Elevation and Freeboard Changes of Ross Sea Ice and Ice Shelf using Icesat

Abstract

In this paper, ICESat (Ice, Cloud, and land Elevation Satellite)-based estimates of sea ice and ice shelf elevation change and the sea ice freeboard changes in the Ross Sea Region of west Antarctic were studied. To process the data, a 3-step procedure was developed. ICESat datasets of three releases were used for the study: laser 1 and laser 2A (2003) and laser 3D (2005). It was found that all of them have suspect elevation data and extensive quality control was needed to remove the suspect points before a bias-free surface could be generated. Also obtained sea-ice freeboard from ICESat elevation data showed unusual negative results. Simple, universe, and ordinary kriging interpolations as well as inverse distance weights interpolation were tested and evaluated. It is found that the ordinary kriging method based on an unknown mean performed the best and achieved the smallest RMS prediction error (< 1 m) based on cross-validation technique. Seasonal elevation changes of the two complete datasets (laser 1 and laser 2A) have been analyzed and results indicate that (1) the mean elevation of sea ice increased about 0.89 m, with a mean increase of 0.55 m for the sea ice/ice shelf transition zone; (2) the mean elevation of the ice sheet decreased about 0.17 m; (3) maximum elevation of sea ice increased 1.13 m; (4) maximum elevation of ice shelf increased 5.36 m; (5) the transitional region has the largest elevation change: ~60 % of the region shown 48 to 20 m decreases, only small portions shown 20- 40 m increases. Inter-annual elevation changes obtained from laser 2A and laser 3D for sea ice indicate that (1) minimum surface elevation has decreased as 5.27 m with increase of mean surface about 0.12 m; (2) maximum surface elevation from Austral spring 2003 to spring 2005 has decreased as 12.69 m. It is suggested that the freeboard directly resulted from ICESat data has large negative values and needs better geoid or mean sea surface.