Ma et al. (2025) Comprehensive drought detection, spatiotemporal variations, and attribution across different agricultural climate zones in Eastern China using a copula-based drought index

Identification

• Journal: Agricultural Water Management
• Year: 2025
• Date: 2025-10-23
• Authors: Y. Ma, Xuan Yu, Mao-Fen Li, Jin Huang, Hailiang Li, Tianze Zhong
• DOI: 10.1016/j.agwat.2025.109902

Research Groups

• Institute of Scientific and Technical Information, Chinese Academy of Tropical Agricultural Sciences/ Key Laboratory of Applied Research on Tropical Crop Information Technology of Hainan Province
• Hainan Tang Huajun Academician Workstation
• School of Geographical Sciences and Tourism, Zhaotong University

Short Summary

This study developed a novel Copula-based Multivariate Standardized Drought Index (MSDI) to assess drought spatiotemporal variations and attribution across different agricultural climate zones in Eastern China from 2001–2020. Findings reveal a general worsening of drought conditions characterized by a southward shift and increased frequency, intensity, and severity of short-term events, with dominant drivers varying regionally.

Objective

• To develop and evaluate a Copula-based Multivariate Standardized Drought Index (MSDI).
• To characterize the spatiotemporal evolution of drought conditions in the Agricultural Comprehensive Areas in Eastern China (ACAEC).
• To trace regional migration trajectories of drought characteristics.
• To identify the spatial heterogeneity of potential drought-driving factors to inform region-specific drought mitigation strategies.

Study Configuration

• Spatial Scale: Agricultural Comprehensive Areas in Eastern China (ACAEC), approximately 2.87 million km², encompassing Northeastern China (NEC), Huang-Huai-Hai (HHH), the Middle and Lower Reaches of the Yangtze River (MLRYR), and Southern China (SC). Data were processed at a 1 km spatial resolution.
• Temporal Scale: 2001–2020, with data analyzed at monthly and annual scales.

Methodology and Data

• Models used:
  • Copula function (for MSDI construction)
  • Multivariate Standardized Drought Index (MSDI)
  • Standardized Precipitation Index (SPI)
  • Standardized Soil Moisture Index (SSI)
  • Run theory (for drought event identification)
  • Mann-Kendall (MK) test
  • Sen’s slope method
  • Center-of-Gravity (CoG) model
  • eXtreme Gradient Boosting (XGBoost)
  • SHapley Additive exPlanations (SHAP)
  • Kolmogorov-Smirnov (K-S) test (for goodness-of-fit)
  • Kriging and Inverse Distance Weighting (IDW) (for spatial interpolation)
• Data sources:
  • Monthly precipitation, temperature, mean atmospheric pressure, sunshine duration, wind speed, and average relative humidity from 1186 meteorological stations (National Meteorological Information Center).
  • Digital Elevation Model (DEM) and 1:1 million vegetation datasets (1 km resolution) from the National Cryosphere Desert Data Center.
  • Atmospheric moisture index, potential evapotranspiration, and root-zone (0–100 cm) soil moisture data from the National Tibetan Plateau Data Center.
  • Solar radiation data from Google Earth Engine.
  • Digital Soil Map from the FAO Soil Data Hub.
  • Disaster loss data from the Meteorological Disaster Statistical Yearbook.
  • Monthly Normalized Difference Vegetation Index (NDVI) data (8 km resolution) from the National Earth System Science Data Center.
  • Gridded dataset of extreme temperature indices for China (Chen et al., 2024).

Main Results

• The Copula-based MSDI successfully integrated meteorological (SPI) and agricultural (SSI) drought signals, showing strong consistency with both indices (Pearson correlation coefficients > 0.75, p < 0.01) and accurately detecting historical drought events with potential for earlier warning.
• Overall, drought conditions in ACAEC worsened from 2001 to 2020, characterized by a southward migration and increasing frequency, intensity, and severity of short-term events.
• Regional variations were significant: Northeastern China (NEC) showed slight improvement, while Huang-Huai-Hai (HHH), the Middle and Lower Reaches of the Yangtze River (MLRYR), and Southern China (SC) experienced intensified drought severity beyond the regional average.
• Major drought events were concentrated in 2001–2003, 2007–2009, 2011–2012, and 2019–2020, often spreading across multiple regions.
• Drought centroids shifted southwestward in NEC and MLRYR, northeastward in HHH, and southeastward in SC.
• Attribution analysis identified region-specific dominant drivers:
  • Solar radiation (SOL) and Daily Temperature Range (DTR) were most influential in NEC and HHH, contributing 35.5 % and 39.1 %, respectively.
  • Relative Humidity (RH), Normalized Difference Vegetation Index (NDVI), and Potential Evapotranspiration (PET) were more prominent in MLRYR and SC, contributing 38.6 % and 50.0 %, respectively.
• Non-monotonic relationships were observed between driving factors and MSDI, with region-specific thresholds (e.g., SOL thresholds of 441.92 MJ/m² in NEC and 418.63 MJ/m² in HHH; DTR thresholds of 11.23 °C in NEC and 9.71 °C in HHH; RH thresholds of 77.22 % in MLRYR and 79.55 % in SC).

Contributions

• Developed a novel Copula-based MSDI framework for comprehensive, large-scale drought detection and assessment across diverse agricultural climate zones in Eastern China.
• Provided a detailed characterization of the spatiotemporal evolution patterns and migration trajectories of drought, addressing existing knowledge gaps.
• Quantified the heterogeneous contributions of various meteorological and environmental factors to drought variability using advanced machine learning (XGBoost and SHAP), revealing non-monotonic relationships and region-specific tipping points.
• Proposed tailored, region-specific drought mitigation strategies based on the identified dominant drivers and quantitative thresholds, aiming to enhance sustainable agriculture, food security, and disaster resilience.

Funding

• Natural Science Foundation of Hainan (322QN369)
• Central Public-interest Scientific Institution Basal Research Fund (NO.1630012025503, NO.1630072024017)
• Philosophy and Social Science Foundation of Hainan Province (HNSK (QN)23–101)
• Hainan Tang Huajun Academician Workstation

Citation

@article{Ma2025Comprehensive,
  author = {Ma, Y. and Yu, Xuan and Li, Mao-Fen and Huang, Jin and Li, Hailiang and Zhong, Tianze},
  title = {Comprehensive drought detection, spatiotemporal variations, and attribution across different agricultural climate zones in Eastern China using a copula-based drought index},
  journal = {Agricultural Water Management},
  year = {2025},
  doi = {10.1016/j.agwat.2025.109902},
  url = {https://doi.org/10.1016/j.agwat.2025.109902}
}