M.R. DRINKWATER
8.6 Algorithm Issues................................................................................ 176
8.6.1 Antarctic Ice-Motion Tracking Performance ........................................ 176
8.6.1.1 SAR Motion Tracking Validation....................................................... 176
8.6.1.2 Time-Series Observations of Ice Floes...................... .......................... 178
8.6.2 Temporal Changes in Ice Characteristics.............................................. 180
8.6.2.1 Summer and Winter SAR Backscatter Characteristics.......................... 182
8.6.2.2 Time-Series Observations of Backscatter........................................... 182
8.6.2.3 Meltponding .................................................................................. 183
8.7 Conclusions........................................................................................ 183
References............................................................................................... 185
8.1
Introduction
Historical data on Antarctic sea ice extent and concentration have traditionally been
derived from visible and near-infrared images acquired by the polar-orbiting
National Oceanic and Atmospheric Agency's (NOAA) meteorological satellites, using
the Advanced Very High Resolution Radiometer (AVHRR), and more recently by the
Defense Meteorological Satellite Program's Operational Linescan System (OLS)
(Mass om 1991). The limitation of these systems is that the majority of energy imparted to the Antarctic sea-ice system is transferred during frequent episodic storm
bursts, caused by fast-moving low pressure systems (McPhee et al. 1996). Since the
Southern Ocean sea-ice cover is completely bounded at its lower latitude limit by open
ocean, these "polar lows" transport large amounts of moisture (contained in warm
air masses) over the outer ice cover. The result is that most, if not all, noteworthy periods of wind- and temperature-driven dynamic changes in the ice cover are accompanied by periods where the region is blanketed by cloud, and when the atmosphere
is inherently more electromagnetically opaque. During storms, the probability with
which the area is cloud covered is extremely high, thereby ruling out use of visible or
near-infrared images as a practical method of monitoring the associated changes in
ice conditions. Instead, Nimbus-7 Scanning Multichannel Microwave Radiometer
(SMMR) and DMSP Special Sensor Microwave/Imager (SSMII) have been the primary
workhorses to build up a microwave record of Antarctic sea-ice characteristics. Similar problems, however, occur in passive microwave retrievals of sea-ice concentration, and the algorithms are called into question during these periods of change. Oelke
(1997) discovers that the influence of water vapor in the atmosphere alone can modify the ice concentration retrievals by fractions exceeding 10%, and that retrievals of
ice concentration must compensate for the atmospheric water vapor and liquid water
contents.
Prior to 1991, the only active microwave radar data to be acquired in the Southern
Ocean were collected during tlIe Shuttle Imaging Radar-B (SIR-B) mission of the space
shuttle in 1984 (Carsey et al. 1986; Martin et al. 1987). Seasat in 1978 acquired no synthetic aperture radar (SAR) data due to the lack of a direct broadcast receiving station
in Antarctica, and due to the brevity of its lifetime (lasting only 3 months long before
a fatal power failure). Despite the brief Arctic success of Seasat, a period of almost 15
years passed before the second radar remote sensing satellite mission was flown. For
8.6 Algorithm Issues................................................................................ 176
8.6.1 Antarctic Ice-Motion Tracking Performance ........................................ 176
8.6.1.1 SAR Motion Tracking Validation....................................................... 176
8.6.1.2 Time-Series Observations of Ice Floes...................... .......................... 178
8.6.2 Temporal Changes in Ice Characteristics.............................................. 180
8.6.2.1 Summer and Winter SAR Backscatter Characteristics.......................... 182
8.6.2.2 Time-Series Observations of Backscatter........................................... 182
8.6.2.3 Meltponding .................................................................................. 183
8.7 Conclusions........................................................................................ 183
References............................................................................................... 185
8.1
Introduction
Historical data on Antarctic sea ice extent and concentration have traditionally been
derived from visible and near-infrared images acquired by the polar-orbiting
National Oceanic and Atmospheric Agency's (NOAA) meteorological satellites, using
the Advanced Very High Resolution Radiometer (AVHRR), and more recently by the
Defense Meteorological Satellite Program's Operational Linescan System (OLS)
(Mass om 1991). The limitation of these systems is that the majority of energy imparted to the Antarctic sea-ice system is transferred during frequent episodic storm
bursts, caused by fast-moving low pressure systems (McPhee et al. 1996). Since the
Southern Ocean sea-ice cover is completely bounded at its lower latitude limit by open
ocean, these "polar lows" transport large amounts of moisture (contained in warm
air masses) over the outer ice cover. The result is that most, if not all, noteworthy periods of wind- and temperature-driven dynamic changes in the ice cover are accompanied by periods where the region is blanketed by cloud, and when the atmosphere
is inherently more electromagnetically opaque. During storms, the probability with
which the area is cloud covered is extremely high, thereby ruling out use of visible or
near-infrared images as a practical method of monitoring the associated changes in
ice conditions. Instead, Nimbus-7 Scanning Multichannel Microwave Radiometer
(SMMR) and DMSP Special Sensor Microwave/Imager (SSMII) have been the primary
workhorses to build up a microwave record of Antarctic sea-ice characteristics. Similar problems, however, occur in passive microwave retrievals of sea-ice concentration, and the algorithms are called into question during these periods of change. Oelke
(1997) discovers that the influence of water vapor in the atmosphere alone can modify the ice concentration retrievals by fractions exceeding 10%, and that retrievals of
ice concentration must compensate for the atmospheric water vapor and liquid water
contents.
Prior to 1991, the only active microwave radar data to be acquired in the Southern
Ocean were collected during tlIe Shuttle Imaging Radar-B (SIR-B) mission of the space
shuttle in 1984 (Carsey et al. 1986; Martin et al. 1987). Seasat in 1978 acquired no synthetic aperture radar (SAR) data due to the lack of a direct broadcast receiving station
in Antarctica, and due to the brevity of its lifetime (lasting only 3 months long before
a fatal power failure). Despite the brief Arctic success of Seasat, a period of almost 15
years passed before the second radar remote sensing satellite mission was flown. For