Barclay Kamb and Hermann Engelhardt (Caltech),
Frank Carsey, Lonne Lane, and Alberto Behar (JPL)

In a recent study of Anarcticas Ice Stream C, a joint NSF-NASA research team has discovered a surprising gap between the base of the ice stream and the rock below. The gap is up to 1.4 m (60 inches) wide, and is filled with water at a pressure nearly high enough to lift the ice stream off its bed. The discovery is significant in relation to the mechanism for rapid movement of the ice streams, which are huge, fast-flowing ice currents within the slow-moving ice sheet that covers most of Antarctica. The mechanism for their rapid movement is being intensively studied because of the possibility that rapid ice-stream flow may cause the ice sheet to disintegrate, resulting in a disastrous rise in world-wide sea level.

The discovery of the gap between ice and bed was made with a new instrument, the JPL/CIT Antarctic Ice Borehole Probe, developed by a team of engineers at NASAs Jet Propulsion Laboratory (JPL) to study the basal zone of the ice streams, where the fast-flow mechanism is believed to operate. The probe carries two video cameras, one looking straight down and the other looking to the side. The probe is lowered down boreholes drilled through the ice to the bottom, and can then observe features and processes in the basal zone. The boreholes, about 1200 m (3600 ft) deep, are drilled with a hot-water jet drill developed by Caltech(CIT). The drill development and the Antarctic field work and interpretation of the probe results were carried out by a Caltech/JPL team with support of the National Science Foundation (NSF) Office of Polar Programs (OPP). The team members are listed at the end of this report.

In the work leading to discovery of the gap, the Caltech/JPL team put the probe down three boreholes about 2 miles apart, in the vicinity of camp Ice Stream C at latitude 82o 22'S, longitude 136o 24'W. In the first borehole, the ice was frozen to the bed, and there was no gap. In the second, the ice was not frozen to the bed, and there was a gap about 1 or 2 inches wide between the bed and the base of the ice. In the third hole, a gap 62 inches wide was found, far exceeding anything previously observed or inferred under the ice streams. This wide gap was recognized from a combination of features seen in the probes videos. As the probe approached the bottom, the side-looking video came upon a sharp horizontal boundary, above which was nearly clear ice containing a scattering of imbedded rock fragments, and below which was slightly turbid water containing small particles in turbulent motion (see image #1 attached to this report). This boundary could be seen also in down-looking images as a bright sharp rim below which the borehole wall did not continue (see image #2). The underside of the ice was seen in side-looking images as a planar surface cut by a pattern of grooves that mark the boundaries between individual ice crystals (image #3). (For technical reasons the side-looking video images (#1 and #3) are upside down, that is, up on the page is down in the borehole.) As the probe was lowered farther, the bottom of the borehole came into focus in the down-looking video, and was made particularly clear by the probe-chainpiled up on the bottom ( image #4). This chain 0.75 m (30 inches) long, hung from the bottom of the probe and served to provide to the down-looking video an advanced indication of the probe's approach to the bottom of the borehole; the chain is visible in image #1. For image #4 the probe was at depth 1064.55 m (3492 ft), and for image #1 it was at depth 1063.25 m (3488 ft). Thus the width of the gap, determined from the combined side-looking and down-looking video observations, was (1064.55-1063.25+0.2)m = 1.5 m (59 inches). (The 0.2 m is the length of the chain not resting on the bottom in image #3 plus the height of the side-looking camera lens above the bottom of the probe.) The horizontal dimensions of the gap were beyond the illumination from the probe's sideward-shining floodlamp, which is thought to be about 0.5 m under the conditions of the water clarity that prevailed in borehole no.3.

The discovery of a substantial gap between the base of the ice and the bed is quite surprising. No such gaps have been previously detected or inferred under the ice streams. Antarctic subglacial lakes (like Lake Vostok) involve a substantial ice-bed gap, of course, but it is on a dimensional scale hundreds to thousands of times greater than that involved here; also, no such lakes have been found in association with ice streams. Theoretical calculations that interpreted earlier (pre-probe) ice-stream observations have assumed the existence of a gap, but the calculationsbased on the rate at which water exits from boreholes and enters a basal conduit systemimplied only a narrow gap, about a tenth of an inch wide. Some lines of reasoning suggest that such a gap is not present in the natural system at all but is opened up in the drilling process when borehole water under a high overpressure is injected into the ice/bed contact and opens up a gap along it. It seems quite unlikely, however, that a gap 1.5 m (59 inches) wide could be formed in this way during the short time (approximately two minutes) during which the high pressure is applied.

Alternatively, such gaps are a known feature of glaciers that move by sliding of the ice over the bed. The basal sliding can cause cavities to open up between the ice and the bed, generally in the lee of protuberances in the bed. The process is called ice-bed separation or basal cavitation. The size of the cavities is controlled by the basal sliding speed, the amplitude of the basal roughness, and the basal water pressure. At high water pressure, slightly below the ice overburden pressure, the cavities are generally thin in the vertical dimension compared to their horizontal dimensions, and when penetrated by the hot-water drill they should appear as a gap between the base of the ice and the bed. This appears to be the situation encountered by the probe in boreholes no. 2 and 3.

The significance of the probes observations of a basal gap is thus that in the neighborhood of boreholes no. 2 and 3 the ice stream moves by basal sliding. This is contrary to a widely held view that the motion is by shear deformation of soft sediment underlying the base of the ice. The mechanical decoupling of the ice from the bed, which is a necessary consequence of the gap, should contribute to rapid ice-stream motion, and it will introduce a significant complication into the ongoing efforts to analyze and model the rapid-flow mechanism.

Another point of significance is the gaps role in the basal water-conduit system. In down-looking video at the bottom of borehole no. 2 the probe observed plumes of turbid water exiting from the bottom of the borehole via the basal gap, proving that the gap was part of the basal water system. (The water was leaving the borehole in response to clean water being pumped into the hole near the top.) In borehole no. 3 the situation was probably similar, producing the observed turbulent motions of water in the gap. The existence of the wide gap shows that the basal water system includes large cavities that can store and release large volumes of water. This must be important in the functioning of the water system and its influence on the basal water pressure, which is probably a key parameter in the ice-stream mechanism.

In addition to observing the basal gaps in boreholes no. 2 and 3, in all three boreholes the probe observed a basal layer of ice loaded with rock debris, which has not been previously recognized in studies of the ice streams. The debris-laden layer was 15.8 m (52 ft) thick in borehole no. 1, 25 m (82 ft) in no. 2, and 11.6 m (38 ft) in no. 3. From their appearance in the video, various types of debris-laden ice could be identified, depending on their content of coarse rock fragments and fine clay particles (see image #5). Except for a relatively rock-poor layer at the base of the ice in borehole no. 3 (see image #2), and except for complex interbedding of the different debris-laden ice types (image #5), the rock and clay contents generally increase downward toward the bed, and may reach 60% rock by volume. This would qualify them as possible products of basal freeze-on, in which water-saturated subglacial rock debris gets frozen on to the base of the ice. Detection of basal freeze-on would be an indication that the ice mass is tending towards becoming frozen to the bed, which could be responsible for the great slowdown of Ice Stream C that occurred 150 years ago. Basal freeze-on operating for 150 years could add about 1.5 m (5 ft) to the thickness of the basal debris-laden ice layer, which is much less than the observed 12-25 m (38-82 ft). Other mechanisms are needed for emplacing rock debris so high above the bed and for distinctions among the various types of debris-laden ice, whose role in ice- stream mechanism is not known.

It may be noted that the fine clay particles and coarse rock fragments seen imbedded in the ice by the borehole probe represent types of micro-environment that have been suggested as natural microbe repositories, whose investigation may be an objective of future research efforts directed toward Lake Vostok and Europa. To study such micro-environments in detail, the resolution and magnification of the borehole probes cameras need to be substantially increased, but the feasibility of such studies appears encouraging from the success of the ice borehole probe in the present work.

The work was made possible by support of the JPL effort by NASA and support of the Caltech effort by NSF. Details of preliminary results including some 70 still images and 30 video clips are posted on web site http://helios.jpl.nasa.gov/~behar/JPLAntIceProbe.html.

The NSF-NASA research team consisted of the following participants (CIT=Caltech):

JPL: Alberto Behar, team leader; Frank Carsey, PI and project manager;
Lonne Lane, Robert Ivlev, Ken Manatt, Kobie Boykins, engineers;
Fabien Nicaise, and Kai Zhu, technician.
CIT: Hermann Engelhardt, science advisor.

CIT: Hermann Engelhardt, supervisor; Robin Bolsey, technician.

JPL: Alberto Behar, probe operator.
CIT: Barclay Kamb, P.I.; Hermann Engelhardt, Co-PI and field operations supervisor;
Robin Bolsey, technician; Stefan Vogel, Shulamit Gordon, Regina Sterr, Michele Koppes, Katherine Batten, Matt Bachmann, and Daniel Adams, field assistants.
NSF/RPS: Sarah Grundlach, camp manager; Alex Papasava, cook; Jim Evans and Richard Grillo, mechanics; Sarah Harvey, general assistant. (RPS=Raytheon Polar Services)

Image #1 (side-looking in borehole no. 3 at depth 1063.25m).Boundary between basal ice and subglacial water-filled gap. The ice, in the lower 3/4 of the image, appears black (clear) with numerous bright spots, which are imbedded rock fragments. The water contains fewer particles. The boundary is the straight horizontal bright line one-fourth of the way down the top of the image. Because the image is upside down (for technical reasons), the top of the image is toward the bottom of the borehole.
Image #2 (down-looking in borehole no. 3 at depth 1062.65m).Base of the ice as seen looking down the borehole from 3 meters above the bed. The ice in the closer part of the borehole wall appears whitish because it is heavily laden with rock debris. Rocks partially melted out of the ice can be seen protruding through the borehole wall. Below the debris-laden zone is a black-appearing, clearer ice layer containing some rock debris. Below this layer is a partially bright, circular ring, whose sharp inner edge is the line of intersection of the borehole wall with the base of the ice. The probe-chain is seen hanging down from the probe in the lower right part of the image.
Image # 3 (side-looking in borehole no. 3 at depth 1063.29m).In the lower one-quarter of this rather dark image one can see the base of the ice above the water-filled basal gap. (Remember that because of image inversion, down is up.) The ice base (or sole) is a flat surface cut by grooves in a pattern somewhat resembling mud cracks. The grooves may be formed by a thermal etching process acting along the grain boundaries between adjacent ice crystals in the polycrystalline mass.
Image #4 (down-looking in borehole no.3 at depth 1064.55m).Bottom of borehole no. 3. The lower part of the probe chain, with a plastic marker at the tip, is curled up lying on the bottom. The bottom consists of small rock fragments mostly 2-10 mm in size. The two bright spots in left and right center result from light scattering by faint turbidity of the water in the gap.
Image #5 (side-looking in borehole no.2 at depth 1224.90m).Interbedding of three debris-laden ice types seen in the borehole wall just above the bed. The light gray layer in the center is type-2 ice. Below it in the image is type-1 ice (black with a scattering of white chunks). Above it is type-3 ice (granular mass of rock debris with a little interstitial black (clear) ice).