8 Satellite Microwave Radar Observations of Antarctic Sea Ice
hemisphere ground stations the Alaska SAR Facility, Fairbanks (and Gatineau, Canada; Kiruna, Sweden; and Troms0, Norway). Together these SAR receiving stations
enable the majority of the northern-hemisphere sea-ice cover to be mapped using SAR,
especially at the latitudinal limit of the orbit where the SAR swaths converge. Contrastingly less coverage is possible of the more extensive Antarctica sea-ice cover due
to the latitudinal extent of the sea ice and the limited operating times of GARS and
Syowa. The result is a high-resolution SAR image database which is at best discontinuous in space and time. Alternative methods have therefore been sought using other
forms of EScat data, in order to fill the temporal active microwave record in space and
time.
8.3.2
Filling the Gaps
The key advantage of the ERS scatterometer is that it operates whenever the SAR is
switched off, continuously retrieving information from the Weddell Sea without the
necessity of a local receiving station. Its wider swath provides more frequent and
broader incidence angle (20° :::; e:::; 60°) coverage in a given location. This low-bit-rate
(LBR) data source is essential to fill in areas of sparse temporal and spatial SAR coverage of the Antarctic ice cover. Until now, the intrinsic low resolution of the raw measurement data is the main reason why scatterometer data have not been used in such
applications. Nevertheless, the approach described by Drinkwater et al. (1993a) produces
weekly average backscatter maps of the entire Southern Ocean sea-ice cover at a resolution higher (approximately 20 km) than present conventional alternatives such as the
SSM/I passive microwave radiometers. EScat images form an uninterrupted sequence
of vv-polarized C-band backscatter (aO EScat) of Antarctic sea ice from the beginning of
the ERS-l mission. Since these backscatter images are based on multiple azimuth and
incidence angle observations, these data provide greater capability for discriminating
ice types and separating of ice from ocean. Coupled with higher-frequency passive
microwave and finer resolution SAR, these data enable a better understanding of the
time sequence and seasonal history of Antarctic sea-ice formation, drift, deformation
and decay. Furthermore, medium-scale-resolution EScat images essentially provide the
large-scale context within which the high resolution SAR images may be interpreted.
8.4
Ice Classification Issues
Recognition of Antarctic sea-ice types in C-band microwave images, as a proxy for ice
thickness, is limited by the lack of a historical sea-ice signature database equivalent to
that existing for the Arctic (Drinkwater 1995a). Only since ERS-l has flown have surface experiments focused on the acquisition of a microwave database for comparison
with satellite measurements. These provide the foundation for preliminary attempts to
recognize and categorize ice types observed in SAR and scatterometer images. Due to
the logistical constraints of working in accessible regions, most of these measurements
have been focused in the Weddell Sea. Results from these demonstrate that Antarctic
ice appears considerably different from its Arctic counterpart, due in part to its relatively higher salinity and younger mean age and thickness (Drinkwater et al. 1993a).
hemisphere ground stations the Alaska SAR Facility, Fairbanks (and Gatineau, Canada; Kiruna, Sweden; and Troms0, Norway). Together these SAR receiving stations
enable the majority of the northern-hemisphere sea-ice cover to be mapped using SAR,
especially at the latitudinal limit of the orbit where the SAR swaths converge. Contrastingly less coverage is possible of the more extensive Antarctica sea-ice cover due
to the latitudinal extent of the sea ice and the limited operating times of GARS and
Syowa. The result is a high-resolution SAR image database which is at best discontinuous in space and time. Alternative methods have therefore been sought using other
forms of EScat data, in order to fill the temporal active microwave record in space and
time.
8.3.2
Filling the Gaps
The key advantage of the ERS scatterometer is that it operates whenever the SAR is
switched off, continuously retrieving information from the Weddell Sea without the
necessity of a local receiving station. Its wider swath provides more frequent and
broader incidence angle (20° :::; e:::; 60°) coverage in a given location. This low-bit-rate
(LBR) data source is essential to fill in areas of sparse temporal and spatial SAR coverage of the Antarctic ice cover. Until now, the intrinsic low resolution of the raw measurement data is the main reason why scatterometer data have not been used in such
applications. Nevertheless, the approach described by Drinkwater et al. (1993a) produces
weekly average backscatter maps of the entire Southern Ocean sea-ice cover at a resolution higher (approximately 20 km) than present conventional alternatives such as the
SSM/I passive microwave radiometers. EScat images form an uninterrupted sequence
of vv-polarized C-band backscatter (aO EScat) of Antarctic sea ice from the beginning of
the ERS-l mission. Since these backscatter images are based on multiple azimuth and
incidence angle observations, these data provide greater capability for discriminating
ice types and separating of ice from ocean. Coupled with higher-frequency passive
microwave and finer resolution SAR, these data enable a better understanding of the
time sequence and seasonal history of Antarctic sea-ice formation, drift, deformation
and decay. Furthermore, medium-scale-resolution EScat images essentially provide the
large-scale context within which the high resolution SAR images may be interpreted.
8.4
Ice Classification Issues
Recognition of Antarctic sea-ice types in C-band microwave images, as a proxy for ice
thickness, is limited by the lack of a historical sea-ice signature database equivalent to
that existing for the Arctic (Drinkwater 1995a). Only since ERS-l has flown have surface experiments focused on the acquisition of a microwave database for comparison
with satellite measurements. These provide the foundation for preliminary attempts to
recognize and categorize ice types observed in SAR and scatterometer images. Due to
the logistical constraints of working in accessible regions, most of these measurements
have been focused in the Weddell Sea. Results from these demonstrate that Antarctic
ice appears considerably different from its Arctic counterpart, due in part to its relatively higher salinity and younger mean age and thickness (Drinkwater et al. 1993a).