Geology Reference
In-Depth Information
Table 10.3
Rate of decrease of ice area and extent in the Arctic presented in km
2
per decade during two periods (1979-2012)
and (1992-2001) using two categories of passive microwave algorithms: low‐frequency algorithms (LFA) and high‐frequency
algorithms (HFA).
1979-2012
1992-2012
Area/Extent
LFA
AA
LFA
HFA
AA
Ice area (km
2
× 10
6
)
0.534-0.573
0.560
0.866-0.975
0.766-0.978
0.903
Ice extend (km
2
× 10
6
)
0.439-0.491
0.470
0.767-0.812
0.758-0.814
0.789
Note
: All algorithms are denoted (AA). Data are adapted from
Ivanova et al.
[2014].
index (NDSI), defined as the difference between reflec-
tance from the visible and near‐infrared channels divided
by their summation:
information on thin ice because it has a major impact on
heat flux from the ocean to the atmosphere. Moreover,
thermal and radiative properties of thin ice change sig-
nificantly in response to meteorological forcing. The
sensible heat flux in thin ice is one or two orders of mag-
nitude larger than that in thick ice [
Maykut,
1978]. Heat
flux through thin ice is proportional to the ice thickness.
This makes accurate estimates of this parameter crucial
for the calculation of heat flux in winter and the energy
balance at the ice surface throughout the year. Due to
the lack of information about ice thickness, the energy
balance processes at the ice surface are usually simpli-
fied in most climate models.
The spatial scale at which ice thickness information is
required varies according to the application. Monitoring
ice thickness at a tactical scale (a few hundreds of meters
or a few kilometers) is important for the safety of marine
navigation, while monitoring ice thickness at a synoptic
scale (tens of kilometers) would be sufficient to improve
the accuracy of the global and regional weather and cli-
mate models. Ice thickness is also needed to compute the
total ice volume and its mass balance in the polar regions.
Estimation of ice thickness from remote sensing obser-
vations is a challenging task. Space‐borne or airborne
remote sensors can “sense” only the emitted or scattered
radiation from the top few millimeters or centimeters of
the ice cover, depending on the wavelength. Therefore,
observations do not usually carry information about the
entire ice thickness. Advances in remote sensing tools and
analysis techniques have allowed better estimates of ice
thickness, indirectly using one of the following three
approaches: (1) Heat and energy balance equations, which
are used with TIR observations; (2) empirical equations
that relate ice thickness and radiometric observations,
which is used with microwave observations; and (3) the
buoyancy law that allows inference of ice thickness from
its freeboard measurements, which is used with altimeter
data. The three approaches are presented in the rest of
this section, but it would be useful to begin with a sum-
mary of the advantages and limitations of each approach.
Band
4
Band
6
NSDI
6
(10.62)
Band
4
Band
The bandwidth of band 4 (VIS) is 545-565 nm and band
6 (NIR) is 1628-1652 nm. The NSDI is used to ensure the
detection of the snow‐covered ice because it has high val-
ues due to the high and low reflectance in the VIS and the
NIR, respectively. In winter, snow‐covered ice is identi-
fied if NDSI ≥ 0.4, the reflectance in band 2 ≥ 0.11, and in
band 4 ≥ 0.10 (band 2 is also in the NIR range with wave-
length 841-876 nm).
The third criterion, which is used to detect sea ice area
during the dark season in the polar regions, is the ice sur-
face temperature. The parameter is calculated using the
version of the split window technique given in equation
(7.55). The coefficients for this equation were generated
using a multilinear regression of surface temperature
estimates from equation (7.27) (using the brightness tem-
perature observations from the MODIS 11
μ
m channel
with an appropriate value of surface emissivity) and the
observed brightness temperature from MAS. A pixel is
identified as being covered with sea ice if the calculated
surface temperature is ≤ 271.4 K. Ice extent maps are
standard products from MODIS.
10.4. Ice ThIckness
Sea ice thickness estimation is important for both
marine operations and climatic studies. Marine opera-
tors are more concerned with thick ice, which consti-
tutes navigational hazards. Therefore, national ice
monitoring services usually make a special effort to pro-
duce accurate information about thick seasonal and
perennial ice. If in doubt, the analyst would assign
thicker or older ice type in order to make the estimate
conservative. Thin ice types (i.e., YI < 30 cm thick) do
not represent obstacles to most ship classes; therefore,
accurate estimates of thin ice thickness is not as crucial
for this purpose as estimates of thick ice. On the other
hand, climate and weather models require detailed
1. Heat and energy balance equations can be used to
estimate ice thickness up to 50 cm from TIR sensors.
Other limitations include the condition of clear sky, which
