检索
E-mail
RSS

作者投稿 主编审稿 专家审稿 远程编辑

热带海洋学报

• • 上一篇    下一篇

基于Earthformer模型的南海海洋溶解氧预测

孟庆存1, 2, 3, 4, 刘金龙1, 2, 罗云1, 2, 黎刚1, 2, 颜文1, 2, 4, 殷建平1, 徐维海1, 2, 3*
  

  1. 1. 热带海洋环境与岛礁生态全国重点实验室, 中国科学院南海海洋研究所, 广东 广州 511458;

    2. 边缘海与大洋地质实验室, 中国科学院南海海洋研究所, 广东 广州 511458;

    3. 三亚海洋生态环境工程研究院, 海南 三亚 572000;

    4. 中国科学院大学, 北京 100049



  • 收稿日期:2025-10-31 修回日期:2026-02-04 接受日期:2026-03-05
  • 通讯作者: 徐维海
  • 基金资助:

    国家自然科学基金(42376079、42176079); 海南省自然科学基金创新研究团队(422CXTD533); 中国科学院南海海洋研究所自主部署项目(SCSIO2023QY05)

Ocean dissolved oxygen forecasting in the South China Sea based on the Earthformer model

MENG Qingcun1, 2, 3, 4, LIU Jinlong1, 2, LUO Yun1, 2, LI Gang1, 2, YAN Wen1, 2, 4, YIN Jianping1, XU Weihai1, 2, 3*   

  1. 1. State Key Laboratory of Tropical Oceanography, South China Sea Institute of Oceanology, Chinese Academy of Sciences, Guangdong Guangzhou 511458, China;

    2. Laboratory of Ocean and Marginal Sea Geology, South China Sea Institute of Oceanography, Chinese Academy of Sciences, Guangdong Guangzhou 511458, China;

    3. Sanya Institute of Ocean Eco-Environmental Engineering, Hainan Sanya, 572000, China;

    4. University of Chinese Academy of Sciences, Beijing 100049, China



  • Received:2025-10-31 Revised:2026-02-04 Accepted:2026-03-05
  • Supported by:

     the National Natural Science Foundation of China (42376079, 42176079); Hainan Provincial Natural Science Foundation of China (422CXTD533); the Special Fund of the South China Sea Institute of Oceanology of the Chinese Academy of Sciences (SCSIO2023QY05)

PDF (PC)

赞

摘要： 因全球持续高温驱动的海洋大范围低氧/缺氧，已成为本世纪全球重点关注的科学问题之一。传统的海洋溶解氧（dissolved oxygen, DO）监测依赖于现场采样，存在成本高、耗时长，且难以实现大范围观测的问题。尽管已有通过传统机器学习方法实现海洋DO数据重建的研究，但针对DO时空演化规律的深度学习预测研究仍存在明显不足，现有模型难以有效捕捉海洋环境的时空动态关联性与异质性。本研究以南海为研究区域，采用专为地学问题设计的Earthformer架构模型，基于1998—2022年的多源遥感与再分析数据，构建了一个包含海表温度、盐度、叶绿素a浓度、海水流速以及溶解氧的综合数据集。基于上述数据集，本研究对Earthformer模型进行了系统的深入训练与微调，实现了通过历史数据来预测未来（12个月）海洋DO时空分布的目标。实验结果表明，优化后的Earthformer模型具备优异的溶解氧预测性能。在测试集上，模型的多个评价指标整体表现突出：结构相似性指数达到0.78左右，反标准化后的均方根误差和平均绝对误差均可控制在3μmol·kg⁻¹以下的较低误差区间。此外，通过消融实验明确了关键超参数对模型预测性能的调控效应。本研究首次验证了Earthformer模型在海洋DO（典型复杂海洋生物地球化学变量）长期时空预测中的适用性与优越性，填补了深度学习模型在该研究领域的应用不足，为数据驱动的海洋低氧/缺氧环境预报提供了新的技术方法与理论支撑，兼具重要科学价值与实践应用潜力。

关键词: 深度学习, Earthformer, 溶解氧预测, 遥感, 南海

Abstract: Driven by continuous global warming, large-scale ocean deoxygenation/hypoxia has emerged as one of the key global scientific issues of this century. Conventional monitoring of marine dissolved oxygen (DO) relies on in-situ sampling, which is constrained by high costs, long time consumption, and challenges in achieving large-scale observations. Although there have been studies on the reconstruction of marine dissolved oxygen data through traditional machine learning methods, there remains a significant deficiency in deep learning-based forecasting research targeting the spatio-temporal evolution patterns of dissolved oxygen, as existing models struggle to effectively capture the spatio-temporal dynamic associations and heterogeneity of the marine environment. Focusing on the South China Sea, this study employs the Earthformer architecture, a model specifically designed for geoscience problems. Based on multi-source remote sensing and reanalysis data from 1998 to 2022, a comprehensive dataset was constructed, including sea surface temperature, salinity, chlorophyll-a concentration, currents, and dissolved oxygen. Using the aforementioned dataset, this study systematically trained and fine-tuned the Earthformer model, achieving the objective of forecasting the future (12 months) spatio-temporal distribution of marine dissolved oxygen from historical data. The experimental results indicate that the optimized Earthformer model possesses excellent predictive performance for dissolved oxygen. On the test set, the model demonstrated outstanding performance across multiple evaluation metrics, achieving a structural similarity index measure (SSIM) of approximately 0.78, while the un-normalized root mean square error (RMSE) and mean absolute error (MAE) were both maintained below 3μmol·kg⁻¹. Furthermore, ablation experiments revealed the impact of key hyperparameters on the model's prediction performance. This study, for the first time, validates the applicability and superiority of the Earthformer model in the long-term spatio-temporal prediction of marine dissolved oxygen (a typical complex marine biogeochemical variable), addressing the insufficiency of deep learning model applications in this research area. It provides new technical methods and theoretical support for data-driven marine hypoxia/anoxia environmental forecasting, thus possessing both significant scientific value and practical application potential.

Key words: Deep Learning, Earthformer, Dissolved Oxygen Prediction, Remote Sensing, South China Sea