您好，欢迎访问上海市农业科学院 机构知识库！
上海市农业科学院机构知识库

全部

  • 全部
  • 标题
  • 学者
  • 机构
  • 关键词

A deep learning approach for SMAP soil moisture downscaling informed by thermal inertia theory

文献类型： 外文期刊

作者： Xu, Mengyuan 1 ; Yang, Haoxuan 2 ; Hu, Annan 3 ; Heng, Lee 4 ; Li, Linyi 1 ; Yao, Ning 5 ; Liu, Gang 6 ;

作者机构： 1.Shanghai Acad Agr Sci, Inst Agr Sci & Technol Informat, Shanghai, Peoples R China

2.Oklahoma State Univ, Dept Nat Resource Ecol & Management, Stillwater, OK 74078 USA

3.Univ Alberta, Land Reclamat & Remediat, Edmonton, AB, Canada

4.Int Atom Energy Agcy IAEA, Joint FAO IAEA Div Nucl Tech Food & Agr, Soil & Water Management & Crop Nutr Subprogramme, Vienna, Austria

5.Northwest Agr & Forestry Univ, Coll Water Resources & Architectural Engn, Yangling, Peoples R China

6.China Agr Univ, Coll Land Sci & Technol, Beijing, Peoples R China

7.Minist Agr & Rural Affairs, Key Lab Intelligent Agr Technol Yangtze River Delt, Shanghai, Peoples R China

关键词： SMAP soil moisture; Spatial downscaling; Physical mechanism; Thermal inertia theory; Deep learning

期刊名称：INTERNATIONAL JOURNAL OF APPLIED EARTH OBSERVATION AND GEOINFORMATION （ 影响因子：8.6； 五年影响因子：8.6 ）

ISSN： 1569-8432

年卷期： 2025 年 136 卷

页码：

收录情况： SCI

摘要： Deep learning (DL) based methods have recently made remarkable progress in remote sensing (RS) soil moisture (SM) retrieval applications. However, their purely "black box" algorithms suffer from a lack of interpretability, whereas methods based solely on physical mechanisms often underperform in complex scenarios. In this study, we attempt to use an SM downscaling approach that integrates thermal inertia (TI) theory with the DenseNet deep network algorithm. This approach provides partial interpretability of the physical mechanisms while utilizing DenseNet's superior nonlinear learning ability and feature reuse capability for downscaling the Soil Moisture Active Passive (SMAP) satellite product, generating daily 1 km x 1 km SM. A comprehensive assessment of the downscaled results using in-situ SM acquired from 264 International Soil Moisture Network (ISMN) sites densely distributed across the continental U.S. indicated that this downscaling approach had an overall high accuracy, with an average unbiased root mean square error (ubRMSE) of 0.048 m3/m3. In addition, the downscaled SM exhibited marked improvement in spatial details over the original SMAP SM maps, providing clearer land surface features. The proposed SM downscaling approach is a valuable attempt to adopt DL methods with more practical physical meaning and more interpretability in the current RS Big Data era.

  • 相关文献
作者其他论文 更多>>
置顶
购物车
反馈

© 京ICP备09089781号-29 主办单位：上海市农业科学院

学术资源： 中国农业科学院 农业专业知识服务系统 农科联盟资源共建共享平台 中国农业科技文献与信息服务平台 中国农业期刊集成服务平台