Discussion On Relationship Between PM_2.5 Concentration Distribution Based On AOD Inversion And Urban Green Space Landscape Pattern

With the improvement of China’s economic level and the rapid development of urbanization in recent years,urban PM_2.5 pollution has become increasingly serious.At the same time,the physical and mental health of urban residents and the sustainable development of the city have been seriously affected.As an important part of the urban system,urban green space has its distribution and spatial structure characteristics,on the one hand,it affects the ecological benefits,on the other hand,it also has an important impact on the urban green space landscape pattern.The concentration distribution of PM_2.5 is affected by the diffusion and deposition capacity of the space environment in which it is located.Therefore,it is of great significance to study the coupling relationship between PM_2.5 and urban green space and landscape pattern to alleviate the pollution of urban PM_2.5.Based on this,this paper takes nine monitoring stations in the main urban area of Chongqing as the research object,and inverts the four seasons PM_2.5 in 2008,2013,and 2018 in this area using methods such as Aerosol Optical Depth（AOD）and multiple regression models.Concentration data,and through multiple linear regression analysis to explore the common impact mechanism of the space monitoring site buffer radius area land use type area ratio,green space landscape pattern.The main conclusions are as follows:（1）Based on the MODIS L1B satellite remote sensing data,the dark pixel algorithm is used to invert the 1KM resolution AOD image of the main city of Chongqing,and the ground automatic solar photometer（CE-318）monitoring results are used to verify the results.It shows that the correlation R²between the two reaches0.834,indicating that its inversion accuracy is good,which proves the feasibility and reliability of inversion of AOD based on satellite remote sensing images.（2）In this paper,the inverse AOD is used as the dependent variable,and the PM_2.5concentration is the independent variable.At the same time,the meteorological data and topographic data are added to establish the AOD-oriented geographic spatial and temporal multiple linear regression model of PM_2.5 concentration,and the estimation The results are compared with a simple linear model for leave-one-out cross-validation.The results show that the model fitting degree R²is improved from 0.517 to 0.806 of the simple linear model,and the model performance is better.（3）In this paper,through fitting analysis of PM_2.5 concentration to land use types,it is found that the coupling relationship between different types of land use patterns and PM_2.5 is obviously different.The proportion of urban green space and agricultural land area is negatively correlated with PM_2.5 mass concentration;the proportion of urban construction land area is positively correlated with PM_2.5 mass concentration,that is,the larger the patch area of non-construction land,the better the PM_2.5 mass concentration.The better.In the high-scale domain（5km×5km-6km×6km）,the correlation is higher than in the mid-scale domain（3km×3km-4km×4km）than low-scale domain（1km×1km-2km×2km）.（4）The coupling relationship between PM2.5 concentration and green landscape pattern has obvious seasonal scale difference.PM2.5 concentrations and green land plaque area ratio（PLAND）,average patch fractal dimension（FRAC-MN）,average（SHAPE-MN）green patch SHAPE index,landscape apartness index（CONTAG）linear regression of four types of green space landscape pattern index in spring,autumn and winter to keep positive correlation,negative correlation trend for the summer,that PM2.5 are sensitive to seasonal changes;On the time scale,the inter-annual correlation between the three has little difference,but different seasons in the same year have some differences;In summer,there was only a significant positive correlation between PM2.5 and PLAND at the scale of 1km×1km,a significant positive correlation between PM2.5 and LPI at the scale of 1km×1km-3km×3km,and a significant negative correlation between PM2.5 and CONTAG at the scale of1km×1km-2km×2km.At the same time,the coupling relationship between the two had the best scale effect,and the change of green space landscape pattern at a specific scale could play the greatest mitigation effect.For the three seasons of spring and autumn,the positive correlation between PLAND and PM2.5 concentration decreased gradually with the increase of scale,while the negative correlation increased gradually.In spring and winter,3km×3km and 5km×5km began to change from positive correlation to negative correlation,and the negative correlation was the strongest in the three seasons at 6km×6km.For the green landscape pattern index,the correlation between FC-MN and PM2.5 concentration in the spring of 2km×2km turns from positive to negative,and has the strongest negative correlation at 6km×6km,and the strongest negative correlation with PM2.5 concentration in the autumn and winter at 6km×6km and4km×4km.Shape-mn had the strongest negative correlation with PM2.5 concentration in 1km×1km in spring and winter and 2km×2km in autumn.CONTAG is 6km×6km in spring and autumn and 5km×5km in winter.