Land Use/Land Cover Classification and Forecasting of South-Western Nigeria Using Backcasting and Machine Learning Techniques

This study assesses historical and projected land use and land cover (LULC) changes in South-Western Nigeria from 1993 to 2040 using MODIS satellite imagery and predictive modelling. Results reveal substantial urban expansion (+2,915.49 km²) and vegetation loss (−8,143.85 km²), with water bodies remaining relatively stable. Linear Regression outperformed Random Forest and Decision Tree models, accurately capturing gradual trends in land transformation. The classification achieved an accuracy of 87.3%–89.6%, providing a reliable basis for regional forecasting. The findings highlight critical environmental and planning challenges, including ecosystem degradation, flood vulnerability, and peri-urban sprawl. Recommendations emphasize integrating high-resolution imagery, nonlinear modelling, and stakeholder-engaged planning frameworks. Hence, this study contributes to sustainable urban management and LULC policy integration in rapidly urbanizing regions.