Papers Online - #151

On the Sea-Ice Conditions Observed by SAR and AVHRR Imagery in the Ross Sea during Spring 1996

by

Adrian Hauser1, Matthew Lythe and Gerd Wendler2
International Centre for Antarctic Information and Research (ICAIR)
Christchurch, New Zealand

Permanent affiliation:
1 Institut für Meteorologie und Klimaforschung, Universität Karlsruhe, Germany
2 Geophysical Institute, University of Alaska, Fairbanks, USA

Abstract

SAR (Synthetic Aperture Radar) imagery from ERS 2 was obtained from the Terra Nova region in the Ross Sea in spring of 1996. Such active microwave images have a high resolution (14 m) and penetrate clouds and darkness. However, the interpretation of the data is not simple, as besides ice thickness, surface roughness, salinity and water content influence the backscatter.

Statistical procedures were developed, which reduced the speckle noise, but obtained the boundaries between different ice types. More than 95 % of the sea ice could be classified in one of four established sea ice categories based on the backscatter spectrum after 10 iterations. Some of the findings are:

  • The size of the Terra Nova Polynya depended strongly on the strength of the off-shore winds, which were measured by Italian Automatic Weather Stations (AWS). It appear that its presents and size is more dependent on atmospheric than on oceanographic forcing.
  • The Terra Nova Polynya is frequently not ice free, but covered, at least in part, by thin, rough sea ice. This could be deduced both, from thermal IR AVHRR as from SAR imagery.
  • The strong katabatic wind flow down the Priestley and Reeves Glaciers, combined in even stronger winds on the Nansen Ice Shelf. Offshore the Nansen Ice Shelf, IR images displayed an area of relatively warm temperatures, previously reported on by the Italian AWS stations.
  • While large changes in the ice types with time were observed in front of the Nansen Ice Shelf and areas further off the coast, sea ice conditions changed little in over a month for protected areas, for example between the Campbell Ice Tongue and Cape Washington. Here, and in other coastal areas, the ice appears to be landfast.
  • Occasionally, the ice classification derived from the SAR images was not conclusive, as the intercomparision with AVHHR visible and IR data showed. For example, thin, newly formed sea ice produced a similar backscatter as floating, freshwater icebergs.

BACK

 

Gerd Wendler
gerd@gi.alaska.edu
Date Last Modified: 4/30/00