Spatiotemporal change analysis of urban land surface component based on V-I-S Model —A case study in Guangzhou

The components of urban surface are diversified and mainly included building area, vegetation, water body and so on. To give urbanization a quantitative analysis in order to describe and forecast some problems, which occurred in the process of urbanization, the analysis of urban land surface components is important and necessary. One of the recent major advances in urban land surface components analysis is V-I-S (vegetation, impervious surface and soil) model, which is presented for urban land surface components change and deemed the urban as the combo of three components: Vegetation, Impervious surface and Soil. Ignoring water surfaces, the V-I-S model included the most fundamental components of urban land surface cover and may reflect the connotation of urban expansion. Remote sensing have developed a variety of approaches to distinguish surface cover through multi-spectral data and has became the most convenient and popular means to map the V-I-S components in urban area for its wide area coverage and regular orbiting period. As the third biggest city in China, Guangzhou has been in a rapid urbanization process during the past two decades. The V-I-S model is implied in the study to quantify the composition of urban land surface with multi-temporal Landsat images between 1990 and 2009 to explore a spatiotemporal change of urban land surface components. The linear spectral mixture analysis method was used to extract three components in V-I-S model at sub-pixel level. The results show that: temporally, V-I-S proportional changes in the time series reveal different patterns of urban development in different period. Spatially, spatial variation analysis in different districts can highlight regional differences in urban sprawl process. The significant difference of urban expansion pattern is analyzed in the study by assessing the changing abundance and area of vegetation, impervious surface and soil from 1990 to 2009.