Jökull - 01.01.2005, Page 8

Wallace S. Broecker
Superimposed on the episodes of climate decline
are prominent 20,000-year cycles (see Figure 5). The
existence of these cycles tips us off as to what drives
glaciations for their timing beautifully matches that
of changes in the distribution among the seasonals of
the heat received from the Sun. These cycles which
involve increases and decreases in the contrast be-
tween winter and summer insolation, are the result of
the precession of the Earth’s spin axis in combination
with its elliptical orbit. In addition to the pacing by
the 20,000-year precession cycles, evidence for an or-
bital influence on climate comes from the timing of
the terminations which brought to an abrupt end each
of the 100,000-year climate cycles. Each of these ter-
minations occurred at a time of maximal contrast be-
tween summer and winter. The match between or-
bital changes and earth climate is so convincing that
it is now universally accepted that glacial cycles were
paced by changes in seasonal contrast. The impor-
tant point for this paper is that when these seasonal-
ity changes are introduced into the same models used
to predict the impact of rising greenhouse gases, the
models barely respond. This tells us that there must
be interactions taking place in the real world (but not
in the models) which greatly amplify the impact of
nudges associated with changes in the distribution of
solar heat among the seasons.
Example 2. Paleoclimatologists were puzzled to
note that while the changes in air temperature and in
atmospheric CO2 content recorded in ice from long
borings in the Antarctic ice cap (see Figure 5) nicely
conform to expectation based on cycles in the Earth’s
orbit, the record from Greenland ice is quite different.
It was, as shown in Figure 6, dominated by abrupt
jumps spaced at millennial intervals. During the last
glacial cycle, about twenty of these back and forth
jumps occurred (Stuiver and Grootes, 2000). They
appear not only in the temperature record, but also in
those of the dust content (Mayewski et al., 1994) of
the ice and of the methane content (Brook et al., 1996)
of air trapped in bubbles in the ice. Because Green-
land’s ice has prominent annual layers, it is possible to
show that each of these transitions occurred with great
rapidity – i.e., they were completed in a few decades.
The mean annual temperature change associated with
these jumps has been rigorously documented to be on
the order of 10◦C (Severinghaus et al., 2001).
This discovery triggered a search for equivalents
in records from elsewhere on the planet. In rapid
succession evidence turned up at widely spaced north
temperate and tropical locales. But at Southern Hemi-
sphere locales, if present, they were muted and shifted
in time. While from the beginning it was postu-
lated that these shifts were triggered by reorganiza-
tions of the large scale conveyor-like circulation in
the Atlantic Ocean (see Figure 7), nearly twenty years
passed before an explanation for their abruptness and
widespread geographic distribution was found. The
key turned out to be a huge amplification resulting
from large changes in the extent of sea ice cover in
the northern Atlantic (Chiang et al., 2003). These
changes occurred in response to a turning on and off
of the Atlantic’s conveyor-like circulation. In its on
mode, the conveyor’s northward flowing upper limb
flooded the Norwegian Sea with water warmed dur-
ing its passage through the tropics. This warmth pre-
vented ice from forming. But, when the conveyor
shut down, this delivery of tropical heat no longer oc-
curred, allowing winter sea ice to cover the northern
Atlantic as far south as Great Britain. This ice would
not only have blocked the transfer of ocean heat to
the atmosphere, but also it would have reflected away
much of the incoming sunlight. As a result, winters in
Europe would have rivaled those which now plague
Siberia.
But it was the search for the connection to the
tropics which posed the greatest challenge. Climate
records from caves in China and from sediments from
the Arabian Sea clearly indicate that the monsoons
were greatly weakened during periods when sea ice
covered the northernAtlantic. Further, the record con-
tained in sediments from the Cariaco Basin just off
the Caribbean coast of South America and that in sta-
lagmites from a cave in southeastern Brazil revealed
that during times when sea ice covered the northern
Atlantic, the belt tropical rainfall was shifted south-
ward. Models endowed with northern Atlantic sea ice
were able to account for both the weakening of the
monsoons and for the southward shift of the tropical
rain belt (Chiang et al., 2003). The latter reflected the
8 JÖKULL No. 55, 2005