Microwave remote sensing of ice in Lake Melville and the Labrador Sea

Results from a joint experiment on microwave remote sensing of ice in the Labrador Sea and Lake Melville during March 1982 are presented. Data from sensors carried on the Canada Centre for Remote Sensing (CCRS) Convair-580, including Ku-band scatterometer and K-band radiometer profiles and
X-, C-
, and
L
-band synthetic aperture radar (SAR) imagery, were analyzed in conjunction with aerial photography and surface measurements carried out from an icebreaker. Several ice types were encountered. The Labrador pack was very rough, consisting of consolidated first-year cakes mostly less than 10 m across with 20 cm of snow cover in highly saline brash. Lake Melville ice had a snow cover of approximately 60 cm on undeformed brackish ice, also 60 cm thick. Groswater Bay had a shore lead mainly covered with nilas. A quantitative analysis of profiler data revealed that ice type and concentration affected backscatter and emissivity values. Class separation was possible using either
Ku
-band backscatter or
K
-band emissivity alone for open water, nilas, lake, and pack ice, but the combination of these measurements provided additional class separation. Comparison with Beaufort Sea results underlined inherent regional differences in ice characteristics. Examination of SAR imagery showed ridges and individual floes within the Labrador pack ice were not detectable with high-resolution (3.0-m)
X
- band SAR, although cake sizes sometimes exceeded 10 m. Some floe detail was apparent at
L
band and at grazing angles. Icebreaker tracks in landfast and pack ice showed very different rates of decay. Imagery from a large tabular iceberg in pack ice showed distinct radar shadows, and high contrasts were obtained with grazing incidence angles. These provided a substantial detection advantage. SAR imagery also proved more effective for monitoring wave decay in pack ice than photography.