中国地理科学 ›› 2019, Vol. 20 ›› Issue (4): 541-552.doi: 10.1007/s11769-019-1054-y

• 论文 •    下一篇

Analyzing Population Density Disparity in China with GIS-automated Regionalization: The Hu Line Revisited

WANG Fahui, LIU Cuiling, XU Yaping   

  1. Department of Geography and Anthropology, Louisiana State University, LA 70803, USA
  • 收稿日期:2019-03-01 修回日期:2018-11-10 出版日期:2019-08-27 发布日期:2019-06-28
  • 通讯作者: WANG Fahui.E-mail:fwang@lsu.edu E-mail:fwang@lsu.edu

Analyzing Population Density Disparity in China with GIS-automated Regionalization: The Hu Line Revisited

WANG Fahui, LIU Cuiling, XU Yaping   

  1. Department of Geography and Anthropology, Louisiana State University, LA 70803, USA
  • Received:2019-03-01 Revised:2018-11-10 Online:2019-08-27 Published:2019-06-28
  • Contact: WANG Fahui.E-mail:fwang@lsu.edu E-mail:fwang@lsu.edu

摘要:

The famous ‘Hu Line’, proposed by Hu Huanyong in 1935, divided China into two regions (southeast and northwest) of comparable area size but drastically different in population. However, the classic Hu Line was derived manually in absence of reliable census data and computational technologies of modern days. It has been subject to criticism of lack of scientific rigor and accuracy. This research uses a GIS-automated regionalization method, termed REDCAP (Regionalization with Dynamically Constrained Agglomerative Clustering and Partitioning), to reconstruct the demarcation line based on the 2010 county-level census data in China. The results show that the logarithmic transformation of population density is a better measure of attributive homogeneity in derived regions than density itself, and produces two regions of nearly identical area size and greater contrast in population. Specifically, the revised Hu Line by Hu Huanyong in 1990 had the southeast region with 94.4% of total population and 42.9% of total land, and our delineation line yields a southeast region with 97.4% population and 50.8% land. Therefore, the population density ratio of the two regions is 27.1 by our line, much higher than the ratio of 22.4 by the Hu Line, and thus outperforms the Hu Line in deriving regions of maximum density contrast with comparable area size. Furthermore, more regions are delineated to further advance our understanding of population distribution disparity in China.

关键词: Hu Line, regional population density disparity, GIS-automated regionalization, REDCAP (Regionalization with Dynamically Constrained Agglomerative Clustering and Partitioning), China

Abstract:

The famous ‘Hu Line’, proposed by Hu Huanyong in 1935, divided China into two regions (southeast and northwest) of comparable area size but drastically different in population. However, the classic Hu Line was derived manually in absence of reliable census data and computational technologies of modern days. It has been subject to criticism of lack of scientific rigor and accuracy. This research uses a GIS-automated regionalization method, termed REDCAP (Regionalization with Dynamically Constrained Agglomerative Clustering and Partitioning), to reconstruct the demarcation line based on the 2010 county-level census data in China. The results show that the logarithmic transformation of population density is a better measure of attributive homogeneity in derived regions than density itself, and produces two regions of nearly identical area size and greater contrast in population. Specifically, the revised Hu Line by Hu Huanyong in 1990 had the southeast region with 94.4% of total population and 42.9% of total land, and our delineation line yields a southeast region with 97.4% population and 50.8% land. Therefore, the population density ratio of the two regions is 27.1 by our line, much higher than the ratio of 22.4 by the Hu Line, and thus outperforms the Hu Line in deriving regions of maximum density contrast with comparable area size. Furthermore, more regions are delineated to further advance our understanding of population distribution disparity in China.

Key words: Hu Line, regional population density disparity, GIS-automated regionalization, REDCAP (Regionalization with Dynamically Constrained Agglomerative Clustering and Partitioning), China