Jahanbakhsh et al. (2025) Monitoring and forecasting agricultural drought in Golestan Province, Iran (2001–2028): an integrated approach using remote sensing and machine learning

Identification

• Journal: Advances in Space Research
• Year: 2025
• Date: 2025-12-01
• Authors: Mahsa Jahanbakhsh, Mehdi Akhoondzadeh
• DOI: 10.1016/j.asr.2025.11.113

Research Groups

• Photogrammetry and Remote Sensing Department, School of Surveying and Geospatial Engineering, College of Engineering, University of Tehran, Tehran, Iran

Short Summary

This study developed an integrated framework using remote sensing and machine learning to monitor and forecast agricultural drought in Golestan Province, Iran. The framework predicts that severe and extreme drought will expand to cover approximately 13,000 square kilometers (62% of the province) by 2028, with croplands and bare lands being most vulnerable.

Objective

• To develop an integrated framework for monitoring and forecasting agricultural drought in Golestan Province, Iran, using remote sensing and machine learning to provide insights for early warning systems and climate-resilient agricultural planning.

Study Configuration

• Spatial Scale: Golestan Province, Iran (area approximately 20,900 km²)
• Temporal Scale: Monthly drought monitoring from 2001 to 2024; drought forecasting for 2028.

Methodology and Data

• Models used: Random Forest (RF) regression, Extreme Gradient Boosting (XGBoost) regression.
• Data sources:
  • MODIS-derived Normalized Difference Vegetation Index (NDVI) and Land Surface Temperature (LST) for Vegetation Health Index (VHI) calculation (via Vegetation Condition Index (VCI) and Temperature Condition Index (TCI)).
  • Supplementary hydroclimatic indicators: Standardized Precipitation Index (SPI) and Evaporative Stress Index (ESI).
  • Sentinel-derived Dynamic World data for land cover cross-analysis.

Main Results

• XGBoost demonstrated superior performance for drought forecasting with an R² of 0.83, a Root Mean Square Error (RMSE) of 0.065, and a Mean Absolute Error (MAE) of 0.041.
• Forecasts indicate that severe and extreme drought conditions are projected to expand to over 13,000 square kilometers, covering approximately 62% of Golestan Province by 2028.
• Croplands and bare lands were identified as the most vulnerable land cover types to the projected drought expansion.

Contributions

• This study presents an effective integrated framework combining satellite data and machine learning for operational agricultural drought monitoring and prediction.
• It provides valuable insights for developing early warning systems and informing climate-resilient agricultural planning in arid and semi-arid regions.
• The research offers a robust methodology for assessing complex and spatially heterogeneous drought patterns, addressing limitations of traditional single-variable or sparse in-situ approaches.

Funding

• Not explicitly mentioned in the provided text.

Citation

@article{Jahanbakhsh2025Monitoring,
  author = {Jahanbakhsh, Mahsa and Akhoondzadeh, Mehdi},
  title = {Monitoring and forecasting agricultural drought in Golestan Province, Iran (2001–2028): an integrated approach using remote sensing and machine learning},
  journal = {Advances in Space Research},
  year = {2025},
  doi = {10.1016/j.asr.2025.11.113},
  url = {https://doi.org/10.1016/j.asr.2025.11.113}
}