Data from two different years from ESA's CryoSat validation experiments CryoVex 2008 and 2007 are analyzed and possibilities for sea ice classification based on a Ku-band radar altimeter waveforms are presented. While the detection of leads is crucial to correctly retrieve the freeboard the knowledge of ice type improves the calculation of ice thickness as the ice density and snow depth vary with ice type.
The waveforms of the radar power echo vary with underlying surface. In our study we used various parameters to describe these waveforms and to classify different surface types. For the beginning we distinguish between first year ice, multi year ice and leads/open water. Parameters which have been analyzed are the leading edge width, trailing edge width, trailing edge slope, the power max and the pulse peakiness.
We find that for lead detection the pulse peakiness and the power max are sufficient to detect more than 90% of the leads, while the rate of false detection is below 10% . As in particular the waveforms of first year ice are less distinct the detection rate of first year ice and multi year ice is lower. We find, however, that with a combination of the trailing edge width and power maximum more than 75 % of the waveforms can be classified correctly.
Author: Marta Zygmuntowska
ESA-CliC-EGU Conference on Earth Observation and Cryospheric Science, Frascati, Italy, November 13-16, 2012