Geology Reference
In-Depth Information
prediction of climate change and its impacts. Climate
models have underestimated the currently observed
thinning of the Arctic sea ice by a factor of 4 on the aver-
age and failed to predict the decrease of its extent [
Rampal
et al
., 2011]. Sea ice albedo measurements on a grid scale
suitable for climate models can be obtained from the
operational gridded albedo products of satellite data
(currently available through NASA from MODIS data
and the European OSI‐SAF products using AVHRR).
Implementing an accurate spatiotemporal sea ice albedo
estimate is also a key factor for successful modeling of
the sea ice state in the polar regions. This should result
in accurate forecasting of the ice extent, concentration,
thickness, motion, and melt pond coverage. Currently,
models usually assume a constant albedo for sea ice
(a very rough approximation) or use climatological
albedo compiled from numerous measurements of
satellite remote sensing. Albedo is usually parameterized
in terms of surface temperature and snow cover fraction
[
Barry
, 1996]. However, it is strongly affected by several
other factors. A need for compiling a database of sea ice
albedo has been identified to fill information gap required
to advance climate models. A comprehensive and instruc-
tive review of optical properties of sea ice and snow is
presented in
Pedersen
[2007]. The work presents physically
based parameterization for albedo from snow‐covered
sea ice. It addresses the sea ice and snow parameters that
affect the albedo and identifies those that need to be
included in modeling the albedo.
Albedo from snow‐covered sea ice varies according to
a few ice and snow parameters. Ice parameters include
age, thickness, surface salinity, subsurface porosity, pres-
ence of frost flowers, and surface melt. Snow parameters
include snow depth, wetness, and metamorphism
[
Warren et al
., 1997]. To some extent, albedo is also
affected by the presence of biogenic material such as
algae as well as sediments that are typically entrained
during ice formation over relatively shallow water. These
sediments cause a reduction in surface albedo [
Light
et al
., 1998].
Albedo from sea ice demonstrates spectral behavior
similar to albedo from snow. It is high at the visible
spectral range (0.4-0.7
μ
m of wavelength) and decreases
rapidly in the NIR until it levels off to a minimum value
throughout the rest of the NIR spectrum. Visible radia-
tion with its shorter wavelength is capable of penetrating
the ice surface to a limited extent. Therefore, visible albedo
becomes sensitive to the composition and microstructure
of ice in the top layer. Visible albedo of bare sea ice has
been shown to vary the most during early stages of ice
formation. Snow‐covered albedo also varies considerably
during cold and warm spells of atmospheric temperatures
because significant changes in snow and ice composition
may take place during these periods.
On the other hand, albedo in the NIR spectral region is
sensitive only to the physical characteristics of the surface
because the attenuation of incident radiation increases
rapidly at infrared wavelengths. From the physical pro-
cesses point of view, differences in the magnitude of sea
ice broadband albedo result mainly from differences in
scattering, while variations in spectral albedo are mainly
due to different absorption at different wavelengths. A
review of albedo data from sea ice is presented in
De
Abreu
[1996] and another review of the radiative transfer
models of sea ice albedo is presented in
Perovich
[1996].
Useful discussions on optical properties of snow and ice,
in general, and albedo in particular are presented in
Lubin
and Massom
[2006]. A review of snow optical properties
is presented in
Dozier
[1989] and
Mobley et al
. [1998]. In
the rest of this section, samples of sea ice albedo data
from selected publications are presented to address the
aforementioned factors that influence this parameter.
It would be appropriate to start this quick review by
presenting monthly averaged data of the incoming and
outgoing radiation over the sea ice cover in the Arctic,
without taking into consideration the details of ice condi-
tions. It is worth repeating that shortwave radiation occu-
pies the optical range between 0.2 and 3.0
μ
m while
longwave radiation occupies the thermal IR range. The
monthly average incoming and outgoing radiation as well
as shortwave albedo data from Arctic ice are presented in
Maykut
[1986] and included in Table 8.7. The shortwave
albedo is the ratio of the outgoing to the incoming short-
wave. The data assert that the incoming solar shortwave
and longwave radiation peak in the summer. While the
longwave radiation continues in winter, the shortwave is
negligible during most of the fall and winter. The annual
total of the incoming longwave radiation is more than
double that of the shortwave.
This is because of the significantly higher longwave incom-
ing radiation compared to the shortwave in the summer. The
highest shortwave albedo occurs in the spring (March-May)
when the ice surface is dry and most likely covered with dry
snow as well. The lowest values occur in summer when snow
and perhaps ice surface melts, forming low‐albedo melt
ponds. The estimate of each albedo value shown in the table
is an average of measurements that comprise a wide range of
ice growth stages with a variety of surface conditions. Note
the positive net radiation during the summer months (May-
September) with a significant peak in July and the negative
values at all other times.
Albedo can be determined using one of the following
three approaches: field measurements, physical‐based mod-
els, or satellite observations. Each approach has its advan-
tages and limitations. The third approach has been most
commonly used lately to map albedo at regional and global
scale. Each approach is addressed separately in the rest of
this section. Early data were mostly obtained from field
