The nature of the band ogives

The band ogives of Bas Glacier d’Arolla first appear at the foot of the icefall. The “light” ogive bands appear as white to light blue, or white to light brown (depending on weather conditions; see below). The “dark” ogive bands appear as a darker shade of blue or brown. The colour difference between light and dark bands does not fully develop until some distance down-glacier of the foot of the icefall, at about ogive 3 (ogive numbers are given in Fig. 1b). This enhanced colour differentiation coincides with the declining relief of the wave ogives. The boundaries between the light and dark bands are diffuse (Fig. 2), a feature apparently noted previously only by Reference LeightonLeighton (1951). Within the light bands there are thinner zones of “dark highly foliated ogive”, with a higher concentration of surface debris. Similarly, there are zones of light ogive within predominantly “dark” ogive bands (Fig. 2).

Fig. 2. Diffuse boundaries of ogives in the lower section of Bas Glacier d’Arolla. Zones of “dark” ogive within dominantly “light ogive” and vice versa, are clearly visible. Width of exposed glacier surface ∼200 m.

Ogive widths are 12–51 m, with a general but irregular decrease in width down-glacier (Fig. 3). The dark bands are 6.6–29.5 m wide, and the light bands 4.4–30 m wide. Thick dark bands do not necessarily coincide with thin white bands and vice versa (Fig. 3).

Fig. 3. Ogive widths, total ogive widths and widths of light and dark zone of each ogive. Ogive numbers correspond to the ogive numbers in Figure 1.

Natural cross-sections of band ogives

A natural cross-section at a gully cut into the first band ogive at the base of the icefall (Fig. 1b) revealed crevasse traces and sedimentary stratification, which was displaced several tens of centimetres on either side of crevasse traces. Within the cross-section provided by the gully, some areas had a higher intensity of crevasse traces than surrounding areas. Two crevasses on the west side of the glacier also revealed oblique cross-sections through dark ogive bands, exposing alternating zones of fine-grained, white ice and folded, sediment-rich dark ice. This sediment-rich ice was traced to accumulations of debris on the glacier surface, as described below.

Debris content and crystal structure of band ogives

One of the most striking differences between the light and dark ogive bands is that the darker bands are characterized by a higher density of foliae (consisting of coarse, bubble-free blue ice) and more surface debris. This surface debris comprises fine sand to mud-sized particles (Fig. 4), with a high content of dark, organic material (particles of grass, moss, leaf, pine needle and insect). When the weathered surface of the ice is removed, no visible debris is left (Fig. 4). Similarly, after heavy rainfall, the glacier is effectively washed clean of this surface debris. However, it is uncertain if low concentrations of debris are actually present within the ice, as no direct measurements were made. On cleaned surfaces, the foliation appears as thin layers of coarse crystalline bubble-poor blue ice (∼4 cm increasing down-glacier to 6 cm crystal diameter) within finer crystalline white ice (0.5 cm increasing down-glacier to 3 cm diameter). The foliae comprising coarse, bubble-poor ice are more common in the dark ogive bands than in the light bands, eventually dominating light and dark bands down-glacier. For example, in ogive 6, the light band consists of an average 25% by width of blue, coarse foliation-type ice, compared with 60% by width in the dark ogive band. This increases to 60% coarse blue ice in the light band and 90% in the dark band of ogive 23. The differences in ice type cause variations in the albedo of the glacier surface. The difference in ice-type concentrations is best observed from a distance after heavy rainfall when superficial debris has been removed; the ogive banding appears as alternating blue and white bands rather than debris-rich dark bands and debris-poor light bands.

Fig. 4. Nature of the foliation in the dark zone of an ogive towards snout of glacier (ogive 20) Here foliae are closely packed, with alternating white and blue bands according to ice type. Foliae are closely spaced and no folds are present. A section of ice at left has been scraped clean of weathered surface ice, exposing clean ice underneath. Debris is fine sand and silt.

Sand and gravel-sized surface debris are observed at two locations: (i) near the west margin of the glacier, in particular in the area around, and down-glacier of, the crevasse field (Fig. 1b), and (ii) at an isolated area ∼150 m up-glacier of the glacier snout. The shape distribution for two of the clast samples collected is presented in Figure 5, along with comparative samples and a RA/C₄₀ plot. This predominantly sub-angular/sub-rounded material emerges in situ from discrete layers of debris-rich ice at the crevassed area on the western margin of the glacier, both in crevasse walls and from foliation exposed by a supraglacial stream. Again, near the glacier snout, thin debris layers intersect the glacier surface. At these locations, there is no coarse debris above the layer, and below there is washed-down debris. This debris differs from the previously described debris in that it contains coarse clasts, has no obvious organic content, and the clay-rich matrix is a bluish/grey colour. Samples of thinly stratified, bubble-free ice were also extracted from the area around the crevasses (Fig. 6). This ice shows fine debris layers between clear ice layers, as well as the occasional inclusion of coarser debris, in this case a sub-rounded clast of 3 cm diameter.

Fig. 5. (a) Clast-roundness data and (b) RA/C₄₀ plot for four sets of 50 clasts taken from glacier. L1 and L2 are from debris-rich foliation-parallel ice exposed in a dark ogive band by crevassing on western margin of the glacier, L3 is from a medial moraine exposed in central part of glacier and L4 is from 1995 rock fall on central part of glacier. Sample locations are shown in Figure 1b, as L1–L4.

Fig. 6. Close-up view of ice removed from part of foliation-rich dark band exposed at surface of glacier, next to crevasses on western margin of the glacier. The ice is finely stratified, with sand and silt in between each layer. In middle-left section of ice sample is an in situ sub-rounded pebble.

Relationship between band ogives and wave ogives

The first recognizable wave ogive at Bas Glacier d’Arolla forms the lowest part of the icefall. The waves can be traced with certainty to ogive 8, after which the glacier surface is smoothed to a uniform slope by ablation. The relationship between the wave ogives and the light and dark band ogives is best observed after heavy rainfall, which “cleans” the glacier of temporary surface debris, exposing the banding. The dark bands of the ogives are then observed to coincide with the relatively steep, down-glacier slope of the wave ogives.

Sedimentary stratification (S₀)

The earliest formed structure found at Bas Glacier d’Arolla is the remnant of sedimentary stratification, recognized on the basis of decimetre-thick layers of dirty coarse granular ice. Stratification comprises laterally continuous (except where cut by crevasses and crevasse traces) alternating thick light and thin dark bands in the gully in ogive 1 (Fig. 1b) at the base of the icefall. The sedimentary stratification, originating in the upper reaches of Glacier du Mont Collon, has survived passage through the icefall.

Folding (F₁)

Where sedimentary stratification (S₀) is first observed in the accessible part of the glacier, above the base of the icefall, it has already been gently folded (F₁) into large open folds with fold axes approximately parallel to ice flow.

Crevasses and crevasse traces (S₁)

During descent through the icefall, the ice becomes heavily crevassed. Aerial photographs indicate that the principal set of crevasses is orientated transverse to flow, with some crevasses also forming parallel to flow. The icefall is highly dynamic; crevasses and crevasse traces form throughout its height. Older structures, notably stratification (S₀), are displaced by up to 30 cm by these fractures, most often on the down-glacier side.

The remnants of these crevasses appear as crevasse traces at the base of the icefall. Most (∼95%) of the crevasse traces appear as coarse-grained recrystallized blue ice, typical of meltwater freezing within crevasses (Reference Hambrey and MüllerHambrey and Müller, 1978). In addition, some crevasse traces (∼5%) are recognized as white, granular ice infills, appearing almost like firn, which may represent the remnants of snow-filled crevasses.

The distribution of crevasses and crevasse traces is non-uniform on two scales:

1. 1. Observations from aerial photography demonstrate areas of more intensely crevassed ice occurring laterally across the icefall, separated by areas of less intense crevassing, on a scale of tens of metres.

2. 2. The gully across ogive 1 exposes zones with a higher population of crevasse traces within relatively crevasse-trace-free ice, on a scale of metres.

The inclination of the crevasse traces partly depends on their position in the ogive. The cross-section through the gully in ogive 1 shows crevasse traces at a shallower angle (typically 65°) on the steep slope above the plateau of ogive l. They are then rotated to a near-vertical angle in the plateau of ogive 1, before declining to a lesser angle again on the steep slope of the down-glacier end of ogive 1.

A few longitudinal crevasse traces were found between the base of the icefall and ogive 4. These longitudinal crevasse traces were always infilled with white, fine-grained firn-like ice, and are progressively folded with distance from the icefall.

Transverse foliation (S₂)

Bas Glacier d’Arolla is dominated by transverse, or arcuate, foliation (Figs 2, 4 and 7). The transverse foliation originates from a combination of transverse crevasse traces and transverse sedimentary stratification. There is a gradual transition from the earlier structures (S₀ and S₁) to later structures (S₂), between the lower reaches of the icefall and ogive 6 (up to which point there is some inconsistency in foliation dip and orientation (Fig. 7)). Except for the presence of ogives, this transition is typical of other glaciers with icefalls (e.g. Griesgletscher, Switzerland (Reference Hambrey and MilnesHambrey and Milnes, 1977)). At the foot of the icefall, foliation is generally widely spaced (e.g. one folia on average every 0.1–0.2 m). By ogive 16, spacing is reduced to one folia per centimetre in some dark ogive zones (Fig. 6). All foliation measured has a strong east–west orientation, and dips to the south (up-glacier). During its journey down the glacier, the foliation becomes progressively less steeply dipping (Fig. 7).

Folding (F₂)

F₂ folds initially have an open style at the foot of the icefall, with fold axes aligned with the ogives, and plunging up to 56°. During travel, the folds become tighter chevron folds, and their fold axes plunge less steeply (Fig. 8). Stereographic projections (Fig. 8) also show that fold axes in the dark ogive zones are generally less steeply plunging than those in the light ogive zones. The fold axes are orientated transverse to flow, and have an axial planar relationship with ogive foliation (S₁). Fold geometry typically changes down-glacier from open chevron folds to tight, almost isoclinal folds and sheath folds (Fig. 9), with limbs merging into the rest of the transverse foliation. No fold axes were recognized down-glacier of ogive 19.

Fig. 9. The progressive nature of folding from (a) open chevron folds at the base of the icefall (taken from ogive 2) to (b) tighter chevron folding (taken from ogive 8), to (c) tight sheath folding (taken from ogive 10), indicated by arrow. Glacier flow is from top to bottom in all three cases.

1. Semi-continuous reflectors at depth (basal reflectors)

Both profiles have a relatively continuous, undulating basal reflector. The reflector is clearer and more continuous on profile B (Fig. 10). On both profiles, this basal reflector steadily deepens from the snout, to a maximum of ∼1000 ns (84 m) on profile A and ∼950 ns (80 m) on profile B, both below ogive 13. The reflector then rises again, to 900 ns (75 m) on profile A and 800 ns (67 m) on profile B, both below ogive 6 (depths are relative to ice surface).

2. Alternating reflection-rich and reflection-poor areas

Near the surface of the profile, a series of alternating reflection-rich areas and reflection-poor areas gives what appears to be a series of ridges and troughs. However, these features do not have a surface expression, and are related to internal structure only; the profiles were not extended to cover the wave ogives. The “ridges”, or reflection-rich sections, are dominated by near-surface reflections and hyperbolae. The distances between the tops of adjacent “ridges” vary, but they typically occur at ∼30 m intervals. In contrast, the “troughs”, or reflection-poor areas, occur where there is some depth of penetration before any reflector is met. These reflection-free areas reach ∼240 ns (20 m) depth below the glacier surface, below which the surrounding reflection-rich areas converge. Towards the snout of the glacier, the separation between reflection-rich and reflection-poor areas becomes less obvious as the reflection-rich areas overlap. Towards the icefall, there is reflection-free ice above the reflection-rich areas, reaching a maximum of 100 ns (8 m).

3. Up-glacier-dipping reflectors

These features are best observed on profile A. They can also be observed in the up-glacier end of profile B. These reflectors dip up-glacier at 30–50° (or up to 7° less after topographic correction). They begin within the main body of the glacier, and do not continue to the glacier surface. Some can be traced for a long distance from within the top of the “reflection-rich sections” to close to the basal reflective layer. Their dip becomes slightly shallower at depth.

4. Down-glacier-dipping reflectors

Down-glacier-dipping reflectors are best observed on profile A, although they can still be observed on profile B. These reflectors are less continuous, typically extending over 20–40 m, at angles of 30–50°. They appear to cross-cut up-glacier-dipping reflectors.

Differences between antennae orientations

The “inline” and “broadside” antennae profiles demonstrate different features. Broadside antennae orientation (Fig. 10, profile A) has enhanced internally dipping features, whereas inline data collection (Fig. 10, profile B) has enhanced the quality of data at depth. In addition, the near-surface “reflection-poor” areas appear deeper and “cleaner” during inline data collection. The differences in data are related to the preferential orientation of structural features within the ice relative to the polarization of the radar signal (Reference NobesNobes, 1999). Reference NobesNobes (1999) found that the basal ice/rock reflection at Franz Josef Glacier, New Zealand, is clearer when antennae are axially orientated with respect to the direction of glacier flow, but obscured by internal reflections when antennae are transversely orientated to flow. The same effect appears to be true for Bas Glacier d’Arolla. Again, as found by Reference NobesNobes (1999), transversely orientated antennae pick out more internal reflectivity than axially orientated antennae. Reference NobesNobes (1999) accounted for increased internal reflectivity with transversely orientated antennae as the product of reflections from crevasses and “crevasse remains extending to great depths”, which are orientated transverse to flow. A comparable structure on Bas Glacier d’Arolla would be the multiple foliation planes, many originating from crevasse traces, which are also perpendicular to glacier flow.

Basal reflectors

The deep semi-continuous reflectors are interpreted as reflection from the base of the glacier, either the ice/bedrock or ice/sediment interface. Care must be taken with this interpretation since some authors (e.g. Reference Arcone, Lawson and DelaneyArcone and others, 1995; Reference Murray, Gooch and StuartMurray and others, 1997) have interpreted basal reflectors as the transition between relatively clean, englacial ice and debris-rich basal ice, which masks the true ice/bedrock interface. Reference Arcone, Lawson and DelaneyArcone and others (1995) defined their basal reflector at Matanuska Glacier, Alaska, U.S.A., as having “numerous hyperbolic diffractions rather than by any continuous reflector”, which is what may be expected from debris-rich ice. By contrast, Reference Nobes, Leary, Hochstein and HenrysNobes and others (1994) interpreted a “clear basal reflection” at Tasman and Mueller Glaciers, New Zealand, as a junction between lodgement till and glacier ice. Both of these studies involved temperate glaciers. Ideally, as demonstrated by Reference Murray, Gooch and StuartMurray and others (1997) on the polythermal surge-type glacier Bakaninbreen, Svalbard, borehole data are needed to determine the true nature of the reflective layer. At Bas Glacier d’Arolla, the reflector is more likely to be the ice/bed interface because it is continuous rather than comprising a number of hyperbolae.

Alternating reflection-rich and reflection-poor areas

The nature and spacing of the contrasting reflection-rich and reflection-poor areas suggests that these features are band ogives. A schematic interpretation of the relationship between the radar data and the band ogives and foliation of the glacier is presented in Figure 11. The reflection-free areas are interpreted as the light ogive bands, whilst the reflection-rich areas are interpreted as coinciding with the dark ogive bands. Although at the time of data collection the precise position of light or dark bands was not known because of snow cover, this inference was tested by locating tracer pebbles, which were left during the winter field season and recovered in the summer. The light bands allow reflection-free penetration, which may be accounted for by the relative lack of foliation within the light bands. In contrast, the foliation-rich dark bands provide ample reflectors because of changes in ice properties across foliation boundaries (Fig. 11).

Fig. 11. Schematic representation of three-dimensional structure for a segment of ice from icefall to snout along central flowline, illustrating structural relationships and development of ogives. The inset boxes provide a schematic interpretation of the relationship between the radar data and the ogives and foliation of the glacier.

The irregular widths of the reflection-rich and reflection-poor areas relate to the irregular width of the band ogives (Fig. 3). Their irregular behaviour towards the snout is explained by the fact that, this close to the snout, the band ogives “merge” into one dirt-covered foliation-rich mass.

Up-glacier-dipping reflectors

The up-glacier-dipping structures dip at an angle similar to (or slightly lower than) the measured foliation (Fig. 7), and thus to the ogives as a whole. One possible explanation is that the up-glacier-dipping reflectors represent reflections from ice types of different properties within a band ogive, such as the boundary between different individual foliae (Fig. 11).These boundaries could be caused by faulting along the foliae. The fact that some of the reflectors can be traced to considerable depth, possibly to the base of the glacier, demonstrates that they are significant features within the ice, and, given their angle of dip, possibly even shear zones.

Down-glacier-dipping reflectors

The origin of the down-glacier-dipping reflectors is unclear, there being no surface expression. Possible explanations include:

1. 1. An additional structural feature (e.g. back-thrusts or Riedel thrusts associated with shearing)

2. 2. Reflection from englacial watercourses (air-filled at the time of survey) as described by Reference Arcone and YankielunArcone and Yankielun (2000)

3. 3. Thin layers of englacial sediments, as described by Reference Murray, Gooch and StuartMurray and others (1997).

Interpretation of the structural evolution of Bas Glacier d’Arolla

The structural evolution of Bas Glacier d’Arolla initially involves the formation of sedimentary stratification (S₀) on Glacier du Mount Collon, which is gently folded. This early folding may occur in response to confining pressure as the ice is fed into the icefall, or it may occur during its descent of the icefall, as the ice moves over bedrock obstacles.

Ice flowing down the icefall is thinned because of a change in gradient and associated higher velocities, and is highly crevassed because the ice is undergoing longitudinal extension. Crevasses and crevasse traces (S₁) provide planes of weaknesses in the ice, which facilitate the faulting of S₀ on its journey down the icefall, resulting in the displacements observed in the gully (in ogive 1, Fig. 1b) at the foot of the icefall.

A reduction in ice velocity at the base of the icefall results in compression, and the progressive folding during phase F₂ of S₀ and S₁ (both transverse and longitudinal crevasses). This takes place from the base of the icefall down-glacier, before eventually transposition results in fold limbs being lost through attenuation within the arcuate foliation (S₂). This progression of folding suggests that the zone of compression extends some distance down-glacier of the foot of the icefall. The arcuate foliation (S₂) also shows evidence of down-glacier compression with the increase in foliae per unit surface area in conjunction with the general decrease in ogive width. To aid the understanding of this sequence of events, a schematic diagram (Fig. 12) is provided, showing the progressive evolution of structures in plan view, including the evolution of the cumulative strain ellipse.

Fig. 12. Schematic diagram showing the progressive evolution of structures in plan view to create the arcuate foliation that constitutes the band ogives at Bas Glacier d’Arolla. The figure represents a parcel of ice moving down the flow centre line. The evolution of the cumulative strain ellipse is also shown.

The up-glacier-dipping reflectors found within the glacier by GPR represent a further structural feature, possibly representing internal transverse foliation, a continuation of the transverse foliation seen at the glacier surface. These planes provide a weakness along which ice can move when under the highly compressive regime at the base of the icefall. A geological similarity would be a series of thrust-faulted blocks stacked upon one another (an imbricate stack), as illustrated in Figure 11. This sort of structural phenomenon would explain how deeper and even basal ice can become exposed at the glacier surface.

The study of the structure of Bas Glacier d’Arolla aids the interpretation of ogive formation, as it has demonstrated that the ice is compressed both at the base of, and down-glacier of, the icefall. The progressive folding of earlier structures explains why the glacier tongue is dominated by arcuate foliation. Also, areas of more intense crevassing may help explain the more intense foliation in the dark bands than in the light bands, and, on a smaller scale, the zones with more intense crevasse traces exposed in the gully at the base of the icefall may help explain why there are zones of “dark” ogive within a predominantly light ogive band (see below).

Origin of alternating light and dark bands

Light and dark colouring of the band ogives is related to differing amounts of dirt cover on the glacier. However, at Bas Glacier d’Arolla, and other glaciers with band ogives, the question arises whether (i) the dirt originates from within the ice, with the dark ogive zones being richer in englacial debris (Reference HaefeliHaefeli, 1951; Reference LeightonLeighton, 1951; Reference PosamentierPosamentier, 1978), or (ii) the dirt is derived from elsewhere and accumulates favourably on some parts of the ogives, i.e. in the troughs of wave ogives (Reference LliboutryLliboutry, 1958; Reference Allen, Kamb, Meier and SharpAllen and others, 1960; Reference King and LewisKing and Lewis, 1961; Reference AthertonAtherton, 1963).

Band-ogive characteristics at Bas Glacier d’Arolla that are relevant to the evaluation of these hypotheses include:

1. (i) After heavy rainfall, the glacier was washed of most surface debris, but the band ogives were still visible when viewed from a distance. This is because the darker ogive bands have more strongly foliated, blue-ice foliation, compared with lighter, foliation-poor bands.

2. (ii) When surface ice is removed, no debris is visible in either light or dark bands (Fig. 4). The presence of debris trapped in surface ice can be explained by differential weathering of ice structures. Dark bands develop a large number of small, dirt-trapping ridges because of variations in albedo and crystal size between the plentiful foliae.

3. (iii) The hypothesis of debris collecting in troughs (cf. Reference King and LewisKing and Lewis, 1961) does not hold for Bas Glacier d’Arolla for various reasons:

   1. a. The wave-ogive topography melts to a uniform sloping surface within a short distance of the base of the icefall (beyond ogive 8)

   2. b. The light and dark colour differences do not fully develop until some distance from the base of the icefall, once the wave topography has been reduced, which disagrees with the hypothesis of dirt collecting at the foot of the icefall (cf.Reference WashburnWashburn, 1935; Reference King and LewisKing and Lewis, 1961)

   3. c. The actual dark bands of the ogives do not strictly coincide with the troughs of the wave ogives, but more with the down-glacier slopes, as observed elsewhere by Reference AthertonAtherton (1963).

4. (iv) Although, as explained in (ii), there is no observable debris content in the ice for most of the glacier, isolated cases occur in which coarse debris is clearly associated with the dark bands of the glacier, in particular in the area around the crevasses on the western margin of the glacier. Here, the surface is littered with sediment (described in section 3) emerging from foliated ice. The RA index and position for these samples on the RA/C₄₀ plot contrasts strongly with rock-fall and medial-moraine debris sources from the same glacier and, when compared with other datasets (e.g. Reference Benn and BallantyneBenn and Ballantyne, 1994; Reference Bennett, Hambrey and HuddartBennett and others, 1997), indicates a subglacial source. Also, ice samples extracted from the dark bands in these areas are clearly sediment-rich (Fig. 6). The debris content of the ice, and its thin layering, suggests it is basal ice which has arrived at the glacier surface.

At Bas Glacier d’Arolla, the overall determinant of whether a band is dark or light is the intensity with which it is foliated, leading to (i) an overall blue appearance of “cleaned” ice in dark ogive zones, and (ii) differential weathering which traps the sediment and thus gives the ice in a dark band an overall darker appearance, whether “cleaned” or not. Although debris was found emerging in situ from dark ogive bands, these represent isolated instances of coarse subglacial material reaching the surface at the glacier margin, and only have a local influence on the overall colouring of the band ogives. Elsewhere on the glacier, aeolian sources must be the dominant source of surface debris.

The highly foliated nature of the dark bands

Widely spaced arcuate foliation becomes more closely spaced within a short distance as a result of compression below the icefall. This close spacing coincides with the enhanced colour differentiation between light and dark bands. Below are two possible reasons why the dark bands contain more intense foliation than the lighter bands.

1. (i) One ogive, consisting of a light and a dark band, actually represents a cross-section through the glacier as it comes down the icefall. In this case, the dark band represents ice at depth, which has been brought to the surface during rotation at the base of the icefall (in a manner consistent with Reference HaefeliHaefeli’s (1951) syncline model discussed in the introduction). Ice at depth, whilst thinned by extensional flow and moving relatively rapidly down an icefall, may be capable of forming a “basal foliation”: a number of discrete shear zones at the base of the ice mass, exposed at the surface as a high-intensity suite of foliae. The highly stratified blue bubble-free ice (Fig. 6), extracted at the crevasses on the western margin of the glacier (described in section 4), is evidence for ice of “deep glacier origin” as suggested by Reference PosamentierPosamentier (1978). By contrast, the light part of the ogive represents the upper part of the parcel of ice as it descends, with fewer foliae, resulting simply from transposition of crevasse traces and the original sedimentary layering.

2. (ii) Areas of more intense foliation may form within the icefall. Indeed, observations both from aerial photography and in the longitudinal cross-section observed at the base of the icefall demonstrate zones of higher crevasse-trace intensity adjacent to zones of lesser intensity. These crevasse traces are a major source of foliation down-glacier. Any particular parcel of ice is subject to a number of compressive and extensional phases as it passes down the irregular and complex mass of the icefall. Thus some areas of ice will be more susceptible than others to crevassing. In particular, once an area is crevassed, it is more likely to be the source of later faulting since crevasse traces can be reactivated when they acquire an appropriate orientation (Reference Hambrey and MüllerHambrey and Müller, 1978). One aspect of this hypothesis is that small-scale variations in the number of crevasse traces also explain why there are dark areas within light bands and vice versa.

Mechanisms for band-ogive formation

The process of band-ogive formation that is closest to fitting the observations from Bas Glacier d’Arolla is a variation on the “reverse faulting with associated drag folding” hypothesis (Reference PosamentierPosamentier, 1978, fig. 2), which in turn is a variation of the “shearing hypothesis” first presented by Reference ChamberlinChamberlin (1895). Here, a three-dimensional model, partly based on that of Posamentier, is presented which is consistent with the observations at Bas Glacier d’Arolla (Fig. 11).

Reference PosamentierPosamentier (1978) explained that blue, bubble-free debris-rich or debris-poor ice could indicate deep glacier ice in contrast to light-coloured bubbly ice, which indicates near-surface ice. Posamentier’s “reverse faulting and associated drag folding” model demonstrates how this deeper ice is brought to the glacier surface by reverse faulting as a result of compression at the base of the icefall. However, the “drag folding” hypothesis is not required for Bas Glacier d’Arolla, as all foliation is up-glacier-dipping. Despite this, Posamentier’s model fits other observations concerning the structure of Bas Glacier d’Arolla: compression results in folding of the original sedimentary stratification and crevasse traces at, and down-glacier of, the base of the icefall. Crevasse traces and sedimentary stratification then progressively become more transverse and transform into the arcuate foliation.

One problem with applying Reference PosamentierPosamentier’s (1978) hypothesis is that clearly defined thrust planes would be expected to crop out at the surface of the glacier, whereas at Bas Glacier d’Arolla there are no well-defined boundaries between the light and dark bands. Reference PosamentierPosamentier’s (1978) relatively simple model may be better applied to Bas Glacier d’Arolla using multiple, smaller-scale faulting and shearing (Fig. 11). Instead of having one thrust event associated with each ogive as suggested by Reference PosamentierPosamentier (1978), the presence of numerous, smaller shears within the main shear zone allows duplicated parts of basally derived ice to reach the surface within the mainly white, upper ice band (also previously suggested by Reference LeightonLeighton, 1951). Multiple shearing also explains (i) the presence of numerous up-glacier-dipping reflectors within the GPR section as representing numerous fault planes or shears rather than a few discrete planes, and (ii) why there are zones of “dark ogive” within “light ogive” and vice versa.

The wider application of a structurally based model of ogive formation

Although the model presented in this paper is based solely on observations from Bas Glacier d’Arolla, it satisfactorily explains published observations made at other ogive sites. Specific examples include the isolated concentrations of (often rounded) debris found within the dark ogive bands at both Morsárjökull (Reference Ives and KingIves and King, 1954) and Svinafellsjökull (Reference Ives and KingKing and Ives, 1954) in Iceland, and at East Twin Glacier, Alaska (Reference LeightonLeighton, 1951). These descriptions resemble those of the isolated debris-rich ice cropping out at Bas Glacier d’Arolla, and could also be interpreted as debris-rich basal ice being transported to a higher elevation within a shear zone. The structural glaciology of Bas Glacier d’Arolla is also similar to other glacier tongues dominated by arcuate foliation below an icefall and progressive folding of earlier structures (e.g. Blue Glacier, Washington, U.S.A. (Reference Allen, Kamb, Meier and SharpAllen and others, 1960), Gulkana Glacier, Alaska (Reference RaganRagan, 1969), and Griesgletscher (Reference Hambrey and MilnesHambrey and Milnes, 1977)).

Future work

Although this paper provides a plausible structurally based model of band-ogive formation, further work is required to address the relationship between band and wave ogives. This requires detailed multi-year investigation of the kinematics of glaciers with ogives. Ice velocity, deformation measurements and cumulative strain determinations should also be directed to determining whether the proposed shear hypothesis is valid, in particular if periodic shearing is an annual event related to variations in glacier flow (Reference HaefeliHaefeli, 1951, Reference Haefeli1957; Reference PosamentierPosamentier, 1978). Seasonal variations in velocity might lead to a change in deformation style in which the summer flow leads to seasonal reverse faulting, whereas winter flow does not (Reference PosamentierPosamentier, 1978). This difference may be accentuated by increased meltwater, resulting in reducing friction in the system during the spring and summer (Reference HaefeliHaefeli, 1957). The most obvious time for such a shift in velocity to trigger shearing would be the “spring event”, which has been observed to create large increases in velocity and changes in basal conditions (Reference Iken and BindschadlerIken and Bindschadler, 1986), for example at nearby Haut Glacier d’Arolla (Reference Harbor, Sharp, Copland, Hubbard, Nienow and MairHarbor and others, 1997; Reference Mair, Nienow, Willis and SharpMair and others, 2001). Such an event may also be linked with wave-ogive formation.

Studies by Reference NyeNye (1958, Reference Nye1959) and Reference WaddingtonWaddington (1986) evidently have shown the way wave ogives form. We suggest that the next step in producing a comprehensive model for ogive formation would be to reconcile the different approaches.