Estimation Of Near-Surface PM_2.5Concentration Of Beijing-Tianjin-Hebei With Multiple Complex Factors

With China's economic growth and continuous development of social construction,air pollution and frequent smog have become a popular phenomenon,which lead to a great impact on the public health and production.According to statistics,more than 70%of urban air quality in China exceeds the standard.Among them,regional air pollution characterized by Particulate Matter 2.5(PM_2.5)has become the most urgent and prominent environmental problem in the Beijing-Tianjin-Hebei urban agglomeration.The monitoring of air pollution monitoring is the first step in completing air pollution research and control.China has established a nationwide air pollutant ground monitoring network.However,the monitoring sites are concentrated in urban areas,and the suburban and rural areas without monitoring stations have become"blind spots"for air quality assessment,pollution prevention and control,forecast warning,and pollution exposure research.The influencing factors of air pollution are complex and the regional differences are obvious.It is difficult to accurately reveal the temporal and spatial distribution and change mechanism of fine particles using only limited observation station observation data.Considering the shortcomings of the traditional ground PM_2.5monitoring network,aerosol optical depth?AOD?products based on satellite remote sensing are provided as an effective way to reveal the distribution of pollutants in the regional atmosphere,transmission paths,distribution of pollution sources,and diffusion dynamics.On the one hand,it can make up for serious uneven coverage of the monitoring site,on the other hand,it can correct the deviations generated by the calculation of the average status of air pollution in areas with and without monitoring sites,and improve our recognition level of air quality in the entire region.The Beijing-Tianjin-Hebei urban agglomeration selected in this article is the biggest air-polluted area in China,and is the focus area of atmospheric quality improving.In order to accurately estimate the PM_2.5.5 concentration of Beijing-Tianjin-Hebei urban agglomeration and analyze the spatial difference and long-term change of PM_2.5concentration,this paper studied the spatiotemporal changes of PM_2.5 concentration data at ground observation stations from 2014 to 2019.Based on the advantages of satellite observations and ground observations,combined with a complex impact factor structure,the near-surface PM_2.5 concentration is accurately estimated.The Beijing-Tianjin-Hebei area estimation model is proposed,and the fine temporal and spatial distribution mechanism of PM_2.5 is further studied to provide a scientific basis and accurate data sources for air quality forecasting.The main research contents of this article are as follows:?1?Based on the site-scaled data,the temporal and spatial characteristics of PM_2.5concentration were analyzed.From the perspective of time characteristics,the PM _2.5concentration of Beijing-Tianjin-Hebei urban agglomeration is divided based on the natural time dimension,and the change principle of PM_2.5 concentration under different time dimensions is analyzed.At the same time,we use wavelet analysis method to mine the change rule of PM_2.5 concentration in long time series.The annual average PM_2.5concentration trend and the overall monthly average PM_2.5 concentration trend of 13cities in Beijing-Tianjin-Hebei urban agglomeration from 2015 to 2019,the daily average PM_2.5 concentration trend,the mid-term PM_2.5 concentration trend and the PM_2.5concentration level in the first year and the end of the year,the significant seasonal change characteristics of PM_2.5 concentration and the horizontal comparison results in each quarter were obtained.From the perspective of spatial characteristics,this paper analyzes the spatial difference of PM_2.5 concentration of monitoring stations in Beijing-Tianjin-Hebei urban agglomeration,and obtains that PM_2.5 concentration shows the most obvious spatial correlation and spatial difference in winter.?2?Considering that the changes in the concentration of PM_2.5 are affected by a variety of natural factors and human factors,in order to distinguish the impact mechanism of the potential factors on the spatial difference and spatial changes of PM_2.5concentration,through the Greedy algorithm,we reasonably select the characteristic variables according to the performance of each potential factor in the modeling process,and explore the effectiveness of each characteristic variable in studying the spatiotemporal changes of PM_2.5 concentration at the precise spatiotemporal scale.The contribution of multivariate factors in the model optimization process is analyzed,and the optimal factor combination of different seasonal characteristic variables in the modeling process is obtained.?3?PM_2.5 concentration has temporal and spatial characteristics.In order to make full use of the spatio-temporal correlation between data,this paper performs a seasonal fit on the PM_2.5 concentration in Beijing-Tianjin-Hebei region based on the Geographically and Temporally Weighted Regression?GTWR?method.We select the optimal feature variable group by greedy algorithm,and use the time and space features to fuse the air monitoring data with the feature variables to ptimize the GTWR model.We added the Geographically Weighted Regression Model?GWR?and Kriging Interpolation Method in this part,and respectively constructed a prediction model based on the PM_2.5 concentration under different seasonal conditions during the study period as a comparison model of the GTWR model.The accuracy of the results of the three models was tested using the cross-check method.The results show that the GTWR method can obtain a better fitting result when it is used to construct a model for estimating the PM _2.5concentration.Also,result show that the model can objectively estimate the PM_2.5.5 mass concentration in different seasons in Beijing-Tianjin-Hebei region from 2015 to 2019.After the deviation correction based on the GTWR model,the seasonal-spatial distribution characteristics of PM_2.5 concentration in Beijing-Tianjin-Hebei region can be obtained,and the importance variable of key variables can be obtained.The method proposed in this paper can make up for the space lack of PM _2.5,provide high-precision PM_2.5 exposure estimation and accurate data support for urban-scale health effects and environmental epidemiological studies.