Kulkarni et al. (2025) Near-global Agro-climatological Drought Monitoring Dataset

Identification

• Journal: Scientific Data
• Year: 2025
• Date: 2025-11-25
• Authors: Sneha Kulkarni, Yohei Sawada
• DOI: 10.1038/s41597-025-06316-7

Research Groups

• Department of Civil Engineering, The University of Tokyo, Tokyo 113-8656, Japan

Short Summary

This study introduces the Near-global Combined Drought Monitoring (NEC-DROMO) dataset, integrating soil moisture, vegetation water content, rainfall, and temperature at a 0.25-degree monthly resolution from 2002-2021, demonstrating superior reliability in capturing global drought patterns compared to traditional indices.

Objective

• To develop and validate a near-global combined agro-climatological drought monitoring dataset (NEC-DROMO) that integrates soil moisture, vegetation water content, rainfall, and temperature using Principal Component Analysis to provide robust and comprehensive drought insights.

Study Configuration

• Spatial Scale: Near-global land area (90.0° to -90.0° latitude and -180.0° to 180.0° longitude), excluding snow-covered and densely forested regions, at a 0.25-degree spatial resolution.
• Temporal Scale: Monthly data spanning from 2002 to 2021, with a data gap from November 2011 to June 2012.

Methodology and Data

• Models used:
  • Principal Component Analysis (PCA) for objective weighting and integration of variables into a Combined Drought Indicator (CDI).
  • Z-score method for standardizing input variables.
  • Coupled Land and Vegetation Data Assimilation System (CLVDAS) (used in ECHLA for SM and VWC).
  • 4D-Var data assimilation system (used in ERA5 for rainfall).
• Data sources:
  • ECoHydrological Land Reanalysis (ECHLA) dataset (for Soil Moisture (0–195 cm), Vegetation Water Content, and Land Surface Temperature).
  • AMSR-E and AMSR2 microwave satellite observations (inputs to ECHLA).
  • ERA5 reanalysis dataset (for Rainfall).
  • Geocoded Disaster Dataset (GDIS) (for validation of socio-economic drought impacts).
  • CHIRPS data (for comparison with Standardized Precipitation Index (SPI)).
  • Normalized Difference Vegetation Index (NDVI) and Standardized Soil Moisture Index (SSI) (for comparison).

Main Results

• The Near-global Combined Drought Monitoring (NEC-DROMO) dataset was developed, providing monthly agro-climatological drought data for 2002–2021 at a 0.25-degree spatial resolution.
• NEC-DROMO integrates Standardized Precipitation Index (SPI), Standardized Soil Moisture Index (SSI), Standardized Temperature Index (STI), and Standardized Vegetation Water Content Index (SVWCI) using Principal Component Analysis (PCA) to objectively assign dynamic, spatially and temporally varying weights.
• Validation against the Geocoded Disaster Dataset (GDIS) showed that NEC-DROMO successfully detected 1,737 out of 1,923 GDIS drought events (Probability of Detection (POD) of 0.903) during the exact event period, improving to 1,813 events (POD of 0.943) with a one-month lead time.
• NEC-DROMO demonstrated superior performance in detecting GDIS events compared to single-variable indices: SPI detected 1,651 (actual) / 1,747 (lead) events, NDVI detected 1,676 (actual) / 1,765 (lead) events, and SSI detected 1,670 (actual) / 1,780 (lead) events.
• PCA-based weights effectively capture regional and seasonal variability, assigning higher importance to factors like temperature in arid regions, soil moisture/rainfall in monsoon regions, and vegetation water content in high-canopy evergreen regions.
• The dataset provides standardized drought index values, with negative values indicating dry conditions and positive values wet spells, classified using McKee et al. (1993) thresholds (e.g., extremely dry ≤ -2.0).

Contributions

• Introduces the first near-global agro-climatological drought monitoring dataset (NEC-DROMO) that comprehensively integrates Vegetation Water Content (VWC) alongside rainfall, soil moisture, and temperature, offering a more direct measure of vegetation stress than traditional indices.
• Employs an objective Principal Component Analysis (PCA) approach to dynamically determine variable weights based on their spatial and temporal contributions, enhancing robustness and adaptability compared to systems using predefined or regionally fixed weights.
• Provides a long-term (2002-2021) near-global dataset at high spatial (0.25 degrees) and monthly temporal resolution, addressing the critical need for comprehensive historical drought records for global analysis, policymaking, and disaster risk assessment.
• Demonstrates superior capability in identifying and reflecting socio-economic drought impacts compared to single-variable drought indicators (SPI, NDVI, SSI), offering a more holistic understanding of drought events.

Funding

• JSPS KAKENHI grants (21H01430, 25H00760, and 24K17352)
• JAXA grant (ER2GWF102 and ER3AMF106)
• Katsu Kimura research award

Citation

@article{Kulkarni2025Nearglobal,
  author = {Kulkarni, Sneha and Sawada, Yohei},
  title = {Near-global Agro-climatological Drought Monitoring Dataset},
  journal = {Scientific Data},
  year = {2025},
  doi = {10.1038/s41597-025-06316-7},
  url = {https://doi.org/10.1038/s41597-025-06316-7}
}