基于多源遥感和POI数据的多尺度相对贫困评估——以滇桂黔石漠化片区为例

Multiscale relative poverty assessment based on multi-source remote sensing and POI data——A case study of rocky desertification area in Yunnan,Guangxi and Guizhou

减轻相对贫困和缩小贫富差距是国家实现共同富裕的关键。准确有效地识别和评估相对贫困已成为政府制定合理的发展政策的重要前提。以滇桂黔石漠化连片特困区为研究对象,基于多个数据来源,从多角度分析相对贫困问题。首先,将夜间灯光、土地利用、DEM、NDVI、社会经济统计和兴趣点(POI)等多源数据结合起来,构建多维(人口、社会、经济、资源、生态、灾害)发展指数模型(CMDI),量化识别相对贫困县;然后,综合运用莫兰指数、空间局部自相关等分析方法,对相对贫困水平时空动态差异进行研究;最后,构建格网尺度的相对贫困空间分布。结果表明:基于CMDI识别出2021年广西石漠化片区不同维度相对贫困状态,其中,多维极贫困县、贫困县和较贫困县占比分别为5.7%、20%和22.9%;基于夜间灯光指数识别的2011年滇桂黔石漠化片区大部分的县域属于极贫困区和贫困区,其中极贫困区占比55%,贫困区占比34%;滇桂黔石漠化片区的多维相对贫困程度在空间上呈现显著的空间正相关性,2016年和2021年均没有“低—低”类型,表明减贫成效显著,由2011年的集中连片分布转变为离散分布;格网尺度的相对贫困空间分布,可以有效展现研究区贫困状况的时空演变。

In 2020,China eliminated absolute poverty and achieved comprehensive poverty alleviation,but the problem of relative poverty continues to exist for a long time.The development of contiguous poverty-stricken areas and the management of relative poverty were the keys to achieving common prosperity for the country.Based on multiple data sources,this paper analyzed the problem of rela-tive poverty from multiple perspectives.Firstly,multi-source data such as night light,land use,DEM,NDVI,socio-economic statistics and points of interest(POI)were combined to construct a multi-dimensional(population,social,economic,resource,ecological and disaster)development index model(CMDI)to quantitatively identify relatively poor counties.Then,the Moran index and spatial local autocorrelation were used to study the spatiotemporal dynamic differences of relative poverty levels.Finally,the spatial distribution of relative poverty at the grid scale was constructed.The results showed that:The relative poverty status of different dimensions in the rocky desertification area of Guangxi in 2021 was identified based on CMDI,among which the multi-dimensional extremely poor coun-ties,poor counties and poorer counties accounted for 5.7%,20%and 22.9%,respectively.In 2011,most of the counties in the rocky desertification area of Yunnan,Guangxi and Guizhou belonged to extremely poor areas and poverty-stricken areas,of which 55%were extremely poor areas and 34%were poverty-stricken areas.The multidimensional relative poverty degree in the rocky desertification area of Yunnan,Guangxi and Guizhou showed a significant spatial positive correlation.The spatial distribution of relative poverty at the grid scale can effectively showed the spatio-temporal evolution of poverty in the study area.