Ice core methodology - Ice Ages and Interglacials: Measurements, Interpretation, and Models - page 130

temperature affect isotopic composition. These include changes in sea-surface
composition and temperature, changes in atmospheric circulation, changes in
cloud temperature, which may be different from changes in surface temperature,
changes in the seasonality of precipitation, and post-depositional isotopic
exchange in the snowpack. Second, all of these factors may vary through time
in such a way that a single, linear relation between d 18 O and T is inappropriate.
Thus, there is strong motivation to seek paleotemperature information that is
entirely independent of isotopic history to calibrate the paleothermometer.''
Temperatures in the upper 10m of an ice sheet are primarily controlled by
mean annual air temperature. Changes in air temperature propagate downward
into the ice by diffusive heat flow and ice flow. The temperature profile through
an ice sheet thus provides a record of past air temperature modified by heat
diffusion and ice flow. The profile is readily measured in a borehole through the
ice. Although the thermal properties of ice and firn are well known, there may be
large uncertainties in ice flow in different areas of ice sheets. The availability of
excellent dating at the Greenland summit, combined with the central location on a
rather stable ice sheet, allows relatively accurate calculation of ice flow effects.
Because of heat diffusion, conversion of the depth-temperature record of a
borehole to a surface temperature history is a complex process and does not neces-
sarily yield a unique result. However, methods have been developed that appear to
yield good estimates of historical surface temperatures from borehole temperature
profiles at benign sites where ice flow is not so problematic. Borehole analyses
have been conducted for the GISP2 and GRIP.
During the summer of 1994, Cuffey et al. (1995) measured temperature in the
3,044m deep GISP2 core hole (filled with liquid) from 70m below the surface to
the base of the ice sheet. At that time, thermal perturbation from drilling had
decayed to less than 0.04 C, so the temperature in the borehole matched the
temperature in the surrounding ice sheet at this accuracy and better. This depth
corresponded to a time span of about 40,000 years.
Heat diffusion damps high-frequency temperature changes as they propagate
from the surface down into the ice sheet. Therefore, the rapid environmental
temperature changes that occurred in the past are damped out in the actual
temperature vs. depth data in the ice sheet. While the isotope record in the core
may contain a record of rapid fluctuations in the past, the subsurface temperature
record only provides a filtered average. Such thermal averaging is more extensive
for older climatic events. Therefore, the borehole analysis utilized a filtered version
of the GISP d 18 O record, as shown in Figure 3.18 . Thus, borehole isotope calibra-
tion is sensitive mainly to the long-term warming from full glacial conditions to
the Holocene and to Holocene temperature changes, but does not reflect the wild
oscillations in the d 18 O record.
Their procedure for estimating A and B in the isotopic paleothermometer was
as follows. They used the filtered GISP2 d 18 O record ( d 18 O vs. age or depth) and
an initial guess for A and B to specify a 100,000-year history of surface environ-
mental temperature. The initial guess for A and B was based on current data for