Urbanization is one of the key indicators of socio-economic development, representing the transition from a rural to an urban society. It involves profound transformations in population distribution, economic structure, social organization, and land use patterns. A quantitative analysis of the spatio-temporal differentiation of urbanization can provide a scientific basis for optimizing its development. With advances in remote sensing technology, nighttime light data has emerged as an innovative research tool, offering a more precise means of monitoring spatial disparities and the agglomeration patterns of economic activities during urbanization.This paper utilizes long-term nighttime light remote sensing data, employing the total nighttime light value and the nighttime light growth difference index as proxy variables to assess urbanization levels. By integrating methods such as center-of-gravity migration, the standard deviation ellipse, the global Moran’s I index, and local spatial autocorrelation, the study quantitatively analyzes the gradient characteristics and spatio-temporal differentiation of urban development across provincial capitals, prefecture-level, county-level, and township-level areas in China over the past 30 years, focusing on contribution rate, growth rate, and center-of-gravity distribution.The results indicate that (1) Contribution Rate: Between 1992 and 2022, the nationwide total nighttime light value maintained steady growth. In the early 1990s, prefecture-level cities exhibited the highest contribution rate, with provincial capitals and townships contributing equally, while county-level cities contributed the least. After 2000, township-level cities surpassed provincial capitals, and around 2004 they exceeded prefecture-level cities, becoming the highest contributing region. In 2007, county-level cities overtook provincial capitals, and by 2022 their contribution was comparable to that of prefecture-level cities; meanwhile, provincial capitals became the lowest contributing tier after 2010. (2) Growth Rate: From 1992 to 2000, provincial capitals and some prefecture-level cities led other urban tiers, with the Beijing-Tianjin-Hebei, Yangtze River Delta, and Pearl River Delta regions being particularly prominent. Between 2000 and 2010, provincial capitals began to approach saturation, while prefecture-level and some county-level cities gradually emerged as the core; these areas in North, East, Central, and South China exhibited medium or higher growth rates. However, in the Northwest, Southwest, and Northeast—where development levels were lower—provincial capitals remained the fastest growing among the four tiers. From 2010 to 2022, township-level cities became predominant, while provincial capitals experienced predominantly “slow-to-low” growth, indicating a gradual shift in the center of urbanization toward townships. (3) Center-of-Gravity Distribution: From 1992 to 2022, China’s urbanization underwent significant changes. The center of urban development exhibited a gradient downward shift, progressively transferring from provincial capitals to prefecture-level, county-level, and finally township-level cities. This migration trajectory is highly consistent with the “core–periphery” theory, and the expanding area of the urbanization ellipse demonstrates a clear expansion effect. In terms of spatial correlation, “high–high” clusters have shown a trend of shifting from the eastern coastal areas toward the northwestern inland. The urbanization growth levels across different tiers exhibit significant positive spatial correlation, with Moran’s I values generally above zero but on a declining trend. From 1992 to 2000, “H-H” clusters were primarily concentrated in provincial capitals and coastal prefecture-level cities; from 2000 to 2010, county-level cities accounted for a larger share of the high-value clusters; and from 2010 to 2022, township areas emerged as new growth poles, while some provincial capitals formed “L-L” clusters with internal urbanization nearing saturation. These findings mark the entry of China’s urbanization into a new phase characterized by a multi-centric and networked development pattern.

Urbanization;nighttime light remote sensing;Spatiotemporal Differentiation;Gradient Characteristics