Xu et al. (2025) Quantifying the Urbanization and Vegetation Greening Effect on Spatiotemporal Continuous Drought Risk Via Nonstationary C-Vine Copula Model

Identification

• Journal: Water Resources Management
• Year: 2025
• Date: 2025-12-20
• Authors: Pengcheng Xu, Dong Wang, Yuankun Wang, Vijay P. Singh
• DOI: 10.1007/s11269-025-04358-5

Research Groups

• College of Hydraulic Science and Engineering, Yangzhou University, Yangzhou, P.R. China
• Key Laboratory of Surficial Geochemistry, Department of Hydrosciences, School of Earth Sciences and Engineering, State Key Laboratory of Pollution Control and Resource Reuse, Ministry of Education, Nanjing University, Nanjing, P.R. China
• School of Water Resources and Hydropower Engineering, North China Electric Power University, Beijing, P.R. China
• Department of Biological and Agricultural Engineering, Zachry Department of Civil & Environmental Engineering, Texas A & M University, College Station, TX, US
• National Water and Energy Center, UAE University, Al Ain, UAE

Short Summary

This study developed a nonstationary C-Vine Copula model to quantify the distinct impacts of urbanization and vegetation greening on multivariate drought risk in humid (Yangtze River Delta) and arid (Weihe River Basin) watersheds, finding urbanization amplifies risk in humid areas while vegetation greening alleviates it in arid regions.

Objective

• How to construct a nonstationary C-Vine Copula model to simulate dynamic multidimensional drought dependent structures?
• Can urbanization and vegetation related factors affect the multidimensional drought risk of a watershed by influencing the trend-caused nonstationarity of multivariate drought events (Intensity-Duration-Area-characterized drought)?
• Is the multivariate drought risk in watersheds with different wet and dry characteristics (i.e., humid and arid areas) affected differently by urbanization and vegetation?

Study Configuration

• Spatial Scale: Yangtze River Delta (YZD, humid watershed) and Weihe River Basin (WRB, arid watershed) in China.
• Temporal Scale: 1981–2022 (42 years).

Methodology and Data

• Models used:
  • Dynamic C-Vine Copula (DCC) model for nonstationary multivariate dependence structure.
  • Three-dimensional clustering algorithm for drought event identification (Intensity, Duration, Area).
  • Time-varying moment model for marginal distributions (Negative Binomial, Poisson, Gamma, Generalized Extreme Value).
  • Linear Regression-Derived Slope Comparative (LRSC) approach for multivariate drought risk attribution.
  • Maximum Likelihood (ML) method for parameter estimation.
  • Corrected Akaike Information Criterion (AICc) and Log-likelihood ratio (LR) test for model selection and nonstationarity testing.
• Data sources:
  • Daily meteorological data: CN05.1 dataset (0.25° spatial resolution).
  • Potential evapotranspiration: Hargreaves equation.
  • Drought index: 3-month-scaled Nonstationary Standardized Precipitation Evapotranspiration Index (NSPEI3).
  • Large-scale climate indices: Southern Oscillation Index (SOI), Arctic Oscillation (AO), Antarctic Oscillation (AAO), Pacific Decadal Oscillation (PDO), Western Pacific Subtropical High Index (WHI).
  • Temperature anomalies: Local Temperature Anomaly (TEMloc), Global Temperature Anomaly (TEMgw) from HadCRUT4 dataset.
  • Urbanization level: Global Artificial Impervious Area (GAIA) dataset (denoted as PISA).
  • Vegetation greening index: Monthly Normalized Difference Vegetation Index (NDVI) from National Earth System Science Data Center.

Main Results

• A 3D clustering technique identified 36 drought events in YZD and 39 in WRB, each lasting over 2 months, between 1981 and 2022.
• Drought centers in YZD have recently shifted towards highly urbanized areas, while in WRB, they correlate with regions of low vegetation cover.
• The Dynamic C-Vine Copula (DCC) model successfully captured diverse nonstationary patterns in multivariate drought distributions.
• In YZD, the urbanization index (PISA), NDVI, and climatic indices (SOI, AO, AAO) were significant for marginal distributions, while only climatic indices (PDO, SOI) influenced dependence structures.
• In WRB, NDVI and specific climatic indices (SOI, PDO, AAO) were the primary nonstationarity drivers. PISA was not a significant driver for dependence structure in either basin.
• The 50-year joint return period isosurface showed a continuous uplift in YZD, indicating a rising multidimensional drought risk.
• The 50-year joint return period isosurface showed a noticeable flattening in WRB, suggesting a decreasing trend in multidimensional drought risk.
• Urbanization amplified multivariate drought risk (MR) in YZD by 10% during 2003–2022, contrasting with a 5% alleviation during 1983–2002. The impact of vegetation greening on MR in YZD was negligible.
• Vegetation greening significantly alleviated MR in WRB, with its amplification factor decreasing from 32% (1983–2002) to 2% (2003–2022), largely attributed to soil and water conservation projects.

Contributions

• Introduces a novel Dynamic C-Vine Copula (DCC) framework that explicitly incorporates covariates (climate change, urbanization, vegetation greening) to model dynamic, nonstationary dependence structures of multidimensional drought events (characterized by Intensity-Duration-Area).
• Provides the first quantitative isolation of the distinct impacts of human-induced urbanization and vegetation changes from natural climate variability on multivariate drought risk.
• Applies a comparative case-control analysis to two contrasting watersheds (humid Yangtze River Delta and arid Weihe River Basin) to offer new, transferable insights into how primary drivers of drought risk differ across hydroclimatic regimes.
• Develops a systematic methodology for attributing changes in multivariate drought risk to specific environmental and anthropogenic drivers, moving beyond simple correlations.

Funding

• National Natural Science Foundation of China (42301026)
• Natural Science Foundation of Jiangsu Province (BK20220589)

Citation

@article{Xu2025Quantifying,
  author = {Xu, Pengcheng and Wang, Dong and Wang, Yuankun and Singh, Vijay P.},
  title = {Quantifying the Urbanization and Vegetation Greening Effect on Spatiotemporal Continuous Drought Risk Via Nonstationary C-Vine Copula Model},
  journal = {Water Resources Management},
  year = {2025},
  doi = {10.1007/s11269-025-04358-5},
  url = {https://doi.org/10.1007/s11269-025-04358-5}
}