Estimating Multiple-Scale GDP Distribution Using Nighttime Light and Spatial Methods

摘要

Nighttime light (NTL) data derived from Visible Infrared Imaging Radiometer Suite (VIIRS), carried by the Suomi National Polar Orbiting Partnership (NPP) satellite has been widely used as an important index of modeling gross domestic product (GDP). Nevertheless, due to the difference of two kinds monthly composite data, which version is better to estimate the multiple-scale GDP distribution needs to be examined. The GDP distribution in China is always accompanied by regional disparity, but previous studies hardly considered the spatial autocorrelation. This research compared the effect of two kinds NPP-VIIRS monthly NTL data based on eigenvector spatial filtering (ESF) regression to enhance the precision of the GDP estimation. The regional NTL values were used with permanent resident population data to estimate 2015 China mainland province, city level GDP and the results were clearly analyzed.