Huang et al. (2026) Upwind terrestrial influences on soil moisture variability across South America

Identification

• Journal: Nature Communications
• Year: 2026
• Date: 2026-07-24
• Authors: Feini Huang, Shijie Jiang, Wei Shangguan, Gustau Camps‐Valls, Alexander J. Winkler, Wantong Li, Grégory Duveiller, Christian Reimers, Markus Reichstein, Yongjiu Dai
• DOI: 10.1038/s41467-026-75637-x

Research Groups

• School of Atmospheric Sciences, Sun Yat-Sen University, Zhuhai, China.
• Department of Biogeochemical Integration, Max Planck Institute for Biogeochemistry, Jena, Germany.
• ELLIS Unit Jena, Jena, Germany.
• Guangdong Province Key Laboratory for Climate Change and Natural Disaster Studies, Sun Yat-Sen University, Zhuhai, China.
• Southern Marine Science and Engineering Guangdong Laboratory (Zhuhai), Zhuhai, China.
• Image Processing Lab (IPL), Universitat de València, Valencia, Spain.
• Department of Environmental Science, Policy and Management, UC Berkeley, Berkeley, CA, USA.
• Climate and Ecosystem Sciences Division, Lawrence Berkeley National Laboratory, Berkeley, CA, USA.

Short Summary

This study develops an observation-driven framework to quantify the remote influence of vegetation on soil moisture variability across South America. The framework integrates atmospheric transport trajectories with deep learning to estimate vegetation's contribution to downwind soil moisture, revealing a systematic continent-scale source-sink architecture and non-linear regimes of vegetation influence.

Objective

• Investigate how upwind terrestrial conditions shape soil moisture variability downwind.
• Quantify the remote influence of vegetation on soil moisture across South America.

Study Configuration

• Spatial Scale: Continental scale (South America).
• Temporal Scale: Monthly time resolution, with a focus on the period 2001-2018.

Methodology and Data

• Models used: Transport-informed model integrating atmospheric transport trajectories with deep learning.
• Data sources: Soil moisture products (SoMo.ml and ESA CCI), atmospheric transport data from ERA5, and leaf area index (LAI) anomalies from satellite observations.

Main Results

• The transport-informed model outperforms local-only and spatial-context models in predicting soil moisture variability across South America.
• Upwind vegetation information contributes 67% of predictive power to the transport-informed model.
• Vegetation's downwind influence operates through distinct, climate-dependent, and non-linear regimes (R1-R4).
• The source-sink architecture reveals a predominantly positive and spatially coherent structure, with Amazonian vegetation dynamics modulating soil moisture regimes hundreds to thousands of kilometers downwind.

Contributions

• This study provides observation-based evidence for the remote influence of vegetation on soil moisture variability across South America.
• The framework developed in this study can be applied to other large-scale forests to investigate comparable teleconnections and hydrological resilience.

Funding

• Not specified.

Citation

@article{Huang2026Upwind,
  author = {Huang, Feini and Jiang, Shijie and Shangguan, Wei and Camps‐Valls, Gustau and Winkler, Alexander J. and Li, Wantong and Duveiller, Grégory and Reimers, Christian and Zhao, Wenli and Reichstein, Markus and Dai, Yongjiu},
  title = {Upwind terrestrial influences on soil moisture variability across South America},
  journal = {Nature Communications},
  year = {2026},
  doi = {10.1038/s41467-026-75637-x},
  url = {https://doi.org/10.1038/s41467-026-75637-x}
}