Integrating Multisource Geographic Big Data to Delineate Urban Growth Boundary: A Case Study of Changsha

• Published in: IEEE journal of selected topics in applied earth observations and remote sensing 2024, Vol.17, p.9018-9036
• Main Authors: Gao, Xing, Xia, Nan, Zhuang, Sudan, Zhao, Xin, Liang, Jiale, Wang, Ziyu, Li, Manchun
• Format: Article
• Language: English
• Subjects: Accuracy; Big Data; Delineation; Ecological evaluation; Economics; Geographic Big Data; Geographical information systems; GIS spatial analysis; Human factors; Indicators; Machine learning; Planning; Quality of life; random forest (RF); Rivers; Sociology; Spatial planning; Surveys; Sustainable development; territorial spatial planning; Urban areas; Urban development; urban growth boundary (UGB); Urban sprawl; Urbanization
• ISSN: 1939-1404; 2151-1535
• DOI: 10.1109/JSTARS.2024.3389503

An urban growth boundary (UGB) is an important policy tool used to control urban sprawl, which can effectively balance the urban construction needs, residents' quality of life, and urban ecological protection. Current studies of UGB delineation and its indicators have paid little attention to human factors, such as human activities and economic vitality, and weights for evaluation indicators have been determined highly subjectively. In response to these problems, this article integrated multisource geographic Big Data to construct a total of 30 natural, human, and ecological evaluation indicators. The GIS technology and machine-learning (ML) approach were combined to determine indicator weights with an officially manually drawn 2035 UGB as the reference, aiming to reduce the subjectivity. The suitability score was then calculated from indicator grading 4-1 and related weights, and high suitability (>2.15) regions were eventually delineated as the UGB. Results showed a delineated UGB of 1528.06 km 2 with an overall accuracy of over 93% and high consistence with reference data for 2035 in the Chinese city of Changsha. The geographic Big Data totally contributed more than 33.72%, which mainly characterized role of human elements, and a 5-6 percentage point reduction in accuracy was found without these data. Compared with existing articles, our delineated UGB had higher accuracy and closer spatial pattern to the reference data, verifying the effectiveness and reasonability of ML-based weight setting approach and index system with geographic Big Data. The proposed method can provide scientific and accurate framework for UGB delineation, which can promote the territorial spatial planning and sustainable urban development.