Examining Forest Disturbance Using Artificial Intelligence and Remote Sensing Techniques

This paper investigates forest disturbance in Ajloun area of Jordan through artificial intelligence and remotesensing during 1986 to 2025. Environmental variables that have been incorporated using satellite imageshave included increasing temperature patterns, changing precipitation patterns, topographic elevation, soilmoisture, and land surface temperature (LST) data and anthropogenic pressure indices to determine thedrivers of forest change. Support Vector Machines and Random Forests are used to map land cover;regression to measure climate vegetation relationships. The accuracy of the supervised classification is 83-90% (Kappa 0.82-0.86), which can be used as a strong foundation in land changes. The forest cover reducedmore or less by 8% of the area, the built-up land grew by approximately 4%. Precipitation and vegetationhealth are strongly correlated: The vegetation cover estimated by NDVI is related to total rainfall (R2 =0.87). Increasing temperature and frequency of drought over the period of study have also increased stressof forest as indicated by increasing LST and decreasing moisture levels in soil when a disturbance occurs.These results are important to highlight the overall effects of climate change and urban sprawl on the forestecosystem and provide important information on monitoring the environment and guiding policy andregional planning with evidence. The integrative approach supports optimization-driven architecturaldesign strategies for sustainable development in climate-sensitive regions.