Localized accumulation basins in the Transantarctic Mountains, fed completely by precipitation on to the site, provide a new avenue for Antarctic ice-core research. These sites are valuable for the recovery of records detailing climatic change, volcanic activity, and changes in atmospheric chemistry for periods extending well into the last glacial period. Since these sites are located within the transitional zone between plateau ice and ocean-ice shelf. they could provide some of the most climatically sensitive records available from Antarctica. Furthermore, unlike those ice cores retrieved from the interior of Antarctica, there are terrestrial records from nearby sites that can be used for comparison (e.g. Reference Denton, Armstrong, Stuiver and TurekianDenton and others, 1971; Reference DrewryDrewry, 1980; Reference Stuiver, Denton, Hughes, Fastook, Denton and HughesStuiver and others, 1981; Mavewski and Goldthwait, 1985).
The Dominion Range (Fig. 1) is the first in a series of planned Transantarctic Mountains ice-core sites (Fig. 1). The Dominion Range is located along the edge of the East Antarctic ice sheet, approximately 500 km from the South Pole and 120 km from the Ross Ice Shelf, at the confluence of Bcardmore and Mill Glaciers (Fig. 2). These glaciers, along with several other outlet glaciers in the Queen Maud Mountains (sub-sector of the Transantarctic Mountains), drain the Titan Dome area of the East Antarctic ice sheet. Approximately half of the Dominion Range (Fig. 2) is ice-free and the average elevation of the range is 2700 m.
Between 20 November and 14 December 1984, a tent camp was operated in the Dominion Range. Due to logistic restraints, all aspects of the study, including reconnaissance, site characterization, and recovery of a 201 m core were undertaken in the same field season. In this paper we present the results of site and core characterization, specifically ice surface and ice thickness, bore-hole temperature, mean annual net accumulation, crystal size, crystal fabric, oxygen-isotope composition, and examples of ice chemistry (Cl−, , MSA), and isotopic composition of trapped gases.
Ice-Surface And Ice-Thickness Measurements
The early pari of the field season was devoted to establishing an optimum site for recovery of an ice core (Fig. 2). Maps, visual observations of ice-surface topography, and the presence of bedrock ridges all validated initial estimates that the Dominion Range ice cover is either entirely separated from or only minimally connected to the East Antarctic ice sheet and hence the site is a catchment for local precipitation. Exposed bedrock ridges flanking the Dominion Range are cavernously weathered. Comparison of the degree of cavernous weathering with that examined in the general region of the Queen Maud Mountains by Reference Mayewski, Goldthwait, Turner and SplettstoesserMayewski and Goldthwait (1985) suggests that ice has not topped these ridges for at least severaltens of thousands of years.
Based on an examination of USGS (1:250 000) topographic maps and a radio echo-sounding survey conducted in the field, the Dominion Range ice mass is divisible into three major drainage basins, referred to as A, B, and C (Fig. 2). The radio-echo survey employed a mono-pulse system (after Reference Watts and IsherwoodWatts and Isherwood, 1978) and was centered primarily over drainage basin C. It included measurements at 42 stations, ten of which were occupied at least twice to test instrument reproducibility, which proved to be less than the error inherent in reading the oscilloscope. Final ice-thickness measurements were determined using Watts and Isherwood’s (1978) relationship with adjustments for density made using measurements from the core. Crevassed areas in the southern section of basin C, lower Vandament Glacier, prevented the recovery of useful radio echo-sounding data from this area.
Drainage basin C surface topography (Fig. 3) is characterized by a general surface slope to the east, thus the major part of the drainage for C discharges through Vandament Glacier. Ice thicknesses in basin C (Fig. 4) range from ≅350 to <50 m with the thickest areas north of the drill site and in the Mount Pennant area. Thinner ice areas are found in the western part of the basin close to the C-A surface ice divide, and the remainder of the basin is characterized by ice depths most commonly in the range 200–300 m. The general gradient of the subglacial topography is east-south-east.
The core site (see Figs 2, 3, and 4) was chosen ≅1.7km km down-flow line from the C-A ice divide to minimize complications due to flow right on the divide and ≅1.7km up-slope from the base camp to minimize the effects of any local chemical contamination from the camp. Although it cannot be demonstrated definitively with the data available, it appears that if any East Antarctic ice penetrates drainage basin C from the Mount Tennant area that this ice would be deflected eastward toward Vandament Glacier and hence away from the drill site. A comparison between ice-surface contours (Fig. 3) and ice-thickness contours (Fig. 4) in the area of the Vandament Glacier flow line suggests that the ice in this area must be strained.
Bore-Hole Temperature And Mean Annual Net Accumulation
Temperature measurements were made at 5 m intervals (Fig. 5) down the entire length of the bore hole using a thermistor system designed by L. Hansen (PICO). Twenty readings at 20 s intervals were made at each depth immediately following a 10 min equilibrium period. Instrument error is ±0.02°C based on duplicate measure-ments. The mean annual temperature, at 10 m depth, is −37.3°C, and the temperature at the base of the core, 201 mbar (m beneath surface), is −31.3°C. Radio echo-sounding results suggest that the glacier at this point is < 230 m thick, hence there is little doubt that the base of this ice mass is frozen to the bed.
Mean annual net accumulation was determined by a combination of measurements, including total ß-activity, 210Pb, 14C, seasonal signatures in anion chemistry, and volcanic horizons. Details concerning the dating of the core appear in Reference Spencer, Mayewski, Lyons, Twickler and GrootesSpencer and others (in press). We report here only the resultant mean annual net accumulation for the upper ≈100 m of the core which is ≈35 kg m−2 a−1.
The mean annual temperature and mean annual accumulation at this site are consistent with relationships presented by Reference GowGow (1968) for a survey of Antarctic sites.
Crystal Size And Fabric
Horizontal and vertical thin sections were cut from core samples taken at depths of 59.5, 70.7, 85.3, 122.7, 143.2, and 190 mbar. Mean crystal size (Fig. 5) within each section was determined from measurements of the long and short axes of individual crystals. Fabrics (Fig. 6) were determined from measurements of c-axis orientations using a Rigsby stage (Reference LangwayLangway, 1958).
Mean crystal cross-section increased from ≈7 mm2 at 59.5 mbar to ≈13 mm2 at 70.7 mbar and then decreased progressively to ≈2mm2 between 70.7 and 143.2 mbar before increasing to ≈7mm2 at 190 mbar. Sections at 59.5 and 70.7 mbar exhibited random fabrics. At 82.3 mbar, a weakly preferred orientation of caxes is evident with local c-axis concentrations of 8% per 1% area of projection. By 122.7 mbar, c-axes group into two maxima located approximately 35° from the vertical, c-axis concentrations as high as 10% per 1% were observed. At 143.2 mbar, a girdle pattern appears with local c-axis concentrations as high as 12% per 1% area. A ring or small girdle fabric is present at 190 mbar.
The marked decrease in size of crystals between 70.7 and 143.2 mbar may indicate the onset of shearing. Moderate development of c-axis fabrics from the same depth interval might support a shearing process; however, a lack of tight single-pole fabrics would indicate that shearing is not yet a dominant process. Crystal coarsening at 190 mbar could signal the onset of recrystallization in the basal layers of ice. However, such a process would tend to be impeded by the generally low temperatures at the site. It was not possible to obtain azimuthally oriented core and this, together with the limited observations of the texture and fabric of the ice, prevent us from developing a unique flow history for this part of the Dominion Range ice field. Notably, the quality of the ice core recovered deteriorated from whole to fractured core interspersed with whole sections from this depth downward. It was not possible in the field, however, to resolve whether the core quality was necessarily due to strained ice or problems with the cutters.
Oxygen Isotopes Of Ice
A continuous δ18OICE profile was obtained for the ice core using 25 cm increments for most of the core and 2–3 cm samples in the sections studied for ice chemistry. δ18OICE is defined here as being equal to ((18O/16O) sample -(18O/16O)SMOW)/(18O/16O)SMOW and SMOW is Standard Mean Ocean Water. δ18OICE for 50 cm averages of the data appear in Figure 7.
For the upper ≈100 m of the core there appears to be a small trend toward less negative δ18OCE values with depth. The sections below ≈100 m are significantly lighter and are marked by a drop of ≈5‰ from ≈100 to 145 mbar followed by a rise of ≈2–3‰. The ≈5.0‰ marked drop is similar to the glacial/interglacial δ changes of 5.4‰ and about 5‰ observed at Dome C (Reference Lorius, Merlivat, Jouzel and PourchetLorius and others, 1979) and at Vostok (Reference LoriusLorius and others, 1985), respectively.
While details of the distribution of major chemical species in the core are left to other papers (e.g. Reference Spencer, Mayewski, Lyons, Twickler and GrootesSpencer and others, in press), it is worth mentioning the marked difference in the distribution of Cl− and SO2- 4in the upper half of the core (2 and 3 cm sampling interval) and one of the few intact sections that could be analyzed from the lower half of the core at 138.0–138.4 mbar (2 cm sampling interval). Marine aerosols and volcanic activity represent the primary sources for both Cl− and SO2- 4 to the Antarctic ice sheet. While volcanic source inputs would be expected to be randomly distributed in the record, differences in marine source input would result in trends in the data series that probably reflect changes in air-mass circulation and/or ocean/ice relationships. Average values of Cl− (≈250 ppb) and (≈300 ppb) in the 138.0–138.4 m section are two to three times those in the upper half of the core. The contrast between lower-level ice, as represented by the 138.0 −138.4 m section, and the upper ≈ 100 m of ice is striking. Although the higher Cl− and and Cl− concentrations in the deeper sections could coincidently be a volcanic layer, none of the volcanic events in the upper half of the core is as high in concentration or as broad in time span. We conclude, therefore, that the deeper section marks a period which differs from upper sections either in marine source intensity, in transport pathway, and/or for some unknown reason.
The upper and lower sections of the core also appear to differ in their concentration of methanesulfonic acid (MSA). MSA is a constituent of marine aerosols which is formed as a result of the atmospheric oxidation of DMS. Variations in MSA concentration in the core reflect changes in the flux of DMS from the oceans, in the patterns of aeolian transport, and/or in precipitation rate (Reference Saigne and LegrandSaigne and Legrand, 1987). A noticeable difference was, however, observed between the four samples measured from an upper section of the core (29–30 mbar; MSA cone. = 2.2 ± 0.2) and five samples measured from a lower core section (138–139 mbar; MSA cone. = 5.7 ± 1.0).
Isotopic Composition Of Trapped Gases
The isotopic composition of trapped O2 and N2 in two sections of the Dominion Range core appear in Table I. The isotopic composition of the 83 mbar samples was similar to the isotopic composition of Recent (< 1500 a B.P.) samples of ice from five different cores taken from Antarctica and Greenland (paper in preparation by T. Sowers and others). The isotopic composition of the 139 mbar samples had δ18OATM(O2) values which were enriched, compared to Recent ice samples, by 1.0 ± 0.13%.
During the transition from glacial to interglacial periods, isotopically light melt water from glaciers was introduced to the oceans, resulting in a decrease in the δl8Owater of sea-water (where δ18Owater = ((H2 18O/H2 16O)water/(H2 16OSMOW) − 1)103). Photosynthesizing organisms utilized the isotopically depleted melt water to form O2 which was mixed into the paleoatmospheric O2reservoir causing the δ18OATM(O2) of air tO form (where) δ18OATM(O2) = ((18O16O)/(16O2)paleo-air/(((18O16O)/(16O2)present-air-day atmosphere) − 1)103. Studies of the trapped gases in the Dome C core have shown that the isotopic composition of 02 trapped in the ice tracks the isotopic composition of sea-water over the past 20 000 years (Reference Bender, Labeyrie, Raynaud and LoriusBender and others, 1985). Since the δ18OATN(O2) is constant throughout the atmosphere, one can use the composition of the O2 in the ice as a chronologic tool. We have used this tool to estimate the ages of two samples from the Dominion Range core.
Analysis of Recent samples of ice show that the trapped gases are enriched in both 18O and 15N relative to the contemporaneous atmosphere (paper in preparation by T. Sowers and others). The enrichment is probably the result of isotopic fractionation as the bubbles are sealed. Because atmospheric N2 has a very long residence time, the δ15N of the atmospheric N2 is believed to have been constant for the last 106 years (where δ15NATM(N2) = ((14N15N/14N2)paleo-air/((14N15N/14N2)present-day air − 1)103). Given this constancy, and the observed δ15N−δ18O relationship for gases trapped in modern ice, one can use the fractionation of the N2 trapped in ice to determine the δ18OATM(O2) of past atmospheres using the following equation:
* Reported data have been corrrected for the isotopic dependence on the elemental composition (paper in preparation by A.B. Kiddon and others). The data are reported relative to present-day-air, 615N.
† Isotopic composition of the contemporaneous atmosphere from which the trapped gases were derived, also reported relative to present-day air.
Knowing the δ18OATM(O2), one can estimate the age of an unknown trapped gas sample by comparing the measured δ18OATM(O2) value with the down-core record of δ18OATM(O2) measured in the Dome C core (Reference Bender, Labeyrie, Raynaud and LoriusBender and others, 1985). Using this δ18OICE(O2) curve, we assign an ice age for the 139 mbar sample of >10ka B.P. This age is expressed as a lower limit for two reasons. First, ice is older than the age of the trapped air (Reference Schwander and StaufferSchwander and Stauffer, 1984). Secondly, the δ18OICE(O2) values for the Dome C record were not converted to δ18OATM(O2) due to lack of δ18NICE(N2) measurements. Converting the Dome C δ18OICE(O2) to δ18OATM(O2) values will shift the inferred δ18OATM(O2) record closer to the present-day air. By converting the Dome C δ18OICE(O2), one would increase the age of the Dominion Range 139 mbar samples.
Summary And Conclusions
The Dominion Range ice-core site is characterized by a mean annual temperature of −37.3°C and a core-base temperature of −31.3 C which is probably close to the basal ice temperature. The mean annual net accumulation is ≈35 kg m−2 a−1.
Dominion Range ice is divisible into three main drainage systems and a site close to the ice divide between two of these drainage systems was chosen for the recovery of a 201 m core. Differences between ice-surface and subglacial gradients in the area of the drill site suggest that some amount of lateral strain is imposed on the ice column. The difference in δ18O)ICE noted from ≈100 to ≈145 mbar in the Dominion Range core is similar tothe glacial/interglacial δ changes observed at Dome C and Vostok. Measurement of δ18OICE(O2) and δ15NICE(N2) of trapped gases indicates that ice at 139 mbar has an age < 10 ka B P.
If as inferred from the measurements of δ18OICE,δ18OICE(O2) and δ15NICE(N2) the upper approximately half of the core column is Holocene in age and the ice below is glacial, then differences in both crystal size and chemical concentration discussed in this text may be more uniquely defined. While decreases in crystal size in the lower ice may be due partly to shear, they may also simply reflect thc cooler temperature of formation present during the glacial period as observed at Dome C (Reference Duval and LoriusDuval and Lorius, 1980). Furthermore, increases in C− and concentrations may be consistent with increases from Holocene to glacial age as measured from the Byrd core (Reference Ragone and FinelliRagone and Finelli, 1972; Reference Cragin, Herron, Langway jr, Klouda and DunbarCragin and others. 1977) and Vostok core (Reference Angelis de, , Legrand, Petit, Barkov, Korotkevitch and KotlyakovAngelis and others. 1984), and the trend in MSA concentration is similar to that observed from Holocene to glacial ice measured at Dome C by Reference Saigne and LegrandSaigne and Legrand (1987).
The Dominion Range ice core provides relatively easy access to the Holocene record in a site that is potentially climatically sensitive. Were the quality of the lower half of the core better, it could also provide a view through the interglacial/glacial transition and into the last glacial period.
Future papers will document details of the Holocene signal in this region.
We should like to thank J.V. James (Glacier Research Group), H. Rufli (Switzerland), and B. Koci (Polar Ice Coring Office) for their help and friendship in the field. PICO ably recovered the 201 m core. L. Hansen (PICO) designed and provided the thermistor system used in this study. Thanks are also due to VXE-6. J. Dibb (Glacier Research Group) conducted the 210Pb and total ß-activity measurements and A. Wilson (University of Arizona) the 14C analysis which were intrumental in the dating of the upper sections of this core. We are greatly indebted to them for their input. L. Preble patiently typed this paper. This research was supported by U.S. National Science Foundation grants DPP-84–00574, DPP-84–11108, and DPP-85–13699.