基于夜间灯光数据的中国人口密度模拟

人口密度网格化比人口密度行政单元化更接近人口的实际分布，而且是实现人口数据与其他社会经济统计数据、资源数据、环境数据复合，提高人口、资源、环境综合管理能力的重要途径之一。选用专门针对亚洲地区开发的DMSP／OLS非辐射定标夜间灯光平均强度遥感数据模拟了中国的人口密度：基于灯光强度信息模拟了灯光区内部的人口密度，基于人口—距离衰减规律和电场叠加理论模拟了灯光区外部的人口密度，是应用DMSP／OLS夜间灯光数据模拟人口密度研究的拓展和深入，研究结果与其他研究基本一致，但所需数据量较少，更适合于大尺度人口密度的快速估算，为决策提供参考依据。结果表明，DMSP／OLS夜间灯光强度数据具有实现人口以及其他社会经济数据空间网格化的巨大潜力。

Modeling Population Density of China in 1998 Based on DMSP/OLS Nighttime Light Image

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.