Zhi-jie et al. (2025) Nonstationary Spatiotemporal Projection of Drought Across Seven Climate Regions of China in the 21st Century Based on a Novel Drought Index

Identification

• Journal: Water
• Year: 2025
• Date: 2025-11-10
• Authors: Yan Zhi-jie, Gengxi Zhang, Huimin Wang, Baojun Zhao
• DOI: 10.3390/w17223206

Research Groups

• College of Hydraulic Science and Engineering, Yangzhou University, Yangzhou, China

Short Summary

This study projects the spatiotemporal evolution of drought across seven climate regions of China in the 21st century using a novel CO2-aware standardized moisture anomaly index (SZI[CO2]) and nonstationary Copula-based approaches. It finds a wetting trend in Northern and Western China, while Central and Southern China are projected to experience drying, with drought characteristics exhibiting strong nonstationarity and higher joint probabilities under high-emission scenarios.

Objective

• To develop a CO2-aware standardized moisture anomaly index (SZI[CO2]) for more accurate drought assessment.
• To employ a three-dimensional drought identification method to extract and analyze the spatiotemporal dynamics of drought duration, area, and severity.
• To utilize Copula-based approaches with time-varying parameters to estimate joint drought probabilities and project the spatiotemporal evolution of droughts under SSP2-4.5 and SSP5-8.5 scenarios across China's climate regions.

Study Configuration

• Spatial Scale: Seven climate regions of China (Northwest, Inner Mongolia, Tibetan Plateau, Northeast, North China, Central China, South China).
• Temporal Scale: 1985–2100 (21st century projections), with CMIP6 historical period 1850–2014 and future scenario period 2015–2100.

Methodology and Data

• Models used:
  • 10 Global Climate Models (GCMs) from Coupled Model Intercomparison Project Phase 6 (CMIP6).
  • Bias Correction and Spatial Downscaling (BCSD) method for GCM output downscaling.
  • CO2-aware standardized moisture anomaly index (SZI[CO2]), based on the Penman–Monteith (PM) formula accounting for CO2 effects on canopy resistance.
  • Three-dimensional drought identification method (based on run theory and spatiotemporal clustering of SZI[CO2] values).
  • Generalized Additive Models for Location, Scale, and Shape (GAMLSS) for nonstationary frequency analysis.
  • Copula-based approaches for multivariate joint probability analysis of drought characteristics.
• Data sources:
  • Observed monthly precipitation and daily maximum/minimum temperature data from the Climate Change Research Center (2472 ground stations).
  • Monthly wind speed, pressure, relative humidity, and radiation data from the ERA5 reanalysis dataset (0.25° × 0.25° resolution, 1940–present).
  • CMIP6 GCM outputs for historical and future scenarios.
  • CO2 concentration outputs from GCMs.

Main Results

• Regional Drought Trends:
  • Northwest China, Inner Mongolia, the Tibetan Plateau, Northeast China, and North China show a significant increasing trend in SZI[CO2] and a decreasing trend in affected drought area, indicating a future wetting trend.
    • SZI[CO2] in Northwest and Inner Mongolia increases at approximately 0.010 per year.
    • SZI[CO2] in the Tibetan Plateau increases at 0.005 per year (SSP2-4.5) and 0.008 per year (SSP5-8.5).
    • Relative Drought Area (RDA) in Northwest China declines fastest: 0.0019 per year (SSP2-4.5) and 0.0015 per year (SSP5-8.5).
  • Central China and South China show a pronounced decline in SZI[CO2] (from ~0.3 in 1980s to ~-0.3 in 2090s, average rate of -0.005 per year) and increasing RDA (e.g., 0.0023 per year under SSP5-8.5), indicating a future drying trend with increased drought frequency, duration, and severity.
  • National-scale SZI[CO2] increases at 0.004 per year (SSP2-4.5) and 0.005 per year (SSP5-8.5), with RDA declining at 0.005 per year.
• Nonstationary Drought Characteristics:
  • Drought characteristics (duration, area, severity) exhibit obvious trend components, confirming nonstationarity.
  • Nonstationary log-normal (LNO) and generalized extreme value (GEV) distributions, particularly the M2 model (time-varying location parameter), are found to be most suitable for modeling drought feature frequency distributions in most regions.
• Joint Drought Probabilities:
  • The joint occurrence probability of drought under the SSP5-8.5 scenario is higher than under SSP2-4.5 for the same drought characteristic values, with notable differences over the Tibetan Plateau, Central China, and South China.
  • For bivariate cases, the "or" joint probability is higher than the "and" joint probability (e.g., in Northwest China under SSP2-4.5, P(D > 4 months ∪ A > 4 × 10^5 km^2) is 21%, while P(D > 4 months ∩ A > 4 × 10^5 km^2) is 13%).
  • For trivariate cases, the paper states that when drought characteristic variables are held at the same fixed values, the joint probability under the "or" case is higher than under the "and" case. However, an example for Northwest China under SSP2-4.5 shows P(D > 3.2 months ∩ A > 3 × 10^5 km^2 ∩ S > 1.2 × 10^6 km^2·month) is 47%, while P(D > 3.2 months ∪ A > 3 × 10^5 km^2 ∪ S > 1.2 × 10^6 km^2·month) is 21%.
  • Ignoring any one drought characteristic is likely to lead to an underestimation of severe drought event probabilities.

Contributions

• Development of a novel CO2-aware standardized moisture anomaly index (SZI[CO2]) that integrates water–energy balance and accounts for CO2-induced physiological effects on vegetation and potential evapotranspiration (PET), leading to more realistic drought projections compared to indices neglecting CO2 effects.
• Application of a three-dimensional spatiotemporal drought identification method to comprehensively capture the continuity and dynamics of drought events (duration, area, severity), addressing limitations of previous studies that often broke spatiotemporal continuity.
• Implementation of a nonstationary framework for multivariate drought frequency analysis using Copula-based approaches with time-varying parameters, providing a more robust assessment of drought risk under intensifying climate change.
• Provides updated and regionally differentiated future drought projections for China, showing a wetting trend in Northern/Western regions and a drying trend in Central/Southern regions, which aligns with recent observational evidence for Northwest China.

Funding

• China Postdoctoral Science Foundation (Grant No. 2024M752711)
• Natural Science Foundation of Jiangsu Province (Grant No. BK20220590)
• Priority Academic Program Development of Jiangsu Higher Education Institutions of China (PAPD)

Citation

@article{Zhijie2025Nonstationary,
  author = {Zhi-jie, Yan and Zhang, Gengxi and Wang, Huimin and Zhao, Baojun},
  title = {Nonstationary Spatiotemporal Projection of Drought Across Seven Climate Regions of China in the 21st Century Based on a Novel Drought Index},
  journal = {Water},
  year = {2025},
  doi = {10.3390/w17223206},
  url = {https://doi.org/10.3390/w17223206}
}