Urban Indicators of China from Radiance-Calibrated Digital DMSP-OLS Nighttime Images

Abstract

A composite of cloud-free radiance-calibrated Defense Meteorological Satellite Program (DMSP)-Operational Linescan System (OLS) nighttime images of China acquired between March 1996 and January– February 1997 in 30-arc-second grids in byte format is evaluated for its usefulness in extracting urban indicators data of China. With the aid of image-processing and Environmental Systems Research Institute’s ArcView GIS software with 3-D Analysis extension, zonal variations of radiance were extracted and three-dimensional models created using Triangulated Irregular Network (TIN) functionality for thirty-five cities. Two variables were extracted from the TIN model: surface area and volume. These were used separately as independent variables in the form of an allometric growth model to estimate the following urban indicators for individual Chinese cities: nonagricultural population, gross domestic product (GDP), built-up area, and electricity consumption. The estimates obtained were checked against data supplied by the Statistical Bureau of China for 1997, the same period during which the DMSP-OLS images were acquired. It was found that urban indicators of acceptable accuracy could be obtained, provided that cities designated as Special Economic Zones (SEZs) were excluded. Volume appeared to be slightly better than surface area in estimating these urban indicators. In all cases, total nonagricultural population for all the cities combined could be very accurately estimated, indicating the usefulness of the radiance-calibrated DMSP-OLS nighttime images to determine the level of urbanization in China. The three-dimensional TIN model portrays an illuminated urban area (IUA) dome, the shape of which is affected by the internal structure of the city. The core of the city invariably exhibits the highest level of light intensity in the DMSP-OLS nighttime image. However, only a very limited amount of land-use information could be extracted. The image was also found to be useful in detecting the special urban phenomenon in China known as the extended metropolitan region, in which rural and urban uses intermixed and small towns developed. In conclusion, despite the low spatial resolution, the radiance-calibrated DMSP-OLS nighttime images are capable of providing useful demographic and socioeconomic data for cities.