Chinese Journal of Polar Research

Chinese Journal of Polar Research ›› 2020, Vol. 32 ›› Issue (3): 301-313.DOI: 10.13679/j.jdyj.20190054

Previous Articles     Next Articles

Comparison and verification of remote sensing sea ice concentration products for Arctic shipping regions

Wang Jian1,2, Qiu Yubao2,3, Xiong Zhenhua2,Yuan Xiping1, Zhou Jingtian4, Huang Lin5, Shi Lijuan2,3   

  1. 1.Land and Resources Engineering College, Kunming University of Science and Technology, Kunming 650093, China;
    2.Laboratory of Digital Earth Science, Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing 100094, China;
    3.Joint Research Center for Arctic Observations, Aerospace Information Research Institute, Chinese Academy of Sciences and Arctic Space Center, Finnish Meteorological Institute (JRC-AO), Sodankyl? 999018, Finland;
    4.Marine College, Dalian Maritime University, Dalian 116026, China;
    5.Geography and Information Engineering College, China University of Geosciences (Wuhan), Wuhan 430074, China
  • Received:2019-09-26 Revised:2020-03-14 Online:2020-09-30 Published:2020-09-30
  • Contact: Yubao QIU

PDF (PC)

1244

Abstract:

We analyzed and evaluated sea ice concentration data from the marginal ice zone in eight publicly available Arctic sea ice concentration products. Mean bias and standard deviations were derived from the products and compared. Means bias values indicate that Bremen/Artist Sea Ice, Bremen/BT(Bootstrap), NSIDC(National Snow And Ice Data Center)/BT and NSIDC/CDR(Climate Data Record)/ BT products are higher than the mean of the eight products for the entire year of the study period, and their mean bias are higher in summer than in winter. Annual mean bias of Hamburger/ASI product is lower than the mean of the eight products, negative in winter and spring, and positively increasing in summer. Annual mean bias of NSIDC/NT(NASA Team), NOAA OI SIC and OSISAF products are negative and lower in summer and autumn. Large standard deviations in summer and autumn are mainly found along the Northeast Passage where ice cover is thin (marginal ice zone). There are large variations in standard deviation in the regions of the East Siberian, Laptev and Kara Seas; values increase from 3% in spring to 10%–15% in winter. Correlating the eight products with MODIS images at a resolution of 25 km, we found the lowest correlation coefficient (0.63) in the NOAA OI SIC product (0.25° resolution), while correlation coefficients for the two products from Bremen are highest (both 0.80). Bremen/BT has the lowest mean deviation (7.11%). Products derived using the ASI algorithm have higher mean bias; mean bias of the Bremen/ASI and Hamburger/ASI products are 14.38% and 14.99%, respectively, and deviations fluctuate greatly in summer and autumn. Standard deviation of Bremen/ASI products is 11.01%, and SD of Hamburger/ASI product is 12.16%. Our research results are important for the application of remote sensing data products to support maritime navigation in marginal ice zones, and for future development of sea ice concentration algorithms.

Key words: Arctic, sea ice concentration, remote sensing products, critical regions for Northeast Passage, product comparison and verification