Modelling multi-density urban expansion using Cellular Automata for Brussels Metropolitan Development Area

Abstract. Our studies aim at modelling and simulating urban expansion scenarios for Brussels capital region, Brabant of Flanders and Wallonia. Thereby we use a non-ordered multinomial logistic regression (MLR) coupled with cellular automata. Our model helps to study the probability for built-up development based on a) impact of different causative factors on expansion process and b) effect of the neighbouring cells on future built-up development. In our study, we have used 100×100 m raster data representing cadastral built-up of our study area. The model is then calibrated using the maps for years 2000–2010 and to simulate 2020. Thereto, simulated 2020 maps has been validated with observed 2020 built-up maps using fuzzy set theory. Our results show that all through our study, zoning or land use policies play an important role for expansion along all the built-up density classes. Besides, slope distance to highways and major cities encourages new urban development commonly known as ‘Urban sprawls’. Distance to main roads, employment opportunities aids to further development from low density to medium density areas. While most of the factors impact negatively to further high density development showing that our area is highly governed by the zoning status in case of densification.