Revisiting the extension of the Brussels urban agglomeration: New methods, new data… new results?

Two recent techniques are here used to delineate the urban agglomeration of Brussels : a fractal methodology for extracting the border between urban and not-urban surfaces, and a network analysis based on modularity applied on two different interaction matrices (telephone fluxes, commuting). In order to define spatial subsets in which relationships between places are more intense. These methods have the advantage of not depending upon a priori defined thresholds : thresholds are fixed by the methods and data themselves (endogenous). These methods have been applied on the entire territory of Belgium (Blondel, Krings and Thomas, 2010 ; Tannier and Thomas, 2012) and are here discussed for Brussels only. Results are further compared to more traditional delineations. Observed differences are mainly observed at the outskirts of the agglomeration but are very important in terms of policy : the urban-rural limit depends upon the criterion (definition, threshold) and method used. The two methods used here have the advantage of not depending upon arbitrary choices.