Ice core methodology - Ice Ages and Interglacials: Measurements, Interpretation, and Models - page 127

experimentally that a decrease in the content 18 O in East Antarctic ice (relative to
the standard mean of sea water) by 1 percent corresponds to a cooling 1.5 C
whereas a 6 percent D shift will correspond to a temperature drop of 1 C.
Using these correlations, one may transform an isotope curve into one of
paleotemperature.''
However, the conversion of d 18 Oor d D measurements to temperature is far
from straightforward due to a variety of confounding factors that may occur.
While isotope ratios provide a qualitative indication of prevailing temperatures in
the past, the quantitative accuracy of such correlations has been challenged. Other
factors that affect isotope ratios during glacial periods include more extensive sea
ice cover during glacial periods, which increases the distance from the moisture
source and changes the isotopic composition of oceans during glacial periods.
Ideally, records of both surface temperature and the isotopic contents of
precipitation should span a relatively long period at any site where an empirical
relationship between d and T is sought. The only polar site for which suitable
data are available is the South Pole station, where temperatures are available from
1957 to 1978 along with well-dated isotopic profiles. Mean annual and maximum
deuterium ratios correlated well with corresponding mean annual and summer
temperatures; however, winter temperature and deuterium minima were poorly
correlated. Detailed curve matching of isotope-depth records from snow pits to
temperature-time records from automated weather stations (AWS) and satellite
temperature sensors provide additional support for the linear relationship of d
with T (Shuman et al., 2001). These short-term comparisons provide some
support for the belief that isotopic ratios of accumulated snow reflect tem-
peratures, especially for comparisons from summer to summer. Some noise is, of
course, present.
In order to use the isotope signal as a paleothermometer, the present day
spatial isotope-surface temperature relationship d ¼ aT þ b defined over a certain
region ( d stands for either d Dor d 18 O of the precipitation and T for the surface
temperature) the isotope-surface temperature slope is generally assumed to hold
in time throughout the region. Present day isotope ratios and temperatures corre-
late very nicely as shown in Figure 3.15 . This does not necessarily validate the
relationship under past conditions, particularly during extensive glaciation.
Jouzel et al. (1997) reviewed the empirical estimates and theoretical models of
the relationship between isotope ratios and temperature. They focused on polar
areas and singled out Greenland where the GRIP and GISP2 drillings allowed
empirical estimates to be obtained from paleothermometry and motivated model-
ing experiments (see Figure 3.17 ). In a more recent review, Miller et al. (2010)
provided the data shown in Figure 3.16 .
Jouzel et al. (2003) studied the relationship between temperature and isotope
ratios in Antarctica. They examined all relevant information, focusing on the East
Antarctic Plateau where both model and empirical isotope-temperature estimates
are available. Based on the evidence presently available they concluded that,
unlike the case of Greenland, the present day dependence of the isotope ratio on