Geology Reference
In-Depth Information
0.9
0.8
2008
0.8
0.7
2008
2009
2010
2009
2010
0.7
0.6
0.6
0.5
0.5
0.4
0.4
0.3
0.3
0.2
0.2
0.1
0.1
0.0
0.0
80
0
50
60
70
-10
-5
0
5
10
15
20
25
10
20
30
40
Days from onset of ponding
Pond coverage (%)
Figure 9.14
(a) Evolution of albedo from surface measurements around and after onset of melting during, and
(b) albedo versus percentage of melt pond coverage. The solid line represents the linear regression of the data.
Data obtained from 3 years of observations of Arctic sea ice near Barrow, Alaska [
Polashenski et al
. 2012., Fig. 21,
with permission from AGU]. (For color detail, please see color plate section).
and Perovich
, 2004].
Polashenski et al
. [2012] conducted
a series of observations on melting FY and land‐fast
Arctic ice near Barrow, Alaska, in 2008, 2009, and 2010
to explore the seasonal evolution of melt pond coverage
and its expected effect on decreasing the surface albedo.
The albedo was measured along transect lines every
2.5 m using an albedometer that integrates albedo over
wavelengths between 300 and 3000 nm. Spatially aver-
aged albedo values calculated from the measurements
are plotted in Figure 9.14. The snow‐covered ice surface
shows stable albedo around 0.75 in winter. A few days
before the onset of ponding albedo starts to drop
sharply. A few days later it reaches a minimum around
0.25 when the pond coverage reaches maximum. Albedo
starts to increase when meltwater begins to percolate
through connective porosity in the ice, exposing a nearly
dry ice surface. Greater variability of albedo is shown
during this period. In 2 weeks the albedo stabilizes
around values between 0.5 and 0.6. The figure shows
also the correlation between the percentage of the pond
coverage and the measured albedo. It is obvious that
the pond coverage is the primary driver of albedo
decrease.
Daily melt pond cover over sea ice in the Beaufort/
Chukchi Sea region was detected using measurements
of surface reflectance from MODIS through the sum-
mer of 2004 [
Tschudi et al
., 2008]. Results show a rapid
increase of ponding areas from 10% to 40% of the total
ice surface during the first 20 days of melt, followed by
fluctuations through the summer. Toward the end of the
summer, a gradual decrease occurs, with the ponding
area reaching 10% in late August. Since MODIS
observations are usually obtained from heterogeneous
footprints, the author used the well‐known linear decom-
position equation to decompose the observation into
components from ice and melt ponds. The daily fraction
of melt pond coverage can then be obtained by solving
a set of such equations:
R
a r
,
a
1
(9.14)
k
i k
i
i k
where
R
k
is the observed surface reflectance from MODIS
channel
k
,
r
ik
represents the typical spectral reflectance
from the surface
i
by channel
k
, and
a
i
is the fractional
coverage of each surface. Four surface types were selected:
melt ponds, open water, snow‐covered ice, and white ice.
The values of
r
ik
were obtained from in situ measurements
of reflectance in June 2004 near Barrow, Alaska. Three
MODIS bands were used: band 1 (620-670 nm), band 2
(841-876 nm), and band 3 (459-479 nm). The observed
reflectance was obtained from the MODIS product
MODO9 [
Vermote et al
., 2002], which accounted for
atmospheric water vapor and the bidirectional reflectance
distribution. Equation (9.14) is solved for the fractional
coverage of each surface. An example of the daily melt
pond fraction is shown in Figure 9.15. The pond fraction
is higher near the coast where the pack ice is separated
from the land‐fast ice and lower inside the pack. The
study used high‐resolution digital imagery acquired by
an aerial aerosonde unmanned vehicle (UAV) to validate
the results.
