中国科学院机构知识库网格
Chinese Academy of Sciences Institutional Repositories Grid
首页 机构 成果 学者
  
  1. CAS IR Grid
  2. 海洋研究所
  3. 中国科学院海洋研究所
  4. 海洋环流与波动重点实验室
A prior-knowledge-integrated downscaling approach for subsurface thermal structure reconstruction in the tropical Indian Ocean

文献类型：期刊论文

作者Feng, Zhongkun1,2,3; Qi, Jifeng2,3,4,6; Xie, Bowen1,2,3,5; Cao, Yuan7; Li, Delei8; Liu, Chuanyu2,3,4,6; Yin, Baoshu2,3,4,6
刊名DEEP-SEA RESEARCH PART II-TOPICAL STUDIES IN OCEANOGRAPHY
出版日期2026-02-01
卷号225页码:14
关键词Downscaling Ocean subsurface thermal structure Deep learning Tropical Indian Ocean
ISSN号0967-0645
DOI10.1016/j.dsr2.2025.105589
通讯作者Qi, Jifeng(jfqi@qdio.ac.cn)
英文摘要Accurately obtaining high-resolution ocean subsurface thermal structure (OSTS) is essential for resolving mesoscale dynamics in the tropical Indian Ocean (TIO), yet observations remain sparse and uneven. We present a Transformer-based neural network model, the Downscaling Vision Transformer (DSVIT), which integrates prior knowledge to reconstruct high-resolution OSTS in the TIO. Inputs include sea surface temperature (SST), absolute dynamic topography (ADT), and wind stress curl (WSC), as well as temporal, geographic, and climatological information. DSVIT enhances a standard Vision Transformer (ViT) with a geographic positional prior and a physics-aware loss that emphasizes thermocline and surface variability. On an independent test set, DSVIT achieves a Root Mean Square Error (RMSE) of 0.29 degrees C and a Coefficient of Determination (R2) of 0.9962 for reconstructed subsurface temperature, outperforming traditional recurrent neural network (RNN) and convolutional neural network (CNN) models. Moreover, a key innovation of this study lies in its novel downscaling strategy, which effectively improves the EN4 subsurface temperature resolution from 1 degrees to 1/4 degrees by altering the input segmentation. Compared with traditional interpolation and assessed against independent high-resolution products, the downscaled outputs exhibit lower RMSE and higher R2, indicating enhanced physical consistency and mesoscale representation. SHapley Additive exPlanations (SHAP) analysis further reveals that climatology and SST are the dominant predictors, followed by ADT. This study provides a novel approach for downscaling OSTS and offers valuable insights for advancing oceanic and climatic research.
WOS关键词IN-SITU ; MODEL OUTPUT ; TEMPERATURE ; CIRCULATION ; SALINITY ; IMPACTS
资助项目National Natural Science Foundation of China[42476024] ; National Natural Science Foundation of China[42176010] ; Strategic Priority Research Program of the Chinese Academy of Sciences[XDA0370301] ; National Key Research and Development Program of China[2022YFF0801400]
WOS研究方向Oceanography
语种英语
WOS记录号WOS:001660736900001
出版者PERGAMON-ELSEVIER SCIENCE LTD
源URL[http://ir.qdio.ac.cn/handle/337002/204565]  
专题海洋研究所_海洋环流与波动重点实验室
通讯作者Qi, Jifeng
作者单位1.Qingdao Univ Sci & Technol, Sch Math & Phys, Qingdao 266061, Peoples R China
2.Chinese Acad Sci, Inst Oceanol, Lab Ocean Circulat & Waves, Qingdao 266071, Peoples R China
3.Chinese Acad Sci, Inst Oceanol, Key Lab Ocean Observat & Forecasting, Qingdao 266071, Peoples R China
4.Qingdao Marine Sci & Technol Ctr, Lab Ocean Dynam & Climate, Qingdao 266237, Peoples R China
5.Qingdao Univ Sci & Technol, Coll Electromech Engn, Qingdao 266061, Peoples R China
6.Univ Chinese Acad Sci, Beijing 100049, Peoples R China
7.Univ Hong Kong, Sch Comp & Data Sci, Hong Kong 999077, Peoples R China
8.Laoshan Lab, Qingdao, Peoples R China
推荐引用方式
GB/T 7714		 Feng, Zhongkun,Qi, Jifeng,Xie, Bowen,et al. A prior-knowledge-integrated downscaling approach for subsurface thermal structure reconstruction in the tropical Indian Ocean[J]. DEEP-SEA RESEARCH PART II-TOPICAL STUDIES IN OCEANOGRAPHY,2026,225:14.
APA Feng, Zhongkun.,Qi, Jifeng.,Xie, Bowen.,Cao, Yuan.,Li, Delei.,...&Yin, Baoshu.(2026).A prior-knowledge-integrated downscaling approach for subsurface thermal structure reconstruction in the tropical Indian Ocean.DEEP-SEA RESEARCH PART II-TOPICAL STUDIES IN OCEANOGRAPHY,225,14.
MLA Feng, Zhongkun,et al."A prior-knowledge-integrated downscaling approach for subsurface thermal structure reconstruction in the tropical Indian Ocean".DEEP-SEA RESEARCH PART II-TOPICAL STUDIES IN OCEANOGRAPHY 225(2026):14.

入库方式： OAI收割

来源：海洋研究所

浏览0
下载0
收藏0
其他版本

除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。