periods; (2) firn models do not apply to glacial conditions of
the Antarctic plateau; (3) another fractionation process
occurs at the closure of the pores. This issue remains
undecided at present.
Synchronization of Two Ice Cores
As explained above, the ice/gas differential in cores with a
low rate of accumulation from the Eastern Antarctic shelf
and during the ice ages is still poorly constrained. An
alternative way to obtain an estimate, for both ice and gas
recordings, is synchronization with a core with a higher rate
of accumulation, wherein the ice/air differential is better
constrained.
Loulergue et al. (2007) and Parrenin et al. (2012) have
applied this method to constrain the Dage of the EDC
(EPICA Dome C) from the EDML (EPICA Dronning Maud
Land) and the TALDICE (Talos Dome Ice Core) cores. Gas
synchronization is based on the rapid variations in methane,
and ice synchronization uses volcanic signatures. So, this
study shows that the firn model, forced with temperature and
accumulation scenarios as for dating ice (Parrenin et al.
2007b), overestimates the Dage at EDC by 500–1000 years,
during the last glacial period. Consequently, the densification mechanism during glacial periods at EDC is poorly
understood and the models need to be improved.
The Dage, during glacial periods, for Antarctic plateau
sites with low accumulation is therefore an open question.
Further studies are needed to clarify this issue.
The Counting of Annual Layers
On the polar ice caps and glaciers, many of the properties of
snow differ depending on whether it accumulates in summer
or winter. For example, in summer, dust is more abundant in
the snow, because during this season, the winds are more
conducive to dust transport towards the poles. The annual
layers can therefore be identified, either visually, or by
chemical analysis or by isotopic analysis. Counting annual
layers is a simple method of dating, provided that the
accumulation of snow is sufficient, so that the stratigraphy is
not destroyed by winds mixing the layers near the surface.
For this reason, the counting of layers is impossible in the
central regions of the Antarctic plateau where the deep
drilling of Vostok, Dome C and Dome F are located, but it is
possible over Greenland and the coastal regions of
Antarctica.
A large project for systematic counting called Greenland
Ice Core Chronology 2005 (GICC05) has been undertaken
by a Danish team at the Niels Bohr Institute in Copenhagen.
It is based on the cores of DYE-3, GRIP and NorthGRIP,
and currently extends over the last 60,000 years (Svensson
et al. 2008). More recently, the WAIS (West Antarctic Ice
Sheet) Divide ice core has been counted back to
31,000 years (Sigl et al. 2016).
Glaciologists use various records to identify annual layers. Where possible, the isotopic variations in the ice (d
18
O
and dD), which are dependent on the temperature at the
moment of the precipitation, provide the most reliable
recording of the changing of the seasons. However, water
molecules diffuse in the form of vapor through the firn, then
more slowly through the ice. This diffusion smooths out the
seasonal isotopic signal until it disappears at a certain depth,
even more rapidly when accumulation is low and the temperature is high. Thus, the seasonal cycle of isotopes is
hardly recognizable on the NorthGRIP core, which has a low
accumulation; it is quite muted in the GRIP core. The
longest sequence on which the seasonal cycle oxygen-18
was used was obtained from the Dye-3 core in Greenland:
67,000 isotopic analyses allowed the dating of the core year
by year over the last 7900 years. Beyond that, the thickness
of annual layers is insufficient and the isotopic diffusion
through the ice makes counting inaccurate.
Other data taken from the content of impurities in Continuous Flow Analysis (CFA), from the Electrical Conductivity Measurement (ECM), from the insoluble dust content,
and from Visual Stratigraphy (VS) complete the isotopic
information when this is available (Fig. 9.3). The CFA
allows the various soluble compounds, such as Na
+
, Ca
2+,
H 2 O 2 , NH
þ
4 , NO
À
3 and SO
2À
4 to be separated. The ECM is a
non-destructive measurement, conducted continuously in the
field, but it only provides information on an amalgamation of
these different soluble compounds. The VS uses the fact that
impurities diffuse the light in the ice. However, this
recording generally shows several peaks in a year and is
therefore not easy to interpret. In Greenland and during the
Holocene, a typical year is characterized by Na
+ (dominated
mainly by marine inputs) showing a peak in late winter.
Spring has a high dust content, high Ca
2+ and low H 2 O 2 .
Summer is characterized by high concentrations of NH
þ
4 ,
NO
À
3 , and sometimes SO
2À
4 . This method, based on data
from CFA, ECM and VS from GRIP and NorthGRIP, was
the one principally used to establish GICC05 in the period
between 7900 and 14,800 years b2k (this notation means
‘years before 2000’) (Rasmussen et al. 2006). In the older
part (14,800–60,000 Years b2k), the method is the same, but
only NorthGRIP data were used (Svensson et al. 2008).
In summary, none of the individual indicators is perfect,
but combined, they permit an annual dating, as long as the
thickness of the layers, which thin out as they sink into the
ice cap, remain sufficient. For GICC05, counting was carried
out by different people and on different cores, and the
independently obtained results were compared so as to
126
F. Parrenin
the Antarctic plateau; (3) another fractionation process
occurs at the closure of the pores. This issue remains
undecided at present.
Synchronization of Two Ice Cores
As explained above, the ice/gas differential in cores with a
low rate of accumulation from the Eastern Antarctic shelf
and during the ice ages is still poorly constrained. An
alternative way to obtain an estimate, for both ice and gas
recordings, is synchronization with a core with a higher rate
of accumulation, wherein the ice/air differential is better
constrained.
Loulergue et al. (2007) and Parrenin et al. (2012) have
applied this method to constrain the Dage of the EDC
(EPICA Dome C) from the EDML (EPICA Dronning Maud
Land) and the TALDICE (Talos Dome Ice Core) cores. Gas
synchronization is based on the rapid variations in methane,
and ice synchronization uses volcanic signatures. So, this
study shows that the firn model, forced with temperature and
accumulation scenarios as for dating ice (Parrenin et al.
2007b), overestimates the Dage at EDC by 500–1000 years,
during the last glacial period. Consequently, the densification mechanism during glacial periods at EDC is poorly
understood and the models need to be improved.
The Dage, during glacial periods, for Antarctic plateau
sites with low accumulation is therefore an open question.
Further studies are needed to clarify this issue.
The Counting of Annual Layers
On the polar ice caps and glaciers, many of the properties of
snow differ depending on whether it accumulates in summer
or winter. For example, in summer, dust is more abundant in
the snow, because during this season, the winds are more
conducive to dust transport towards the poles. The annual
layers can therefore be identified, either visually, or by
chemical analysis or by isotopic analysis. Counting annual
layers is a simple method of dating, provided that the
accumulation of snow is sufficient, so that the stratigraphy is
not destroyed by winds mixing the layers near the surface.
For this reason, the counting of layers is impossible in the
central regions of the Antarctic plateau where the deep
drilling of Vostok, Dome C and Dome F are located, but it is
possible over Greenland and the coastal regions of
Antarctica.
A large project for systematic counting called Greenland
Ice Core Chronology 2005 (GICC05) has been undertaken
by a Danish team at the Niels Bohr Institute in Copenhagen.
It is based on the cores of DYE-3, GRIP and NorthGRIP,
and currently extends over the last 60,000 years (Svensson
et al. 2008). More recently, the WAIS (West Antarctic Ice
Sheet) Divide ice core has been counted back to
31,000 years (Sigl et al. 2016).
Glaciologists use various records to identify annual layers. Where possible, the isotopic variations in the ice (d
18
O
and dD), which are dependent on the temperature at the
moment of the precipitation, provide the most reliable
recording of the changing of the seasons. However, water
molecules diffuse in the form of vapor through the firn, then
more slowly through the ice. This diffusion smooths out the
seasonal isotopic signal until it disappears at a certain depth,
even more rapidly when accumulation is low and the temperature is high. Thus, the seasonal cycle of isotopes is
hardly recognizable on the NorthGRIP core, which has a low
accumulation; it is quite muted in the GRIP core. The
longest sequence on which the seasonal cycle oxygen-18
was used was obtained from the Dye-3 core in Greenland:
67,000 isotopic analyses allowed the dating of the core year
by year over the last 7900 years. Beyond that, the thickness
of annual layers is insufficient and the isotopic diffusion
through the ice makes counting inaccurate.
Other data taken from the content of impurities in Continuous Flow Analysis (CFA), from the Electrical Conductivity Measurement (ECM), from the insoluble dust content,
and from Visual Stratigraphy (VS) complete the isotopic
information when this is available (Fig. 9.3). The CFA
allows the various soluble compounds, such as Na
+
, Ca
2+,
H 2 O 2 , NH
þ
4 , NO
À
3 and SO
2À
4 to be separated. The ECM is a
non-destructive measurement, conducted continuously in the
field, but it only provides information on an amalgamation of
these different soluble compounds. The VS uses the fact that
impurities diffuse the light in the ice. However, this
recording generally shows several peaks in a year and is
therefore not easy to interpret. In Greenland and during the
Holocene, a typical year is characterized by Na
+ (dominated
mainly by marine inputs) showing a peak in late winter.
Spring has a high dust content, high Ca
2+ and low H 2 O 2 .
Summer is characterized by high concentrations of NH
þ
4 ,
NO
À
3 , and sometimes SO
2À
4 . This method, based on data
from CFA, ECM and VS from GRIP and NorthGRIP, was
the one principally used to establish GICC05 in the period
between 7900 and 14,800 years b2k (this notation means
‘years before 2000’) (Rasmussen et al. 2006). In the older
part (14,800–60,000 Years b2k), the method is the same, but
only NorthGRIP data were used (Svensson et al. 2008).
In summary, none of the individual indicators is perfect,
but combined, they permit an annual dating, as long as the
thickness of the layers, which thin out as they sink into the
ice cap, remain sufficient. For GICC05, counting was carried
out by different people and on different cores, and the
independently obtained results were compared so as to
126
F. Parrenin