地球科学进展 ›› 2025, Vol. 40 ›› Issue (8): 778 -794. doi: 10.11867/j.issn.1001-8166.2025.065

综述与评述 上一篇    下一篇

中国极地冰冻圈人工智能技术应用的进展与展望
唐学远()   
  1. 1. 自然资源部极地科学重点实验室,中国极地研究中心,上海 200136
    2. 极地生态与气候变化教育部 重点实验室,上海交通大学 海洋学院,上海 200030
    3. 复旦大学 极地海—冰—气系统与天气 气候教育部重点实验室,复旦大学 大气与海洋科学系/大气科学研究院,上海 200438
    4. 浙江大学 海洋研究院,浙江 舟山 316021
  • 收稿日期:2025-06-26 修回日期:2025-07-18 出版日期:2025-08-10
  • 基金资助:
    国家自然科学基金项目(42276257); 2024年度上海市东方英才计划拔尖项目(BJKJ2024035)

Advances and Prospects of Artificial Intelligence in China’s Polar Cryosphere Research

Xueyuan TANG()   

  1. 1. Key Laboratory of Polar Science of Ministry of Natural Resources (MNR), Polar Research Institute of China, Shanghai 200136, China
    2. Key Laboratory of Polar Ecosystem and Climate Change, Ministry of Education, School of Oceanography, Shanghai Jiao Tong University, Shanghai 200030, China
    3. Key Laboratory of Polar Atmosphere-Ocean-Ice System for Weather and Climate, Ministry of Education, Department of Atmospheric and Oceanic Sciences & Institute of Atmospheric Sciences, Fudan University, Shanghai 200438, China
    4. Ocean Academy, Zhejiang University, Zhoushan Zhejiang 316021, China
  • Received:2025-06-26 Revised:2025-07-18 Online:2025-08-10 Published:2025-10-20
  • Supported by:
    the National Natural Science Foundation of China(42276257); Top-notch Project under the 2024 Shanghai Oriental Talents Program(BJKJ2024035)

在全球变暖加速激活极地冰冻圈临界点的背景下,由于观测稀疏、模型局限和认知不足等限制,在理解极地冰冻圈演变状态及预测其趋势时面临巨大挑战。人工智能成为高效挖掘海量极地数据、弥补认知鸿沟的强大工具。中国极地冰冻圈人工智能研究虽起步较晚但发展迅速,在数据驱动建模、特征自动提取以及多源融合等方面展现出巨大潜力,近年来在以下方向取得了重要进展:①海冰预测:开发纯数据驱动深度学习模型(如SICNet和SIPNet),显著提升周至季节尺度海冰密集度预报能力,部分超越传统模型;发展多种方法(如改进U-Net和PMDRnet等)用于海冰类型识别、水道提取及影像分辨率提升。②冰盖水文:应用随机森林(RF)和神经网络(BP)估算冰盖融化和识别表面湖;创新运用改进U-Net高精度自动提取冰盖/冰架水体,克服传统方法的局限。③冰下系统:提出基于变分自编码与无监督聚类的新方法,实现对冰雷达观测数据中冰下湖体的自动识别。④冰裂隙识别:应用改进U-Net及变体(如ResUNet)从合成孔径雷达/光学影像中自动提取冰架裂隙特征。⑤冰层结构与地形:探索深度学习(如EisNet)自动提取冰雷达内部等时层与基岩界面,致力于解决训练标签的人工依赖瓶颈。⑥其他应用:冰盖质量平衡重建(SVM/BPNN)、辐射平衡数据集(RF)、气温反演(RF/DNN)、冰底通道识别、冰川地震分类和GPS插补等。考虑到当前仍面临模型可解释性和物理机制融合不足、标注数据稀缺以及区域泛化能力有限等问题,未来需深化物理约束人工智能模型、发展多模态学习、提升鲁棒性与可解释性,并加强国际合作与数据共享,以精准刻画极地变化,服务全球气候应对与风险评估。

Amidst the accelerating activation of polar cryosphere tipping points due to global warming, significant challenges must be overcome to understand their state and changes, including sparse observations, insufficient physical knowledge, and limitations of traditional model simulations. Artificial Intelligence (AI) provides a powerful tool for efficiently extracting information from vast polar datasets and bridging cognitive gaps. This paper summarizes notable progress by Chinese researchers in AI applications for the polar cryosphere: Sea ice forecasting: purely data-driven deep learning models (e.g., SICNet, SIPNet) have been developed, significantly improving weekly, monthly, and seasonal-scale forecasts of Arctic/Antarctic sea ice concentration. Some models incorporate physical constraints and outperform traditional dynamical and statistical models. Various methods (e.g., improved U-Net, EW-Net, SAC-Net, and PMDRnet) have been proposed for sea ice type identification, lead extraction, sea ice thickness relationship modeling, and enhancing the spatial resolution of passive microwave imagery. Ice sheet surface hydrology: Applied Random Forest (RF) and BP neural networks were applied to estimate surface melt of the Greenland Ice Sheet and identify supraglacial lakes. An improved U-Net model was used to automatically extract surface water bodies of the Antarctic ice sheet/ice shelf with high accuracy, thereby overcoming the limitations of traditional NDWI methods. Subglacial systems: A novel method based on Variational Autoencoders (VAE) and unsupervised clustering was used to automatically detect subglacial lakes from ice-penetrating radar data, thereby improving efficiency and accuracy. Crevass identification: Improved U-Net and its variants (e.g., ResUNet) were applied to automatically extract surface crevasse distributions on Antarctic ice shelves from SAR and optical imagery. Ice stratigraphy and topography: Deep learning (e.g., EisNet and ST-SOLOv2) was employed to automatically extract internal isochronous layers and bedrock interfaces from radargrams, aiming to solve this long-standing manual bottleneck. Other applications include: Mass balance reconstruction of covered ice sheets (fusing multi-source data with SVM/BPNN), radiation balance dataset construction (RF), near-surface air temperature inversion (RF/DNN), ice shelf basal channel identification (improved U-Net), intelligent classification of glacial seismic events (autoencoders and Gaussian mixture models), GPS data interpolation, tropospheric delay modeling, and identification of geological structures. Although Chinese polar cryosphere AI research began relatively late, it has developed rapidly and yielded fruitful results, demonstrating significant potential in data-driven modeling, automated feature extraction, and multisource information fusion. Current challenges include model interpretability, insufficient integration of physical mechanisms, scarcity of high-quality labeled data, and limited generalization ability in complex regions. Future efforts should focus on developing physically constrained AI models, advancing multimodal learning, enhancing model robustness and interpretability, and strengthening international collaboration and data-sharing to more accurately characterize polar cryosphere changes and support global climate response and risk assessment.

中图分类号: 

图1 SIPNet模型结构 (a)包括输入、编码器、解码器和输出的模型结构,编码器模块包括残差网络模块和最大池化层,解码器由残差网络模块、上采样和串联层组成;(b)ResNet TSAM 块的结构(据参考文献[41]修改)。
Fig. 1 SIPNet model Structure (a) The model structure includes input, encoder, decoder, and output. The encoder consists of a residual network module and a max-pooling layer, while the decoder is composed of residual network modules, upsampling, and concatenation layers. (b) The structure of the ResNet TSAM block (modified after reference [41]).
图2 深度学习模型对 Sentinel-2 图像(aceg)自动提取水体结果(bdfh35
Fig. 2 Deep learning model for automatic extraction of water bodies from Sentinel-2 imagesacegand the results of extractionbdfh35
图3 冰下湖检测的流程(据参考文献[64]修改) (a)单道冰底反射特征提取;(b)变分自动编码器(VAE);(c)冰底反射降维特征聚类;(d)冰下湖标记。
Fig. 3 Flowchart of subglacial lake detectionmodified after reference64]) (a) Single-pulse ice bottom reflection feature extraction; (b) Variational AutoEncoder (VAE); (c) Dimensionality reduction and feature clustering of ice bottom reflections; (d) Subglacial lake labeling.
图4 EisNet拾取雷达图像内部等时层与基岩界面的各模块应用工作流程15
Fig. 4 Workflow of EisNet various modules for picking internal isochrones and bedrock interfaces from radar images15
表1 中国在极地冰冻圈AI研究中的典型模型及性能对比
Table 1 Comparison of typical models and performance in AI research of the polar cryosphere in China
类别及特征 模型 应用中的异同对比 输入 输出 实际表现 优缺点 改进方向
卷积神经网络(CNN)相关模型:基于卷积操作,通过提取局部特征进行分类、分割,适用于图像数据 CNN 卷积神经网络,常用于图像识别,适合处理极地环境中的遥感图像 图像数据和遥感数据 图像分类结果 在海冰图像分类、裂隙识别 等任务中表现 优异 优点:图像处理能力强,适应性强 缺点:计算资源消耗大,对数据标注要求高 减少模型训练对数据标注的依赖,优化计算资源消耗
U-Net 用于图像分割,广泛应用于医学图像和遥感图像分割 图像数据 分割图像 在极地冰盖、海冰分割中应用广泛,效果显著 优点:图像分割性能强 缺点:训练数据要求高, 模型较大 优化计算效率,减少训练数据需求
ResUNet 结合残差网络与U-Net,用于图像分割和语义分割任务 图像数据 分割图像 在极地冰盖、海冰图像分割中表现突出 优点:图像分割精度高,适应性强 缺点:模型较为复杂,训练数据要求高 优化模型结构,减少训练数据需求
EisNet 基于合成数据训练的卷积神经网络,用于极地冰盖的自动识别 合成数据和遥感数据 冰盖内部结构和基岩界面识别结果 在冰盖结构提取中表现优越,能够有效抑制噪声干扰 优点:高效且稳定,适应性强 缺点:对训练数据要求高,可能出现过拟合 优化训练策略,提高模型的泛化能力
循环神经网络(RNN)相关模型:关注时间序列数据的建模,能够处理数据中的时间依赖关系 LSTM 长短期记忆网络,用于时间序列预测 时间序列数据 预测结果 在时间序列预测中,尤其是海冰预报中表现良好 优点:适合时序数据,能够处理长期依赖关系 缺点:训练时间长,参数选择困难 提高训练效率,优化模型的长期依赖处理能力
ConvLSTM 结合卷积神经网络与长短期记忆网络(LSTM),用于时空数据的预测 时间序列数据、空间数据 时序预测结果 在预测海冰面积、范围等方面表现良好 优点:能够处理时序数据,适合海冰动态预测 缺点:对训练数据量要求高,训练时间长 优化模型的时间效率,减少训练数据需求
BPNN 反向传播神经网络,广泛用于回归和分类任务 特征数据、标签数据 分类结果、回归结果 在冰盖底部融水和冰下湖的预测任务中表现出色 优点:高效且灵活 缺点:容易过拟合,需要较长的训练时间 增加正则化技术,减少过拟合
类别及特征 模型 应用中的异同对比 输入 输出 实际表现 优缺点 改进方向
生成模型与自编码器:主要用于数据生成、压缩、特征提取等无监督学习和生成建模任务 VAE 变分自编码器,用于生成数据和异常检测 原始数据和潜在变量数据 重建数据和潜在 变量 在异常检测任务中具有较好的效果 优点:适合生成任务和异常检测 缺点:模型训练较慢,生成质量受限 改进生成数据的质量,提高训练速度
自编码器 用于数据降维与特征提取 高维数据 降维后的特征 在数据降维和特征提取方面表现良好 优点:能够有效降低数据维度 缺点:对噪声敏感 提高对噪声的鲁棒性,优化降维效果
高斯混合模型 用于数据的概率建模,适用于密度估计与聚类任务 数据集(带有潜在分布) 聚类结果和概率 分布 在海冰分类与模式识别中有较好的表现 优点:处理复杂数据分布 缺点:参数选择困难,计算复杂 改进模型的拟合速度,简化参数调整过程
基于统计学原理的模型:适用于有标签数据的分类、回归等任务,尤其在数据较少或存在不确定性的情况下表现较好 SVM 支持向量机,用于分类和回归问题,特别适用于高维 数据 特征数据

结果

分类结果和回归

在极地冰盖的分类任务中表现出色,尤其在高维空间中 优点:适合高维数据,性能稳定 缺点:训练时间长,参数调整困难 提高训练效率,优化参数调节方法
贝叶斯 基于概率推理的模型,适用于不确定性较高的数据环境 输入数据(带不确定性)

结果

概率分布和预测

在不确定性较大的极地环境数据中有较好的表现 优点:处理不确定性较强的 数据 缺点:推理过程计算复杂 提高推理速度,简化计算过程
集成学习模型:通过集成多个弱学习器来提高整体模型的表现,适用于各种监督学习任务,尤其在大规模数据集上表现良好 KNN K近邻算法,用于分类与回归任务 特征向量数据 分类结果和回归 预测 在小样本数据的分类任务中表现较好,但对大数据集性能较差 优点:实现简单,容易理解 缺点:对大规模数据集的计算时间较长 增强对大规模数据集的适 应性
RF 随机森林,集成学习方法,适用于分类和回归任务 特征数据和标签 数据

结果

分类结果和回归

在海冰数据分类任务中有较好效果,表现 稳定 优点:适用于高维数据,鲁棒性强 缺点:模型较大,计算资源消耗较高 优化模型的内存占用和计算效率
KVFC 基于向量场共识的K近邻方法,用于处理动态变化数据 时空数据 预测结果和模式 识别 在海冰动态变化的监测中表现优异 优点:处理时空数据具有 优势 缺点:对大数据集计算时间较长 提高计算效率,减少对大数据集的训练时间
物理模型与数据驱动模型:将物理模型的知识与数据驱动的学习方法结合起来,能够在复杂系统建模中提供更具解释性和准确性的预测 PMDRnet 用于数据驱动的极地海冰建模 海冰数据 模型预测结果 在极地海冰建模中的应用较为成功 优点:适应性强,预测精度高 缺点:对输入数据质量要求 较高 改进对不同数据源的适应性,提升模型鲁 棒性
神经网络与信号处理结合的模型 小波神经网络 结合小波变换与神经网络,用于信号处理和模式识别 信号数据 信号特征提取结果 在极地冰盖底部地形的提取中,有较好的表现 优点:能处理信号的多尺度 特征 缺点:对噪声敏感,计算复杂 提高抗噪能力,减少计算复 杂度
其他模型:用于处理一些特殊任务 SICNet 基于深度学习的海冰密度预测网络,适用于季节尺度海冰预报 海冰图像和气象 数据 海冰密度和海冰 范围 优化海冰密度预报效果,适用于大规模数据处理 优点:处理大数据集,预测精度高 缺点:对数据噪声敏感,处理速度较慢 改进算法的计算效率,提高对复杂环境的适应能力
SIPNet 结合物理约束的深度学习网络,能够预测海冰变化和融化情况 物理约束数据和遥感数据 海冰厚度和融化量 成功提高次季节尺度的海冰预报能力 优点:能结合物理约束,表现稳定 缺点:计算复杂度较高,需要大量数据训练 提高计算效率和数据处理 速度
RA 回归分析,用于建立自变量与因变量之间的关系 模型 自变量 数据 因变量预测值 在简单预测任务中应用广泛,适合线性关系的情况 优点:实现简单,计算快速 缺点:只能处理线性关系,无法捕捉复杂模式 增加对非线性关系的建模 能力
[1]
HANNA E PATTYN F NAVARRO F, et al. Mass balance of the ice sheets and glaciers-progress since AR5 and challenges[J]. Earth-Science Reviews2020, 201. DOI: 10.1016/j.earscirev.2019.102976 .
[2]
MASSON-DELMOTTE V ZHAI P PIRANI A,et al. Climate Change 2021-the physical science basis: working group I to the sixth assessment report of the intergovernmental panel on climate change[EB/OL]. [2025-06-17].
[3]
RIGNOT E MOUGINOT J SCHEUCHL B, et al. Four decades of Antarctic Ice Sheet mass balance from 1979-2017[J]. Proceedings of the National Academy of Sciences of the United States of America2019116(4): 1 095-1 103.
[4]
STEINER D WALTER A ZUMBÜHL H J. The application of a non-linear back-propagation neural network to study the mass balance of Grosse Aletschgletscher, Switzerland[J]. Journal of Glaciology200551(173): 313-323.
[5]
BOLIBAR J RABATEL A GOUTTEVIN I, et al. A deep learning reconstruction of mass balance series for all glaciers in the French Alps: 1967-2015[J]. Earth System Science Data202012(3): 1 973-1 983.
[6]
STEINER D PAULING A NUSSBAUMER S U, et al. Sensitivity of European glaciers to precipitation and temperature-two case studies[J]. Climatic Change200890(4): 413-441.
[7]
ANDERSSON T R HOSKING J S PÉREZ-ORTIZ M, et al. Seasonal Arctic sea ice forecasting with probabilistic deep learning[J]. Nature Communications2021, 12. DOI:10.1038/s41467-021-25257-4 .
[8]
VERJANS V ROBEL A. Accelerating subglacial hydrology for ice sheet models with deep learning methods[J]. Geophysical Research Letters202451(2). DOI: 10.1029/2023GL105281 .
[9]
WERDER M A HEWITT I J SCHOOF C G, et al. Modeling channelized and distributed subglacial drainage in two dimensions[J]. Journal of Geophysical Research: Earth Surface2013118(4): 2 140-2 158.
[10]
SURAWY-STEPNEY T HOGG A E CORNFORD S L, et al. Mapping Antarctic crevasses and their evolution with deep learning applied to satellite radar imagery[J]. The Cryosphere202317(10): 4 421-4 445.
[11]
TOLLENAAR V ZEKOLLARI H LHERMITTE S, et al. Unexplored Antarctic meteorite collection sites revealed through machine learning[J]. Science Advances20228(4). DOI: 10.1126/sciadv.abj8138 .
[12]
XU M Z CRANDALL D J FOX G C, et al. Automatic estimation of ice bottom surfaces from radar imagery[C]// 2017 IEEE International Conference on Image Processing (ICIP). Beijing, China: IEEE, 2017: 340-344.
[13]
RAHNEMOONFAR M FOX G C YARI M, et al. Automatic ice surface and bottom boundaries estimation in radar imagery based on level-set approach[J]. IEEE Transactions on Geoscience and Remote Sensing201755(9): 5 115-5 122.
[14]
LEONG W J HORGAN H J. DeepBedMap: a deep neural network for resolving the bed topography of Antarctica[J]. The Cryosphere202014(11): 3 687-3 705.
[15]
DONG S TANG X Y GUO J X, et al. EisNet: extracting bedrock and internal layers from radiostratigraphy of ice sheets with machine learning[J]. IEEE Transactions on Geoscience and Remote Sensing2022, 60. DOI: 10.1109/TGRS.2021.3136648 .
[16]
VARSHNEY D RAHNEMOONFAR M YARI M, et al. Deep learning on airborne radar echograms for tracing snow accumulation layers of the Greenland Ice Sheet[J]. Remote Sensing202113(14). DOI:10.3390/rs13142707 .
[17]
YARI M RAHNEMOONFAR M PADEN J. Multi-scale and temporal transfer learning for automatic tracking of internal ice layers[C]// IGARSS 2020-2020 IEEE international geoscience and remote sensing symposium. Waikoloa, HI, USA. IEEE, 2020: 6 934-6 937.
[18]
GIFFORD C M AGAH A. Subglacial water presence classification from polar radar data[J]. Engineering Applications of Artificial Intelligence201225(4): 853-868.
[19]
ILISEI A M BRUZZONE L. A system for the automatic classification of ice sheet subsurface targets in radar sounder data[J]. IEEE Transactions on Geoscience and Remote Sensing201553(6): 3 260-3 277.
[20]
ILISEI A M KHODADADZADEH M FERRO A, et al. An automatic method for subglacial lake detection in ice sheet radar sounder data[J]. IEEE Transactions on Geoscience and Remote Sensing201957(6): 3 252-3 270.
[21]
RAHNEMOONFAR M YARI M PADEN J, et al. Deep multi-scale learning for automatic tracking of internal layers of ice in radar data[J]. Journal of Glaciology202167(261): 39-48.
[22]
LAI C Y KINGSLAKE J WEARING M G, et al. Vulnerability of Antarctica’s ice shelves to meltwater-driven fracture[J]. Nature2020584(7 822): 574-578.
[23]
GONÇALVES B C LYNCH H J. Fine-scale sea ice segmentation for high-resolution satellite imagery with weakly-supervised CNNs[J]. Remote Sensing202113(18). DOI: 10.3390/rs13183562 .
[24]
WANG Y R LI X M. Arctic sea ice cover data from spaceborne synthetic aperture radar by deep learning[J]. Earth System Science Data202113(6): 2 723-2 742.
[25]
LÖSING M EBBING J. Predicting geothermal heat flow in Antarctica with a machine learning approach[J]. Journal of Geophysical Research: Solid Earth2021126(6). DOI: 10.1029/2020JB021499 .
[26]
HU Z Y KUIPERS M P LHERMITTE S, et al. Improving surface melt estimation over the Antarctic Ice Sheet using deep learning: a proof of concept over the Larsen Ice Shelf[J]. The Cryosphere202115(12): 5 639-5 658.
[27]
CHENG D HAYES W LAROUR E, et al. Calving front machine (CALFIN): glacial termini dataset and automated deep learning extraction method for Greenland, 1972-2019[J]. The Cryosphere202115(3): 1 663-1 675.
[28]
BRAAKMANN-FOLGMANN A SHEPHERD A HOGG D, et al. Mapping the extent of giant Antarctic icebergs with deep learning[J]. The Cryosphere202317(11): 4 675-4 690.
[29]
LI L AITKEN A R A LINDSAY M D, et al. Sedimentary basins reduce stability of Antarctic ice streams through groundwater feedbacks[J]. Nature Geoscience202215: 645-650.
[30]
ROSIER S H R BULL C Y S WOO W L, et al. Predicting ocean-induced ice-shelf melt rates using deep learning[J]. The Cryosphere202317(2): 499-518.
[31]
JOUVET G CORDONNIER G KIM B, et al. Deep learning speeds up ice flow modelling by several orders of magnitude[J]. Journal of Glaciology202268(270): 651-664.
[32]
RONNEBERGER O FISCHER P BROX T. U-Net: Convolutional networks for biomedical image segmentation[C]// Medical image computing and computer-assisted intervention-MICCAI 2015. Cham: Springer International Publishing, 2015: 234-241.
[33]
MOHAJERANI Y WOOD M VELICOGNA I, et al. Detection of glacier calving margins with convolutional neural networks: a case study[J]. Remote Sensing201911(1). DOI: 10.3390/rs11010074 .
[34]
RADHAKRISHNAN K SCOTT K A CLAUSI D A. Sea ice concentration estimation: using passive microwave and SAR data with a U-Net and curriculum learning[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing202114: 5 339-5 351.
[35]
NIU L H TANG X Y YANG S H, et al. Detection of Antarctic surface meltwater using Sentinel-2 remote sensing images via U-Net with attention blocks: a case study over the Amery Ice Shelf[J]. IEEE Transactions on Geoscience and Remote Sensing2023, 61. DOI: 10.1109/TGRS.2023.3275076 .
[36]
de RODA H S LHERMITTE S BOLIBAR J, et al. A high-resolution record of surface melt on Antarctic ice shelves using multi-source remote sensing data and deep learning[J]. Remote Sensing of Environment2024, 301. DOI: 10.1016/j.rse.2023.113950 .
[37]
SIMMONDS I. Comparing and contrasting the behaviour of Arctic and Antarctic sea ice over the 35 year period 1979-2013[J]. Annals of Glaciology201556(69): 18-28.
[38]
SHEN X Y KE C Q WANG Q M, et al. Assessment of Arctic sea ice thickness estimates from ICESat-2 using IceBird airborne measurements[J]. IEEE Transactions on Geoscience and Remote Sensing202159(5): 3 764-3 775.
[39]
REN Y B LI X F. Predicting the daily sea ice concentration on a subseasonal scale of the Pan-Arctic during the melting season by a deep learning model[J]. IEEE Transactions on Geoscience and Remote Sensing2023, 61. DOI: 10.1109/TGRS.2023.3279089 .
[40]
ZAMPIERI L GOESSLING H F JUNG T. Predictability of Antarctic sea ice edge on subseasonal time scales[J]. Geophysical Research Letters201946(16): 9 719-9 727.
[41]
WANG Y H YUAN X J REN Y B, et al. Subseasonal prediction of regional Antarctic sea ice by a deep learning model[J]. Geophysical Research Letters202350(17). DOI: 10.1029/2023GL104347 .
[42]
DONG X R NIE Y F WANG J F, et al. Deep learning shows promise for seasonal prediction of Antarctic sea ice in a rapid decline scenario[J]. Advances in Atmospheric Sciences202441(8): 1 569-1 573.
[43]
LIANG Z Y PANG X P JI Q, et al. An entropy-weighted network for polar sea ice open lead detection from Sentinel-1 SAR images[J]. IEEE Transactions on Geoscience and Remote Sensing2022, 60. DOI: 10.1109/TGRS.2022.3169892 .
[44]
WANG Zhihao KE Changqing. Identification of melt pond on sea ice based on deep learning technology[J]. Remote Sensing Information202237(6): 85-93.
王智豪, 柯长青. 基于深度学习的海冰融池识别[J]. 遥感信息202237(6): 85-93.
[45]
LI H L KE C Q ZHU Q H, et al. A deep learning approach to retrieve cold-season snow depth over Arctic sea ice from AMSR2 measurements[J]. Remote Sensing of Environment2022, 269. DOI: 10.1016/j.rse.2021.112840 .
[46]
LIANG Z Y JI Q PANG X P, et al. Estimation of daily Arctic winter sea ice thickness from thermodynamic parameters using a Self-Attention Convolutional Neural Network[J]. Remote Sensing202315(7). DOI: 10.3390/rs15071887 .
[47]
HE B ZHAO X CHEN Y, et al. Application of feature tracking using K-Nearest-Neighbor Vector Field Consensus in sea ice tracking[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing202215: 4 326-4 336.
[48]
LIU X M FENG T T SHEN X F, et al. PMDRnet: a progressive multiscale deformable residual network for multi-image super-resolution of AMSR2 Arctic sea ice images[J]. IEEE Transactions on Geoscience and Remote Sensing2022, 60. DOI: 10.1109/TGRS.2022.3151623 .
[49]
FENG T T JIANG P LIU X M, et al. Applications of deep learning-based super-resolution networks for AMSR2 Arctic sea ice images[J]. Remote Sensing202315(22). DOI: 10.3390/rs15225401 .
[50]
FENG T T LIU X M LI R X. Super-resolution-aided sea ice concentration estimation from AMSR2 images by encoder-decoder networks with atrous convolution[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing202316: 962-973.
[51]
LIU X M FENG T T YANG Y S, et al. Characterization of north Greenland polynyas with super-resolved passive microwave sea ice concentration[J]. GIScience & Remote Sensing202461(1). DOI: 10.1080/15481603.2023.2300222 .
[52]
SHANG X Y CHENG X ZHENG L, et al. Decadal changes in Greenland ice sheet firn aquifers from radar scatterometer[J]. Remote Sensing202214(9). DOI: 10.3390/rs14092134 .
[53]
ZHENG L CHENG X SHANG X Y, et al. Greenland ice sheet daily surface melt flux observed from space[J]. Geophysical Research Letters202249(6). DOI: 10.1029/2021GL096690 .
[54]
HU J J HUANG H B CHI Z H, et al. Distribution and evolution of supraglacial lakes in Greenland during the 2016-2018 melt seasons[J]. Remote Sensing202214(1). DOI: 10.3390/rs14010055 .
[55]
PENG Boyang ZHOU Chunxia ZHU Dongyu, et al. Extraction and area change analysis of supraglacial lakes in Greenland ice sheet using U-Net model[J]. Geomatics and Information Science of Wuhan University202449(9): 1 621-1 629.
彭博洋, 周春霞, 朱冬雨, 等. 利用U-Net的格陵兰冰盖冰面湖提取和面积变化分析[J]. 武汉大学学报(信息科学版)202449(9): 1 621-1 629.
[56]
ZIRIZZOTTI A CAFARELLA L BASKARADAS J A, et al. Dry-wet bedrock interface detection by radio echo sounding measurements[J]. IEEE Transactions on Geoscience and Remote Sensing201048(5): 2 343-2 348.
[57]
LINDZEY L E BEEM L H YOUNG D A, et al. Aerogeophysical characterization of an active subglacial lake system in the David Glacier catchment, Antarctica[J]. The Cryosphere202014: 2 217-2 233.
[58]
BENTLEY C R LORD N LIU C. Radar reflections reveal a wet bed beneath stagnant Ice Stream C and a frozen bed beneath ridge BC, West Antarctica[J]. Journal of Glaciology199844(146): 149-156.
[59]
SCHROEDER D M BLANKENSHIP D D YOUNG D A. Evidence for a water system transition beneath Thwaites Glacier, West Antarctica[J]. Proceedings of the National Academy of Sciences of the United States of America2013110(30): 12 225-12 228.
[60]
WRIGHT A SIEGERT M. A fourth inventory of Antarctic subglacial lakes[J]. Antarctic Science201224(6): 659-664.
[61]
LIVINGSTONE S J LI Y RUTISHAUSER A, et al. Subglacial lakes and their changing role in a warming climate[J]. Nature Reviews Earth & Environment20223(2): 106-124.
[62]
HILLS B H SIEGFRIED M R SCHROEDER D M. Entrained water in basal ice suppresses radar bed‐echo power at active subglacial lakes[J]. Geophysical Research Letters202451(13). DOI: 10.1029/2024GL109248 .
[63]
WANG H TANG X Y XIAO E Z, et al. Basal melt patterns around the deep ice core drilling site in the Dome A region from Ice-Penetrating Radar measurements[J]. Remote Sensing202315(7). DOI: 10.3390/rs15071726 .
[64]
DONG S FU L TANG X Y, et al. Deep clustering in subglacial radar reflectance reveals subglacial lakes[J]. The Cryosphere202418(3): 1 241-1 257.
[65]
COLGAN W RAJARAM H ABDALATI W, et al. Glacier crevasses: observations, models, and mass balance implications[J]. Reviews of Geophysics201654(1): 119-161.
[66]
IZEBOUD M LHERMITTE S. Damage detection on Antarctic ice shelves using the normalised radon transform[J]. Remote Sensing of Environment2023, 284. DOI: 10.1016/j.rse.2022.113359
[67]
ZHAO J J LIANG S LI X W, et al. Detection of surface crevasses over Antarctic ice shelves using SAR imagery and deep learning method[J]. Remote Sensing202214(3). DOI: 10.3390/rs14030487 .
[68]
PANG A LIANG Q LI W J, et al. The distribution and evolution of surface fractures on pan-Antarctic ice shelves[J]. International Journal of Digital Earth202316(1): 3 295-3 320.
[69]
FUJITA S MAENO H URATSUKA S, et al. Nature of radio echo layering in the Antarctic ice sheet detected by a two‐frequency experiment[J]. Journal of Geophysical Research: Solid Earth1999104(B6): 13 013-13 024.
[70]
SIME L C HINDMARSH R C A CORR H. Automated processing to derive dip angles of englacial radar reflectors in ice sheets[J]. Journal of Glaciology201157(202): 260-266.
[71]
XIONG S MULLER J P CARO C R. A new method for automatically tracing englacial layers from MCoRDS data in NW Greenland[J]. Remote Sensing201810(1). DOI: 10.3390/rs10010043 .
[72]
DELF R SCHROEDER D M CURTIS A, et al. A comparison of automated approaches to extracting englacial-layer geometry from radar data across ice sheets[J]. Annals of Glaciology202061(81): 234-241.
[73]
PANTON C. Automated mapping of local layer slope and tracing of internal layers in radio echograms[J]. Annals of Glaciology201455(67): 71-77.
[74]
BINGHAM R G EISEN O KARLSSON N B, et al. AntArchitecture[R/OL]. [2025-06-17]. The Scientific Committee on Antarctic Research (SCAR), 2021.
[75]
TANG X Y DONG S LUO K, et al. Noise removal and feature extraction in airborne radar sounding data of ice sheets[J]. Remote Sensing202214(2). DOI: 10.3390/rs14020399 .
[76]
PENG C Y ZHENG L LIANG Q, et al. ST-SOLOv2: tracing depth hoar layers in Antarctic ice sheet from airborne radar echograms with deep learning[J]. IEEE Transactions on Geoscience and Remote Sensing2024, 62. DOI: 10.1109/TGRS.2024.3480698 .
[77]
SHI Z Y WANG Z M ZHANG B J, et al. Bridging the spatiotemporal ice sheet mass change data gap between GRACE and GRACE-FO in Greenland using machine learning method[J]. Journal of Hydrology2024, 629. DOI: 10.1016/j.jhydrol.2024.130622 .
[78]
ZENG Z L WANG Z M DING M H, et al. Estimation and long-term trend analysis of surface solar radiation in Antarctica: a case study of Zhongshan station[J]. Advances in Atmospheric Sciences202138(9): 1 497-1 509.
[79]
PANG X P LIU C ZHAO X, et al. Application of machine learning for simulation of air temperature at Dome A[J]. Remote Sensing202214(4). DOI: 10.3390/rs14041045 .
[80]
HU Y L LI Z F FU L, et al. Environment‐modulated glacial seismicity near Dålk Glacier in east Antarctica revealed by deep clustering[J]. Journal of Geophysical Research: Earth Surface2024129(4). DOI: 10.1029/2023JF007593 .
[81]
SONG X Y WANG Z M LIANG J C, et al. Automatic extraction of the basal channel based on neural network[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing202215: 5 013-5 023.
[82]
CHEN G X KUSKY T LUO L, et al. Hadean tectonics: insights from machine learning[J]. Geology202351(8): 718-722.
[83]
ZHANG S K GONG L ZENG Q, et al. Imputation of GPS coordinate time series using missForest[J]. Remote Sensing202113(12). DOI: 10.3390/rs13122312 .
[84]
ZHANG Q C LI F ZHANG S K, et al. Modeling and forecasting the GPS zenith troposphere delay in west Antarctica based on different blind source separation methods and deep learning[J]. Sensors202020(8). DOI: 10.3390/s20082343 .
[85]
RAN Youhua LI Xin CHE Tao,et al. Recent progress and emerging frontiers in China’s cryosphere remote sensing research[J]. National Remote Sensing Bulletin202529(6):1 831-1 847.
冉有华, 李新, 车涛, 等. 中国冰冻圈遥感近期研究进展与若干前沿问题探讨[J]. 遥感学报202529(6): 1 831-1 847.
[86]
CHENG Xiao CHEN Zhuogi HUI Fengming, et al. Status and outlook of China’s space-based and airborne remote sensing systems for polar regions[J]. Science and Technology Foresight2022(2): 183-197.
程晓, 陈卓奇, 惠凤鸣, 等. 中国极地空天基遥感观测现状与展望[J]. 前瞻科技2022(2): 183-197.
[87]
HAO Tong WANG Xiaofeng FENG Tiantian, et al. Intelligent and multi-scale surveying of key areas and processes of the Earth system[J]. Acta Geodaetica et Cartographica Sinica202150(8): 1 084-1 095.
郝彤, 王晓峰, 冯甜甜, 等. 地球系统多尺度关键区域与关键过程的智能化测绘[J]. 测绘学报202150(8): 1 084-1 095.
[1] 刘定洲, 刘虹辰, 张锦昌, 张江阳, 周鹏程, 郭鑫, 李璐, 黄彦铭. 基于U-Net卷积神经网络的海底磁异常条带自动识别方法研究——以沙茨基海隆为例[J]. 地球科学进展, 2025, 40(8): 864-876.
[2] 周杰, 巨能攀, 张燕, 田华兵, 何朝阳. 泥石流视频图像跟踪检测方法研究[J]. 地球科学进展, 2025, 40(4): 388-400.
[3] 王凡, 张旭东, 任沂斌, 刘颖洁, 王浩宇, 李晓峰. 基于人工智能技术的海洋智能预报研究进展[J]. 地球科学进展, 2025, 40(2): 111-125.
[4] 张井勇. 面向碳中和的气候变化预估与风险研究新框架[J]. 地球科学进展, 2025, 40(1): 15-20.
[5] 龙清云, 张通, 王彻, 车涛, 效存德. 基于机器学习和动力过程的南极冰架崩解特征分析[J]. 地球科学进展, 2024, 39(8): 837-846.
[6] 张宇翱, 张旭, 昝金波, 方小敏. 冰期指数法模拟北半球冰盖演化的不确定性研究[J]. 地球科学进展, 2023, 38(6): 619-630.
[7] 张永垂, 陈诗尧, 王宁, 汪浩笛, 周林, 王本洪. 全球业务化海洋预报系统进展[J]. 地球科学进展, 2022, 37(4): 344-357.
[8] 屈猛, 赵羲, 庞小平, 雷瑞波. 北极冰间水道区域的物理过程和遥感观测研究进展[J]. 地球科学进展, 2022, 37(4): 382-391.
[9] 马雷鸣. 天气预报中的人工智能技术进展[J]. 地球科学进展, 2020, 35(6): 551-560.
[10] 吴延俊, 赵进平. 欧亚海盆大西洋水输运过程及热释放研究进展[J]. 地球科学进展, 2020, 35(3): 231-245.
[11] 胡利民,石学法,叶君,张钰莹. 北极东西伯利亚陆架沉积有机碳的源汇过程研究进展[J]. 地球科学进展, 2020, 35(10): 1073-1086.
[12] 效存德,陈卓奇,江利明,丁明虎,窦挺峰. 格陵兰冰盖监测、模拟及气候影响研究[J]. 地球科学进展, 2019, 34(8): 781-786.
[13] 王汝建, 肖文申, 章陶亮, 聂森艳. 极地地质钻探研究进展与展望[J]. 地球科学进展, 2017, 32(12): 1236-1244.
[14] 聂红涛, 王蕊, 赵伟, 罗晓凡, 祁第, 鹿有余, 张远辉, 魏皓. 北冰洋太平洋扇区碳循环变化机制研究面临的关键科学问题与挑战[J]. 地球科学进展, 2017, 32(10): 1084-1092.
[15] 汪燕敏, 祁第, 陈立奇. 南大洋酸化指标——海水文石饱和度变异的研究进展[J]. 地球科学进展, 2016, 31(4): 357-364.
阅读次数
全文


摘要