Zhao et al. (2025) A Hybrid Modeling Approach for Improved Simulation of Thermal‐Hydrological Dynamics in Active Layer on the Qinghai‐Tibet Plateau

⚠️ Warning: This summary was generated from the abstract only, as the full text was not available.

Identification

• Journal: Water Resources Research
• Year: 2025
• Date: 2025-11-27
• Authors: Yi Zhao, Zhuotong Nan, Hailong Ji, Yuhong Chen, Minyue Ou, Dongfeng Li
• DOI: 10.1029/2025wr040288

Research Groups

Not explicitly mentioned in the abstract.

Short Summary

This study presents a novel hybrid modeling approach combining the SHAW model with random forest-corrected Noah LSM simulations to accurately model active layer dynamics on the Qinghai-Tibet Plateau, significantly outperforming existing models in simulating soil temperature and moisture.

Objective

• To develop and evaluate a novel hybrid modeling approach that enables accurate simulation of active layer dynamics in data-sparse permafrost regions, specifically on the Qinghai-Tibet Plateau, by integrating advanced physical process representation with robust lower boundary conditions.

Study Configuration

• Spatial Scale: Seven permafrost sites on the Qinghai-Tibet Plateau for evaluation; framework designed for large-scale simulation of permafrost dynamics.
• Temporal Scale: Continuous simulation of active layer dynamics, including freezing periods, over an unspecified duration.

Methodology and Data

• Models used: Simultaneous Heat and Water (SHAW) model, Noah Land Surface Model (LSM), Random Forest (RF) algorithm.
• Data sources: Observational data from seven permafrost sites on the Qinghai-Tibet Plateau for evaluation. Forcing data generation from Noah LSM simulations, corrected by random forest, to address data-sparse regions.

Main Results

• The hybrid modeling approach significantly outperformed both the standalone Noah LSM and traditional SHAW configurations in simulating active layer temperature and moisture.
• For testing data, the hybrid approach achieved higher average Nash-Sutcliffe efficiency (NSE) values for soil temperature (0.81) compared to the Noah LSM (0.69).
• For testing data, the hybrid approach achieved higher average Nash-Sutcliffe efficiency (NSE) values for soil moisture (0.35) compared to the Noah LSM (0.17).
• The hybrid approach effectively corrected key biases of the Noah LSM, which tends to overestimate soil temperature and unfrozen water content during freezing periods.

Contributions

• Introduction of a novel hybrid modeling framework that integrates a physically advanced model (SHAW) with a large-scale land surface model (Noah LSM), enhanced by machine learning (random forest) for robust lower boundary conditions.
• Provides a robust framework specifically designed for large-scale simulation of permafrost dynamics in data-sparse regions, addressing a critical limitation of existing models.
• Demonstrates significant improvements in simulating active layer temperature and moisture compared to standalone models, particularly in correcting biases during freezing periods.
• Offers direct implications for assessing environmental change, infrastructure risks, and carbon emissions linked to permafrost degradation.

Funding

Not explicitly mentioned in the abstract.

Citation

@article{Zhao2025Hybrid,
  author = {Zhao, Yi and Nan, Zhuotong and Ji, Hailong and Chen, Yuhong and Ou, Minyue and Li, Dongfeng},
  title = {A Hybrid Modeling Approach for Improved Simulation of Thermal‐Hydrological Dynamics in Active Layer on the Qinghai‐Tibet Plateau},
  journal = {Water Resources Research},
  year = {2025},
  doi = {10.1029/2025wr040288},
  url = {https://doi.org/10.1029/2025wr040288}
}