关键词: 夜间灯光数据, 城市扩张, 空间自相关, 参数最优地理探测器

Abstract: To study the spatiotemporal changes in urban built-up areas in China over the past 20 years from 2002 to 2021, promoting the urbanization process, and socially sustainable development in China, urban units in China were selected as the research objects. DMSP-OLOS data from 2002 and NPP-VIIRS night light data from 2021 were used, combined with indicators such as expansion speed, standard deviation ellipse, and spatial autocorrelation. Quantitative analysis of the spatial distribution characteristics in built-up areas is elaborated in two aspects: quantity structure and spatial distribution. Some factors are used as independent variables such as GDP difference, population difference, elevation, road network density, temperature difference, and rainfall difference. Expansion speed is used as the dependent variable. The driving force is quantitatively analyzed by parameter optimal geographic detector. 1) In the past 20 years, the expansion speed of urban built-up areas in our country has been mainly in medium and low speed, The annual expansion rate of the built-up areas in the eastern and northern coastal areas is higher than the national average. The middle reaches of the Yellow River and the southern coast are comparable to the national average, while the middle reaches of the Yangtze River, Southwest China, Northeast China and Northwest China are significantly lower than the national average. The eastern coastal areas have the fastest expansion speed and the smallest difference in urban development, while the center of gravity of China′s built-up areas moves 100.22 kilometers from the eastern coastal areas to the northwest. 2) The expansion speed of the built-up area has spatial heterogeneity, reflecting the accumulation of cold and hot spots. Cold spot areas are distributed in the northern and central regions of China and exhibit low-low clustering and high-low clustering, while hot spot areas are distributed in the southeastern coastal areas of China and exhibit high-high clustering and low-high clustering; 3) GDP,population are the main factors affecting the development of built-up areas, with the explanatory power of 0.6395, 0.4965, respectively. The interaction between various factors is either double factor enhancement or non-linear enhancement, with strong interaction between socioeconomic factors, and the strongest interaction between GDP and population factors. The explanatory power of 1.1361 shows a double-factor enhancement.

Key words: nighttime lighting data, urban sprawl, spatial autocorrelation, parameter optimal geographic detector