Introduction
During the 1981–82 field season, eight submarine ice cliffs were discovered along the western side of McMurdo Sound (Fig. 1). Some of these features were quite large (greater than 300 m long and in excess of 30 m tall). A description of their locations, profiles and dimensions has already been published (Reference StocktonStockton, 1983). Three of the ice cliffs (Miers Valley, Cape Chocolate, and Strand Moraines) were obviously the exposed faces of ice cores in adjacent moraines or glaciers intruding into the Sound. The existence of ice cores in the moraines bordering the southern parts of the west side of McMurdo Sound has been known since the early explorations (Reference David, Priestley, Taylor and GoddardDavid and Priestley, 1914; Reference TaylorTaylor, 1914, Reference Taylor1922; Reference DebenhamDebenham, 1921; Reference HealyHealy, 1975). The presence of the ice cliffs at Explorers Cove is more difficult to account for, because ice cores have not been reported on the Sound north of Blue Glacier (Fig. 1), except for a small piece of ice buried in a beach at Marble Point (Reference NicholsNichols, 1961).
Two uniquely different mechanisms may explain the presence of the ice cliffs at Explorers Cove. These ice cliffs may be composed of glacial ice, implying the existence of previously undescribed ice cores in the adjacent coastal moraines. Alternatively, the ice cliffs in question could have frozen directly from sea-water. It should be possible to distinguish between these two hypotheses through stable isotopic analyses of the ice cliff, because glacial ice which is formed from atmospheric precipitation is depleted in deuterium and 18O relative to ice frozen directly from sea-water (e.g. Reference Gow and EpsteinGow and Epstein, 1972; Reference MorganMorgan, 1972).
In this paper, we present the results of a stable isotope analysis of cores taken from an ice cliff in Explorers Cove. The data provide a basis for determining the mechanism of formation of this particular cliff and possibly of the other cliffs in Explorers Cove and elsewhere (e.g. the southern boundary of New Harbour) that are not associated with identified ice cores in moraines or with glaciers (Reference StocktonStockton, 1983).
Locality Description
Explorers Cove is at the mouth of Taylor Dry Valley (Fig. 1). The mean annual surface temperature at Explorers Cove is approximately −25°C (Reference WeyantWeyant, 1967). The shoreline is composed entirely of glacial moraine. Four ice cliffs have been located in the cove. The ice cliff which was selected for coring is approximately 70 m wide and 6 m tall (Reference StocktonStockton, 1983). It is located on the south side of the cove at approximately lat. 76° 33' 48″ S., long. 163° 29' 17″ E. This cliff was selected because it was the most easily accessible of the four in the cove.
Materials and Methods
Ice cores were taken from the cliff and the overlying sea ice with a SIPRE corer on 8 January 1982. Two vertical cores were taken, both 40 cm shoreward from the face of the cliff and 2 m apart, midway along the length of the cliff. The first core was taken to sample the cliff systematically and the second in an attempt to get deep samples. Segments of the first core were retained at approximately every meter from both the sea ice and the ice cliff (Fig. 2). (There was a gap of approximately 50 cm between the sea ice and the ice cliff.) Segments deeper than 305 cm into the ice cliff could not be recovered for technical reasons. Attempts to recover parts of an hypothesized ice core from the moraine in the vicinity of the ice cliff (Fig. 2) proved unsuccessful because it was not possible to drill to any significant depth (greater than 0.5 m) through the moraine due to the abrasive properties of the permafrost-immobilized sediments.
In order to obtain a sample of known glacial origin, a part of the ice core in the moraine at Cape Chocolate (Fig. 1) was exposed by digging (the ice core is buried under about 20 cm of sediment). Pieces were then broken out with a shovel.
The core segments and pieces of the Cape Chocolate ice core were stored frozen, wrapped in aluminum foil. The segments were melted in beakers which were tightly covered with aluminum foil. Melting took place at approximately 20°C and was accomplished in 4–6 h. After melting, the water samples were stored in Nalgene LPER bottles and shipped back to the United States for analysis.
The hydrogen- and oxygen-isotope ratios of the water samples were determined by standard techniques (Reference Bigeleisen, Bigeleisen, Perlman and ProsserBigeleisen and others, 1952; Reference Epstein and MayedaEpstein and Mayeda, 1953). The isotope ratios are reported in the δ notation, where
The standard in both cases is Standard Mean Ocean Water (SMOW).
The salinities of the water melted from the samples were determined with an induction salinometer.
Results
The results of the hydrogen- and oxygen-isotope ratio determinations and the salinity measurements are given in Table I. The isotope ratios determined for two international water standards, SLAP (accepted δD and δ18O values of −428 ‰ and −54.5 ‰) and SMOW (δ0 and δ18O values 0 ‰ and 0.0‰), are also given in Table I. The δ0 and δ180 values of the samples obtained from the ice cliff at Explorers Cove are plotted against the vertical profile of the core in Figure 2
The isotope ratios and salinities of the upper parts of the ice cliff are essentially identical to those of the overlying sea ice. The deepest cliff segments, at 290–300 cm in the first core and at 290–305 cm in the second core, have δD and δ18O values lower than those of the sea ice. The salinities of these two samples are not significantly different from those of the other samples in the cliff or from those of the sea-ice samples. The sample of glacial ice from Cape Chocolate has more negative δD and δ18O values and lower salinity than any of the samples from the sea ice or ice cliff at Explorers Cove.
Discussion
With respect to the alternative hypotheses of glacial or sea-water origin for the ice cliff, the analysis of the samples suggests that the upper parts of the ice-cliff face were frozen directly from sea-water. The four ice-cliff samples down to 215 cm into the cliff had a mean δD value of +5 ± 1 ‰, a mean δ18O value of +1.2 ± 0.1 ‰, and a mean salinity of 1.826 ± 0.344 ‰. These values are consistent with the values we obtained for the sea-ice samples and with the values that have been obtained for other ice samples frozen directly from sea-water (Reference Gow and EpsteinGow and Epstein, 1972; Reference MorganMorgan, 1972). Thus, it would appear that these parts of the ice cliff are not evidence of ice cores in the adjacent moraine but are simply expressions of submarine permafrost.
The two deepest samples (290–305 cm depth) of the ice cliff have significantly lower δD and δ18O values than the upper ice-cliff samples and the sea ice. The mean δD value is −24 ‰ and the mean δ18O value is −2.6 ‰. These values are less negative than those we found for the glacial ice from the moraine at Cape Chocolate (Table I) and published values for other glacial ice samples, such as Koettlitz Glacier, which has δD and δ18O values of about −27 ‰ and −36 ‰ (Reference Gow and EpsteinGow and Epstein, 1972), and Taylor Glacier, which has δ18O values ranging from −34.0 to −44.2 ‰ (Reference Stuiver, Stuiver, Yang, Denton, Kellog and McGinnisStuiver and others, 1981). Reference Stuiver, Stuiver, Yang, Denton, Kellog and McGinnisStuiver and others (1981) published δ18O values for a number of glaciers and ice cores in moraines around McMurdo Sound. The moraine values are of particular interest here. The δ18O values of samples from the west side of the Sound vary considerably, with some close to those of sea ice (e.g. Garwood Valley with values of +1.4 and +3.9 ‰), while others are highly variable and/or clearly not ice derived from sea-water, such as Blue Glacier (ranging from −41.3 to +2.3 ‰), Hobbs Glacier (−28.0 to −1.1 ‰), and Miers Glacier (−31.6 to −27.9 ‰). The precise location relative to the coastline of these samples is not known but these data serve to illustrate that the values for the deepest parts of the ice cliff at Explorers Cove are within the range covered by other moraine ice cores in McMurdo Sound. It would be interesting to know whether or not the less negative δ18O values observed by Reference Stuiver, Stuiver, Yang, Denton, Kellog and McGinnisStuiver and others (1981) were for samples taken from coastal moraine ice cores and the more negative values from moraines adjacent to glaciers.
The more negative δD and δ18O values we observed for the deepest sections of the ice cliff may be caused by the mixing of large amounts of ice frozen directly from sea-water (with δD and δ18O values of about +5 ‰ and +1 ‰ and salinities of about 2 ‰) with small amounts of glacial ice, with more negative δD and δl8O values (possibly less than −300 ‰ and −40 ‰, respectively), and with salinities less than 0.05 ‰ (Reference Gow and EpsteinGow and Epstein, 1972; Reference MorganMorgan, 1972; Reference Stuiver, Stuiver, Yang, Denton, Kellog and McGinnisStuiver and others, 1981). Mixing could have occurred due to the fortuitous orientation of the core segments at an interface between glacial ice and ice frozen directly from sea-water, or might have resulted from processes such as solid-state diffusion or partial melting followed by refreezing. Our data do not permit us to distinguish among these possibilities.
The simplest interpretation of the results presented here is that the upper parts of the ice-cliff face were frozen directly from sea-water while the deepest parts of the cliff probably include remnants of glacial ice. Whether or not the upper parts of the ice-cliff face is a veneer of sea ice or a much larger piece of glacial ice is presently not known, nor is there any information available to test this possibility. Attempts to investigate the shoreward extent of the ice cliff were unsuccessful. There may be an extensive ice core behind the face of the cliff, since circular and irregularly shaped, steep-sided ponds were observed in the intertidal flats behind two of the ice cliffs. These ponds are suggestive of sub-surface melting of an ice mass in the moraine (paper in preparation by DeLaca and Stockton).
Although the analysis presented here applies directly to a single ice cliff in Explorers Cove, it is possible that all of the ice cliffs in the cove and possibly that near Herbertson Glacier (Fig. 1) share similar characteristics and origins.
Conclusions
Stable isotope ratios and salinities of water samples obtained from two cores demonstrate that the upper seaward parts of a submarine ice cliff in Explorers Cove have frozen directly from sea-water due to the underwater effects of permafrost. The origins of the lower parts of the ice-cliff face are more difficult to interpret but appear to be partially glacial in origin, suggesting that there may be ice cores in the shore moraines of Explorers Cove. If the shoreline contains ice cores, the coastline of McMurdo Sound is more extensively ice-cored than previously known.
Acknowledgements
We thank G. H. Denton for discussions at McMurdo Station, R. B. Sharman and T. L. Roberts for collecting the ice cores, and K. Crocker, H. Ajie and D. Winter for performing the isotopic analyses. This research was supported by N.S.F. grants DPP80-034320 (T.E.D.) and ATM79-24581 (M.J.D.).