A significant difficulty in urban studies is obtaining urban areas. Nighttime light (NTL) data provide efficient approaches to map urban areas. Previous methods have utilized visual particularities of cities with ancillary data to obtain the optimal thresholds. How cities behave differently from rural areas should be considered. A Zipf's law-based method is proposed to bootstrap the optimal threshold based on the statistical properties of a Zipf's law model on continuous thresholds at the country scale. In our method, the Zipf's law model is utilized to quantify fractal, self-organized, and agglomeration behaviors of cities. The three-phase cluster dynamics are discovered and the abrupt transition between Phase 1 and Phase 2 denotes the rural-urban demarcation point. The urban areas are derived by the proposed method from the Suomi National Polar-Orbiting Partnership Visible Infrared Imaging Radiometer Suite (NPP-VIIRS) NTL data in 2013 in China. An accuracy assessment is conducted to compare it with the GlobeLand30-2010 data and the overall accuracy has directly confirmed the effectiveness of the method. The validation using point of interest (POI) data verifies that the urban areas show strong responses to social interactions and production with R 2 values of 0.91 and 0.92, respectively, implying that the city areas extracted by our method can be a proxy for human activities. Comparisons with existing methods validate the effectiveness and high degree of automation of the proposed method in mapping urban areas at the country level. According to our analyses, the Zipf's law-based method shows great potential to provide a universal criterion to map urban areas from the perspective of the behaviors of urban systems without ancillary data, and a valuable tool for spatial and temporal urban studies.

No abstract