| 格陵兰冰盖次表面湖多源遥感监测——以格陵兰中西部流域为例 |
| |
| 引用本文： | 李岚静, 陈卓奇, 郑雷, 程晓. 2022. 格陵兰冰盖次表面湖多源遥感监测——以格陵兰中西部流域为例. 地球物理学报, 65(10): 3818-3828, doi: 10.6038/cjg2022P0558 |
| |
| 作者姓名： | 李岚静 陈卓奇 郑雷 程晓 |
| |
| 作者单位： | 中山大学测绘科学与技术学院,广东珠海 519082;中山大学测绘科学与技术学院,广东珠海 519082;南方海洋科学与工程广东省实验室(珠海),广东珠海 519082;中国高校极地联合研究中心,北京 100875 |
| |
| 基金项目： | 国家重点研发计划项目(2019YFC1509104), 国家自然科学基金项目(42006192), 广东省基础与应用基础研究基金(2021B1515020032), 南方海洋科学与工程广东省实验室(珠海)创新群体项目(311021008)资助 |
| |
| 摘 要： | 格陵兰冰盖在夏季会发生剧烈融化，融水在低洼处汇集形成冰面湖，储存了大量的融水，在冰盖的水文系统中起着至关重要的作用，是格陵兰冰盖质量平衡的重要组成部分.近期有研究发现大量的冰面湖在冬季不会完全冻结，而是掩埋在雪层或冰层下以液态水形式存在，形成冰盖次表面湖.冰盖次表面湖对格陵兰冰盖稳定性和物质平衡会产生重要的影响.由于次表面湖存在于冰盖表层之下难以通过可见光影像进行提取和分析，次表面湖的监测成为次表面湖研究的难点之一.本研究提出一种利用Landsat8和Sentinel-1数据自动提取次表面湖范围的方法.该方法先利用可见光影像提取夏季冰面湖的范围对SAR影像进行掩膜，再根据冬天水体和冰面的后向散射具有可分性的原理，通过Rosin阈值分割算法，提取得到2018—2019年冬季格陵兰中西部流域的次表面湖并对其分布进行分析.在研究区内选择10个测试区域，利用该方法对区域内的次表面湖进行目视解译，并对算法自动提取次表面湖结果进行精度验证.结果表明，本研究提出的自动提取算法Kappa系数为0.85.基于该方法，本研究在2018年格陵兰中西部流域共提取夏季冰面湖的面积为102.28 km2.约43.09%的夏季冰面湖不会完全冻结，从而在冬季形成次表面湖(面积为44.07 km2).
|
| 关 键 词： | 格陵兰冰盖 次表面湖 Sentinel-1 Rosin算法 |
| 收稿时间： | 2021-08-03 |
| 修稿时间： | 2022-01-07 |
Extraction of Greenland Ice Sheet buried lakes using multi-source remote sensing data: With application to the Central West basin of Greenland |
| |
| LI LanJing, CHEN ZhuoQi, ZHENG Lei, CHENG Xiao. 2022. Extraction of Greenland Ice Sheet buried lakes using multi-source remote sensing data: With application to the Central West basin of Greenland. Chinese Journal of Geophysics (in Chinese), 65(10): 3818-3828, doi: 10.6038/cjg2022P0558 |
| |
| Authors: | LI LanJing CHEN ZhuoQi ZHENG Lei CHENG Xiao |
| |
| Affiliation: | 1. School of Geospatial Engineering and Science, Sun Yat-sen University, Zhuhai Guangdong 519082, China; 2. Southern Marine Science and Engineering Guangdong Laboratory (Zhuhai), Zhuhai Guangdong 519082, China; 3. University Corporation for Polar Research, Beijing 100875, China |
| |
| Abstract: | The Greenland Ice Sheet melts violently in summer, and supraglacial lakes appear in low-lying regions where the meltwater converges. The monitoring of supraglacial lakes is important as they play a key role in modulating the hydrological system and the Greenland Ice Sheet′s mass balance. Recent studies show that numerous supraglacial lakes in the Greenland Ice Sheet do not refreeze completely in winter, but existing in the form of liquid water under snow or ice, which are referred as buried lakes. These buried lakes has the potential to impact the stability and mass balance of Greenland Ice Sheet. However, it is difficult to extract and analyse these buried lakes by using the visible images, because they exist under the surface of the Greenland Ice Sheet. In this study, we proposed a method to automatically extract the buried lakes by using both the Landsat8 and Sentinel-1 images. This method uses visible images to extract supraglacial lakes in summer to mask the SAR images, and then adopts the Rosin Threshold Segmentation Algorithm to extract buried lakes in the Central West basin of Greenland in the winter of 2018—2019. In addition, we analysed the distribution of these buried lakes by separating the water and ice based on the backscatter coefficient. We extracted the buried lakes in selected 10 test areas with our method and compared the results with those derived from visual interpretation. The results show that our method achieves good performance with the Kappa coefficient being up to 0.85. Moreover, the area of supraglacial lakes is 102.28 km2 in the Central West basin of Greenland in summer of 2018. About 43.09% of supraglacial lakes occurred in summer do not refreeze completely and becomes the buried lakes (44.07 km2) in winter. |
| |
| Keywords: | Greenland Ice Sheet Buried lakes Sentinel-1 Rosin Algorithm |
| 本文献已被 万方数据 等数据库收录！ |
| 点击此处可从《地球物理学报》浏览原始摘要信息 |
|
点击此处可从《地球物理学报》下载全文 |
|
