Spatial-temporal analysis of industrial heat and productivity in China

• Published in: Applied geography (Sevenoaks) Vol. 138; p. 102618
• Main Authors: Lai, Jianbo, Zhu, Jun, Chai, JinChuan, Xu, Bingli
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.01.2022
• Subjects: DBSCAN; Industrial heat source Identification; Logistics; Spatial pattern; VNP14IMG; DBSCAN; VNP14IMG; Spatial pattern; Industrial heat source Identification; Logistics
• ISSN: 0143-6228; 1873-7730
• DOI: 10.1016/j.apgeog.2021.102618

China is currently in a key period of transformation and upgrading for its industrial infrastructure, and it is important to investigate the quantity, spatial distribution and production status of heat source industries to ensure effective supervision by environmental protection departments and to optimize the layout for regional industrial planning. In this study, based on the 2013–2019 Visible Infrared Imaging Radiometer Suite (VIIRS) I-band 375-m active fire data (VNP14IMG) product, a model combining density-based spatial clustering of applications with noise (DBSCAN) and logistic regression (LR) was adopted to identify industrial heat source objects in China, and the spatiotemporal patterns and trends of industrial objects at different levels were analyzed. The results indicate the following: (1) the VNP14IMG-based DBSCAN-LR identification model feasibly identifies industrial heat sources, and the accuracy of cross-validation is 93% and that of typical cities reaches as high as 92%; (2) The number of industrial heat source objects (NIHSO) in China from 2013 to 2019 exhibits a general downward trend, with the NIHSO values from 2013 to 2019 being 2673, 2608, 2374, 2273, 2284, 2317, and 2269, respectively; (3) the spatial pattern of regional industrial development is extremely unbalanced and can be characterized as being "dense in the north and sparse in the south". The industrial focus has shifted southwest, reflecting advances in the central and western regions achieved through industrial transfer; (4) the global spatial autocorrelation of industrial distribution is positive, and the correlation increases over time. The local correlation shows that the geographic concentration characteristics of “high-high” agglomeration in the north and “low-low” agglomeration in the west. •The VNP14IMG data and DBSCAN-LR method attain strong feasibility and robustness in the identification of large-scale IHSO.•The NIHSO in China has declined, but the production capacities of most heavy industries is increasing.•Long time series fire hotspot data can reveal the whole life cycle of heat source factory.•Objectively analyzed the change trend of the spatial pattern of China's industry.