Abrupt Climate Changes During the Holocene Across North America From Pollen and Paleolimnological Records - Abrupt Climate Change: Mechanisms, Patterns, and Impacts - page 176

Alder is abundant in the forest-tundra, and thus can probably
tolerate more cold conditions than spruce or poplar. However,
this example also illustrates another aspect to be considered
when interpreting pollen records. Just before 9 ka, poplar
populations decreased in abundance and remained low for
the remainder of the Holocene, until the European distur-
bance of the past 400 years. In this case, it was suggested that
competition from other trees may have been responsible
for the decrease. Analyses such as these indicate the rapid
response of tree populations to climate changes. Although
the changes may be complex and not always intuitive, they
provide good examples of the rapid nature of climate transi-
tions, as well as the response.
Viau et al. [2006] estimated the climate of North Amer-
ica during the past 14 kyr using over 750 pollen records
that comprised over 30,000 samples dated by approximately
2500 radiocarbon dates extracted from the NAPD from
across North America. After the July temperature was
estimated using the MAT for each pollen sample, data
were interpolated to uniform time intervals, and a mean
temperature curve for North America was computed. This
continental-scale reconstruction quanti

large-scale paleoclimate reconstructions. Comparison of pol-
len records from lakes in Labrador and Quebec with pollen
and dino agellate records in offshore marine cores showed

broad-scale coherence [Sawada et al., 1999]. Pollen have
been studied in ice cores [Bourgeois et al., 2000], permitting
a direct comparison of the ice core record with vegetation
change. Indeed, a high-resolution study of seasonal ice layers
in the Agassiz ice core has shown that pollen are transported
to the ice cap in the same season as they are liberated and
thus faithfully record the phenology of the major plant
groups [Bourgeois, 2001; Gajewski, 2006]. Studies such as
this can be used to directly relate climate variability recon-
structed from terrestrial and marine or ice caps, as the pollen
are a common proxy found in all systems.
When well-dated records of abrupt changes in the past
are available, they can be compared and lend insight into
potential climate forcing mechanisms. A spectral analysis
of the reconstruction of Viau et al. [2006] found a major
peak around 1100 years (Table 1). A similar peak has been
identified in several studies, including the Greenland Ice
Core Project 18 O record during the Holocene [Schulz and
Paul, 2002], North Atlantic IRD records [Bond et al., 2001]
and North Atlantic circulation patterns [Chapman and
Shackleton, 2000], among others. A similar periodicity was
identi

ed temperature varia-
tions of several timescales during the past 14 kyr (Figure 3).
Temperatures increased during the late glacial, and maxi-
mum Holocene temperatures occurred between 6000 and
3000 cal years B.P. in North America. Millennial-scale
temperature variations, with a range of ±0.2°C and sepa-
rated by abrupt transitions, are superimposed on the
Milankovitch-scale variability. The dominant frequency of
the July temperature variability is ~1100 years (as opposed
to the ~1500 year cycle found during the last glacial
period), with abrupt transitions between climate states. In
addition, there was a scale interaction, with the millennial-
scale variability being more pronounced during both the
late glacial warming and the late Holocene cooling. A
comparable study was performed using European pollen
data [Davis et al., 2003].
In a related study, Gajewski et al., [2006] showed synchro-
nous abrupt vegetation transitions during the Holocene be-
tween both North American and European continents. In this
study, two independent data sets from Europe were used to
test the robustness of the results. Major vegetation transitions
averaging ~1100 years intervals between both continents
during the Holocene were also found to be synchronous with
North Atlantic ice-rafted detritus (IRD) events [Bond et al.,
2001].

ed in high-resolution records from central United
States [Overpeck, 1987], Scotland [Langdon et al., 2003],
and Alaska [Hu et al., 2003]. Bond et al. [2001] have
shown a cross-spectral coherence of their IRD record with
cosmogenic nuclide records ( 14 Cand 10 Be, proxies for solar
variability), at the 900 to 1100 year frequency bands for the
Holocene.

Table 1. Dominant Peaks in Spectral Analyses of Several Time
Series From the North Atlantic Region and North America

Periodicities
10 2 Time
Scale

10 3 Time
Scale Reference

Cross-spectral ice-rafted detritus
(IRD)/ 14 C[Bond et al., 2001]

400 - 500 900 - 1100

Cross-spectral IRD/ 10 Be
[Bond et al., 2001]

400 - 500 900 - 1100

North Atlantic IRD
[Bond et al., 1997, 2001]

~1350

GRIP/GISP2 18 O[Schulz and Paul, 2002] 900

THC/NADW 13 C[Chapman and
Shackleton, 2000]

550 1000

North America [Viau et al., 2006] 1150

3.3. Relating Terrestrial to Ice Core and Marine Records

Subarctic Alaska [Hu et al., 2003] 950

Midwest United States [Overpeck, 1987] 1100

Studies of pollen in marine cores [e.g., Heusser and
Morley, 1985; Lezine and Dene

Southeast Scotland [Langdon et al., 2003] 1100

e, 1997] have been used for