Mapping hierarchical urban boundaries for global urban settlements

Urban boundary is the spatial basis for urban statistics and urban planning. However, the city definition by which the urban boundary is determined is not comparable among different countries. A globally consistent delimitation of hierarchical urban boundaries (HUBs) is rare but urgently needed. In this research, we proposed a model to consistently identify HUBs according to physical urban entity. The main quantitative basis of our model is the density, size and spatial relationship of the artificial impervious surface. We produced a vector dataset of global HUBs (named GHUB) based on 30-meter resolution remote sensing products. Cross-product comparison shows the effectiveness and advantages of our model. We furtherly, for the first time, investigated and compared the physical form and spatial characteristics of global urban settlements through the same lens supported by GHUB. We have identified 10,242 urban settlements with an area above five square kilometers, 39.18% of them are located within 100 km of the coastline. Urban settlements are not the “reinforced concrete forest” as commonly thought. The global average proportion of open space within urban settlements is 41.51%. The total area of urban open space in the world is 279.47 thousand square kilometers. Notably, we verified the Rank-size rule of urban settlement is universal, meanwhile, the verification results of different urban systems in 47 countries show that the Zipf’s exponent varies from country to country. Our work can help to consistently investigate global urban settlements without being affected by various definitions of cities. The GHUB can not only be applied to various urban studies but also can support global affairs such as the Sustainable Development Goal 11 due to its international comparability.