Monitoring and modeling urban expansion-A spatially explicit and multi-scale perspective

In the context of promoting new urbanization in China, urban expansion has been the subject of a consolidated line of research in the past two decades. In this study, we integrate remote sensing, a geographical information system, and spatial analysis techniques to monitor and model urban expansion with a spatially explicit and multi-scale perspective in Wuhan, the megacity in central China. We first extract urban built-up land from medium to high resolution images from 1995 to 2010 and analyze the expansion dynamics at the parcel level. For the exploration of the driving mechanisms underlying urban expansion, 20 explanatory variables are then categorized into three groups: characteristics, density, and proximity. A kernel window is then utilized to filter the extracted urban built-up land map at multiple scales. Moran's I is later used to test the spatial autocorrelation in the percentage of urban built-up land area and the residuals. Finally, a spatial lag model and a spatial error model are applied to explore the causal factors of urban built-up land. It is revealed that the area of transportation land in Wuhan has increased tremendously, and urban built-up land is less scattered at the micro-scales. Regardless of the scale or model, housing density and gross domestic product (GDP) are positively correlated with the urban built-up land area, whereas the influence of other factors is shown to vary along with the scale or model. The results also confirm the superiority of the spatial regression models, and better fitting is produced with the increase in the scale. In the future, it is anticipated that the interpretation of remote sensing images and spatial analysis techniques will be optimized for better manifestation of intra-urban spatial interaction. In this process, the scale effect is an issue that should not be neglected in future studies of urban expansion.