AI-driven spatial statistical modeling and smart ecological restoration framework for Pistacia atlantica in Saudi Arabia

This study presents an AI-driven spatial modeling and Smart Ecological Restoration Framework for Pistacia atlantica in Saudi Arabia, integrating Geographic Information Systems (GIS), remote sensing, exploratory factor analysis (EFA), binary logistic regression, and the Dynamic Restoration Priority Index (DRPI). Environmental variables, including rainfall, temperature, NDVI, elevation, desertification severity, and grazing density, were examined across 13 Saudi administrative regions to assess environmental suitability and restoration priorities. Rainfall (β = 0.68, p < 0.01) and NDVI (β = 0.74, p < 0.001) were positively associated with species occurrence, whereas desertification severity (β = − 0.59, p < 0.01) showed a negative association. The logistic regression model achieved an overall classification accuracy of 84.3%; however, this result should be interpreted cautiously given the limited regional sample size. Spatial suitability analysis classified approximately 30% of the study regions as highly suitable, while Asir, Al Bahah, and Ha’il recorded the highest restoration-priority scores of 0.88, 0.84, and 0.71, respectively. Field observations of five naturally occurring Pistacia atlantica trees in Ha’il provided preliminary evidence of spatial agreement with the predicted suitability areas. The proposed Smart Ecological Restoration Framework provides a transparent environmental decision-support approach for restoration prioritization, afforestation planning, and sustainable ecological management, while offering a scalable basis for future integration of additional environmental, socioeconomic, and climate-related criteria.