A study of periurban areas land uses based on their fractal dimension: the case of three periurban municipalities of Buenos Aires.

The urbanization process has accelerated in the last fifty years, particularly in Latin American cities such as Buenos Aires. In the face of this situation, the local government efforts turn insufficient for planning a proper urban expansion. In this context, remote sensing provides information for studying the process of soil occupation. This paper proposes a processing chain that allows the extraction of land use maps for periurban areas, as well as the study of impervious surfaces variability, over a time span of thirty years. The methods used for this analysis include spectral linear unmixing and supervised classification of multispectral images (EO1 Hyperion, Landsat, SPOT). Subsequently, predominant urbanization types in periurban areas were identified based on their fractal dimension (D). D value allows the identification of the compaction degree of an area, hence it is considered a useful tool for prioritizing interventions and planning territorial expansion.