Modeling Population Density of China in 1998 Based on DMSP/OLS Nighttime Light Image
基于夜间灯光数据的中国人口密度模拟

Spatial distribution of population density is crucial for analyzing the relationship among economic growth, environment protection and resource utilization. In this study, population density of China in 1998 at 1-km resolution grids was simulated by integrating DMSP/OLS non-radiance calibrated nighttime light image, SPOT/VEGETATION 10-day maximum NDVI data, population census data and vector data of county boundary. Not only the population density in light patches but also that out of them was estimated in four types of areas. For each area, in light patches, the model for population density estimation was developed based on the significant correlation between light intensity and population, and in "dark area", the models for population density estimation were developed based on Coulomb's law and field superposing theory. Compared with the existed methods for spatializing population density, our method is simpler and more cost saving. The result of the study is consistent with those of other researches on the whole, but the spatial difference is more distinct and the information is richer. The maximum population density simulated is 41096 persons/km2 and average population density at inhabitable area of China is 189 persons/km2. It indicates that the 1-km resolution non-radiance calibrated DMSP/OLS nighttime lights image has the potential to provide population density estimation at 1-km grids.