Ph.D student Yang Ju and Assistant Professor Iryna Dronova in Landscape Architecture & Environmental Planning, and their co-authors, Qin Ma, PhD candidate at University California, Merced, and Xiang Zhang, research associate at Nanjing University, recently published a paper using satellite-sensed night-time light images and machine learning to detect fine-scale urbanization trends over time. The paper, titled ‘Analysis of urbanization dynamics in mainland China using pixel-based nighttime light trajectories from 1992 to 2013’, is recently published in International Journal of Remote Sensing’s special issue, remote sensing of night-time light, featuring the application and methodological innovations of night-time light remote sensing.
Understanding how urbanization progresses over both space and time has several important implications in planning and management. Night-time light is suitable for such task as the data reflects three major aspects of urbanization: urban land-cover expansion, population growth, and economic activity intensification. This paper developed a framework to use night-time light time-series trajectories to identify major typologies of urbanization trends at 1 km by 1 km resolution. With several machine learning techniques, authors found five major trends– stable, high and low level steady growth, acceleration, and fluctuation, in mainland China between 1992 and 2013.
The paper focuses on one trend, acceleration, as it is a potential indicator for aggressive urbanization. The authors identified several spatial clusters of cities with high concentration of the acceleration trend. One cluster is in the Yangtze River Delta whose growth is facilitated by rural industries. The other is in the Inner Mongolian region where urbanization is stimulated by its rich natural resources and land-centered development policy.
Ju and Prof. Dronova concluded with the promising future of using night-time light time-series to monitor urbanization, as this approach provides a low-cost, timely, and large-scale understanding about how urbanization happens over space and time. Ju is looking forward to applying such approach to other topics, such as understanding urban exposure to climate change, and assessing environmental disaster impact and recovery.
Read the journal article in full here.