Wilson Piedmont Glacier and distributaries

Wilson Piedmont Glacier, stretching for 54 km between New Harbour and Granite Harbour (Fig. 1), is the largest of a series of ice piedmonts which border the west side of the Ross Sea in southern Victoria Land. Data obtained from three north-south R/E flights along this glacier and two flights, each which traverse it along Victoria and Wright Lower distributary glaciers, are summarized in Figures 3 and 4A and B; flight lines are located on Figure 1. Figure 4C and D of Bowers Piedmont Glacier to the south and Oates Piedmont Glacier to the north (see also Figures 1 and 6) are the results of single flights.

Fig. 3.

Subglacial contour map of Wilson Piedmont Glacier superimposed on U.S. Geological Survey topographic map. Location of data used, which includes profiles 4A and B, is shown in Figure 1.

Glacier thickness and bed topography. Wilson Piedmont Glacier averages between about 250 and 350 m in thickness, with a maximum of 500 m east of the divide below Victoria Lower Glacier. This is thicker than Bowers Piedmont Glacier but similar to thicknesses suggested by the short, Oates Piedmont Glacier profile.

Subglacial elevation contours (Fig. 3) show that Wilson Piedmont overlies an irregular, seaward-sloping surface between 200 and –200 m elevation which is partitioned by the coastal ends of three east-west glacial troughs (Fig. 1). The trough of Debenham Glacier (once a major distributary of Mackay Glacier) is the most marked and extends more than 260 m below sea-level. In contrast, Wright Valley reaches up from below sea-level on the coast to about 270 m above sea-level (Fig. 4B) under the snout of Wright Lower Glacier. Because the snout rests on an unknown thickness of drift, the bedrock threshold must lie below this point but probably above the 130 m divide at 11 km on Figure 4B. Like Debenham Glacier, the sub-ice profile of the Wright Lower Glacier ice stream is more compatible with repeated outlet glaciation than that of lower Victoria Valley. The latter displays a distinct threshold reaching to about 670 m elevation (Fig. 4A; 34 km) that may be a remnant of marked alpine glacier erosion preceding outlet glaciation (see Reference TaylorTaylor, 1922, p. 76).

Sub-piedmont bedrock thresholds and late Cenozoic history. Radio-echo sounding definition of subglacial thresholds of lower Wright and Victoria valleys helps to reconcile some apparent major differences in the tectonic and glacial histories of the ice-free valleys and it complements recent detailed studies in the area as follows. Reference WebbWebb (1972) and Reference McSaveney and McSaveneyMcSaveney and McSaveney (1972) provided convincing evidence, in the form of fossiliferous marine deposits, that Wright Valley was a fjord in Tertiary time and hence submerged by an amount at least equal to or greater than its threshold level at the east end. Preliminary work by the McSaveneys and by R. E. Behling and myself suggests that marine deposits may prove to be widespread in the western part of Wright Valley up to 200 300 m elevation with a few-possible occurrences to about 600 m (see Reference Bull, Webb and BalkemaBull and Webb, 1972; Reference Calkin, Fahey and ThompsonCalkin, 1974). Thus the bedrock threshold (130–270 m) under Wilson Piedmont Glacier was probably not at a critical elevation for marine invasion as assumed in the past (see Reference BullBull, 1972; Reference Bull, Webb and BalkemaBull and Webb. 1972). However, it marked the point of subsequent valley emergence. Furthermore, in the adjacent Victoria Valley system, lack of obvious evidence for marine deposits (Reference Calkin and CraryCalkin, 1971[a], Reference Calkin, Fahey and Thompson1974) may suggest that the marine submergence in this area was limited to below about 670 m, which is the elevation of the eastern threshold shown in Figure 4A. Alternatively, submergence of the Victoria Valley system was blocked by glacial ice (Reference McSaveney and McSaveneyMcSavency and McSaveney, 1972)’

Fig. 4.

Radio-echo profiles extending east-west across the Wilson Piedmont (A and B) from the sea to the ice-free valleys and north-south along the Bowers (C) and Oates (D) piedmont glaciers. Profiles A-C are located in Figure 1, D in Figure 6. Dashed lines (B) represent data from adjacent R/E run; dotted lines (A and B) are inferred from topographic map.

Differences in number, magnitude and apparent source of glacial advances which have occurred at the east ends of the ice-free valleys have been a puzzling result of glacial geo-morphologic mapping in these valleys. Correlation of these advances is important since they have been ascribed to grounding of the Ross Ice Shelf and in turn provisionally to lowered eustatic sea-levels and Northern Hemisphere glaciations (Reference Calkin and CraryCalkin, 1971 [a]; Reference Denton, Denton, Armstrong and StuiverDenton and others, 1971). Although problems of inter-valley correlation still exist, available information on thresholds now suggests that the following numbers of advances are reasonable. Four westward axial invasions from the Ross Sea occurred into Taylor Valley (Reference Denton, Denton, Armstrong and StuiverDenton and others, 1971), where there is no eastern threshold, while only three advances occurred over the bedrock threshold under Wright Lower Glacier from the east (Reference BehlingBehling, 1972; Reference Calkin, Fahey and ThompsonCalkin, 1974). The high threshold at 670 m under Victoria Lower Glacier (Fig. 4A) may have entirely prevented westward flow of ice from the Ross Sea area into this valley system. Although there were west-moving glacial advances at the east end of Victoria Valley that have been provisionally ascribed to the grounded Ross Ice Shelf (Reference Calkin and CraryCalkin, 1971[a]), no evidence of direct ice movement from the Ross Sea has been found. Therefore, their source may have been local alpine glaciers.

Mackay, Ferrar and Taylor outlet glaciers

Figure 5A, B and C are R/E profiles along Mackay Glacier, lower Ferrar-upper Taylor Glacier system, and upper Ferrar Glacier, respectively (Fig. 1). Substantial surface clutter echoes from intense crevassing have prevented a complete bedrock profile of Mackay Glacier (Fig. 5A) from being obtained. However, runs up Ferrar and upper Taylor Glaciers are much more complete and probably more accurate. Figure 5B and C represent composites of two R/E runs each. General descriptions of these outlet glaciers have been given by Reference TaylorTaylor (1922), Reference Wright and PriestleyWright and Priestley (1922), Reference Hamilton and HayesHamilton and Hays (1961), and Reference Gunn and WarrenGunn and Warren (1962).

Glacier thickness and bed topography. The average thicknesses of Taylor, Ferrar and Mackay Glaciers shown by the R/E profiles are similar, varying respectively from about 450 to 550 m through much of their lengths. However, the thickness at the grounding line (sudden glacier thinning and leveling of the surface) of Mackay Glacier (Fig. 5A) is much greater than the other two glaciers, partly reflecting its greater activity. All three glaciers are substantially smaller than when they cut their valleys during an earlier, full-bodied stage of the ice sheet.

The profiles show innumerable bottom irregularities, i.e. steps and adjacent depressions, similar to those throughout the ice-free valleys. Some of the smaller depressions or irregularities may not be either closed and/or so large; they may be due to the varying path of the bottom-echo return relative to the path of the deepest part of the valley cross-section or thalweg. Other explanations are discussed below.

Origin of subglacial steps. Reference TaylorTaylor (1922) believed that many of the irregularities on the floors and walls of Taylor Valley and Ferrar Valley represented parts of headwardly eroded cirques or short valleys overwhelmed by the outlet glaciers. However, many exposed steps, thresholds and depressions have been related to outcrops of the dolerite sills by field mapping. The sill-controlled steps most frequently occur in forms referred to as “pseudocirques” by Reference Gunn and WarrenGunn and Warren (1962). They believed these to have been formed by selective headward retreat of dolerite scarps once breached by concentrated vertical erosion from the outlet glaciers. This origin appears applicable in part to the Mackay, Ferrar and Taylor Valley profiles. Steps and lows are particularly pronounced under Taylor and Ferrar Glaciers just down-valley from where two dolerite sills converge to form very thick units, i.e. below Finger Mountain at 69 km on Figure 5B and also at the west edge of the Kukri Hills at 58 and 43 km. Another pronounced step in the valleys coincides with the eastern edge of the dolerite sill exposure at 29 km.

Fig. 5.

Radio-echo profiles along: A, Mackay Glacier and ice-sheet margin above the valleys; B, lower Ferrar–upper Taylor Glaciers; and C, upper Ferrar Glacier. Profiles are located in Figure 1. Blanks in base show missing data, dashed lines represent data, from adjacent R/E run. Positions of tributary glaciers are indicated on profile B.

The steps and lows also correspond in position with incoming tributary glaciers as well as being controlled partly by the dolerite sills as shown by Figures 1 and 5B. The pronounced basin east of the Ferrar–Taylor Glaciers divide (Fig. 5B; 43 km) is analogous in position with the deep basins suggested by the Mackay Glacier records (Fig. 5A) as well as with the basins of Lakes Vida and Vanda immediately down-valley from major tributary junctions in Victoria and Wright Valleys, respectively (Fig. 1).

Mountain crest thresholds and glacier regimen. The basic control on intrusion of the plateau ice to the mountain valleys, including the Victoria, Wright and lower Taylor ice-free areas, must be the delicate balance maintained between the elevation of the adjacent ice sheet and that of the resistant dolerite sills or thresholds which field studies suggest underlie their heads or form the mountain crestal cols (Reference Gunn and WarrenGunn and Warren, 1962). This balance in turn has been controlled in the past by several factors including climate and by tectonic eustatic and ice-sheet movements as well (Reference Bull, Webb and BalkemaBull and Webb, 1972). Geologic studies have shown the effect of these sills at the heads of the ice-free valleys (Reference Calkin and CraryCalkin, 1971[a]), where parts of the sills are exposed, and the R/E records confirm the limiting effects on these and on the surrounding glacierized valleys (Figs. 2 and 5A).

At the head of Taylor Glacier (Fig. 5B; 115 km) the subglacial surface rises to about 1 900 m and the glacier thins to 100 m. Ferrar Glacier, on the other hand, heads in the Skelton Névé (Fig. 5C; 45 km) over a col where the ice stream is 400 m thick. Above The Portal (Fig. 1) on the ice sheet near the same latitude as the Taylor Glacier threshold, the sub-surface divide reaches only to 1 800 m (Fig. 5C; 90 km) and ice coming down from the ice sheet is again 400 m thick. Thus Ferrar Glacier reaches the sea, albeit just barely, and Taylor Glacier does not. This is despite the fact that Taylor Glacier is fed by about one-third of the upper Ferrar Glacier discharge via Windy Gully and across the divide of Figure 5B (at 53 km), and also has a wide and direct opening to the ice sheet.

Although the traverse over Mackay Glacier did not extend directly west on to the ice sheet, adjacent data (Fig. 6) indicate that the threshold across the mountain crest is broad and lower than that across Fcrrar or Taylor Glaciers, and hence Mackay Glacier is more active. Summer movement of up to 0.85 m d⁻¹ has been measured on the Mackay Glacier ice tongue (Reference TaylorTaylor, 1922) compared to approximate velocities of 0.04 and 0.08 m d⁻¹ on sleeper sections of lower Ferrar and lower Taylor Glaciers, respectively (Reference Wright and PriestleyWright and Priestley, 1922; Reference Hamilton and HayesHamilton and Hays, 1961). Radio-echo sounding records indicate that thresholds are not present at the inland ends of Mulock Glacier (Fig. 6) or Byrd Glacier 100 km farther south; therefore, these glaciers fill their valleys and provide some of the largest discharges to the Ross Ice Shelf that come from the ice sheet of east Antarctica.

The relatively thin connections between ice sheet and valley glaciers in the ice-free valley area, particularly to Taylor Glacier, indicate that local precipitation may play an important part in any current ice-tongue movement. Studies by Reference Wilson and CraryWilson and Crary (1961), for example, suggest that Skelton Glacier with a movement of 0.24 m d⁻¹ at its snout receives negligible ice supply from the plateau because of high sill thresholds (such as below The Portal; Figs. 1 and 5C) and it is nourished almost entirely by direct precipitation.

Victoria Land ice plateau margin

The records. The subglacial as well as ice-sheet surface topography immediately west ol the ice-free valleys on the west flank of the Transantarctic Mountains is given by Figures 6 and 7. These figures were compiled from data of 11 east–west and two north–south radio-echo flights along with information from the oversnow geophysical traverses. The oversnow data are located along profile line 71–73.4 (F–F’) and the two adjoining segments to the west in Figure 6; it was assembled from Reference WeihauptWeihaupt (1961), Reference RobinsonRobinson (1962) and Reference CraryCrary (1963). Limitations of detail in the contour map of Figure 6 are due to previously described problems in R/E accuracy, navigation and the broad spacing of the flights. In addition, contours depend on absolute elevation (a.s.l.) determined by aircraft altimetry. This may be in error locally by up to 200 m in areas beyond topographic map coverage, although elevations are tied inland to oversnow-traverse elevations as corrected by Reference CraryCrary (1963) which may be accurate to ±50 m.

Fig. 6.

Subglacial contour map of the west flank of the Transantarctic Mountains between Mawson and Mulock Glaciers. Straight lines show location of data and profiles of Figure 7. Profile line F and two adjoining lines to the west locate oversnow traverse data.

Fig. 7.

Radio-echo profiles 7A–E and oversnow traverse profile (F) across the west flank of the Transantarctic Mountains it the ice-free valley area. Profile F was assembled for geophysical data (largely gravity and magnetic) of Reference WeihauptWeihaupt (1961), Reference RobinsonRobinson (1962) and Reference CraryCrary (1963). Profiles are aligned in their respective positions along long. 159° E. Profiles A–F are located in Figure 6, and C–F are also located in Figure 1.

The profiles shown in Figure 7 give more accurate ice thickness than can be interpolated from the contour map in addition to their much greater detail. Data along cast–west traverse lines which include profiles B, C and 1) of Figure 7 may be up to 100 m more in error than those along the other traverses. Malfunction of the flight recorder on these three traverses required that the aircraft terrain clearance used to compute ice thicknesses be taken from a few widely spaced radar height readings.

Subglacial topography. The R/E results show that some of the general assumptions of the cross-sectional shape of the Transantarctic Mountains held since the first geologic exploration (Reference FerrarFerrar, 1907; Reference David and PriestleyDavid and Priestley, 1914) are correct; that is, what may be characterized as the main mountain mass is about 170 km wide and with a decrease in bedrock elevations westward somewhat similar to that on the exposed east flank. Without more closely spaced traverses, it will never be certain just how similar the two sides are; however, it is clear that most of the subglacial western flank recorded by the R/E system does not show in detail a single pronounced west-facing scarp or even a simple descending series of west-facing escarpments as has often been assumed since that time. An exception may occur above Skelton Glacier (Fig. 1), where the oversnow-traverse profile (Fig. 7F; 50 km; see also Reference CraryCrary, 1963, fig. 25a) does indicate a rather simple and steep westward descent of about 700 m between long. 157° and 158° E. Although the changes in topography along this ground traverse correspond very generally with those of the R/E profiles adjacent to the north (Figs. 6 and 7E), effective seismic depth control on the traverse was not obtained (Reference CraryCrary, 1963). Therefore, the true west slope here may be much more irregular than indicated and structural conclusions partly based on this profile such as that by Calkin and Nichols (1972) need re-evaluation.

The R/E profiles are characterized by a series of isolated subglacial mountains, particularly plateau-like or dissected tablelands with gently west-sloping tops, flanked by very steep slopes on at least east and west sides. These show relief of 400–900 m and are separated similarly by gently west-sloping valley or broader lowland surfaces 1–20 km wide. Most profiles are too far apart (about 30 km) to establish whether there are also north- or south-facing escarpments. The average inclination of the major slopes, which appear near vertical on the scale of Figure 7, probably vary between 20° and 55° as interpolated between R/E readings on near-horizontal parts of the top and bottom. The common gentle west-sloping surfaces vary in steepness, but profiles A, C and E show many broad lows with slopes of about 0.01 or somewhat less than 1°.

The crestal upland area is relatively narrow, being about 50 km wide between the 1 000 m contours. This is less than one-quarter of the width of the upland in the Beardmore Glacier area (Drewry, Reference Drewry1972[b]). Bed elevations uniformly reach down westward along the study area to at least 300 m above sea-level 60–70 km west of the crest, giving the mountains a gross symmetrical appearance; however, in detail, they are slightly skewed westward. Several plateaux project more than 900 m above the basins beyond the 300 m elevation contour and the broad open areas below sea-level of the Wilkes Subglacial Basin occur more than 100–150 km from the mountain divide (Fig. 6). The most pronounced and interesting of the subglacial plateau(x) occur in this area (Fig. 7A) strung out east–west between long. 154° and 157° E. Traverses here (Fig. 6) are close enough to show that these are probably terminated by very steep slopes on the north side as well as to the east and west. With this major exception, the grain of the topography is north-south.

Origin of subglacial topography. Structural effect. Faulting and differentially tilted fault blocks, accentuated by sub-horizontal surfaces of the thick Beacon Supergroup sandstones and the Ferrar dolerite sill units, have been the major structural elements defined by field studies throughout the Transantarctic Mountains (Reference Grindley and LairdGrindlcy and Laird, 1969; Reference McGregor and WadeMcGregor and Wade, 1969; Reference WarrenWarren, 1969; Reference RobinsonRobinson, unpublished). These faults consist of a complex of longitudinal and transverse faults related to the Victoria orogeny of Tertiary and Quaternary time (Reference Gunn and WarrenGunn and Warren, 1962; Reference WebbWebb, 1972; Reference MaigkovMaigkov, 1973). The early concept of Reference David and PriestleyDavid and Priestley (1914) that the mountains are bounded by two major normal fault systems, or that they are part of a “Great Antarctic horst” (Reference TaylorTaylor, 1922; Reference GouldGould, 1935) has not been supported by most previous geophysical work (Reference McGinnis and MontgomeryMcGinnis and Montgomery, 1972; Reference SmithsonSmithson, 1972; Reference RobinsonRobinson, unpublished) nor by the R/E data given here. Alternative anticlinal or monoclinal tectonic hypotheses for this area (Reference HamiltonHamilton, 1960, Reference Hamilton1965; Reference KingKing, 1966) have similarly had little definitive support.

Considering the pervasive occurrence of block faulting, it is not surprising that this pattern appears to be repeated on the submerged flank of the mountains. Analysis by Reference DrewryDrewry (1972[b]) of R/E data from the Beardmore Glacier area to the south has confirmed a multi-scarped decent of the inland mountain flank there which is similar to that described above for the present study area. Drewry has attributed this topography to westward-tilted fault blocks affected by varying amounts of erosion.

Many of the steep slopes suggested by the subglacial profiles and the contour map, particularly those aligned north-south or near parallel to the coast at about long 157° E. and farther westward, appear to be most easily explained by large faults. In addition, the isolation of the high mountains or plateaux, such as that series revealed by Figure 7A, may be related to transverse or east–west fault lines. For example, a fault delineated along the northern margin of the Mackay Glacier valley (Reference WarrenWarren, 1969) and/or the low immediately inland, if real, could be an eastern extension of such a system. However, there is no indication of a single major normal fault system bounding the inland side of the Transantarctic Mountains here. The general westward slope of the summits on the western mountain flank, as well as the gentle, near-uniform, westward-tilted plateau-like surfaces (high or low) seen in the profiles (Fig. 7) can be reproduced crudely by projection of the regional westward-dipping Beacon Supergroup rocks and Ferrar dolerite sills which crop out at the mountain crest. Whether these slopes are directly due to differential tilting during periods of block faulting and isostatic movement or are also due in part to a distinct folding episode(s), as suggested by the tectonic hypotheses of Reference HamiltonHamilton (1965) or Reference KingKing (1966), is not known.

Glacial effect. Erosional features, formed by alpine glaciers before build-up of the ice sheet of east Antarctica must also account for some of the texture, particularly the broadened longitudinal or transverse valleys suggested by Figure 6 and the subglacial profiles of Figure 7 (Reference Selby and WilsonSelby and Wilson, 1971; Reference Calkin, Fahey and ThompsonCalkin, 1974). In addition, inland-flowing valley glaciers nourished from successively diminished local glacier-covered plateaux (Reference Gunn and WarrenGunn and Warren, 1962) may have contributed to the build-up of the ice sheet in this area (Reference Bull, Bull, McKelvey and WebbBull and others, 1962). Reference DrewryDrewry (1972[a], Reference Drewry[b]) has shown evidence from the R/E data for inland-trending glacial valleys in the Queen Maud Range.

The broad contour interval of the sub-surface data here precludes definition of most glacial features but some large inland-trending valleys of possible glacial origin may be represented in the areas of profiles A and C of Figures 6 and 7 near the heads of Mackay and Mulock Glaciers. The apparent back-to-back alignment of east- and west-sloping valleys here is similar to that in the Queen Maud Mountains (Reference DrewryDrewry, 1972[b]) and may be evidence of the local plateau-glacier initiation of opposing glacial valleys. However, the narrowness of the mountains in this area suggests that most glacial valleys should be very small compared to those described by Drewry for the southern Transantarctic Mountains, and hence the examples cited above may be suspect.

The erosional effect of eastward movement of the ice sheet over the mountains during the outlet-glacier-cutting phase is of uncertain importance. Local accumulations of wet-based till believed to have been deposited during an early, more extensive (full-bodied) stage of the ice sheet have been described from nunataks near the heads of Ferrar and Mackay Glaciers (Mayewski, in press); thus some erosion must have occurred on the west mountain flank.