| PM2.5 is directly related to the environment and climate change and is a major factor in the rise of atmospheric temperatures.It is a general term for fine particles with a particle size of less than 2.5μm.However,due to its extremely small particle size,it is usually suspended in the air and very easy to enter the respiratory tract through the nasal mucosa,leading to a series of respiratory tract diseases and causing certain harm to human health.The presence of so many fine particles in the atmosphere is directly linked to vehicle emissions and the burning of fossil fuels in cities.With the development of society,more and more people are paying more and more attention to PM2.5 in their cities or communities.Although the traditional air quality monitoring network can continuously monitor pollutant concentration,its accuracy and convenience still need to be further improved.This study attempts to develop and establish a PM2.5 regression model based on mobile monitoring technology,which can greatly improve the monitoring accuracy of fine particulate matter concentration in small scale space,which provides a new idea for the study of PM2.5 concentration prediction in different regions.In view of the representation of the spatial distribution of Guangzhou Higher Education Mega Center,this research area was determined to be carried out here.The observation of near-surface atmospheric PM2.5 concentration was carried out through the method of moving measurement,and the temporal and spatial variation characteristics of this area were also analyzed.Accordingly,by referring to the information and data of major land types,climatic conditions,traffic network distribution map and population density and other related factors,we analyzed the relationship between the spatial and temporal variation characteristics of atmospheric PM2.5 concentration near the surface of Guangzhou Higher Education Mega Center.On this basis,using the integrated learning algorithm model as the research method,we simulated the temporal and spatial distribution pattern of near-surface PM2.5 concentration in the study area based on the mobile measured data of near-surface PM2.5 concentration obtained by a portable laser aerosol spectrometer in August 2020,December 2020 and January2021.Its spatial characteristics and influential factors were also discussed.The main conclusions of this study were as follows:（1）Combined with the data of the median PM2.5 concentration obtained from the mobile measured route;it was analyzed that the concentration in winter(32.78±11.89μg·m-3)was higher than that in summer(19.51±1.83μg·m-3).Winter morning study area on the spatial distribution of the inner ring line PM2.5 concentrations below the Central Line,the road cross the confluence and PM2.5 concentrations was seen near expressway entrance area,the higher the overall performance for traffic road network density,the higher the concentrations of PM2.5,presented a growth trend,central region winter weekday morning PM2.5 median average concentration was highest(40.11±2.99μg·m-3),the outer ring the lowest average PM2.5 concentration value in area near river(21.94±2.00μg·m-3),This difference in distribution was mainly caused by the regional commercial catering source emissions and road network.Influenced by the emission of heavy diesel vehicles near the construction site,there was a certain correlation between the median average concentration of PM2.5 and traffic flow.（2）The spatial distribution of PM2.5 concentration mainly through Inverse Distance weighted（Inverse short Weighting）method and kriging(kriging interpolation method,for simulating the study area,and found that the outer ring near the artificial land of PM2.5concentrations in winter was low,the inner ring was higher,the main reason was local PM2.5emissions increased,and high concentrations of PM2.5 area mainly concentrated in the North-East population density and higher living areas and business districts into the island by the inland area,the former was because of the large dining PM2.5 emissions by sources of emissions,The latter was mainly affected by the air pollution caused by transport vehicles entering the island.Moreover,the area was the main road leading to Cheung Chau Island.In addition,the spatial simulation distribution of PM2.5 concentration in Guangzhou had obvious seasonal differences.In winter,the PM2.5 concentration in Guangzhou was significantly affected by the urban background.In summer,due to the overall atmospheric diffusion conditions,the spatial distribution of PM2.5 concentration was more uniform.（3）The multiple regression prediction model established based on the mobile measured PM2.5 data and other characteristic data had a high prediction accuracy,with a determination coefficient of 0.986,an average absolute error of 0.83,a relative error percentage of 2.81%,and a root mean square error of 1.42.It was an excellent model that can be applied to the prediction of fine particulate matter in Guangzhou.Most of the influencing variables selected into the multi-factor regression model were traffic road variables,population variables,time variables and physical geographical variables,which had a significant impact on the near-surface atmospheric PM2.5 concentration in Guangzhou.The near-surface PM2.5 movement measurement and multi-factor modeling study in urban areas were helpful to further explore the variation characteristics and influence mechanism of PM2.5 in urban areas of China,and to quickly predict the temporal and spatial variation of PM2.5 concentration in a certain area.It could provide strong theoretical support and a scientific basis for the treatment of the ecological natural environment and man-made air pollutants in a larger range. |
PDF Full Text Request