The Classification Of Arctic Sea Ice Using Multi-source Microwave Remote Sensing Data

Sea ice is an important component of the cryosphere,and plays a vital role on global climate change.It not only affects the global ocean circulation,but also limits the energy exchange between the ocean and atmosphere.In addition,the freezing and drifting of sea ice has a devastating impact on marine production activities such as shipping and offshore petroleum exploration.Therefore,the study of sea ice monitoring is very meaningful,and the monitoring of sea ice types is helpful to maritime shipping and climate research.Microwave remote sensing is widely used in sea ice monitoring because it is not affected by clouds and light.Accordingly,multi-source microwave remote sensing data are used to observe the types of Arctic sea ice in this paper,and we focusing on the temporal and spatial characteristics of sea ice types at different scales.The temporal and spatial characteristics of sea ice type were observed by using SAR data in the Canadian Arctic Archipelago from May 2017 to April 2018.According to the difference of texture feature and backscatter value between first-year ice and multiyear ice on SAR image,SVM is used to classify sea ice.The experimental results show that,even in summer,the northwest of study area is covered with sea ice,which can not navigate.However,the northeast and south of study area are almost no sea ice in August.The distribution characteristics of multiyear ice in research area are as follows: for the west of 90°W,the distribution and shape of multiyear ice are basically unchanged,mainly distributed in the northeast side of Victoria Island and the Parry Islands;for the east of 90°W,multiyear ice has been moving eastward.Our results in this chapter are in good agreement with the ice charts released by CIS,and provide further information about the distribution details of multiyear ice.Microwave Radiometer and Scatterometer data can cover the Arctic and have higher temporal resolution,which were used to explore the temporal and spatial variations of first-year ice and multiyear ice over the Arctic from 2002 to 2018.According to the difference of brightness temperature and backscatter value between first-year ice and multiyear ice,MLE is used to classify sea ice.The results show that multi-year ice mainly distributes in the central area of Arctic Basin,and it increases during the ice freezing period and decreases during ice melting period,which is similar to first-year ice.But,the increase of multi-year ice is much less than that of first-year ice when sea ice freezes.During the study period,Arctic multi-year ice decreased significantly in the Beaufort Sea,the Eastern Siberian Sea and the Laptev Sea,and reached the first-lowest value and second-lowest value in 2011-2013 and 2008,respectively.Finally,the accuracy of the experimental results in this paper is verified by comparing with several data.