Spatial Analysis Of Pm10, The Primary Pollutant Of Atmosphere In Gansu Province In The Summer Of 2016

With the emission of ultra-fine particles in various industrial processes and the sec-ondary formation of ultra-fine particles and aerosols in the atmosphere, the inhalable par-ticulate matter (PM 10) in the atmosphere has gradually become the primary air pollutant in large and medium cities. And it seriously affects the urban ecological environment and the health of residents. It is of great value and practical significance to discuss the spatial distribution of PM 10 concentration in a region and the pattern of aggregation. In this paper, Gansu Province as the study area, PM10 as the research object, we intend to use the data of six air indicators (PM10,PM2.5,CO,NO2,O3, SO2)on 33 air pollution monitoring stations in Gansu at the summer of 2016 to perform the research. Firstly, the importance of the influence of the other five air parameters on PM 10 was determined by the machine learning algorithm based on the mean impact value (MIV) of Elman neural network. PM2.5, which have the greatest influence on PM10, was screened out, and the statistical methods correlation coefficients and scatter plots were used to verify the cor-rectness of the results. Then, PM 10 of the study area was interpolated by the ordinary kriging method and the co-kriging method with the selected PM2.5 as the assistant vari-able. The accuracies of the two interpolation models were compared by cross-validation.The experimental results show that the co-Kriging interpolation method is superior to the ordinary Kriging method in interpolating precision. So the whole study area is in-terpolated by co-kriging to know the spatial distribution of PM10. Finally,the clustering analysis was used for PM 10 concentration to recognize the high concentration region, and the 75-day PM 10 spatial distribution interpolated by co-Kriging was used to find out the geographical location of PM 10 concentration maxima for every day. These spatial loca-tions of these maxima are all plotted in the study area for spatial point pattern analysis.The results show that the distribution of PM10 in summer is mainly concentrated in the area with large urban population density, and strong aggregation effects exist.