The MERS research group has been actively involved in polar research
in the southern hemisphere. As part of these studies Seasat scatterometer
(SASS), ERS-1/2 scatterometer (Escat), NASA Scatterometer (NSCAT),
and SeaWinds data have been processed with the Scatterometer Image
Reconstruction (SIR) resolution enhancement algorithm.
While satellite microwave radar scatterometers were originally
designed to measure winds over the ocean from space, they can also
be useful in polar ice studies. Using the Scatterometer Image Reconstruction
(SIR) resolution enhancement algorithm we have generated a time
series of radar images of the Antarctic region. New methods for
mapping the spatial extent and ice type have been developed for
use with NASA Scatterometer (NSCAT) and SeaWinds data. The results
enable study of the dynamics of the sea-ice sheet and permit multidecadal
studies of change by comparison with previous scatterometers. The
broad coverage, dual-pol measurements, and high resolution of NSCAT
and SeaWinds yield very high quality images.
Microwave radar mitigates the need for optimal meteorological
conditions and solar illumination which can hamper optical sensors.
The radar scattering signal provides insight into characteristics
of the ice which can not be inferred from optical images. In particular,
the microwave images depend on both surface roughness and the electrical
properties, which vary for different ice types, enabling the retrieval
of ice characteristics from the radar data. Scatterometers observe
the ice from a variety of incidence angles, further enhancing the
utility of the radar data. Knowledge of the ice characteristics
is of crucial import in modeling the interaction of the ocean and
atmosphere in the polar regions and in evaluating the Earth's heat
balance. Sea-ice cover also influences the production of Antarctic
Bottom Water, a crucial factor in global ocean circulation.
SeaWinds Antarctic Studies
Unlike previous fan-beam instruments, SeaWinds makes sigma-0 measurements
at only two discrete incidence angles, each with a different polarization.
SeaWinds make measurements at a greater variety of azimuth angles
than previous sensors. A variety of studies
are underway. An operational automated algorithm for SeaWinds ice
extent mapping has been generated
NSCAT Antarctic Studies
NSCAT makes measurements at both vertical and horizontal electrical
polarization at a variety of azimuth angles. Sea-ice has a nearly
isotropic response while the ocean exhibits a very directional response
which is ordinarily used to determine the direction of the wind
blowing over the surface. While the sea-ice evolves on time scales
of days to weeks, winds over the ocean can change on hourly time
scales. Using images of the vertical and horizontal radar response
and the temporal variation in the radar response over a several
day period, the spatial extent of the sea-ice can be mapped. The
resulting ice edge compares favorably with passive microwave ice
maps but has higher spatial resolution and precision (Remund and
Time Series (95K gif) shows the seasonal recession and growth
of the sea-ice around Antarctica. A sample 256x256 pixel
NSCAT Movie (900K animated gif) was generated from the orignal
1940x1940 SIRF images. The images illustrated here show the incidence
angle normalized radar backscatter (denoted `A'). The original Ku-band
images have approximately 4.5 km resolution with images every three
days. The images have been masked to show only sea-ice and the Antarctic
continent. Masking was done using only NSCAT data and a newly developed
algorithm. The ice edge corresponds closely with the SSM/I-derived
NSIDC 30% ice concentration contour. Longer version of this movie
are available as (animated
gif 4.4 MB), (.avi
4.4 MB), (.mov
[quiktime] 1.6 MB), or (.m1v
[mpeg bitstream] 2.2 MB).
Images (290K gif) shows fine resoluione NSCAT A and B images
of both the Arctic and Antarctic. Sea-ice edge masking for these
particular images was done using SSM/I data though the NSCAT-only
algorithm works well in both hemispheres.
In this sample time series and movie, the Antarctic continent
is clearly visible in the center of the images and changes relatively
little (except along the coast) while the sea-ice sheet dramatically
moves. A large (100x60 km) white iceburg which broke of the Thwaites
ice tongue is visible on image left. (It appears to vanish when
the ice-sheet around it melts and the algorithm to remove the ocean
part of the image also removes the iceburg--the massive burg is
still there, however.) Movies of ice-masked enhanced resolution
NSCAT Arctic images are available as (animated
gif 8.8 MB), (.avi
4.3 MB), (.mov
[quiktime] 1.0 MB), or (.m1v
[mpeg bitstream] 1.1 MB).
Multisensor studies involving SSM/I, NSCAT, and ERS-1/2 scatterometer
data suggest that the microwave response exhibits a wind-dependent
signature which enables evaluation of the wind direction (Long and
Drinkwater, 2000). A pdf
version of this paper is available for download. Because of
the small sized figure reproductions in the printed paper, full-size
electronic versions of the figures are provided (these are copyrighted).
ERS-1/2 Antarctic Studies
While NSCAT has wider, more frequent coverage, better resolution,
and dual-polarization capability, data from the ERS-1 and ERS-2
AMI scatterometers (EScat) is also use for ice sheet studies because
of its long time history. We have made time series images of EScat
data and done studies on azimuth modulation of C-band data over
Recent evidence for warming in the vicinity of the Antarctic Peninsula
has raised concerns about the stability of the large ice shelves
in that region. The recent past has seen calvings resulting in the
disintegration of the King George and the Larsen Ice Shelves. Time
series information provided by the EScat instrument shows the value
for long-term monitoring of such delicate regions. An
ERS-1 Antarctic Peninsula time series (152K) shows the inter-annual
variability in melt, clearly visible as the brightness modulation.
Melt is recognized as the lowest backscatter coefficients, and a
distinct melt front is visible in 1992. In 1993 the entire Larsen
ice shelf is melting, while the subsequent 1994 melt season has
minimal melting. The following melt seasons in 1995 and 1996 show
similar characteristics to 1992 and 1993, respectively. These data
indicate that the summers of 1993 and 1995 were both extensive in
time and space, perhaps contributing to the large iceberg calvings
in those years. These images are being used to chart the regional
distribution of seasonal melting on the ice shelves and the continental
An early global
ERS-1 Colorized time series (275K)
An early ERS-1
Weddell Sea Image (255K)
See the paper "Greenland Observed" below.
or Greenland Studies
The SIR Algorithm
The Scatterometer Image Reconstruction (SIR) algorithm was developed
to generate enhanced resolution scatterometer imagry from raw sigma-0
For fan-beam scatterometers SIR generates images of 'A' and 'B'
which are related to the normalized radar cross section sigma-0
where theta is the incidence angle of the measurements. For SeaWinds
only 'A' images are produced. SIR has also been applied to SSM/I
data. The resolution of the resulting images is as fine as ~4-5
km for SASS and NSCAT, ~ 2-6 km for SeaWinds, and 25-30 km for ERS-1/2.
Previous techniques were limited to the intrinsic resolution (typically
50 km) of the scatterometer.
Some Related Papers from the MERS Group
Variation in Microwave Scatterometer and Radiometer
Data Over Antarctica
D.G. Long and M.R. Drinkwater, IEEE Trans. Geosc. Remote Sens.,
Vol. 38, No. 4, pp. 1857-1870, Aug. 2000.
Extent Mapping Using Ku-Band Scatterometer
Q.P. Remund and D.G. Long, Journal of Geophysical Research,
Vol. 104, No. C5, pp. 11515-11527, 2000.
at High Resolution by the Seasat-A Scatterometer
D.G. Long and M.R. Drinkwater, Journal of Glaciology, Vol.
32, No. 2, pp. 213-220, 1994.
of Spaceborne Scatterometer Data
D.G. Long, P. Hardin and P. Whiting, IEEE Trans. Geosc. Remote
Sens., Vol. 31, No. 3, pp. 700-715, May 1993.
For a more extensive list see the MERS Bibliography