Commercial land use change and growth processes – An assessment of retail location in Lisbon, Portugal, 1995–2020

Abstract

Land use modelling is essential to understanding future land use and change dynamics. Analyses of land use and change using Cellular Automata (CA) have been frequent in the literature since the latter were introduced to geography in the 1970s. In this paper, the statistical subsections – city blocks – of a previous analysis of Lisbon's commercial structure are used as the cells of a CA. A multi-layer perceptron (MLP) models the transition potential of those cells between 1995 and 2020, considering the impact of street centrality, subway development and other proxies of urban growth. The results confirm that commercial growth and change can be simulated as a function of street centrality and that using Local integration and Choice as drivers of a CA may help improve predictions of commercial growth and change. Moreover, the CA results reinforce that street centrality positively affects street commerce, especially when combined with access to transit and amenities. While these results can already inform planning practice and policymaking, the framework presented in this paper may be replicated to simulate ​future commercial land use, thus helping planners and policymakers prepare for change.