Feng et al. (2026) Transferring soil moisture estimation skills to evapotranspiration and streamflow modeling through remote sensing data assimilation

Identification

• Journal: Remote Sensing of Environment
• Year: 2026
• Date: 2026-01-31
• Authors: Huihui Feng, Jianhong Zhou, Zhiyong Wu, Jianzhi Dong, Long Zhao, Luca Brocca, Hai He
• DOI: 10.1016/j.rse.2026.115274

Research Groups

• Chongqing Jinfo Mountain Karst Ecosystem National Observation and Research Station, School of Geographical Sciences, Southwest University, Chongqing 400715, China
• National Key Laboratory of Water Disaster Prevention, College of Hydrology and Water Resources, Hohai University, Nanjing 210098, China
• School of Earth System Science, Tianjin University, Tianjin 300072, China
• Research Institute for Geo-Hydrological Protection, National Research Council, Perugia 06128, Italy

Short Summary

This study introduces an improved soil moisture (SM) data assimilation (DA) framework that first calibrates the coupling strengths between SM and hydrological fluxes (evapotranspiration and runoff) within a land surface model using remote sensing (RS) data. Subsequently, RS SM retrievals are assimilated into the calibrated model, significantly enhancing the simulation accuracy of SM, evapotranspiration, and streamflow, particularly in (sub-)humid regions.

Objective

• To develop an improved soil moisture data assimilation framework that optimizes the coupling strengths between soil moisture and hydrological fluxes within a land surface model to enhance the transferability of soil moisture estimation skills to evapotranspiration and streamflow modeling.

Study Configuration

• Spatial Scale: Regional (improvements especially evident in (sub-)humid regions)
• Temporal Scale: Continuous time series

Methodology and Data

• Models used: Variable Infiltration Capacity (VIC) model, Ensemble Kalman Filter (EnKF)
• Data sources: Remote sensing (RS) soil moisture (SM) retrievals

Main Results

• The developed SM DA framework enhanced DA efficiency, increasing SM correlation from 0.45 to 0.49.
• Hydrological flux simulations were significantly improved: evapotranspiration (ET) correlation increased from 0.77 to 0.80, and the Nash-Sutcliffe efficiency (NSE) for streamflow improved from 0.21 to 0.71, relative to the default VIC scheme.
• These improvements were particularly pronounced in (sub-)humid regions, where the VIC model's saturation-excess runoff generation mechanism is well-suited.

Contributions

• Introduces a novel SM DA framework that explicitly incorporates the optimization of land surface model (LSM) coupling strengths between soil moisture and hydrological fluxes (SM-ET and SM-runoff) using remote sensing data prior to assimilation.
• Demonstrates that calibrating these coupling strengths significantly enhances the effectiveness of SM data assimilation in improving estimates of key hydrological fluxes like evapotranspiration and streamflow, addressing a persistent limitation in existing SM DA systems.
• Provides a promising pathway to improve hydrological flux simulations through land data assimilation by focusing on the physical consistency and representation of land surface processes within LSMs.

Funding

Not specified in the provided text.

Citation

@article{Feng2026Transferring,
  author = {Feng, Huihui and Zhou, Jianhong and Wu, Zhiyong and Dong, Jianzhi and Zhao, Long and Brocca, Luca and He, Hai},
  title = {Transferring soil moisture estimation skills to evapotranspiration and streamflow modeling through remote sensing data assimilation},
  journal = {Remote Sensing of Environment},
  year = {2026},
  doi = {10.1016/j.rse.2026.115274},
  url = {https://doi.org/10.1016/j.rse.2026.115274}
}