Research On Remote Sensing Inversion Method Of PM2.5 In HuBei Province Based On Geographically Weighted Regression Model

In recent years,with the rapid development of China’s economic construction,the excessive consumption of energy and the discharge of various high-intensity atmospheric pollutants far exceed the carrying capacity of resources and environment.The air pollution caused by human activities is increasing day by day,seriously destroying the balance of the ecological environment,and the problem of air pollution is becoming more and more serious.PM_2.5 is one of the important components of atmospheric pollutants.It not only has a serious negative impact on climate change and human health,but also brings huge economic losses.Therefore,it is very important to master the distribution and changing rules ofPM_2.5 comprehensively and effectively for urban environmental governance in China.However,due to the late establishment of thePM_2.5 ground monitoring network in China and the uneven distribution of sites in space,there are certain limitations in the large-scalePM_2.5 concentration monitoring.Compared with traditional ground monitoring,retrieving aerosol optical thickness by satellite remote sensing technology and estimating near-groundPM_2.5concentration is a new technology developed in recent years.It has strong macroscopical characteristics and can quickly obtain the spatial and temporal dynamic changes of surface pollutants.It can not only make up for the shortcomings of ground observation,but also provide a larger spatial scale.The temporal and spatial distribution ofPM_2.5 concentration can be used for long-term continuous and stable monitoring of atmospheric pollution,providing more abundant information ofPM_2.5 on time scale,providing an important data basis for the prevention and treatment ofPM_2.5 pollution and the assessment of its impact on human health.Based on the above background,this study takes Hubei Province as the research area,using high temporal resolution Himawari-8 AOD data,combined with meteorological data and land use data,establishes a geographically weighted regression-based estimation model ofPM_2.5 concentration,and analyses the hourly temporal variation trend and spatial distribution characteristics ofPM_2.5 concentration near the surface in Hubei Province.On this basis,the influence and reasons of meteorological factors on the change ofPM_2.5 concentration are studied.The main contents and conclusions are as follows:（1）Through the correlation analysis of meteorological factors andPM_2.5concentration in Hubei Province,it is found that there is a significant positive correlation betweenPM_2.5 concentration and atmospheric pressure and relative humidity,with correlation coefficients of 0.502 and 0.418,respectively;there is a weak negative correlation betweenPM_2.5 concentration and atmospheric temperature,with correlation coefficient of-0.231;PM_2.5 concentration and wind speed showing a significant negative correlation.The correlation coefficient is-0.515,which shows that atmospheric pressure,relative humidity and wind speed are the main meteorological factors affecting the change ofPM_2.5 concentration.（2）Regional validation of Himawari-8 AOD products was carried out using the Global Aerosol Automated Observation Network（AERONET）.By comparing the AOD data of Hefei and Shouxian two AERONET ground stations with Himawari-8AOD data,the observation results show that the overall validation results² of the two sites are greater than 0.75,showing a high correlation,which proves that Himawari-8 AOD data is also applicable in China and can be used in the study of statistical modeling ofPM_2.5-.（3）In view of the obvious advantages of geographically weighted regression model in the construction ofPM_2.5-model in a large area,the model ofPM_2.5-in Hubei Province was established by GWR,and compared with multivariate linear regression model,the results show that the modeling effect of geographically weighted regression model is better than that of multivariate linear regression model.Generally speaking,the method of inversion ofPM_2.5 concentration by satellite remote sensing can provide more spatial and temporal distribution ofPM_2.5concentration and more abundant information ofPM_2.5 concentration on time scale.It can effectively compensate for the shortcomings of ground monitoring stations and provide important data support for the prevention and control ofPM_2.5 pollution and the assessment of human health.