Measurements of oxygen-isotope ratio in cores from polar ice sheets have provided detailed longterm records of past fluctuations in temperature. Cores in which annual layers can be identified also contain a record of past precipitation rates provided that one can calculate the total vertical strain to which each layer has been subjected since it was deposited at the surface. Because this is difficult, few such records have been published so far.
Reference NyeNye (1963) proposed a method based on the assumption that the vertical strain-rate along any vertical line in the ice was uniform at any instant and that there was no basal melting. The first assumption is invalid and the method gives implausible results in the cases in which we have used it. Reference Hammer, Clausen, Dansgaard, Gundestrup, Johnsen and ReehReeh and others (1978) obtained continuous records of precipitation, extending back to 600 AD in one case, from three cores in Greenland. They also assumed that the vertical strain-rate did not vary with depth, but only down to the maximum depth of measurement. Because this was less than one-quarter of the ice thickness, their assumption is probably satisfactory. They also assumed that the strain-rate had not changed over the period of their records; this is also reasonable.
Reference Hammer, Clausen, Dansgaard, Gundestrup, Johnsen and ReehHammer and others (1978) have measured annual layer thicknesses at about 35 points in the core from Camp Century, Greenland, and Koerner (unpublished) has made similar measurements on a core from the Devon Island ice cap. Because these measurements extend to within 100 m of bedrock at Camp Century and 15 m at Devon Island, an assumption of uniform strainrate is certainly incorrect. We assumed that the horizontal velocity at Camp Century was given by the usual formula for laminar flow and obtained the vertical strain-rate from the condition of incompressibility. On the assumptions of (1) steady-state and (2) no change in accumulation rate and ice thickness for about 35 km upstream, the total vertical strain at any depth was obtained by integrating the strain-rate from the surface to that depth. Because the Devon Island bore hole is within three ice thicknesses of the ice divide, we used a finite element model, developed from a model of Raymond (in press), to calculate total vertical strain as a function of depth.
The Camp Century record can be divided into four parts, according to date:
-
(a) 0–3000 BP. Accumulation rate approximately equal to its present value. None of the ten individual measurements differs by more than 5% from the present value.
-
(b) 3000–4000 BP. Accumulation rate about 10% greater than at present.
-
(c) 4000–6000 BP. Mean accumulation rate about same as at present but larger variations than in the first period.
-
(d) 7000–8500 BP. Accumulation about 10% less than at present.
Because ice thickness and velocity at Camp Century during the last glaciation are unknown, the total vertical strain cannot be calculated. So we have not used the measurements of Reference Hammer, Clausen, Dansgaard, Gundestrup, Johnsen and ReehHammer and others (1978) in the ice-age ice.
For Devon Island, comparison of measured layer thicknesses with those calculated from the steadystate model suggests that (a) the accumulation rate has not changed significantly for the past 1 500 a and (b) it was greater than it is at present for the 3 500 a before that.
Reference ImbrieImbrie (1972) and Reference DansgaardDansgaard and others (1982) have dated the Camp Century core by comparing features of the oxygen isotope record with corresponding features in the climatic record in ocean cores, while Reference MörnerMörner (1972) made a similar comparison with the glacial geological record in North America and northwest Europe. The authors of these three studies then calculated variations in accumulation rate during the last glaciation from the relative lengths of core sections between horizons of known age. This method is unreliable because it takes no account of ice deformation. Moreover, there is little correlation between the results of these studies.
