| The situation of PM2.5 pollution in China is still very serious.In 2017,the average annual concentration of PM2.5 in China was 43 μg/m3,which was still far from the second class standard 35μg/m3 in China’s Ambient air quality standard(GB3095-2012).Serious PM2.5 pollution will aggravate the morbidity and mortality of cardiovascular and cerebrovascular diseases.It is particularly critical for environmental epidemiological studies to conduct the relevant population fine exposure study and obtain the PM2.5 concentration exposure response factor.However,as the air quality monitoring network in China started relatively late,the distribution of monitoring stations was uneven and scarce.As far as the national scope is concerned,there are only three monitoring stations on an area of 20 000 square kilometers.In theory,the monitoring concentration of a monitoring station can only represent the pollution situation in the surrounding local area,but not the large-scale pollution exposure level.The LUR model is an effective method to simulate the spatial distribution of air pollutants developed from European research in the 1990s.It can be used to simulate the concentration map of air pollutants with high resolution.Compared with other methods to simulate the spatial distribution of air pollutants,the LUR model is not only relatively simple in data acquisition and modeling process,but also has high accuracy in predicting the concentration of pollutants in unknown areas.It has become the most widely used method to simulate the spatial distribution of air pollutants in recent 20 years.However,most of the current LUR models are modeled by multiple linear regression models(MLR).In essence,the simulation of the spatial distribution of air pollutants is a spatial model,and the influence of geographical location on the regression process cannot and should not be ignored.Therefore,this study optimized the LUR model,used the geographically weighted regression model(GWR)to carry on the spatial modeling,based on the local weighted regression instead of the global regression technology,to improve the accuracy of the regression model.It would provide a basis for the epidemiological study of fine population exposure and air pollution.In this study,the average annual PM2.5 concentration of 95 environmental air quality monitoring stations in Jiangsu Province in 2015 was taken as dependent variable,and the land use type,population density,meteorological conditions,road traffic and key pollution sources around the monitoring site that affect PM2.5 concentration and its spatial distribution were used as independent variables.Through correlation analysis,buffer analysis,regression analysis and other methods,the LUR model adapted to the spatial distribution of PM2.5 concentration in Jiangsu Province was established.The GWR model was used to model the model in order to improve the goodness of fit and the accuracy of prediction,and to obtain a high resolution PM2.5 pollution map.The results show that the GWR model is superior to the MLR model in both the goodness-of-fit and the residual spatial autocorrelation.The R2 of GWR-LUR model is 0.50,the R2 of MLR-LUR model is 0.33,the former is about 50%higher than that of the latter.The residual Moran’s I of the GWR-LUR model is 0.011,and the Moran’s I of the MLR-LUR model is 0.271.The residuals of the former show spatial random distribution,while the latter show a certain positive spatial correlation.The GWR model also reduces the spatial autocorrelation of the residuals.The PM2.5 pollution map of Jiangsu Province with 1km resolution simulated by GWR-LUR model can effectively identify pollution hot spots and low pollution values.The pollution situation of the densely populated urban areas and the surrounding areas with more pollution sources is obviously more serious.It is indicated that the key pollution sources are still the main sources of PM2.5 pollution in Jiangsu Province,and the areas with intensive crowd activities will not only produce PM2.5 pollution,but also will not be easy to spread pollutants,and form pollution hot spots.Areas with more water areas,such as Luoma Lake,Hongze Lake,Gaoyou Lake and Taihu Lake,have a much lower concentration value.It is also found that the concentration value of nearly 3/4 areas in the PM2.5 pollution map of Jiangsu Province in 2015 is lower than that of the average PM2.5 concentration of 57.9μg/m3 in Jiangsu Province in 2015,which is obtained only by the average monitoring concentration of the monitoring stations.It is shown that the actual pollution situation in Jiangsu Province will be exaggerated to a certain extent by simply averaging the concentration values of monitoring stations to indicate the pollution situation in Jiangsu Province.This is worthy of further consideration in the assessment of ambient air quality.Finally,in order to put forward regional control measures for PM2.5 pollution in Jiangsu Province,the correlation analysis of influencing factors of PM2.5 pollution was carried out in South Jiangsu,Central Jiangsu and North Jiangsu respectively.The results show that the most significant influencing factor in South Jiangsu is the key pollution source.Therefore,in the prevention and control of PM2.5 pollution,the more important measures and energy should be put on those factories and enterprises with more pollutant emissions.It is encouraged to actively upgrade industries,eliminate backward production capacity,and let factories and enterprises to enter the park.The most important factor that affects the PM2.5 pollution situation in North Jiangsu is the local meteorological conditions,which indicates that it is particularly important to carry out early warning and emergency response of heavy pollution weather in North Jiangsu.The environmental protection department should strengthen the joint response with the meteorological department,issue the regional heavy pollution weather warning in time,and organize various cities to actively carry out the regional emergency response.Due to the lack of selection of significant influencing factors in the Central Jiangsu,the focus of work should be placed on the comprehensive control of "scattered and disorderly pollution" enterprises,speed up the relocation and renovation of pollution in urban areas and around sensitive buildings,and strive to concentrate on the industrial parks or industrial agglomeration areas in the urban fringe areas as soon as possible. |
PDF Full Text Request