Understanding sea ice melting via functional data analysis

Abstract

We have addressed the problem of sea ice extent (SIE) melting for Arctic and Southern Ocean. The 'satellite passive microwave remote sensing data' for daily SIE over a year has been considered as a smooth continuous function, called functional data. The analysis of the mean function of SIE over the different year-blocks and 95% bootstrap confidence bands for Arctic Ocean shows a statistically significant drop in SIE. Additional evidence of SIE melting in the Arctic Ocean is provided through phase plane analysis. During the summer, the SIE is observed about 30% less in the current year-block than that of the first year-block.