Amiranipour et al. (2025) Meteorological and agricultural drought assessments using satellite imagery and machine learning models

Identification

• Journal: Advances in Space Research
• Year: 2025
• Date: 2025-10-10
• Authors: Mahshid Amiranipour, Mohammad Najafzadeh, Sedigheh Mohamadi
• DOI: 10.1016/j.asr.2025.10.014

Research Groups

• Department of Water Engineering, Faculty of Civil and Surveying Engineering, Graduate University of Advanced Technology, Kerman, Iran
• Department of Ecology, Institute of Environmental Sciences, Graduate University of Advanced Technology, Kerman, Iran

Short Summary

This study comprehensively assessed meteorological and agricultural droughts and their interrelationship across Iran using satellite imagery and machine learning, demonstrating robust predictive accuracy for drought early warning and management.

Objective

• To provide a comprehensive assessment of meteorological and agricultural droughts and their interrelationship across Iran, and to evaluate the predictability of agricultural drought indices using climatic and other drought indices with machine learning.

Study Configuration

• Spatial Scale: Country-wide across Iran
• Temporal Scale: Multi-temporal analysis of drought indices, considering varying temporal scales for correlation analysis.

Methodology and Data

• Models used: Machine learning models (Random Forest Regression, Clustering algorithm), Sequential Feature Selection (SFS)
• Data sources: Satellite imagery, Standardized Precipitation Index (SPI), Standardized Precipitation Evapotranspiration Index (SPEI), Vegetation Condition Index (VCI), Temperature Condition Index (TCI), Evaporative Drought Index (EDI), Soil Moisture Deficit Index (SMDI)

Main Results

• Meteorological drought events occurred concurrently across much of Iran, with regional variations in intensity.
• Pearson correlation between SPI and other indices generally increased with longer temporal scales.
• SPEI exhibited the highest correlation with SPI among the evaluated indices.
• Strongest correlations between meteorological and agricultural drought indices were observed in western, southwestern, and some northern regions of Iran.
• EDI and SPI were identified as the most influential predictors across all target indices.
• EDI was most critical for predicting vegetation (VCI) and soil moisture (SMDI), while SPI dominated for TCI and EDI estimation.
• Vegetation- and temperature-based indicators showed lagged responses to precipitation deficits.
• The developed framework achieved robust predictive accuracy, effectively capturing nonlinear interactions.

Contributions

• Provides a comprehensive assessment of both meteorological and agricultural droughts and their interrelationship across Iran.
• Evaluates the predictability of agricultural drought indices using a combination of climatic and other drought indices with machine learning and Sequential Feature Selection (SFS).
• Advances beyond conventional composite index approaches by integrating multiple drought indices with machine learning to capture complex nonlinear interactions.
• Offers actionable insights for drought early warning systems, agricultural planning, and water resource management in drought-prone regions.

Funding

Not explicitly mentioned in the provided text.

Citation

@article{Amiranipour2025Meteorological,
  author = {Amiranipour, Mahshid and Najafzadeh, Mohammad and Mohamadi, Sedigheh},
  title = {Meteorological and agricultural drought assessments using satellite imagery and machine learning models},
  journal = {Advances in Space Research},
  year = {2025},
  doi = {10.1016/j.asr.2025.10.014},
  url = {https://doi.org/10.1016/j.asr.2025.10.014}
}