数字农科院2.0

Sub-daily global vegetation optical depth reconstruction from SMAP using 3D partial convolutional networks

文献类型： 外文期刊

作者： Kou, Jixiang;Wei, Zushuai;Guan, Linjie;Yang, Beibei;Leng, Pei;Zhang, Jianxin;Zhao, Tianjie;Meng, Lingkui

作者机构：

关键词： Vegetation optical depth;3D partial convolutional neural network;spatial gap-filling;spatiotemporal reconstruction;daily seamless

期刊名称： INTERNATIONAL JOURNAL OF REMOTE SENSING

ISSN： 0143-1161

年卷期： 2026 年

页码：

收录情况： SCIE（2025版） ; ; EI（2025版）

摘要： Vegetation Optical Depth (VOD) observed by spaceborne microwave radiometers is closely related to vegetation density and moisture content. Time-series remotely sensed VOD have been widely used in global and regional vegetation change research. However, the impact of satellite orbital gaps leads to significant information loss, resulting in temporally discontinuous and spatially incomplete VOD time-series data, which limits further applications of the data. To address this issue, this paper proposes a VOD reconstruction model based on a 3D Partial Convolutional Neural Network (3DPCNN). The 3DPCNN model ful ly leverages the spatiotemporal variations of VOD, utilizing the advantages of partial convolutional networks in handling complex nonlinear problems to spatially fill gaps in ascending and descending VOD data from the SMAP satellite, yielding a seamless global twice-daily VOD product from 2015 to 2022. Results show that the reconstruction effectively restores missing VOD values while maintaining spatial variability. In simulated missing experiments, the reconstructed VOD preserves the spatial variation characteristics of the original data, achieving an unbiased root mean square difference (ubRMSD) between 0.014 and 0.041, a correlation coefficient (R) from 0.943 to 0.991, and a Bias ranging from 0.0005 to -0.0250 when compared with the original SMAP VOD. The seamless, twice-daily global VOD data produced in this study captures vegetation moisture diurnal variation, reproducing the spatiotemporal distribution of vegetation water-related characteristics globally, supporting a deeper understanding of physiological response mechanisms of vegetation in various regions under drought conditions.

分类号：

  • 相关文献
作者其他论文 更多>>