Introduction - Sea Ice: Physics and Remote Sensing - page 22

Preface of the notable topic on Antarctic sea ice [ Jeffries ,
1998], the editor refers to a few statements made by Lyn
Lewis and Willy Weeks in a report published in 1971 by
the Scientific Committee on Antarctic Research. One
statement underlines the thought that prevailed up to
the 1950s that Antarctic sea ice is probably not any dif-
ferent than Arctic ice, so why bother studying it? But
realizing that this may not be true to some extent, the
authors of the report added: “It is clear that future work
will depend critically on the logistics available to allow
surface observations beyond the fast ice edge at all sea-
sons of the year. Of almost equal importance will be the
development of instruments and recording equipment
suited for use in the polar environment” ( Lewis and
Weeks , 1971). The identification of that need was made
one year before the launch of the first U.S. space‐borne
microwave instrument for Earth observations; the
Electrically Scanning Microwave Radiometer (ESMR).
The foresight of the authors was true. ESMR was a
milestone instrument for sea ice observations in the
polar regions, following the success of previous satellites
that used optical sensors. A brief synopsis of satellite
remote sensing for sea ice applications is presented in
section 7.2.
Satellite data of sea ice represent one of the longest
Earth observation records from space. Since the early
1970s many different satellite Earth observation sensors
were developed and used effectively in routine surveil-
lance of sea ice. Sensors operate in different bands of
the electromagnetic spectrum: optical, infrared, and
microwave (passive and active) (see section 7. 1).
Information from these bands complement each other.
Different sensors provide information at different
scales. The basic premise of sea ice parameter retrieval
from remote sensing data is the significant difference
between physical and radiometric properties of sea ice
and open water; namely physical temperature, reflectiv-
ity, and microwave emissivity. Optical sensors discrimi-
nate between sea ice and open-water based on their
contrast in albedo. Thermal infrared sensors use the
difference in the physical temperatures (ice surface is
usually colder than water surface). Passive microwave
sensors use the difference between ice and water micro-
wave emissivity. Radar sensors use the difference
between backscatter from the ice and the water surface,
although they both occupy a wide range of values and
they overlap.
Many sea ice parameter retrieval algorithms have
been developed using observations from different sen-
sors. Examples include surface temperature, retrieved
from thermal infrared sensors; ice concentration and
extent from passive microwave; ice types from passive
and active microwave; and surface features, ice drift,
and deformation from radar imagery data. The two

most commonly used categories of sensors are active
microwave (radar), which produces imagery data at
fine resolutions of a few tens of meters, and passive
microwave, which produces images at a coarse resolu-
tion of a few kilometer or tens of kilometer. Radar
imagery generates information at tactical scale (a few
tens or hundreds of kilometers of an imaged scene)
primarily to support ship navigation though it has
been used also to produce synoptic views of ice motion
and deformation in the polar regions (section 10.7).
Passive microwave generates information at a synoptic
scale (a few thousands of kilometers of an imaged
scene), although it has been used to produce ice con-
centration maps at medium‐resolution scales of a few
kilometers.
Some of these algorithms have been used to generate
useful records of ice concentration, extend, surface tem-
perature, thickness (with some limitations), motion, and
deformation. The information is used to support marine
operational tasks as well as climate‐related studies.
However, the use of remote sensing data in operational
sea ice monitoring programs still depends on visual anal-
ysis of satellite imagery data. The reason is the strict
requirement of the operational analysis for robustness.
Algorithms may not be reliable under all possible ice con-
ditions, but the operational environment requires nothing
less than reliability. Visual analysis of the data is certainly
subjective, but it incorporates many factors that cannot
be incorporated in a quantitative algorithm (e.g., climatic
information, recent history of the ice field, records of
meteorological data, in addition to other heuristic rules
used by expert ice analysts). The approach of visual anal-
ysis of satellite images to retrieve operational sea ice
information will probably continue for many years until a
coincident multisensor data system is developed to
acquire multichannel optical, thermal, and microwave
data simultaneously.
In addition to the valuable information offered by
remote sensing about sea ice, another unnoticeable
positive development has been brought by this tech-
nology to the sea ice community. That is the engage-
ment of many researchers and operators who come
from the remote sensing systems and applications in
the sea ice properties and physical processes. Many
researchers who spend their career in developing hard-
ware instruments or software methodologies for
remote sensing data collection and analysis find them-
selves, at one point or another, involved in testing or
validating their ideas using sea ice. Some of those peo-
ple have later developed interest in sea ice as a phe-
nomenon and pursued serious research work in this
field. This happened to scientists who came from the
fields of electrical or civil engineering, mathematics,
earth sciences, physics or geography.