Hu et al. (2026) Atmospheric dryness and flash drought severity drive the shifts of different flash drought types into agricultural droughts

Identification

• Journal: Agricultural Water Management
• Year: 2026
• Date: 2026-01-06
• Authors: Chen Hu, Dunxian She, Liping Zhang, Gangsheng Wang, Zhaoxia Jing, Si Hong, Jun Xia
• DOI: 10.1016/j.agwat.2025.110109

Research Groups

• State Key Laboratory of Water Resources Engineering and Management, Wuhan University, Wuhan, PR China
• Institute for Water-Carbon Cycles and Carbon Neutrality, Wuhan University, Wuhan, PR China
• Changjiang Water Resources Protection Institute, Wuhan, PR China

Short Summary

This study developed an integrated framework to investigate the propagation pathways and underlying mechanisms of meteorological, soil, and evaporative flash droughts into agricultural droughts in the Middle and Lower Reaches of the Yangtze River Basin from 2000 to 2022. It found strong causal relationships with varying propagation times and identified flash drought severity, precipitation, and vapor pressure deficit as dominant drivers with specific triggering thresholds.

Objective

• To recognize the relationships, propagation time, and propagation processes between meteorological, soil, evaporative flash droughts and agricultural droughts.
• To identify the primary drivers influencing the propagation from different types of flash droughts to agricultural droughts.
• To estimate threshold values of the dominant drivers that trigger propagation for different flash drought types.

Study Configuration

• Spatial Scale: Middle and Lower Reaches of the Yangtze River Basin (MLRYRB), covering approximately 0.8 million square kilometers, located between 24°30’–34°20’N and 106°09’–122°25’N.
• Temporal Scale: 2000 to 2022.

Methodology and Data

• Models used:
  • Integrated framework combining Convergent Cross Mapping (CCM), Random Forest model (with Shapley Additive Explanations - SHAP), and a Copula-based Bayesian approach.
  • Drought identification indices: Standardized Precipitation Evaporative Index (SPEI), Soil Moisture Percentile (SMP), Standardized Evaporative Stress Ratio (SESR), and Standardized Soil Moisture Index (SSI).
  • Penman-Monteith equation for potential evapotranspiration (PET) calculation.
  • Statistical distributions (log-logistic, gamma, Generalized Extreme Value, Generalized Pareto, Kernel Density) and copula functions (Normal, Clayton, Frank, Gumbel, Student’s t) for data fitting and joint probability modeling.
  • Kolmogorov-Smirnov (K-S) test, Akaike Information Criterion (AIC), and Bayesian Information Criterion (BIC) for model selection.
• Data sources:
  • European Centre for Medium-Range Weather Forecasts Reanalysis 5 (ERA5) dataset: Hourly meteorological variables (precipitation, temperature, shortwave radiation, 2 m dewpoint temperature, surface pressure, 10 m wind speed) at 0.25° × 0.25° spatial resolution.
  • Global Land Evaporation Amsterdam Model (GLEAM) v3.8a: Daily evapotranspiration and root-zone soil moisture (0–100 cm).
  • In-situ observations for dataset validation.

Main Results

• All three flash drought types (meteorological, soil, evaporative) showed strong causal relationships with agricultural droughts, with area-averaged Convergent Cross Mapping (CCM) skills (ρ) exceeding 0.5, and over 0.95 for soil flash droughts.
• The average propagation time from flash droughts to agricultural droughts ranged from 36.8 to 48.8 days. Meteorological flash droughts had the shortest propagation time (average 36.8 days), while evaporative flash droughts had the longest (average 48.8 days).
• Soil flash droughts exhibited the highest number of propagation events, highest sensitivity (Tr1 = 0.42), and highest drought translation rate (Tr2 = 0.2) to agricultural droughts.
• Drought conditions generally intensified and prolonged during propagation, with duration and severity propagation rates typically less than 1.
• Dominant drivers and triggering thresholds for propagation:
  • Meteorological flash droughts: Precipitation was the most critical driver (areal average threshold: 14.3 ± 7.6 mm), followed by flash drought severity (5.0 ± 1.3) and temperature (14.8 ± 4.7 °C).
  • Evaporative flash droughts: Vapor Pressure Deficit (VPD) was the most influential factor (areal average threshold: 7.8 ± 2.3 hPa), followed by precipitation (14.0 ± 7 mm) and severity (4.7 ± 1.2).
  • Soil flash droughts: Flash drought severity was the dominant factor (areal average threshold: 11.2 ± 2.3), with duration being the second most important.

Contributions

• Developed and applied an integrated, robust framework combining advanced causal inference (CCM), machine learning (Random Forest with SHAP), and probabilistic modeling (copula-based Bayesian approach) to analyze flash drought propagation.
• Provided a novel, comprehensive understanding of the causal relationships, propagation times, and specific mechanisms by which different types of flash droughts transition into agricultural droughts.
• Quantified the dominant climatic and flash drought characteristics that drive propagation, identifying critical threshold values for these drivers.
• Highlighted distinct mechanisms governing the initiation versus the development (severity and duration) of propagated agricultural droughts.
• Offered valuable insights for improving drought forecasting, risk assessment, and the development of targeted, effective early warning systems and adaptive agricultural water management strategies.

Funding

• National Natural Science Foundation of China (grant no. 52309032)
• China Postdoctoral Science Foundation (grant no. 2025T180083)
• Wuhan Natural Science Foundation Special Zone Project (grant no. 2024040701010035)

Citation

@article{Hu2026Atmospheric,
  author = {Hu, Chen and She, Dunxian and Zhang, Liping and Wang, Gangsheng and Jing, Zhaoxia and Hong, Si and Xia, Jun},
  title = {Atmospheric dryness and flash drought severity drive the shifts of different flash drought types into agricultural droughts},
  journal = {Agricultural Water Management},
  year = {2026},
  doi = {10.1016/j.agwat.2025.110109},
  url = {https://doi.org/10.1016/j.agwat.2025.110109}
}