Based on their drilling experience and hot-water drilling records, B. Kamb and H. Engelhardt suspected that a debrisrich basal ice layer existed beneath West Antarctic ice streams (personal communication from Reference Engelhardt, Kamb and BolseyKamb and Engelhardt, 2000). Several attempts to recover ice from the base of West Antarctic ice streams were made in the 1990s using the Caltech hot-water ice-coring drill (Reference Engelhardt, Kamb and BolseyEngelhardt and others, 2000). During the 2000/01 Antarctic field season, we utilized an ice borehole camera system (Reference Carsey, Behar, Lane, Realmuto and EngelhardtCarsey and others, 2002) to obtain the first in situ live-stream investigations, and we confirmed the existence of a sediment-laden basal ice layer beneath Kamb Ice Stream. We observed the basal ice layer to be 10–14 m thick, accreted to the base of the ice sheet by the freeze-on of subglacial meltwater (Reference VogelVogel and others, 2005). The sediment incorporated onto the basal ice in this process consisted of particles ranging in size from clays to pebbles and cobbles. Sediment content ranged from individual dispersed sediment clots (5 mm in diameter, mainly consisting of clay and silt) to layers of thick frozenon sediment (well-mixed diamictide). The acquisition of a short section (70 mm core diameter) of clear basal ice with a single layer of dispersed sediment clots (Fig. 1) from the glacial/basal ice transition confirmed the nature of the basal ice. However, our attempts to recover ice with a higher sediment content covering the subsequent portion of the 10 m thick basal ice layer were unsuccessful. After detailed inspection of the ice corer it was discovered that the core catcher had jammed. Sediment melted out from the basal ice had moved into the V-shaped gap between the core catcher and the slot in which the core catcher was mounted. Due to the geometry of this slot, a few sand-size sediment particles were sufficient to prevent the core catcher from engaging the core.

Fig. 1. Ice from the top of the basal ice layer at Kamb Ice Stream showing thin highly dispersed banded debris layers.

Video observations show that the 4 m long section of basal ice that we had attempted to drill was characterized by an increase in sediment content. Individual particles, mainly sediment clots up to several mm in diameter, changed to cm thick sediment layers, which were followed by thicker laminated sediment layers (Reference VogelVogel and others, 2005). In the process of hot-water drilling, sediment is transported upward with the drilling fluid, cleaning the cuttings from the drill head. As the gap between the borehole wall and the corer widens at the core-catcher window, drill fluid slows down, allowing individual particles to settle into the gap. This method is commonly used in conventional rock drilling to obtain drill cutting samples during non-coring drill operations. However, in our case, particles settling in the core-catcher window became entrapped in the core catcher. A higher sediment content in the ice increases the sediment load and exacerbates the effect. A simple solution to prevent settling of sediment into the core-catcher window is to cover the outside of the core-catcher window with a shield. However, particles can still enter from the inside, either from the winnowing of sediment ahead of the drill head or from melting of the outer layer of the ice core, necessitating a review of the core catcher.

In our drill, the core catcher closes towards the inside (Fig. 2a). In this configuration, the gap between the core catcher and the core-catcher wall tapers in the direction in which the core catcher engages (towards the inside of the core barrel). In the open position, the geometry of the core-catcher holder allows particles larger than the gap on the inside of the core barrel to move into the slot between the core catcher and the core-catcher wall. When the core catcher now tries to engage, it moves towards the center of the corer, jamming larger particles into the narrowing gap (Fig. 3b). Based entirely on the geometry of the core-catcher holder and independent of the size, strength and geometry of the core catcher, this in turn prevents the core catcher from engaging, thus preventing recovery of the ice core.

Fig. 2. Schematic cross-section V-V′ showing core-catcher holder configuration: (a) V-shaped slot converging and narrowing to the inside; (b) straight edge slot; (c) V-shaped slot opening to the inside and outside; (d) vertical cross-section showing moving of core catcher during engaging in core. Stopper mechanism not shown. For details on Caltech hot-water ice-coring drill see Engelhardt and others (2005).

Fig 3. Illustration of (a) natural freeing of sediment from an open slot, and (b) jamming sediment into a V-shaped corner, dependent on the movement direction of the core catcher or any other moving mechanical part.

Inspection of the core-catcher holder indicates that a change in geometry could provide a simple and robust solution to prevent the entrapment of particles. If the geometry of the core-catcher holder is changed from a V-shaped slot opening to the outside (Fig. 2a), or a flat slot (Fig. 2b), to a V shape opening to the inside (Fig. 2c) of the core barrel, particles will be moved out of the widening core-catcher gap rather than trapped in a narrowing gap (Fig. 3). This will allow the core catcher to fully engage so that an ice core, even of irregular shape, can be recovered. Unfortunately, time constraints in the field did not allow modification and testing of our new core-catcher design.

Theoretical evaluation of drilling into sediment-laden ice suggests that other mechanical components could also be affected by the same principle of sediment entrapment when drilling into sediment-laden ice. So, for example, an engaged core catcher could, in the disengaging process, trap sediment particles in the outside narrowing gap. Modifying the core-catcher holder geometry to open inwards as well as outwards may prevent such entrapment in the case of an open core-catcher window (Fig. 2c). Covering the core-catcher window with a shield (as discussed above) would trap such sediment within the core-catcher window. While significant sediment content might prevent further drilling progress within a run, it is less likely that it would prevent recovery of an already drilled ice core. A further problem area may be the geometry of the resting points of the engaged core catcher (for simplicity not shown in Fig. 2; see fig. 2 of Reference Engelhardt, Kamb and BolseyEngelhardt and others (2000) for details). Here sediment particles could be trapped between the core catcher and the base of the slot, again preventing the core catcher from partially or fully engaging. In addition, the geometry of other openings, allowing the entrapment of particles and disabling mechanical components of the core catcher or other parts of the coring system, should be reviewed in the design and/or redesign process of ice-coring systems for drilling into sediment-laden ice.