Urban development boundary delineation supported by spatiotemporal big data: A case study of Changsha
Delineating urban development boundaries is a crucial measure for strengthening spatial development control and preventing the disorderly spread of urban areas. The land use expansion simulation model is a key tool for delineating these boundaries. However, previous work has primarily relied on land use data for boundary delineation, insufficiently reflecting human activities in the process. Additionally, there has been a lack of characterization of human economic activities such as business, healthcare, and education due to data and technological limitations. With urban development, human activities generate massive spatiotemporal big data, which contains rich information about human activities and socioeconomic conditions. Many scholars have begun using multi-source spatiotemporal data to depict human activities from a "human-centric" perspective. The rise of social sensing and the continuous progress in population spatial research have provided new opportunities for land use simulation and urban development boundary delineation research.In this study, multiple sources of spatiotemporal big data, such as POI data, check-in data, and population spatialization data, are collected and combined with land use simulation models to delineate urban development boundaries. The study focuses on Changsha, selecting Landsat images from 2000, 2010, and 2020. The random forest algorithm is used to classify land use and its changes in Changsha. Natural factors such as terrain, climate, soil, and vegetation, as well as social and economic factors such as population density, residential activity space distribution, education, and transportation infrastructure, are combined to analyze the driving factors for land use change. The CLUMondo model is used to simulate the land use change and delineate urban development boundary of Changsha City for 2035 based on current land use data and land use service demand. The overall accuracy of the model simulation is greater than 90％. The results show: (1)In 2035, the central urban area of Changsha will have a total of 1339 km2 of construction land. The urban development boundary in the central urban area is 1207 km2, with which 1157 km2 of construction land within the development boundary, accounting for 95.86％. (2)From 2020 to 2035, the eastern and western directions of the study area maintain relatively high expansion intensity, with significant agglomerative expansion in the east direction and diffusive development along the Xiangjiang River in the northwest direction, extending into the interior of Wangcheng District. (3)The agglomeration effect of construction land in the study area is significant, showing high consistency between population density, business activity intensity, and boundary simulation results. This paper integrates multi-source spatiotemporal big data into urban development boundary delineation research, aiming to reflect socioeconomic characteristics and "human" features at a more refined scale. This approach provides valuable reference and learning opportunities for simulating urban development boundary expansion in the era of big data.
POIbig dataimage classificationDMSP/OLSland useCLUMondourban land usespatiotemporal evolution
NETL
NSTL
万方数据
