黄河流域经济差异的空间多尺度统计建模

Spatial Multi-level Statistical Modeling of Economic Disparities in the Yellow River Basin

区域经济差异的研究多聚焦于单一空间尺度(县、市、省)或将不同空间尺度进行主观割裂,难以准确刻画经济活动在不同空间尺度上的系统差异以及估算经济差异的影响因素。本文提出一个空间多尺度统计建模框架,系统地对不同尺度的经济活动同时进行建模,并提供一个基于模型的区域经济差异分解方法。本文应用1992年至2018年的夜间灯光数据作为社会经济水平的代理变量,以黄河流域为案例区域,分析了经济活动差异在栅格、县、市、省四个尺度上的相对大小及其时间动态演化,基于拓展的Shapley Values方法,量化了自然、区域创新能力和初始发展条件对黄河流域经济发展的相对贡献。结果表明:黄河流域区域经济差异是一个动态变化过程,栅格尺度的差异持续上升,县级和省级尺度的差异呈现波动下降趋势,市级尺度的差异保持相对稳定的趋势。自然因素对黄河流域经济发展的贡献度最大,约占整体模型解释力的75%;区域创新能力贡献较低,仅为0.1%。

Previous studies on regional economic disparities tended to focus on a single spatial scale(province/county)analysis,despite that the nature of economic disparity and data under study was usually multi-scale.In this paper,we proposed a spatial multi-level statistical modelling framework to investigate economic development at multiple scales simultaneously.They include scales of province,city,county and fine-resolution grids in the Yellow River Basin.With economic activity intensity measured by the nighttime light data from 1992 to 2018,we found that grid-scale economic disparity continued to rise during the study period,whilst county-and province-scale variabilities tend to decrease gradually.By extending the classic Shapley value method to the multi-level statistical model,relative contributions of various factors were further quantified.Overall,the contribution of natural factors to regional economic variability in the Yellow River Basin is the largest,accounting for roughly75%of the growth disparity explained,followed by initial condition factors.