Wu et al. (2026) Definition- and dataset-dependent characterization of global compound drought–heat extremes

Identification

• Journal: Weather and Climate Extremes
• Year: 2026
• Date: 2026-11-01
• Authors: Lixia Wu, Shuiqing Yin, Zihui Zhao, Bin He, Jim Hall
• DOI: 10.1016/j.wace.2026.100972

Research Groups

• State Key Laboratory of Earth Surface Processes and Hazards Risk Governance (ESPHR), Faculty of Geographical Science, Beijing Normal University, Beijing, People’s Republic of China
• Environmental Change Institute, University of Oxford, Oxford, UK

Short Summary

This study reveals that the characterization of global compound drought–heat extremes is highly sensitive to both event definition and dataset choice, with biases in climate-variable dependence playing a crucial role in dataset-related uncertainties.

Objective

• To examine how monthly anomaly-based and heatwave-based definitions of compound drought–heat extremes differ in their inferred spatial patterns, event characteristics, and historical trends.
• To assess how ERA5 reanalysis and the bias-corrected WFDE5 dataset represent these characteristics relative to observation-based references.
• To determine the extent to which dataset-related differences can be explained by biases in temperature, precipitation, and their inter-variable dependence.

Study Configuration

• Spatial Scale: Global land (excluding Antarctica), analyzed on a 1° grid, with regional statistics for 44 IPCC AR6 land reference regions.
• Temporal Scale: Event characteristics evaluated over 1980-2014; long-term trends in spatial extent analyzed over 1980-2019. Monthly resolution for Compound Dry-Hot Events (CDHEs) and daily resolution for heatwaves within Compound Drought–Heatwave Events (CDHWs). Warm season defined as May–September in the Northern Hemisphere and November–March in the Southern Hemisphere.

Methodology and Data

• Models used:
  • Compound Dry-Hot Events (CDHEs): Monthly Standardized Precipitation Index (SPI) ≤ -1 and Standardized Temperature Index (STI) ≥ 1.
  • Compound Drought–Heatwave Events (CDHWs): Monthly SPI ≤ -1 combined with daily heatwaves (at least three consecutive days with daily temperature exceeding the 90th percentile of the 1980–2014 baseline).
  • Theil–Sen median slope estimator for trend analysis.
• Data sources:
  • Observation-based: Climatic Research Unit gridded Time Series (CRU TS v4.07), Berkeley Earth Surface Temperature (BEST), Global Precipitation Climatology Centre (GPCC). A composite of BEST and GPCC (BEST+GPCC) was used as a reference.
  • Reanalysis: ECMWF Reanalysis v5 (ERA5).
  • Bias-corrected: WATCH Forcing Data methodology applied to ERA5 (WFDE5).

Main Results

• Event definition fundamentally shapes compound extreme characteristics: CDHEs are concentrated in the tropics and show significant global expansion (3.2 to 4.4 percentage points per decade), while CDHWs extend into Northern Hemisphere mid-latitudes and exhibit contrasting regional trends without significant global expansion.
• Dataset choice introduces substantial uncertainty: ERA5 overestimates CDHE number by 30% and intensity by 28%, and CDHW number by 11% and intensity by 19% compared to observation-based references.
• WFDE5 significantly reduces biases for CDHEs (e.g., number bias reduced to -10.0%, intensity bias to -13.2%) but provides limited improvement for CDHWs, particularly underestimating their frequency and maximum duration (e.g., -18.2% for frequency).
• Biases are linked to climate-variable dependence: ERA5 exaggerates the negative precipitation–temperature dependence, leading to an overestimation of drought–heat co-occurrence and CDHEs. WFDE5's improved performance for CDHEs is attributed to its better representation of this dependence.
• WFDE5's limited improvement for CDHWs is due to its monthly bias correction not adequately addressing the underestimated temporal persistence of daily temperature extremes in ERA5.

Contributions

• Provides a systematic comparison highlighting the distinct impacts of event definition and dataset choice on the characterization of compound drought–heat extremes.
• Demonstrates the critical role of inter-variable dependence (precipitation–temperature correlation) in explaining dataset-dependent biases, beyond the accuracy of individual climate variables.
• Offers crucial guidance for climate-dataset evaluation, emphasizing the need to assess temporal persistence and inter-variable dependence, and for developing more effective bias-correction strategies for compound-event applications.

Funding

• National Natural Science Foundation of China (grant no. W2412145)
• National Key Research and Development Program of China (Grant No. 2020YFF0304401)
• China Scholarship Council (Grant No. 202206045003)

Citation

@article{Wu2026Definition,
  author = {Wu, Lixia and Yin, Shuiqing and Zhao, Zihui and He, Bin and Hall, Jim},
  title = {Definition- and dataset-dependent characterization of global compound drought–heat extremes},
  journal = {Weather and Climate Extremes},
  year = {2026},
  doi = {10.1016/j.wace.2026.100972},
  url = {https://doi.org/10.1016/j.wace.2026.100972}
}