Integrating CA-Markov Modeling, machine learning classifiers, and spatial metrics for urban sprawl analysis in Naqamte City, Ethiopia

Urban sprawl poses a significant challenge to sustainable development, especially in rapidly growing secondary cities of developing countries. This study investigates the spatial and temporal dynamics of urban expansion in Naqamte City, Ethiopia, from 2005 to 2025, using machine learning classifiers—Support Vector Machine (SVM) and Random Forest (RF)—and geospatial technologies. Multi-temporal Landsat imagery was employed to analyze land use/land cover (LULC) changes, revealing a dramatic increase of over 300% in built-up areas, primarily at the cost of forests and bare land. RF outperformed SVM in classification accuracy, demonstrating higher overall accuracy and Kappa coefficients. Urban growth patterns were marked by increased fragmentation, as indicated by higher Shannon entropy and compactness index values. Socio-economic drivers, such as rapid population growth, rural-to-urban migration, rising income levels, and infrastructure development, have significantly contributed to the unregulated expansion, leading to the conversion of forested lands. Predictive modeling for 2035, using the Cellular Automata-Markov (CA-Markov) model, forecasts continued urban expansion under various scenarios. The Business-as-Usual (BAU) scenario predicts a 40% increase in built-up areas, while the Compact Development scenario estimates a 17% rise, emphasizing sustainable growth. The Infrastructure-Led Expansion scenario anticipates a 30% increase. This study highlights the importance of integrating machine learning and remote sensing for accurate urban growth analysis, offering data-driven insights for evidence-based urban planning and sprawl management in Naqamte and similar cities.