Abstract
Impervious surface area is a key factor affecting urbanization and urban environmental quality. It is of great significance to analysis timely and accurately the dynamic changes of impervious surface for urban development planning. In this study, we use a comprehensive method to extract the time series data on the impervious surface area (ISA) from the multi-temporal Landsat remote sensing images with a high overall accuracy of 90%. The processes and mechanisms of urban expansion at different political administration and direction level in the Nanjing metropolitan area are investigated by using the comprehensive classification method consisting of minimum noise fraction, linear spectral mixture analysis, spectral index, and decision tree classifiers. The expansion of Nanjing is examined by using various ISA indexes and concentric regression analyses. Results indicate that the overall classification accuracy of ISA is higher than 90%. The ISA in Nanjing has dramatically increased in the past three decades from 427.36 km2 to 1780.21 km2 and with a high expansion rate of 0.48 from 2000 to 2005. The city sprawls from monocentric to urban core with multiple subcenters in a concentric structure, and the geometric gravity center of construction land moves southward annually. The stages of urbanization in different district levels and the dynamic changes in different direction levels are influenced by the topographic and economic factors.
Highlights
Development of a city during rapid urbanization is often at the expense of natural landscape pervious surfaces which are replaced with an impervious surface composed of mixed clay, asphalt, stone, and metal [1]
The accuracy assessment of the confusion matrix showed that the overall accuracy of the Impervious surface area (ISA) data in Nanjing was more than 90% in 2010 and 2017 (Table 2)
We examined the relationship between Gross Domestic Product (GDP) and ISA by exploring regression models of 11 districts in Nanjing
Summary
Development of a city during rapid urbanization is often at the expense of natural landscape pervious surfaces which are replaced with an impervious surface composed of mixed clay, asphalt, stone, and metal [1]. Urban land cover extraction from RS images results in poor accuracy because an urban land is usually a mosaic of different land covers, such as plants, soils, buildings, and water [18,19]. Impervious surface area (ISA) is an essential component of urban landscapes and is the artificial surface that cannot be penetrated by water because of the composition of man-made construction materials in structures, such as road and building roofs [18]. ISA can truly reflect the expansion of urban construction land and can be extracted from RS images [13]
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
