Sea Ice Variability in 2001 and 2002 in the Antarctic and
B/A Seas

Josefino C. Comiso

Laboratory for Hydrospheric Processes, Code 971

NASA Goddard Space Flight Center

Greenbelt, MD 20771

email: comiso@joey.gsfc.nasa.gov

GLOBEC Meeting, Hilton Hotel

9-11 December 2002

Scientific Motivations:

The polar regions may provide the earliest signal of a climate change because of feedbacks between ice, ocean and atmosphere.

The Antarctic Peninsula region appears to be anomalously warm climatologically compared to the rest of the continent.

The entire Antarctic sea ice cover has been observed to be increasing at less than 1%/decade while the Bellingshausen/Amundsen Seas region has been declining at an anomalously large negative rate.

Correlation of SO indices with B/A ice cover appears to be very strong.

A rapidly retreating Arctic perennial ice

The perennial sea ice cover has been retreating at the rate of -9 %/decade

There is a 90% level of confidence that the trend is between -5% and -12%.

The summer ice surface temperature has also been on the rise at 1.2 K/decade.

Ref: Comiso, J., A rapidly declining sea ice cover, Geophys. Res. Let., 29(20), 1956, 2002.

Trends in the Antarctic Sea Ice

Trends in extent and ice area are 0.3 and 0.8 %/decade, respectively.

Difference in the trends for extent and area are associated with variations in average ice concentration.

Trends in the Bellingshausen/ Amundsen Sea ice cover

Despite slight recovery, the B/A region is still declining at a rapid rate of -7% and -6% per decade for ice extent and area, respectively.

Sea ice in this region is highly variable.

Seasonal ice cover in 2000

The multiyear ice cover in the Bellingshausen Sea was reported by Jacobs and Comiso (1993 &1997) to be declining fast.

In the 1990s and 2000s, the region was basically covered by seasonal ice.

Seasonal Ice Cover in 2001

The Western Weddell was ice free at the southern and coastal regions for the 2^nd time in two decades.

The Ross Sea polynya was very different with extensive coastal areas becoming ice free.

Seasonal Ice Cover in 2002

For the first time in the satellite era, the Northern tip of the Antarctic Penninsula was ice free in January thru March.

The multiyear ice cover in the B/A seas was very minimal

Climatological Ice Cover

The climatological ice cover is quite different from those in 2001 and 2002.

The summer months show concentration of ice in the Western Weddell and B/A regions.

The winter months show basically a circumpolar ice cover.

Monthly Anomalies in 2000

The monthly anomalies show alternating patterns of growth and retreat along the pheriphery of the ice cover.

Such pattern has been associated with the ACW.

Wave number can be inferred and is generally wave 3.

Monthly anomalies are very different in
2001 and 2002

Ice Concentrations in the B/A region in June 2001

Freeze up at the Marguerite Bay did not occur until late June.

The ice cover along the western part of the Antarctic Peninsula are generally loss/new ice.

Ice Concentrations in the B/A region in July 2001

The western tip of the Antarctic peninsula is the last region to undergo freeze-up in winter.

The SST in the general area has been shown to be anomalously warm.

Ice Concentrations in the B/A region in November 2001

In 2001, the decay of sea ice in the B/A region occurs in early November and starts at the northernmost margins of the ice cover.

Ice melt at the coastal areas are a lot slower than at the northern ice edges.

B/A climatological ice cover compared with 2000-2002.

The seasonal ice cover for the different years are very different.

The ice cover in 2001 appears to be the least extensive of the three while that for 2000 appears to be closest to normal.

AMSR-E Sensor Characteristics

Freq. (GHZ)      6.9    10.65    18.7     23.8    36.5    89

Polarization      H&V     H&V       H&V      H&V      H&V   H&V

IFOV (km)       75x43   51x30    27x16    31x18    14x8    6x4

Sens (K)             0.3       0.6         0.6        0.6        0.6      1.1

Int. time(msec)   2.6       2.6         2.6         2.6        2.6     1.3

Beamwidth (^o)     2.2      1.4         0.8         0.9        0.4    0.18

Swath width   –   1445 km

Satellite altitude – 705 km

Satellite TB intercomparison

AMSR and SSMI are highly correlated with the σ being about 1 K within the ice pack and higher in the open ocean.

Significant variance in ocean regions is mainly due to mismatchs in time of observation.

Sensor TB and IC spatial consistency

Differences in TBs are mainly in open ocean regions where weather effects are apparent.

The changes are mainly caused by differences in revisit times over the polar regions.

Despite bias and a slight change in TB calibration, the derived ice concentrations are basically identical.

Deviations at
the Ice edge

High resolution data provide a better definition of the ice edge.

With AMSR data, all channels provide consistent ice edge information.

Some discrepancies between AMSR and SSM/I IC ice edge location is observed.

Ice edge studies

SSM/I ice edge differs from channel to channel showing the effect of different resolution.

AMSR ice edge differs from that of SSM/I by about 12 km mainly due to the difference in resolution.

Trend studies require a proper matching of ice edges between sensors.

Ocean Mask with SSM/I and AMSR

Patterns for open ocean data are similar but are more defined with AMSR.

With AMSR, a 10% ice edge is much easier to consistently obtain despite varying weather conditions than with SSM/I data.

Comparison of Techniques

The use of the 6 GHz data provides a good baseline for validating retrieved ice concentrations.

The V1836 technique yields similar results to that of V0636.

The combined V1836 & VH36 technique accounts for new ice distributions but overestimates IC.

MODIS data on July 23, 2002
This data provide good validation for the actual spatial distribution of the ice cover.
Full picture Enlarged version

Summary and Conclusions

The Antarctic sea ice cover is generally stable but some areas are largely varying.

The ice cover at the Bellingshausen/Amundsen Seas is declining rapidly but much of the decline is compensated by increases in the ice cover at the Ross Sea.

The ice cover in 2001 and 2002 are very different not only in the B/A region but at the entire hemisphere.

AMSR is an excellent successor to SSM/I.

Advantages of AMSR includes: (a) More accurate ice concentration and better definition of ice edges – because of higher resolution and more frequency channels; (b) Wider swath and smaller gap around the North Pole; (c) Improved masking of ice free ocean; and (d) Improved masking of ice free land/ocean boundaries.

Some disagreements between sensors are apparent but may be largely due to resolution differences and side lobe effects.

Co-registered and coincident AMSR and MODIS data in tandem will provide complementary and more accurate information about the ice cover AMSR can be used to assess the accuracy of historical passive microwave data on sea ice.

The validation of sea ice products from satellite data is very important

End
of Presentation


	Josefino C. Comiso
	Laboratory for Hydrospheric Processes, Code 971
	NASA Goddard Space Flight Center
	Greenbelt, MD 20771
	email: comiso@joey.gsfc.nasa.gov

	GLOBEC Meeting, Hilton Hotel
	9-11 December 2002


	The polar regions may provide the earliest signal of a climate change because of feedbacks between ice, ocean and atmosphere.
	The Antarctic Peninsula region appears to be anomalously warm climatologically compared to the rest of the continent.
	The entire Antarctic sea ice cover has been observed to be increasing at less than 1%/decade while the Bellingshausen/Amundsen Seas region has been declining at an anomalously large negative rate.
	Correlation of SO indices with B/A ice cover appears to be very strong.


	The perennial sea ice cover has been retreating at the rate of -9 %/decade
	There is a 90% level of confidence that the trend is between -5% and -12%.
	The summer ice surface temperature has also been on the rise at 1.2 K/decade.
	Ref: Comiso, J., A rapidly declining sea ice cover, Geophys. Res. Let., 29(20), 1956, 2002.


	Trends in extent and ice area are 0.3 and 0.8 %/decade, respectively.
	Difference in the trends for extent and area are associated with variations in average ice concentration.


	Despite slight recovery, the B/A region is still declining at a rapid rate of -7% and -6% per decade for ice extent and area, respectively.
	Sea ice in this region is highly variable.


	The multiyear ice cover in the Bellingshausen Sea was reported by Jacobs and Comiso (1993 &1997) to be declining fast.
	In the 1990s and 2000s, the region was basically covered by seasonal ice.


	The Western Weddell was ice free at the southern and coastal regions for the 2^nd time in two decades.
	The Ross Sea polynya was very different with extensive coastal areas becoming ice free.


	For the first time in the satellite era, the Northern tip of the Antarctic Penninsula was ice free in January thru March.
	The multiyear ice cover in the B/A seas was very minimal


	The climatological ice cover is quite different from those in 2001 and 2002.
	The summer months show concentration of ice in the Western Weddell and B/A regions.
	The winter months show basically a circumpolar ice cover.


	The monthly anomalies show alternating patterns of growth and retreat along the pheriphery of the ice cover.
	Such pattern has been associated with the ACW.
	Wave number can be inferred and is generally wave 3.


	Freeze up at the Marguerite Bay did not occur until late June.
	The ice cover along the western part of the Antarctic Peninsula are generally loss/new ice.


	The western tip of the Antarctic peninsula is the last region to undergo freeze-up in winter.
	The SST in the general area has been shown to be anomalously warm.


	In 2001, the decay of sea ice in the B/A region occurs in early November and starts at the northernmost margins of the ice cover.
	Ice melt at the coastal areas are a lot slower than at the northern ice edges.


	The seasonal ice cover for the different years are very different.
	The ice cover in 2001 appears to be the least extensive of the three while that for 2000 appears to be closest to normal.


	Freq. (GHZ) 6.9 10.65 18.7 23.8 36.5 89
	Polarization H&V H&V H&V H&V H&V H&V
	IFOV (km) 75x43 51x30 27x16 31x18 14x8 6x4
	Sens (K) 0.3 0.6 0.6 0.6 0.6 1.1
	Int. time(msec) 2.6 2.6 2.6 2.6 2.6 1.3
	Beamwidth (^o) 2.2 1.4 0.8 0.9 0.4 0.18
	Swath width – 1445 km
	Satellite altitude – 705 km


	AMSR and SSMI are highly correlated with the σ being about 1 K within the ice pack and higher in the open ocean.
	Significant variance in ocean regions is mainly due to mismatchs in time of observation.


	Differences in TBs are mainly in open ocean regions where weather effects are apparent.
	The changes are mainly caused by differences in revisit times over the polar regions.
	Despite bias and a slight change in TB calibration, the derived ice concentrations are basically identical.


	High resolution data provide a better definition of the ice edge.
	With AMSR data, all channels provide consistent ice edge information.
	Some discrepancies between AMSR and SSM/I IC ice edge location is observed.


	SSM/I ice edge differs from channel to channel showing the effect of different resolution.
	AMSR ice edge differs from that of SSM/I by about 12 km mainly due to the difference in resolution.
	Trend studies require a proper matching of ice edges between sensors.


	Patterns for open ocean data are similar but are more defined with AMSR.
	With AMSR, a 10% ice edge is much easier to consistently obtain despite varying weather conditions than with SSM/I data.


	The use of the 6 GHz data provides a good baseline for validating retrieved ice concentrations.
	The V1836 technique yields similar results to that of V0636.
	The combined V1836 & VH36 technique accounts for new ice distributions but overestimates IC.


	The Antarctic sea ice cover is generally stable but some areas are largely varying.
	The ice cover at the Bellingshausen/Amundsen Seas is declining rapidly but much of the decline is compensated by increases in the ice cover at the Ross Sea.
	The ice cover in 2001 and 2002 are very different not only in the B/A region but at the entire hemisphere.
	AMSR is an excellent successor to SSM/I.
	Advantages of AMSR includes: (a) More accurate ice concentration and better definition of ice edges – because of higher resolution and more frequency channels; (b) Wider swath and smaller gap around the North Pole; (c) Improved masking of ice free ocean; and (d) Improved masking of ice free land/ocean boundaries.
	Some disagreements between sensors are apparent but may be largely due to resolution differences and side lobe effects.
	Co-registered and coincident AMSR and MODIS data in tandem will provide complementary and more accurate information about the ice cover AMSR can be used to assess the accuracy of historical passive microwave data on sea ice.
	The validation of sea ice products from satellite data is very important