Introduction
One of the diffigulties in interpreting oxygen-isotope data from deep cores is in making allowance for the different place of deposition of the lower layers. If the direction and velocity of the ice flow is known, a first-order correction to the core data can be made by considering the variations in the present-day surface δ 18O. For more sophisticated studies, however, changes in the surface isotope ratio which will have occurred with changes to the ice sheet also have to be considered and this presents considerable difficulty.
The prominent change in the δ 18O values found in the deep core data around 10000 years b.p. (see e.g. (Reference Johnsen, Johnsen, Dansgaard, Clausen and LangwayJohnsen and others 1972)) is rather too large to be solely due to temperature and probably also reflects a general lowering of the ice sheet surface in conjunction with a retreat of the ice margin. Both the lowering and the reduction in area will reduce the depletion of 18O in the surface snow but at present it is not easy to assign magnitudes to the changes.
The main idea behind this collection and presentation of data is to give a broad-scale picture of the surface δ 18O for comparison with core δ 18O profiles. A second use is that future examinations of the compiled data will enable a better understanding of the variations in the depletion in 18O in Antarctic precipitation in terms of the temperature, elevation, continentality, etc. of the particular location.
The experimentally observed relationship between δ 18O and temperature arises because of the decrease in the saturation vapour pressure of water with temperature. A cooling air mass must precipitate water to avoid supersaturation and as this water will be enriched (c. + 8‰) in 18O with respect to the vapour, the vapour will therefore become progressively depleted in 18O. The mechanism, complicated by non-equilibrium transfers of water vapour, is affected by the path the air mass takes for instance over land, sea, or ice, its temperature and rate of temperature change, its turbulence, and other factors most of which are climatic variables and thus presumably will have varied in the past.
It is envisaged that the data presented here in Figs 1 and 2 and Table I may be used to obtain empirical values for the fractionation processes which affect precipitation falling on different parts of Antarctica and hence can be used to adjust deep core δ 18O data for such factors as changes in ice-sheet elevation and area.
Fig. 1. Mean annual surface values of δ18O in parts per thousand obtained from measurements in Antarctica.
Fig. 2. Contours of constant mean annual surface values of δ18O in parts per thousand derived from the point values in Figure 1.
Table I. Antarctic ice sheet surface data
Sample and Data Collection
The large seasonal δ 18O variability means that some form of averaging is required to obtain a value which approximates the long-term mean. Where deep bore-hole analysis data are available, this is effectively done by averaging values from the top 50 m or so of core. Where only shallow cores are obtained, e.g. by a hand auger or where pit samples are used, an even number of seasons’ accumulation will probably not be collected and, due to the relatively small number of seasons represented, the δ 18O value will be biased away from the long-term mean.
We can obtain some idea of the magnitudes involved by examining a core from the Law Dome summit for which some detailed measurements have been made. (See (Reference Budd and MorganBudd and Morgan, 1977)). This location is probably representative of one of the more difficult places to sample because the high annual accumulation and the general lack of wind (for Antarctica) allows the large seasonal δ variation to be well preserved in amplitude. The accumulation is 0.6 Mg m−2a−1 and the range of isotope values is 8‰ at 100 m depth, and probably about 12‰ near the surface. Assuming roughly equal amounts of summer and winter precipitation, the deviation in the δ value from the mean if an exactly odd number of seasons are sampled is:
To have a sample within 1‰ of the long-term in such a location requires that at least six seasons (i.e. three years) accumulation be sampled. This could cover a depth of 4 m in the near-coastal areas, but will be much less inland, where the accumulation rate is much smaller.
Some of the data used may be marginal in respect of sampling depth. Where possible the depth covered by the sample is noted in the listing of data.
The measurements are given as δ 18O with respect to SMOW, expressed as parts per mil deviation (see (Reference CraigCraig, 1961)). Where a number of measurements are closely spaced, e.g. on the Law Dome, the number of points has been reduced by averaging to give a reasonable spacing on the map (Fig. 1). The second map shows contours of constant δ 18O values (Fig. 2).
Acknowledgements
I would like to acknowledge the work done by various A.N.A.R.E. (Australian National Antarctic Research Expedition) personnel in collecting samples while on traverse and also in particular the assistance of T. H. Jacka and J. Birch in analysing samples for δ 18O and collating data.
