122
L.M. Mitnik and V.A. Dubina
Fig. 7.7 Spiral ice eddy in
the Southern Okhotsk Sea on
Envisat ASAR image taken at
00:30 UTC (a) and on Terra
MODIS visible image taken
at 00:50 UTC (b) on 8
February 2003; 1 marks the
ice-free area
SAR, visible and passive microwave data decreases the ambiguity of interpretation.
This is evident from a comparison of the SAR images with the AMSR-E T B s and
MODIS visible images. Both eddies have higher T B s values at 89 GHz in comparison with the open sea as seen in (Fig. 7.6b, d). The spatial resolution of T B s data is
4 km × 6 km. In the visible image, having a spatial resolution of 250 m × 250 m,
the open sea has a dark tone even at high winds (Fig. 7.7b) as opposite to SAR
images where the increase in wind speed causes the increase in radar backscatter
(Fig. 7.7a). The brightness of sea ice depends on its thickness (type) and also on the
presence of snow. So, grease ice and dark nilas (ice which is under 5 cm in thickness
and is very dark in colour) have very low brightness contrast, as opposite for example to grey-white or white ice. The concurrent analysis of SAR and visible images
(Fig. 7.7) gives a better idea of ice types and structure of ice eddies at different
scales. Moreover Envisat ASAR, Terra MODIS and Aqua AMSR-E images form an
extended time series, which allows to study over time the surface circulation in this
dynamic area under influence of both oceanic and atmospheric factors.
7.3.3 Usage of Multiple SAR Sensors
The analysis of co-located ERS-2 SAR and Envisat ASAR images, acquired with
time difference of about 30 min, allows the estimation of the velocity of the SAR
signatures caused by both oceanic (currents, IWs, etc.) and atmospheric dynamic
phenomena (Gade et al., 2003; Dubina and Mitnik, 2005; Liu and Hsu, 2009).
Two SAR images, consisting of several successive frames and covering the
Northern Japan Sea and the Southern Okhotsk Sea, were used. The first was taken
by ASAR at 12:11–12:14 UTC, in wide swath mode, and the second by ERS-2
SAR at 12:40 UTC, on 4 October 2003. To estimate the speed V, the original highresolution ERS-2 SAR image was down-sampled to 75 m pixel size with the use of
bi-cubic interpolation. The resampled image was warped onto the ASAR original
image (75 m × 75 m pixel size) by triangulation technique. The resulting pixels of
the warped ERS SAR image were derived using cubic convolution.
L.M. Mitnik and V.A. Dubina
Fig. 7.7 Spiral ice eddy in
the Southern Okhotsk Sea on
Envisat ASAR image taken at
00:30 UTC (a) and on Terra
MODIS visible image taken
at 00:50 UTC (b) on 8
February 2003; 1 marks the
ice-free area
SAR, visible and passive microwave data decreases the ambiguity of interpretation.
This is evident from a comparison of the SAR images with the AMSR-E T B s and
MODIS visible images. Both eddies have higher T B s values at 89 GHz in comparison with the open sea as seen in (Fig. 7.6b, d). The spatial resolution of T B s data is
4 km × 6 km. In the visible image, having a spatial resolution of 250 m × 250 m,
the open sea has a dark tone even at high winds (Fig. 7.7b) as opposite to SAR
images where the increase in wind speed causes the increase in radar backscatter
(Fig. 7.7a). The brightness of sea ice depends on its thickness (type) and also on the
presence of snow. So, grease ice and dark nilas (ice which is under 5 cm in thickness
and is very dark in colour) have very low brightness contrast, as opposite for example to grey-white or white ice. The concurrent analysis of SAR and visible images
(Fig. 7.7) gives a better idea of ice types and structure of ice eddies at different
scales. Moreover Envisat ASAR, Terra MODIS and Aqua AMSR-E images form an
extended time series, which allows to study over time the surface circulation in this
dynamic area under influence of both oceanic and atmospheric factors.
7.3.3 Usage of Multiple SAR Sensors
The analysis of co-located ERS-2 SAR and Envisat ASAR images, acquired with
time difference of about 30 min, allows the estimation of the velocity of the SAR
signatures caused by both oceanic (currents, IWs, etc.) and atmospheric dynamic
phenomena (Gade et al., 2003; Dubina and Mitnik, 2005; Liu and Hsu, 2009).
Two SAR images, consisting of several successive frames and covering the
Northern Japan Sea and the Southern Okhotsk Sea, were used. The first was taken
by ASAR at 12:11–12:14 UTC, in wide swath mode, and the second by ERS-2
SAR at 12:40 UTC, on 4 October 2003. To estimate the speed V, the original highresolution ERS-2 SAR image was down-sampled to 75 m pixel size with the use of
bi-cubic interpolation. The resampled image was warped onto the ASAR original
image (75 m × 75 m pixel size) by triangulation technique. The resulting pixels of
the warped ERS SAR image were derived using cubic convolution.