Geoscience Reference
In-Depth Information
While the time series of the monthly AO index values
shown in Figure 9e does not indicate a consistent anom-
aly, this does not preclude the presence of strong cyclonic
circulation within the Arctic. In fact, strong low pressure
throughout the central Arctic is found for the simulation
years 556-558 and is illustrated for year 556 in Figure 10.
Sustained negative wintertime sea level pressure anomalies
of greater than 12 hPa are shown for the central Arctic, and
significant negative anomalies even extend to the middle lat-
itudes near the Ural Mountains in eastern Russia. The effects
on the sea ice coverage may be seen in the ice thickness con-
tours and ice velocity vectors shown in Plate 5. For the year
556, a notable absence of the Beaufort Gyre in the central
Arctic may be seen, similar to the year 451 event as shown in
Plate 4a. Sea ice under these conditions does not recirculate
within the central Arctic; rather, it is seen to be compressed
along the Canadian Arctic Archipelago or to exit directly
through Fram Strait. Plate 5 indicates anomalously thick ice
of greater than 5 m along the canadian Arctic Archipelago,
suggesting a compaction of the sea ice into those locations.
In Figure 9c, it is seen that the recovery period is marked
by a period of low sea ice export, which extends for several
years to year 561. In Plate 5 (right), the pattern of the Beau-
fort Gyre is found to be reestablished after the anomaly, and
thicker sea ice may be seen to be extending into the central
Arctic Basin.
Although the three events and their recoveries are hetero-
geneous, they highlight mechanisms that have been iden-
tified with sea ice loss in the contemporary Arctic. These
include the influence of the prevailing wintertime atmo-
spheric circulation in producing enhanced ice export and
the preconditioning of the Arctic through volume losses to
produce large extent anomalies at a later time [e.g.,
Rigor
et al.
, 2002]. The second event is important in demonstrat-
ing enhanced losses through surface radiative fluxes during
the melt season [e.g.,
Francis et al.
, 2005]. Finally, the third
event denotes a coupled model simulation of the scenario
proposed by
Holloway and Sou
[2002] in showing an extent
anomaly in the absence of a volume anomaly. Such a condi-
tion requires a sustained, concurrent atmospheric forcing to
produce the extent anomaly, and recovery is found to occur
immediately after the forcing is removed.
Over the extended period of the simulation, several of the
climate variables examined here show strong correlations
with the modeled September sea ice cover; however, these
variables are not directly involved in the remarkable mini-
mum ice cover events that have been examined. Shown in
Figure 11 is the time series of September sea ice extent in
comparison with the seasonal AO index and with the North
Atlantic oceanic poleward heat transport. Atmospheric cir-
culation is found in this study to be a key initiator of the first
and third events. In the first event, the anomalous AO index
occurred in February and march of the year of the anomaly,
while the seasonal average of the index was unremarkable.
In the case of third event, low-pressure anomalies extended
well beyond the high latitudes, thus obfuscating the canoni-
cal AO index. It may be seen in Figure 11 that there is a
large amount of high-frequency variability in the seasonal
AO index but that there is a correlation with sea ice extent at
low frequencies. A cross-spectral analysis indicates a broad
spectrum of coherence for periods from 29 through 58 years
that is significant at the 99.9% confidence level. Similarly,
the oceanic heat transport shows a correlation with sea ice
extent at low frequencies. The ocean heat flux and Septem-
ber sea ice extent show significant coherence centered at a
period of 25 years. As may be seen in the Figure 11, how-
ever, the global ocean does not appear to play a direct role in
influencing the anomalous events examined here. This is in
marked contrast with simulations using 21st century forcing
scenarios in which the ocean plays a more direct role [
Hol-
land et al.
, 2006a].
An important question arising from this study is the interpre-
tation of the presence of such events and their rapid recovery
in a control simulation, assuming an absolute validity of the
model in reproducing the Arctic climate (a tenuous assump-
tion). One may speculate that such events are evidently part
of the internal, natural climate variability and that evidence
4. DIScuSSION
In this study, three events of extraordinary Northern
Hemisphere minimum sea ice cover in an extended control
simulation of a coupled climate model are investigated. The
events are episodic and have differing forcing characteristics.
A pervasive characteristic of these events is the absence of
the climatological Beaufort Gyre, which serves to recirculate
multiyear sea ice within the Arctic Basin and restricts export.
The first two events are associated with ice volume anomalies
that were established by anomalous export and melting in the
decade prior to each event. In each case, the extent anomaly
recovered to normal conditions very rapidly, while the ice
volume anomaly continued for an extended period of time.
In the case of the second event, a minimum volume anomaly
was established 10 years after the minimum sea ice extent
event, while complete recovery of Northern Hemisphere sea
ice volume to normal conditions was an additional 10 years
later. In the third event, anomalously low pressure in the cen-
tral Arctic over a 3-year period produced a redistribution of
the sea ice cover, resulting in an ice extent anomaly but not
a sea ice volume anomaly. The ice extent subsequently re-
turned to normal conditions in the presence of anticyclonic
circulation forced by high atmospheric pressure.
Search WWH ::
Custom Search