Delineation of Central Business Districts in mega city regions using remotely sensed data

Abstract

Central Business Districts (CBDs) are an important urban structural type (UST), and an apparent structural feature of many large cities. CBD locations play a decisive role in the spatial arrangement of functions and exposures within cities. However, while past research underscores the importance of their spatial detection, delineation and cartographic representation, the definitions used are mostly functional and qualitative. Objective pre-defined methods/thresholds for the semi-automated spatial classification of CBDs, based on a quantitative approach, do not yet exist. This paper presents a conceptual framework to define the CBD using physical and morphological parameters, and tests the approach using 3-D city models of three European test sites (Canary Wharf in London, La Defense in Paris, and Levent in Istanbul). From these case studies, we develop a transferable method to detect and delineate CBDs over larger areas from a combination of Cartosat-1 digital surface models and multispectral Landsat ETM+ imagery. Applying this wide-area method to the entire extents of the three European megacities of London, Paris and Istanbul, we detect CBDs with a user accuracy of 75.7% and spatially delineate them with overall accuracies of 82.9%. Finally, we apply spatial metrics to analyze and compare the location and distribution of CBDs across the three mega cities, finding many similarities between London and Paris, but showing that Istanbul features a more complex urban footprint, and a different spatial CBD pattern.