Authors

X. Zhou; N. B. Chang;S. Li

Comments

Authors: contact us about adding a copy of your work at STARS@ucf.edu

Abbreviated Journal Title

J. Environ. Inform.

Keywords

Coastal sea ice; feature space image; multiple modes; SAR; single-mode; Environmental Sciences

Abstract

Application of SAR images to detect the onset of breakup and dynamic decay process of coastal sea ice was investigated, using a time series of RADARSAT-1 C-band scanning SAR images of the areas around Seward Peninsula coastal region, Alaska, including Bering Strait and Norton Sound. Satellite image analyses mainly include a time series feature space images composed of radar backscattering coefficient versus look angle derived from RADARSAT-1 amplitude images covering both coastal sea and land area. It was found that the feature space images have a single mode before the decay and breakup of the coastal sea ice, and have multiple modes when large-scale open water first appears until the coastal sea ice disappears completely. The mode patterns in the feature space images can then be determined by the water surface roughness caused by the wind field. The single-mode phenomenon before ice melting is thought to be due to a similarity of the backscattering of snow cover on sea ice and land surface due to the common coverage by snow and ice. The appearance of multiple modes is due to the differences in backscattering of sea ice surface at various decaying stages or open water with different wave roughness due to wind effect, and the land surface with snow and ice cover with different melting conditions once the melting or decay starts and proceeds. The feature space images reported here can be used as a use tool in detection of decay of coastal sea ice and land fast ice in the field of sea ice dynamics as the imaging radar technology is weather independent, providing information of coastal ice on a regular basis.

Journal Title

Journal of Environmental Informatics

Volume

10

Issue/Number

1

Publication Date

1-1-2007

Document Type

Article

Language

English

First Page

37

Last Page

46

WOS Identifier

WOS:000207621700006

ISSN

1726-2135

Share

COinS