Vol. 6, No. 2, 101-114, 2007

Abstract
Urban remote sensing research and approaches to modelling residential mobility focus predominantly on growth patterns. In this paper, the phenomenon of extreme urban decline, named 'shrinkage', is scrutinised. The different characteristics of urban decline are illuminated using a multi-scale approach. Selected patterns of the spatial growth and shrinkage are first calculated by means of satellite imagery for the City of Leipzig, Germany. Here, Landsat data for 1994 and 2005 provide information regarding different phases of urban land use dynamics, thereby revealing a pattern of spatial expansion into the peri-urban surroundings. In addition, potential drivers of this detected pattern are investigated through analysis of municipal statistical data, at the local district level, providing evidence that urban growth in general and particularly shrinkage are results of population fluxes and migration. Because urban shrinkage can be found in both the central and peripheral parts of Leipzig City, an even more detailed scale, using a very high resolution (VHR) colour-infrared data set has then been integrated with the local district data, in order to achieve detailed information on intra-urban differentiation of both urban structure and fabric. Finally, using predictor variables such as fertility, life expectancy, migration and residential preferences, a prototype model approach is presented that analyses recent patterns of residential use and the related building vacancies that characterise the housing sector of a shrinking city.

ISSN 1729-3782