Abstract
Inconsistent measurements of city-size and a lack of time-series information on urban socioeconomic development have hindered determining whether China's city-size distribution (CD) follows a Pareto distribution according to multiple perspectives. This article has attempted to evaluate China's CD based on the defense meteorological satellite program- operational line-scan system (DMSP-OLS) nighttime light data in terms of socioeconomic size (SS) and spatial size (SC). First, city size was defined from the DMSP-OLS data. Then, whether China's CD followed a Pareto distribution was evaluated from different perspectives. The results show that China's CD from 1995 to 2015 presents a flat distribution trend; the flat distribution trend of the SC is more obvious than that of the SS; “borrowed size” has become an important reason for the flat trend of China's CD; and residential suburbanization, transportation cost reductions, local government policies, and land finances could effectively explain the CD differences in the flat trends between the SS and SC. This article offers an effective means for quantifying and comparing CD in long time series at a large scale (e.g., national scale or regional scale) and provides a scientific decision basis for governments to build a reasonable CD system in China.
Highlights
C ITY-SIZE distribution (CD) has consistently remained a topic of interest in urban geography and economicsManuscript received February 27, 2021; revised March 31, 2021 and April 15, 2021; accepted May 10, 2021
Because the Pareto regression is sensitive to city samples, the top 50, 100, 150, 200, and all cities were selected for the regression test
From the perspectives of changes of time and city samples, the coefficients show an increasing trend, which indicates that the socioeconomic size (SS) presents a flat distribution trend
Summary
C ITY-SIZE distribution (CD) has consistently remained a topic of interest in urban geography and economicsManuscript received February 27, 2021; revised March 31, 2021 and April 15, 2021; accepted May 10, 2021. This article has supplementary downloadable material available at https://ieeexplore.ieee.org, provided by the authors. The Pareto distribution indicates that the regression coefficient between logarithmic city ranks and their logarithmic city sizes yields an approximate constant [7], [8]. Zipf [9] proved that the Pareto distribution coefficient of CD is close to 1 and is typically called Zipf’s law. If the Pareto distribution coefficient is not equal to 1, it is considered to deviate from Zipf’s law [10]. When the coefficient is greater than 1, this indicates that the regional CD is more uniform than that of Zipf, which is commonly called a flat distribution. By comparing and analyzing the Pareto distribution coefficient, we can trace and identify the uniformity and rationality of the CD system
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
