The capabilities of remote sensing to derive urban location factors for probability-based spatial growth analysis

Kurzfassung

Urbanization is arguably the most dramatic form of irreversible land transformation. Though urbanization is a worldwide phenomenon, it is especially prevalent in India, where urban areas have experienced an unprecedented rate of growth over the last 30 years (UN, 2007). This paper focuses on the capabilities of remote sensing to identify and derive spatial urban location factors which influence future urban growth. We utilize multitemporal remotely sensed data sets from Landsat and TerraSAR-X sensors as well as a digital elevation model (DEM) from the Shuttle Radar Topography Mission (SRTM). The land cover of the test site, the highly dynamic incipient mega city of Hyderabad in India, was classified and a change detection analysis was performed to monitor the dimension and the spatial configuration of urban growth since the 1970s. The results of the change detection as well as the DEM serve as basis to derive and develop spatial location factors influencing urban growth. Parameters like the slope, the major street network, continuous intra-urban open spaces, main direction of growth, etc. were calculated. Furthermore external data sets on locations of commercial centers, airports, etc. were integrated. Based on regional theory for every single parameter a specific hypothesis was stated. For example: We assumed that high slope gradients have a lower probability for future settlements or that new commercial centers have a positive influence for future settling. In addition, results from a comparative study of the 12 largest Indian cities (Taubenböck et al., 2009), like saturation effects for built-up density, were integrated as additional information. An approach combining all urban location factors for the metropolitan area of Hyderabad was developed to identify areas that are theoretically highly probable for future settlements. The approach was applied to the spatial physical extension of the urban area of 2001, the so called urban footprint. Accuracy was assessed for predicted areas of urban growth comparing the result to the actual urban footprint acquired in 2009. The results of the method basically showed high probabilities for those areas which actually have experienced growth, but the limitations of the approach revealed low absolute accuracy. This is due to the manifold parameters having an impact on spatial growth – e.g. socio-economic, physical, demographic or political parameters – which could not be derived using remotely sensed data. Thus, the method basically enables location study to differentiate between preferred and unlikely areas of future urbanization.