Zhao et al. (2026) Utilizing Multi-Source Remote Sensing Data and the CGAN to Identify Key Drought Factors Influencing Maize Across Distinct Phenological Stages

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

Identification

• Journal: Remote Sensing
• Year: 2026
• Date: 2026-04-03
• Authors: Hui Zhao, Jifu Guo, Jing Jiang, Funian ZHAO, Xiaoyong Yang
• DOI: 10.3390/rs18071085

Research Groups

Not explicitly mentioned in the provided text.

Short Summary

This study integrates an improved Conditional Generative Adversarial Network (CGAN) with SHapley Additive exPlanations (SHAP) and multi-source remote sensing data to accurately identify the dominant environmental factors driving maize drought stress at different growth stages in rain-fed Northwest China, revealing a dynamic evolution of these factors across phenological stages.

Objective

• To accurately identify the dominant environmental factors that drive drought stress at different growth stages of maize in the rain-fed agricultural region of Northwest China, by quantifying the nonlinear interactions among multiple environmental factors.

Study Configuration

• Spatial Scale: Rain-fed agricultural region of Northwest China.
• Temporal Scale: Different maize phenological stages (seedling, jointing–tasseling, and maturity stages).

Methodology and Data

• Models used: Improved Conditional Generative Adversarial Network (CGAN) for high-precision drought severity estimation; SHapley Additive exPlanations (SHAP) for interpretability and quantitative analysis of dominant factors.
• Data sources: Multi-source remote sensing data, including:
  • Temperature (maximum, minimum, mean: Tmax, Tmin, Tmean)
  • Precipitation (P)
  • Evapotranspiration (ET)
  • Soil moisture at 0–0.1 m (SM0–10)
  • Soil moisture at 0.1–0.4 m (SM10–40)
  • Solar-induced chlorophyll fluorescence (maximum, minimum, mean: SIFmax, SIFmin, SIFmean)

Main Results

• The CGAN model demonstrated excellent nonlinear modeling capability under small samples, achieving high coefficients of determination (R²) for maize drought severity estimation:
  • Seedling stage: R² = 0.963
  • Jointing–tasseling stage: R² = 0.972
  • Maturity stage: R² = 0.979
• SHAP analysis revealed a clear dynamic evolution of dominant factors across phenological stages:
  • Seedling stage: Dominated by Evapotranspiration (ET), reflecting the primary role of surface water–heat balance.
  • Jointing–tasseling stage: Co-dominated by ET, topsoil moisture (SM0–10), and minimum SIF, indicating intensified crop transpiration and physiological stress under meteorological drought.
  • Maturity stage: Shifted to an absolute dominance centered on mean temperature (Tmean), highlighting the critical impact of heat stress.

Contributions

• Provides a data-driven quantitative perspective for understanding maize drought mechanisms.
• Offers a scientific basis for formulating differentiated drought management strategies tailored for different maize growth stages.
• Demonstrates the potential and effectiveness of integrating CGAN with SHAP for agricultural remote sensing and drought attribution research, particularly in data-scarce regions.

Funding

Not explicitly mentioned in the provided text.

Citation

@article{Zhao2026Utilizing,
  author = {Zhao, Hui and Guo, Jifu and Jiang, Jing and ZHAO, Funian and Yang, Xiaoyong},
  title = {Utilizing Multi-Source Remote Sensing Data and the CGAN to Identify Key Drought Factors Influencing Maize Across Distinct Phenological Stages},
  journal = {Remote Sensing},
  year = {2026},
  doi = {10.3390/rs18071085},
  url = {https://doi.org/10.3390/rs18071085}
}