基于多源数据的福建省人口数据空间化研究

Spatialization of Population Data for Fujian Province Based on Multi-source Data

针对人口统计数据无法精细直观反映人口真实的空间分布状况的问题,该文以福建省为例,对其2015年的人口数据进行空间化。以多源数据为基础数据,在县级尺度上采用了空间回归模型构建福建省2015年福建省500 m人口空间分布数据,在乡镇尺度上对其空间化结果进行精度验证并与WorldPop、中国公里网格人口分布数据集比较。结果表明:NPP/VIIRS夜间灯光、路网、DEM、土地利用数据与人口具有较强的相关性,有足够能力模拟人口的空间分布;对于2015年福建省人口数据空间化,空间误差模型比空间滞后模型的回归拟合效果更好;人口数据空间化结果精度比较高,在空间上能精细展现2015年福建省的人口分布状况;人口高值区主要集中在县城所在地,人口呈现出主城区高、四周低的空间分布格局。

Aiming at the problem that traditional census data can’t reflect the real spatial distribution of population visually,this paper took Fujian Province as a case study to conduct spatialization of population data for it in2015. A 500 m spatial population distribution data for Fujian Province in 2015 was established based on multisource data by spatial regression at county scale. The spatialization result accuracy of the population distribution data was verified at township scale and compared with the WorldPop and China population grid distribution datasets. The results show that NPP/VIIRS nighttime light,road,DEM and land use had high correlation with the population and could be used to simulate the population spatial distribution. The spatial error model was more effective than the spatial lag model in fitting the population for Fujian Province in 2015. The result of spatialization of population data produced was quite accurate,and it could show the fine population distribution for Fujian province in 2015 from space. High population areas were mainly concentrated in the county seat. The population presented a spatial distribution pattern that the population in the main urban area was high and the population in the surrounding area was low.