Estimating Ground-level PM_2.5 In The 5 Provinces Along The Huaihe River Based On Remote Sensing

PM_2.5 is one of the main atmospheric pollutant in China,especially in industrial developed and densely-populated urban areas,has caused serious harm to the environment and human health.Large-scale and dynamic ground-level PM_2.5monitoring provides important basic data for current atmospheric environment management,and it has become a hot issue in atmospheric environment research.The ground-based monitoring site can directly obtain PM_2.5.5 mass concentration,which is an important data source for PM_2.5 monitoring in China.Although there are more than 1600 ground-based monitoring sites,the distribution of them is severely uneven,and it's difficult to reflect the PM_2.5 concentration in the whole region.Satellite remote sensing has the advantages of real-time observation,continuous,and large coverage,which can greatly make up the spatial distribution and spatial coverage of PM_2.5 ground-based monitoring site.Using aerosol optical depth obtained from remote sensing to construct statistical model is an effective way to indirectly obtain spatial continuous PM_2.5.However,the conversion relationship between AOD and PM_2.5 tends to be uncertain in terms of time,space and modeling parameters.There is no unified and universal transformation model.This study mainly focuses on the two aspects of“modeling parameters”and“space”.Taking the five provinces along the Huaihe River?Henan Province,Hubei Province,Anhui Province,Jiangsu Province,Shandong Province?as a research area,which are densely populated,industrial concentration,and have severe respiratory tract chronic disease.This study aims to explore an effective method for estimating PM_2.5 using satellite remote sensing technology.Based on the research results of previous studies,this research mainly conducts research from the following aspects and obtains relevant results:?1?In terms of the meteorological parameters of the model,this study examines more meteorological factors on the basis of conventional ground-based meteorological observations.This is to incorporate the meteorological factors that characterize the enrichment of particulate matter in the near-surface layer into the estimation model of PM_2.5.The results showed‘?T'which means temperature difference between 850hpa and surface,can better reflect the diffusion conditions of local pollution discharge,and have significant relationship with PM_2.5.Incorporating?T to PM_2.5 estimation model can improve the its fitting results.?2?As for parameter AOD in the model,this study explores the controversial problem whether the vertical correction or relative humidity correction of AOD can improve PM_2.5 estimation model.The results show that AOD to be corrected or not and the correction methods vary from place to place,which also have significant impacts on the result of PM_2.5 estimation.But how to choose a better AOD correction form still needs to be based on statistical results.?3?In terms of space,considering that AOD inversion is greatly affected by environmental conditions such as cloud,therefore coverage of AOD is low.This study combines Terra MODIS AOD?MOD04?and Aqua MODIS AOD?MYD04?products through data fusion to improve the spatial coverage of AOD.The results show that spatial coverage of fused AOD is 30.2%higher than MOD04 and 57.8%higher than MYD04,and the fused AOD are better consistent with AOD from ground-based AERONET site observation,and R has increased from 0.832 and 0.859 to 0.880,respectively.?4?Considering that the five provinces along the Huaihe River are vast,the emission of pollutants vary greatly from place to place,so it is necessary to divide the area according to the characteristics of PM_2.5 pollution.However,the commonly used administrative district-based zoning method is too subjective and the scientific basis is not strong.Therefore,based on the time series of the ground-based monitoring PM_2.5in the study area,this study uses the K-means method to cluster the sites and then make it an objective partition basis.Results have shown that it is reasonable to divide the whole study area into three sub-areas,and the results of the zoning showed that the zoning boundary has broken the traditional administrative boundary.After partitioning,the effect of the model is effectively improved.The method is simple and easy to implement,and has certain scalability.