Predictive modelling of land use land cover dynamics for a coastal urban city in Brazil

Better urban planning depends on assessing how land use and land cover (LULC) have evolved in recent decades and what the prospects are for change in the future. Cities are the result of various factors interacting, and land configuration directly influences rainwater management, thermal comfort, and urban mobility, among other aspects. This study analysed changes in LULC in the municipality of Recife, Brazil, between 1990 and 2080. The study used predictive modelling to support territorial management strategies. The methodology entailed the integration of data on LULC from the MapBiomas project, followed by the implementation of a model based on cellular automata and artificial neural networks. The model was executed using the MOLUSCE plugin within the QGIS software framework, and the validation showed a percentage of correction equal to 89.08 % and kappa coefficients greater than 0.80, indicating excellent predictive performance. The results clearly indicate a significant advance in urbanization, driven by factors such as demographic expansion and economic development in the region, resulting in the conversion of pasture areas and native vegetation. The area of forest cover initially declined, then stabilized after 2050, likely due to conservation policies and spatial limitations on deforestation. Mangroves and water bodies remained relatively stable despite growing urban pressure. Research clearly highlights the need for public policies aimed at sustainable urban planning, the implementation of green infrastructure, and the strengthening of environmental legislation. Future projections make it clear that preventive actions are essential to minimize climate and socio-environmental impacts while promoting resilient urban development.