Análise histórica e projecção da expansão urbana do Município de Maputo usando técnicas de geoprocessamento e aprendizado de máquina

Abstract

The rapid urbanization in the Municipality of Maputo has generated significant transformations in land use and land cover, demanding tools that support the understanding of its dynamics and the prediction of future scenarios. The general objective of this study was to analyze the urban expansion of the Municipality of Maputo between 1984 and 2004, with projections for 2044, using Geoprocessing techniques combined with Machine Learning methods. Satellite images from the Landsat series (1984, 2004, and 2024) were used, classified through the Random Forest algorithm in the R environment, with subsequent validation based on the confusion matrix and Kappa index. The overall classification accuracy was high: 94.3% (1984), 96.1% (2004), and 92.21% in the 2024 simulation, with corresponding Kappa indices of 0.8737, 0.9143, and 0.8852. The 2024 simulation, performed with the MOLUSCE plugin in QGIS, served as the basis for validating the methodology. For the 2044 projection, an Artificial Neural Network approach was employed, parameterized with a learning rate of 0.002, 10 hidden layers, and 1000 iterations, considering conditioning variables such as slope, aspect, distance to built-up areas, roads, and water bodies. The results revealed a continuous pattern of built-up area expansion, increasing from 6,046 ha in 1984 to 29,325 ha in 2024, with a forecast to reach 38,760 ha in 2044. This growth occurred mainly over areas previously covered by vegetation and bare soil. Maputo shows signs of urban saturation, while KaTembe emerges as a new growth vector, driven by infrastructure projects such as the Maputo–KaTembe Bridge. KaNyaka, although still preserving larger vegetation cover, also exhibits signs of anthropogenic pressure. The validated projection allows for a better understanding of territorial occupation trends and offers a practical tool to support land-use planning decisions. Based on these results, it is possible to identify priority areas for intervention and planning, aiming for a more balanced and sustainable urban expansion