Aghelpour et al. (2025) Re-constructing and projecting vegetation coverage area variations: A numerical approach based on MRI-ESM2.0 climatic datasets

Identification

• Journal: Advances in Space Research
• Year: 2025
• Date: 2025-09-10
• Authors: Pouya Aghelpour, A A Sabziparvar, Vahid Varshavian
• DOI: 10.1016/j.asr.2025.09.019

Research Groups

Department of Water Engineering, Faculty of Agriculture, Bu-Ali Sina University, Hamedan, Iran

Short Summary

This study numerically models and predicts vegetation coverage area (VCA) in the mountainous Zagros region of Iran using machine learning and CMIP6 climatic data. It successfully reconstructs past VCA and projects a mild increasing trend for future VCA, particularly under the SSP585 climate scenario.

Objective

• To numerically model and predict vegetation coverage area (VCA) based on climatic factors for the mountainous Zagros region in Iran.

Study Configuration

• Spatial Scale: Mountainous Zagros region, Iran (integrated across the entire region).
• Temporal Scale:
  • VCA data extraction: Monthly, 2000–2024.
  • Meteorological data: Monthly.
  • VCA reconstruction: 1950–1999.
  • VCA projection: Post-2015, specifically 2015–2024 (validation), 2025–2050 (near future), and 2051–2080 (distant future).

Methodology and Data

• Models used:
  • Climate Model: MRI-ESM2.0 (part of CMIP6).
  • Machine Learning Models: Multi-Layer Perceptron (MLP), Support Vector Machine (SVM).
• Data sources:
  • Vegetation Coverage Area (VCA): Derived from NDVI (Normalized Difference Vegetation Index) using MODIS sensor imagery.
  • Meteorological data: CMIP6 (MRI-ESM2.0 model), including 19 variables such as temperature, precipitation, solar radiation, relative humidity, air pressure, and wind speed.

Main Results

• Cross-correlation tests indicated that meteorological variables showed higher correlations with VCA variations at temporal lags, with radiation, temperature, and relative humidity demonstrating the strongest impacts.
• Both MLP and SVM models showed satisfactory performance in estimating VCA for the 2000–2014 period, with SVM performing particularly well.
• For the 2015–2024 validation period, the SVM model under the SSP585 climate scenario achieved the best alignment with MODIS-derived VCA (R² = 0.8074 and NRMSE = 0.124).
• Projections for the near future (2025–2050) and distant future (2051–2080) indicated a mild increasing trend in VCA for the Zagros region.
• The increasing trend in VCA was relatively stronger under the SSP585 scenario, especially for the 2051–2080 period.

Contributions

• Developed a robust numerical approach for reconstructing and projecting VCA variations using machine learning and CMIP6 climatic datasets.
• Provides a valuable tool for natural resource and agricultural planners to define environmental strategies for the Zagros region's future.
• Holds significant research value for application in other vegetation-covered regions globally.

Funding

Not explicitly mentioned in the provided text.

Citation

@article{Aghelpour2025Reconstructing,
  author = {Aghelpour, Pouya and Sabziparvar, A A and Varshavian, Vahid},
  title = {Re-constructing and projecting vegetation coverage area variations: A numerical approach based on MRI-ESM2.0 climatic datasets},
  journal = {Advances in Space Research},
  year = {2025},
  doi = {10.1016/j.asr.2025.09.019},
  url = {https://doi.org/10.1016/j.asr.2025.09.019}
}