rotation speed, penetration speed, stress load, temperatures, pressure, angle, depth, duration... Hole deformation is monitored across the runs (diameter change) but also across the years (deformation and angle change due to ice flow) thanks to a special 'logger' sent down instead of the normal drill.
Right: The drill being put into a vertical position before being sent back down.
Getting the drill stuck is an ever present risk. If it happens very near the surface, it is sometimes possible to make an external hole to reach it, for instance with a bulldozer. If somewhat below, the drill can be split, part of it abandoned and a new hole restarted nearby. If the accident happens deeper, the loss of time to restart the operation can mean several years lost on the project, so all is tried to free the drill: pulling to the elastic limit of the cable, pouring warmer and denser fluid inside the hole to try and melt the ice locking it in place, shaking it, etc...
Left: Positioning the drill before sending it down.
Temperature in the hole goes down sharply (in summer) in the first 10 meters to -53°C (the average temperature at Dome C), but then gradually goes up: -45°C at -1000m, -45°C at -1000m, -29°C at -2000m, -8°C at -3000m and finally -3.5°C where the drill was last stopped. Note that at such high pressure, ice melts below 0°C, so the risk of hitting wet slush ice near bedrock is high. This temperature increase is due to geothermal flux coming from the center of the planet. The problem with bringing back warm ice through colder layers of ice above, is that the whole drill freezes shut on its way up, and need to be warmed up in a temperature controlled drill-fluid bath for a while before dismantling and core recovery is possible.
Left: Cores waiting to be processed are placed in shelves at the entrance of the cold lab.
Right: Cutting a core with a band saw.
In order to determine the structure of the last few meters above bedrock, seismic reflection was used: explosives were triggered in shallow holes below the surface and a microphone was placed at the then bottom of the hole. The incoming sound was recorded, and then the reflected sound. Two separate layer were found, one corresponding to the bedrock itself, but another one attributed to a layer of liquid water two meters thick above the rock. The presence of as many as several hundred lakes under the Antarctic ice shelf has been known since detailed radar surveys in the last decade, but since then, other studies have confirmed the presence of rivers connecting those lakes, so a large part of Antarctica actually has water underneath the ice.
Left: Detail on a core cutting process.
Right: An ice core, split lengthwise so as to get access to the clean center. The bigger part stays in Dome C for reference while the slice is sent to european laboratories for detailed analysis.
Left: A bunch of ice cores ready for pre-processing.
Right: Display of a couple thousand years worth of climatic data. At those depths, one meter of ice can hold 500 years of climate information.
Right: The cores as well as their wrappings have detailed information on them: location, depth and time of extraction, up and down directions, estimated size of the missing parts on top and bottom, etc...
Left: Inger putting the cores into cold storage boxes after wrapping them up with details written on
On December 21st 2004, drilling was stopped at a depth of 3270.70m, just about 6 meters above the 2m thick water layer in order to avoid risking getting the drill stuck in slush. Mission accomplished for the Epica team after 8 years of field work. Champagne time.
Left: The Epica project has several buildings and tents at the summer camp. Besides the tall drilling tent into which most of the other pictures are taken, there's a cold lab used to work on ice cores (see above), a warm lab not used anymore and this workshop tent where most mechanical equipment is stored.
Right: For the bedrock reach success, the Epica team throws a party in their workshop tent. The special treat, besides a large bottle of Champagne, is ice cubes taken from one of the deepest cores. They tinkles as the bubbles of super-compressed air are released from their tremendous pressure as the ice melts. Only drawback: it still tastes like drill fluid, i.e. kerosene.
Left: After the party. In the next few days the Epica people depart, taking their precious ice samples with them and leaving the deepest ice hole in the world. They are going to start a new hole at Talos Dome, but will be back regularly to check the hole for deformation, indicating some possible shearing ice flows.
Particular thanks to Laurent Augustin (Epica leader), Olivier Cattani, Inger and the rest of the Epica team for their explications and willingness to pose for pictures. Epica was the biggest science project of the first 8 years of activity at Dome C. Now that the deep drilling project is finished, activity is shifting to surface glaciology to calibrate the cores and also onto newer big projects... But there's still a lot of work going on in labs in Europe before all the Epica data is fully analyzed... Yes, you can access the data and play with it to check your own doomsday global warming scenario if you wish.