Simulation Of PM_2.5 Concentration Spatio-temporal Changes Based On LUR Model:A Case Study Of Ganzhou City

The construction of ecological civilization is an important part of the cause of socialism with Chinese characteristics,but in the process of urbanization,people pay more attention to economic development than to environmental protection,resulting in increasingly serious environmental pollution.As an intuitive manifestation of environmental pollution,air pollution has received widespread attention.As one of the sources of atmospheric pollution,PM_2.5has a small content in the atmosphere,but it has caused extremely serious harm to human health and the atmospheric environment.Therefore,in-depth exploration of the temporal and spatial distribution of PM_2.5will provide important evidence for reducing human exposure intensity and health risks,as well as formulating effective pollution prevention measures,and has extremely important significance.Based on the PM_2.5concentration data and land use data,road data,population data,and meteorological factor data from the air monitoring station in Ganzhou City,this paper explores the spatial and temporal characteristics of land use structure and PM_2.5changes in Ganzhou City.Meanwhile,analyze the correlation between various factors and PM_2.5concentration,screen appropriate data,and establish land use regression models based on multiple linear regression,geographic weighted regression,and support vector regression methods to simulate the spatial distribution of PM_2.5.The main research conclusions are as follows:（1）From 2015 to 2020,the average annual PM_2.5concentration in Ganzhou showed a trend of first increasing and then decreasing.In 2017,the PM_2.5concentration reached a peak.Overall,from 2015 to 2020,the PM_2.5concentration decreased by 38.17%.The trend of monthly average concentration has obvious peaks and troughs every year,with peaks generally from November to January and troughs generally from June to August.The concentration is highest in winter,lowest in summer,and intermediate in spring and autumn.From the perspective of spatial distribution,the PM_2.5concentration in Ganzhou City is characterized by a high concentration in the middle and a low concentration around Zhanggong District.Through correlation analysis,the impact of various factors on the temporal and spatial distribution of PM_2.5is obtained.Factors that have a significant positive correlation with PM_2.5concentration include artificial land surface,water body,population,roads,boundary layer height,temperature,and wind speed,while factors that have a significant negative impact on PM_2.5concentration include cultivated land,forests,atmospheric pressure,and humidity.（2）The LUR model based on multiple linear regression uses forests and main roads,boundary layer height,and wind speed as modeling variables.The adjusted R²of the model is0.855,which has a good fitting effect.The RMSE value is 0.635,and also performs well in terms of accuracy.The results of the seasonal LUR model show that the explanatory variables for each season are inconsistent,indicating that the PM_2.5influencing factors are different in different seasons.（3）The LUR model based on a geographic weighted regression model uses modeling variables such as forests and major roads.The global adjustment R²of the model is 0.823,with a high degree of fit,and the RMSE is 0.663,with a slightly lower accuracy than the multiple linear regression model.In the LUR model based on support vector regression,forests,main roads,boundary layer height,temperature,atmospheric pressure,wind speed,and relative humidity are all explanatory variables of PM_2.5.The adjusted R²of the model training set reached 0.9,higher than the other two models,with the best fit,but the RMSE was 2.08,with lower accuracy than the other two models.The spatial distribution of each model is slightly different,but the common ground is that the central and southern parts of Nankang District and Zhanggong District,and a small area in the north of Xinfeng County all belong to the region with the highest PM_2.5concentration in the three models,while the PM_2.5concentration in most areas of Chongyi County,Dayu County,and Shicheng County is the lowest in the city.