海南岛GDP与夜间灯光指数之间的关系模型研究

GDP simulation of Hainan Island based on NPP/VIIRS night light data

本文提出结合NPP/VIIRS夜间灯光数据和统计年鉴数据模拟海南岛GDP的思路。通过提取夜间灯光总强度(TNL)、区域平均灯强度(I)和综合夜间灯光强度(CNLI)3种夜间灯光指数,分别构建2012—2018年灯光指数与GDP的回归模型,利用2019年GDP进行预测验证,遴选出最佳的回归模型,并利用其构建海南岛GDP空间化密度图。结果表明:(1)通过模型对比,TNL指数的线性拟合模型是最佳的,海南岛2019年GDP预测值与真实值的相对误差仅为-0.71%,更适合海南岛GDP的模拟;(2)海南岛GDP的空间分布整体呈沿海高、内陆低,并随时间由沿海区域向内陆扩展的趋势。

In order to reflect the economic law more quickly,in real time and objectively,the idea of combining NPP/VIIRS night light data and statistical data to simulate the GDP of 18 cities and counties in Hainan Island is proposed.By extracting three night light indexes of total night light intensity(TNL),regionally average light intensity(I)and comprehensive night light intensity(CNLI),the regression models of three light indexes and GDP from 2012 to 2018 are firstly constructed.Then,the best regression model is selected by using the 2019 GDP predicted value.Finally,the spatial density map of GDP in Hainan Island is drawn according to the model.The results show that:(1)Through model comparison,the linear fitting model of TNL index and GDP is the best,and the relative error between the predicted value of Hainan Island’s GDP in 2019 and the actual value is-0.71%,which is more suitable for the simulation of Hainan Island’s GDP;(2)The overall spatial distribution of GDP in Hainan Island is high in coastal areas and low in inland areas,and tends to expand from coastal areas to inland areas with time.