Polar Climate Change as Observed Mostly from Space

1
Polar Climate Change as Observed (Mostly) from
Space
The Cryosphere, the Atmosphere, and Their
Interactions
Jeff KeyCenter for Satellite Applications and
Research (STAR)NOAA/NESDIS at UW/CIMSSMadison,
Wisconsin
Acknowledgements Much of this material was
provided by Mark Anderson (U. Nebraska), Anne
Walker (Met. Service. Canada), Walt Meier (U.
Colorado), Xuanji Wang (U. Wisconsin), Yinghui
Liu (U. Wisconsin), Jennifer Francis (Rutgers U.)
and Axel Schweiger (U. Washington).
2
2007 was a record year

• 552 billion tons of ice melted this summer from
  the Greenland ice sheet. That's nearly quadruple
  the amount that melted just 15 years ago. It's an
  amount of water that could cover Washington,
  D.C., a half-mile deep.
• The surface area of summer sea ice floating in
  the Arctic Ocean this summer was nearly 23
  percent below the previous record. The dwindling
  sea ice already has affected wildlife, with 6,000
  walruses coming ashore in northwest Alaska in
  October for the first time in recorded history.
• Another first the Northwest Passage was open to
  navigation.
• The remaining Arctic sea ice is unusually thin,
  another record. That makes it more likely to melt
  in future summers. The overall volume of ice is
  half of 2004's total.
• Alaska's frozen permafrost is warming, not quite
  thawing yet. But temperature measurements 66 feet
  deep in the frozen soil rose nearly four-tenths
  of a degree from 2006 to 2007.
• Surface temperatures in the Arctic Ocean this
  summer were the highest in 77 years of
  record-keeping, with some places 8 degrees
  Fahrenheit above normal.

3
(No Transcript)
4

• Arctic vs Antarctic
• Ocean surrounded by land vs elevated land
  surrounded by ocean
• Polar bears vs penguins

5
Sea Ice
6
2007 Lowest Arctic Ice Extent on Record!
Arctic Sea Ice Extent on 19 Sept 2007
Northwest passage fully navigable nearly open
7
Passive microwave-derived (SMMR / SSM/I) sea ice
extent in September, 1979 through 2006. The
linear trend is approximately -9 per decade
(through 2006).
8
Mean Sea Ice Draft (from submarine sonar)
Melt Onset over the Arctic Ocean (based on
passive microwave data)
Julian Day
R20.50
9
Trends in Lake Ice Break-up and River Run-Off
An historical survey of rivers and lakes in the
Northern Hemisphere during the past 150 years
shows they have been thawing earlier and freezing
later each year. And river run-off/discharge
has increased.
10
Snow Cover
Passive Microwave-Derived Snow Water Equivalent
(SWE)
11
NA snow cover was generally above average in the
70s and 80s, and below average from the late 80s
on. Overall, NA snow cover has declined by about
10 since 1972.
12
Ice Sheet and Glaciers
13
Greenland Ice Sheet Melt
Melt extent and duration (number of days) from
QuikSCAT, 2000-2004 (a-e) (From Wang and Sharp,
2006). Greenland ice sheet melt area has
increased 16 between 1979 and 2002.
News Flash! 2007 has been a record melt year for
Greenland.
14
(from satellite and aircraft altimeters)
15
Satellite gravity provides direct measurements of
ice sheet mass change. Data from the GRACE
(Gravity Recovery and Climate Experiment)
satellite show that the Greenland Ice Sheet is
losing mass at an equivalent rate of sea level
rise of about 0.54 mm/yr.
16
Antarctica Warming
Breakup of the Larsen B ice shelf, January
through March 2002. (MODIS)
17
Antarctica Cooling?
Antarctic sea ice extent is at a record high in
2007.
18
Antarctica Warming and Cooling
Some parts of the Antarctic surface have been
cooling, some warming. The magnitudes of the
trends are not large. The mid-troposphere has
been warming at the rate of 0.5-0.7 degrees C per
decade (not shown). Cooling may be the result of
ocean cooling (2003-2005) and changes in the wind
field. Increasing ice sheet thickness in some
areas could result from increasing snowfall in
response to higher temperatures. Changes in the
polar vortex also play a role.
19
Glaciers
Not only has glacier length generally decreased,
global glacier mass has also (decreasing length
and thinning).
20
Above USGS photos of the South Cascade Glacier
(Washington) show dramatic retreat. Below Muir
Glacier in Glacier Bay on August 13, 1941 (left,
by William O. Field) and on August 31, 2004,
(right, by Bruce F. Molnia, USGS).
21
Permafrost Temperature across the Canadian Arctic
Western Arctic
High Arctic
Alert 15 m depth
Trend 1994-2004 0.07C/yr
22
The Atmosphere andAtmosphere-Ice Interactions
23
Surface Temperature and Cloud Amount from AVHRR
Winter Surface Temperature Trend APP-x
Arctic cloud cover has been increasing in summer
but decreasing in winter, with no significant
trend annually. However, the net result is an
increasing in cloud cooling and a decrease in
cloud warming. Arctic surface temperature has
generally be increasing, but decreasing in the
central Arctic during winter.
24
Low-Level Temperature Inversion Strength from
HIRS, 1980-1996
Trends in inversion strength are not everywhere
correlated with trends in surface temperature,
implying an influence of heat advection in the
lower troposphere.
Monthly trend of clear-sky inversion strength
(K/year) in January and winter (DJF), 1980-1996,
using a 2-channel statistical method.
25
Trends in Thermal Winds from TOVS, 1979-2001
and Trends in Heat Advection
Decadal trends in poleward advective heating in
the 500-300 hPa layer (K/day/decade) during
winter.
26
Possible Causes of Decreasing Wintertime Clouds
Correlation Between Cloud Cover Everywhere and
Cloud Cover in Pie Slice
APP-x Cloud Cover Trend
ERA-40 Cloud Cover Trend
TOVS Path-P Moisture Convergence Trend
Cyclone Frequency Difference
The reduction in moisture convergence results in
a decrease in cloud formation due to weakening
cyclone activity. Reduced cloud amount over the
pie slice leads to decreased cloud cover over
the entire central Arctic because less cloud is
advected to other regions.
1990-2000 minus 1982-1999
27
Long-Term Cloud-Ice Relationships
Ice concentration
Low cloud
Static stability
Middle cloud
Autumn 1980-2001 anomalies for years with values
more (left) or less (right) than 0.5 SD from the
mean. Based on ERA-40, but TOVS Path-P verifies
the patterns.
Changes in cloud cover can be explained by a
decrease in static stability and a deepening of
the atmospheric boundary layer resulting from the
removal of ice, the increase in surface
temperature, and the corresponding reduction in
stability.
28
Cloud-Ice Relationships Summer of 2007
An examination of cloud cover, sea ice extent,
and sea level pressure during the record minimum
ice extent of 2007 indicates that cloud cover is
largely, but not exclusively driven by
large-scale circulation. Local processes over
open water also play a role, but perhaps not a
large one. 1st row cloud fraction monthly
anomalies from MODIS () for 2007 as compared to
the mean 2002-2007 period 2nd row sea level
pressure monthly anomalies from NCEP-NCAR (hPa)
for 2007 versus the 2002-2007 time period 3rd
row sea ice concentration monthly anomalies from
AMSR-E () for the same period.
29
What does the future hold? (Its not so easy to
predict)
30
Supplemental Slides
31
Arctic Oscillation
Positive phase
Negative phase
The strong positive pattern between 1989 and 1995
may have flushed older, thicker ice out of the
Arctic, leaving the region with younger, thinner
ice that was more prone to summer melting. So the
acceleration in the sea ice decline since the mid
1990s may have been triggered by the strongly
positive AO mode during the preceding years, and
continued as a result of rising temperatures.
32
Clues to Sea Ice Variability Using AVHRR and TOVS
Percentages of variance (y-axis) in anomalies of
sea ice maximum retreat explained by anomalies in
zonal wind (U, black), meridional wind (V, blue),
downwelling longwave flux (DLF, green), and the
convergence of advected sensible heat (ADV, red)
in each peripheral sea of the Arctic Ocean. The
bars represent explained variance at lags of 0,
10, 25, 50, and 80 days, where the ice edge
anomaly lags the forcing anomaly.
33
Satellite Climatology Major Findings
The loss of Arctic perennial ice cover is almost
10 per decade. The relative roles of advection
and radiation vary by region. North America
snow cover has decreased by about 10 since 1972.
Melt onset over the Arctic Ocean is about 10
days earlier now than in 1980. Glacier and ice
sheet mass balance has decreased over the last
few decades. Greenland ice sheet melt area has
increased by 16 over the last 20 years. The
Arctic has been cooling at the surface during the
winter in the central Arctic, but warming at
other times of the year, particularly over land.
The surface albedo has decreased, especially
during the autumn months. Cloud amount has been
decreasing during the winter but increasing in
spring and summer. If Arctic cloud cover had
not been changing this way, surface temperatures
would probably have risen at an even greater rate
than what has been observed. Decreases in sea
ice extent and albedo from surface warming
modulate the increasing cloud cooling effect,
resulting in little or no change in the radiation
budget. The decreasing trend in winter surface
temperature over the central Arctic cannot be
explained solely by large-scale atmospheric
circulation changes. Changes in summer albedo
over Alaska correlate with a lengthening of the
snow-free season that has increased atmospheric
heating locally by 3 W/m-2/decade. Current trends
in shrub and tree expansion could further amplify
this by 2-7 times.
34
Satellite Climatology Major Findings, cont.
There is a strong coupling between changes in
surface temperature and changes in inversion
strength, but trends in some areas may be a
result of advection aloft rather than
warming/cooling at the surface. The location
and strength of the polar vortex has changed over
the past two decades, affecting the movement of
heat and moisture into/out of the Arctic.
Wintertime moisture convergence has decreased
significantly over the Nansen Basin and parts of
the Barents and Kara Seas. The reduction in
moisture convergence results in a decrease in
cloud formation due to weakening cyclonic
activity in that region. Reduced cloud amount
over that area leads to decreased cloud cover
over the entire central Arctic because less cloud
is advected to other regions. Cloud cover
variability near the sea ice margins is strongly
linked to sea ice variability. Sea ice retreat
is linked to a decrease in low level cloud amount
and a simultaneous increase in middle-level
clouds. Changes in cloud cover can be explained
by a decrease in static stability and a deepening
of the atmospheric boundary layer resulting from
the removal of ice and the increase in surface
temperature. A method was developed to assess
the influence of changes in Arctic cloud cover on
the surface temperature trend, allowing for a
more robust diagnosis of causes for surface
warming or cooling. The relative contribution of
trends in cloud cover to trends in the all-sky
surface temperature varies by season. Our
newly-compiled 20 year record of tropospheric
winds from AVHRR has been used to identify
deficiencies in the polar wind field in one of
the major reanalysis products (ERA-40). We have
shown where, meteorologically, the largest
differences between the satellite and ERA-40
winds occur.