Construction of a Cities Evolution Tree, with Applications

China has been experiencing urbanization at an unprecedented rate over the last two decades. This study analyzes the impact of urbanization on land occupation. Cities are clustered by their functions and development stages, which is illustrated by a cluster tree, a dynamic tree that depicts the evolution of cities. The evolution tree in one year is used to predict the state of a city in a future time period. Another application of the evolution tree is to predict urban-type relevant phenomena, such as urban occupation. It is found that comprehensive cities, business cities, and manufacturing cities have higher urban expansion rates than tourist cities, with a few exceptions that focus on both industry and tourism. Meanwhile, the speed and extent of city land growth are dominated by industrialization stages and economic patterns, as well as leap-development. The methodology presented in this study is especially suitable for identifying transition paths of a stochastic process in a complex dataset of 253 cities in China.