Harraf et al. (2026) Climate change and crop production in North Africa: Insights from machine learning models

Identification

• Journal: Journal of Environmental Management
• Year: 2026
• Date: 2026-01-01
• Authors: Raja Ben Harraf, Manal Mhada, Nabil El Moçayd
• DOI: 10.1016/j.jenvman.2025.128451

Research Groups

• AgrobioSciences (AgBS), College of Agriculture & Environmental Sciences (CAES), Mohammed VI Polytechnic University, Ben Guerir, Morocco
• International Water Research Institute (IWRI), College of Agriculture & Environmental Sciences (CAES), Mohammed VI Polytechnic University, Ben Guerir, Morocco

Short Summary

This study uses machine learning models to project the impact of future temperature increases and precipitation declines on major crop yields across North Africa. The findings indicate overall yield reductions for several crops but highlight the resilience of maize and sorghum, especially under irrigated conditions.

Objective

• To evaluate and project the impact of future temperature and precipitation variations on major crop yields in North Africa using machine learning models, thereby informing proactive agricultural planning and adaptation strategies.

Study Configuration

• Spatial Scale: North Africa, specifically Morocco, Algeria, Tunisia, Libya, and Egypt.
• Temporal Scale:
  • Historical baseline: 1981–2014
  • Future projections:
    • Near-term: 2015–2050
    • Mid-term: 2051–2080
    • Long-term: 2081–2100

Methodology and Data

• Models used: Multiple Linear Regression (MLR), Random Forest (RF), and eXtreme Gradient Boosting (XGBoost). The best-performing model was selected for future yield projections.
• Data sources:
  • Climatic conditions: ERA5 reanalysis data.
  • Crop yield data: FAO crop yield data.
  • Future climate projections: SSP2-4.5 and SSP5-8.5 scenarios from the CMIP6 dataset.

Main Results

• Projections indicate an overall increase in temperature and a decline in precipitation across North Africa.
• These climatic changes are projected to lead to yield reductions for several major crops.
• Some crops, such as maize and sorghum, demonstrate greater resilience to climate change, particularly when grown under irrigated conditions.
• The study identifies the most vulnerable crop species to climate change and highlights more resilient alternatives for the region.

Contributions

• Enhances the understanding of climate–yield interactions in semi-arid regions, specifically North Africa.
• Demonstrates the efficacy and value of data-driven machine learning models for long-term agricultural planning and climate impact assessment.
• Provides actionable insights for policymakers, researchers, and agricultural stakeholders to identify climate-resilient crops, optimize land use, and develop region-specific adaptation strategies.

Funding

No funding information was provided in the excerpt.

Citation

@article{Harraf2026Climate,
  author = {Harraf, Raja Ben and Mhada, Manal and Moçayd, Nabil El},
  title = {Climate change and crop production in North Africa: Insights from machine learning models},
  journal = {Journal of Environmental Management},
  year = {2026},
  doi = {10.1016/j.jenvman.2025.128451},
  url = {https://doi.org/10.1016/j.jenvman.2025.128451}
}