| Abstract: |
Urban expansion represents a critical challenge for sustainable development, necessitating advanced analytical frameworks to monitor ecological transformations and environmental service provision. This review synthesizes current research on Geospatial Artificial Intelligence (GeoAI) applications for spatio-temporal urban growth assessment and ecosystem service evaluation. The study examines machine learning algorithms, deep learning architectures, and remote sensing technologies employed for land use/land cover change detection, urban morphology analysis, and environmental impact assessment. Through comprehensive literature analysis, we identify Random Forest, Convolutional Neural Networks, and ensemble methods achieving accuracies exceeding 90% in urban classification tasks. Results demonstrate GeoAI's capacity to integrate multi-source geospatial data for real-time monitoring, revealing significant urban expansion patterns globally, with developing nations experiencing 200-300% growth in built-up areas over two decades. Discussion highlights trade-offs between urbanization and ecosystem services, emphasizing vegetation loss, urban heat island intensification, and biodiversity decline. The review concludes that GeoAI frameworks provide robust, scalable solutions for sustainable urban planning, though challenges persist regarding model interpretability, data integration, and ethical considerations in deployment. |
