基于遥感反演的武汉市空气污染物与植被覆盖度相关性研究

Study on Correlation Between Air Pollutants and Vegetation Coverage in Wuhan Based on Remote Sensing Inversion

空气污染问题是众多国家在经历城镇化的快速发展过程中,难以回避的问题。以快速城镇化的武汉为研究对象,通过遥感反演大气气溶胶光学厚度、解译土地覆盖类型,进而基于像元二分模型反演归一化植被指数得到植被覆盖度后,分别在500m×500m、1km×1km、2km×2km、3km×3km的网格尺度上,探讨大气气溶胶光学厚度与植被覆盖度的相关关系。结果表明:4种尺度下,都在0.01水平上呈现极明显负相关,且pearson系数较接近。进而得到4种网格尺度的回归模型作为植被覆盖度与大气气溶胶光学厚度的预测模型,揭示了两者的初步变化规律。进一步在1km网格尺度下分析发现,植被覆盖度≤10%和> 45%时,两者在0.01水平极显著负相关。植被覆盖度≤10%的区域与水体及其周边地区和城乡建设用地具有高度的空间分布一致性,污染整体较严重;植被覆盖度> 45%的区域主要分布在森林、疏林草地、农业用地等植被覆盖较好的生态用地区域,污染程度以中度、轻度和优良为主,说明增加植被覆盖对缓解空气污染具有重要意义。通过遥感反演获取综合反映空气污染物的大气气溶胶光学厚度,并且为更精确的反映植被覆盖度,在归一化植被指数的基础上进一步运用改进后的像元二分模型进行植被覆盖度的遥感反演。从而探究大气气溶胶光学厚度与植被覆盖度之间的相关性规律以及空间分布规律,从缓解综合空气污染物的角度为城乡规划设计提供定量化的参考。

Air pollution is an unavoidable problem in the rapid development of urbanization in many countries.The paper takes Wuhan,a city of rapid urbanization,as the object of study.Through remote sensing,it has inversion of the atmospheric aerosol optical depth,interpretation of land cover types and,based on the pixel binary model,inversion of NDVI to obtain the vegetation coverage.Then,it explores the correlation between AOD and vegetation cover on the grid scales of 500m×500m,1km×1km,2km×2km,and 3km×3km.The results show that under the four scales,they all have extremely negative correlation at the 0.01 level,with close pearson coefficient.Then,it applies four kinds of grid-scale regression models as vegetation coverage and AOD prediction models,to reveal initial changes of the two.Further analysis at a grid scale of 1 km shows that when the vegetation coverage rate is≤10%and >45%,there is a significant negative correlation between the two levels at the 0.01 level.The area with vegetation coverage rate≤10%is highly spatially consistent with the water body and its surrounding areas and urban and rural construction land.The overall pollution is serious.The area with vegetation coverage rate >45%is mainly distributed in forests,open forests,and agricultural land.In ecological land areas where vegetation coverage is good,the pollution is mainly moderate,light,and fine,indicating that increasing vegetation cover is of great significance in alleviating air pollution.In this study,the atmospheric aerosol optical depth comprehensively reflecting air pollutants is obtained by remote sensing inversion,and to inverse the vegetation coverage rate more accurately,the improved pixel bipartite model is used based on NDVI.This paper explores the correlation law and spatial distribution law between AOD and VCR,and provides a quantitative reference for urban and rural planning and design from the perspective of mitigating comprehensive air pollutants.