Dai et al. (2026) Identification of Key Factors Driving Dissolved Oxygen in Riparian Aquifers Through Deep Learning‐Assisted Global Sensitivity Analysis

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

Identification

• Journal: Water Resources Research
• Year: 2026
• Date: 2026-01-30
• Authors: Heng Dai, Yijie Yang, Fangqiang Zhang, Alberto Guadagnini, Jing Yang, Xiaochuang Bu, Liang Wang, Songhu Yuan, Ming Ye
• DOI: 10.1029/2025wr041884

Research Groups

Not specified in the abstract.

Short Summary

This study employs a global sensitivity analysis (GSA) framework, enhanced with deep learning surrogate models, to identify the dominant physical and biogeochemical controls on dissolved oxygen (DO) dynamics in riparian aquifers, revealing that river stage dynamics are the primary drivers.

Objective

• To identify the dominant physical and biogeochemical controls on dissolved oxygen (DO) dynamics in riparian aquifers.

Study Configuration

• Spatial Scale: Riparian aquifer system (high-resolution model).
• Temporal Scale: Dynamics of river stage fluctuations (period and amplitude).

Methodology and Data

• Models used: Coupled flow and transport models; Global Sensitivity Analysis (GSA) framework integrating Bayesian network-based and variance-based methods; Deep learning surrogate models (multi-layer perceptrons, convolutional neural networks).
• Data sources: Not explicitly specified in the abstract, but a high-resolution model was used.

Main Results

• River stage dynamics (period and amplitude of water level fluctuations) are identified as the primary drivers of dissolved oxygen (DO) supply to the aquifer system.
• Hydraulic conductivity, riverine DO concentration, and the maximum DO reaction rate exhibit important but localized effects, influencing specific transport pathways including river water infiltration, entrapped air dissolution, and diffusion through the unsaturated zone.
• Parameters such as porosity, longitudinal dispersion, and van Genuchten soil parameters show negligible influence on DO dynamics.

Contributions

• Development and application of a comprehensive Global Sensitivity Analysis (GSA) framework integrating Bayesian network-based and variance-based methods with deep learning surrogate models for efficient evaluation of complex environmental systems.
• Identification of dominant physical and biogeochemical controls on dissolved oxygen (DO) dynamics in riparian aquifers, specifically highlighting the primary role of river stage dynamics.
• Provides guidance for model simplification and diagnosis in complex environmental systems.

Funding

Not specified in the abstract.

Citation

@article{Dai2026Identification,
  author = {Dai, Heng and Yang, Yijie and Zhang, Fangqiang and Guadagnini, Alberto and Yang, Jing and Bu, Xiaochuang and Wang, Liang and Yuan, Songhu and Ye, Ming},
  title = {Identification of Key Factors Driving Dissolved Oxygen in Riparian Aquifers Through Deep Learning‐Assisted Global Sensitivity Analysis},
  journal = {Water Resources Research},
  year = {2026},
  doi = {10.1029/2025wr041884},
  url = {https://doi.org/10.1029/2025wr041884}
}