Analysis of Urbanization in China by Remotely-sensed Data

China has been undergoing rapid urbanization since the "open door" policy in 1978. The fast process of urbanization has significantly influenced its society, economy, and environment. To quantify and describe this process, various indicators of urbanization, which are usually extracted from statistical yearbook and census data, have been adopted. However, these data sets are usually inconsistent, problematic, and cannot depict the spatial information of urbanization. Therefore, remote sensing images are usually employed as complementary datasets.;However, the existing studies remain insufficient for understanding how the information of urbanization can be extracted from remote sensing imagery and used properly, especially for China, where urbanization has unique characteristics. Therefore, this thesis aims to explore the usage of remote sensing techniques to further observe the process of urbanization and from four different perspectives.;From the perspective of socioeconomic development, we analyze the spatial-temporal relationship between indicators derived from NTL images and the socio-economic indicators of urbanization. The results of the analysis indicate that the summed lights in a city can represent the overall level of urbanization and that the average light of lit-up areas can reflect the density of urbanization. Meanwhile, when the amount of NTL approaches saturation, it becomes a less sensitive reflection of the level of urbanization. The proper NTL indicator has then been utilized on the analysis of urbanization in China's cities during the last 20 years, and the results reveal that the cities with political and geographical advantages have higher levels of urban development. Meanwhile, the cities in metropolitan areas and the Shandong province have undergone a more compact urbanization process than some inland cities.;Second, from the perspective of urban expansion, we extract the time series of urban built-up areas at the city level via a newly-proposed thresholding technique on the calibrated time series of NTL images. The threshold for extracting built-up areas has been found to vary across different cities and years, and it has high correlations with the level of economic development. We then analyze the urban expansion in Chinese cities based on prefecture cities in three provinces of south China. The results indicate that urban expansion occurred in all cities from 1992 to 2010, especially in coastal cities, capital cities, and cities in special economic zones.;Third, from the perspective of spatial pattern evolution, we explored how to quantify the urban spatial pattern, and use it to study the evolution of metropolitan areas. We compare the discrepancies of various remote sensing images in describing spatial patterns and combine the landscape metrics, Pareto distribution, and gradient analysis to explore growth type, distribution, and reaction of cities in metropolitan areas. Moreover, based on the comparison of the spatial patterns among three of the largest metropolitan areas of China during the last twenty years, we find that the driving force and growth type vary over metropolitan areas and that each area has its own regional characteristics.;Fourth, from the perspective of urban sprawl, we introduced two sets of indicators, which can measure urban sprawl both as a certain spatial pattern of urban development and as a type of urban growth, that quantify urban sprawl based on NTL imagery. The results present the degree of urban sprawl in various metropolitan areas in China.;Overall, this thesis extends our understanding on how to use information derived from remote sensing as a proxy for studies on urbanization. Moreover, urbanization in China is scrutinized by remote sensing indicators.