极地研究

极地研究 ›› 2023, Vol. 35 ›› Issue (4): 517-532.DOI: 10.13679/j.jdyj.20220426

• 研究论文 • 上一篇    下一篇

基于海冰密集度的南极威德尔冰间湖监测与评价

邝慧妍1,2  华健聪1,2  叶玉芳1,2  程晓1,2  惠凤鸣1,2  陈卓奇1,2   

  1. 1中山大学测绘科学与技术学院, 广东 珠海519082;
    2南方海洋科学与工程广东省实验室(珠海), 广东 珠海519082
  • 出版日期:2023-12-30 发布日期:2024-01-10
  • 通讯作者: 叶玉芳, E-mail: yeyf8@mail.sysu.edu.cn
  • 作者简介:邝慧妍, 女, 1998年生。硕士研究生, 主要从事极地遥感、极地海冰与气候变化研究。E-mail: kuanghy5@mail2.sysu. edu.cn
  • 基金资助:
    国家自然科学基金青年项目(42106225)、广东省自然科学基金面上项目(2022A1515011545)和南方海洋科学与工程广东省实验室(珠海)创新团队建设项目(311021008)资助

Detection and evaluation of the Weddell Polynya based on sea ice  concentration products

Kuang Huiyan1,2, Hua Jiancong1,2, Ye Yufang1,2, Cheng Xiao1,2, Hui Fengming1,2,  Chen Zhuoqi1,2    

  1. 1School of Geospatial Engineering and Science, Sun Yat-sen University, Zhuhai 519082, China;
    2Southern Marine Science and Engineering Guangdong Laboratory (Zhuhai), Zhuhai 519082, China
  • Online:2023-12-30 Published:2024-01-10

PDF (PC)

424

摘要： 海冰密集度产品在冰间湖的监测与研究中应用广泛。本文使用8种典型的被动微波遥感海冰密集度产品(NSIDC-BT-25km、NSIDC-NT2-25km、NSIDC-NT2-12.5km、NSIDC-NTBT-25km、EUMETSAT-BTBR- 25km、EUMETSAT-BTBR-10km、UH-ASI-12.5km和UB-ASI-6.25km)以及5种常用的海冰密集度阈值(15%、40%、50%、60%和70%)对南极威德尔海2016—2017年出现的冰间湖进行监测, 并使用形态学后处理操作对监测结果进行优化; 在此基础上, 对比不同阈值条件、海冰密集度反演算法以及空间分辨率差异对冰间湖面积和范围的影响, 并进一步探究形态学操作对监测结果的影响。结果表明: NSIDC-NTBT-25km产品对阈值的敏感性最高, NASA Team 2(NT2)算法反演的海冰密集度产品对阈值的敏感性最低, 并且其监测的冰间湖面积和范围相较于Bootstrap(BT)和ARTIST Sea Ice(ASI)算法产品的监测结果整体偏小; 高空间分辨率产品监测到的冰间湖开放时间更早, 面积和范围更大, 持续时间也更长; 空间分辨率对NT2算法产品影响最小, 对ASI算法产品影响最大。形态学后处理操作虽对监测结果有一定影响, 但与反演算法和空间分辨率所带来的影响相比均要小。

关键词:

微波遥感, 海冰密集度, 威德尔冰间湖, 阈值法, 形态学处理

Abstract: Sea ice concentration (SIC) products have been widely used in the detection of polynyas. Eight typical SIC products retrieved from passive microwave remote sensing data (NSIDC-BT-25km, NSIDC-NT2-25km, NSIDC-NT2-12.5km, NSIDC-NTBT-25km, EUMETSAT-BTBR-25km, EUMETSAT-BTBR-10km, UH-ASI- 12.5km and UB-ASI-6.25km) were adopted to detect the Weddell Polynya openings in 2016 and 2017 using five SIC thresholds (15%, 40%, 50%, 60% and 70%) separately. Morphological post-processing operations were then applied to optimize the results. On this basis, the effects of different SIC thresholds, retrieval algorithms, and spatial resolutions of the SIC products on the area/extent of polynya were compared, and the influence of morphological post-processing operations on the results was further explored. The results showed that NSIDC-NTBT-25km demonstrated the highest sensitivity to thresholds, while NASA Team 2 (NT2) algorithm-retrieved SIC products showed the lowest sensitivity and the area/extent of polynya detected from which were generally smaller than those detected by BT (Bootstrap) and ASI (ARTIST Sea Ice) products. High spatial resolution SIC products were more likely to derive polynya with earlier opening times and longer duration, and the area/extent of polynya detected from these were also larger as well. Additionally, spatial resolution exerted the least impact on NT2-retrieved products and had the greatest impact on ASI-retrieved products. Morphological post-processing operations showed some influence on the detected results, yet the effect was less than that of the retrieval algorithms and spatial resolutions.

Key words: microwave remote sensing, sea ice concentration, Weddell Polynya, threshold, morphological processing