Quantitative Simulation Of Spatio-Temporal Variations Of PM_2.5 Concentration In Beijing-Tianjin-Hebei Region

In recent years,China's economy has developed with high speed,but behind the rapid development of the economy is the extensive model development pattern,which makes a lot of energy resources consumption,especially the burning of fossil fuels to make atmospheric pollutants emissions increased,leading to the serious problem of air pollution in our country.As one of China's major economically developed regions,Beijing-Tianjin-Hebei region is the most serious haze pollution area in our country,and has aroused widespread concern in the community.PM_2.5 has also become the primary air pollutant in most areas of our country,seriously affecting human health and people's production and living.Therefore,it is need to be accurate and comprehensive understanding the temporal and spatial distribution of PM_2.5concentration,with a view to provide a scientific foundation for identifying the temporal and spatial distribution of regional air pollution sources and the temporal and spatial variation of atmospheric diffusion conditions,it also provides a scientific basis for the regional air pollution control.The research shows that satellite remote sensing has the unparalleled advantage in air pollution monitoring and plays an important role in the study of regional atmospheric particulates concentration.The aerosol optical thickness retrieved by satellite remote sensing can characterize the number of aerosol particles in the atmosphere,which can reflect the pollution status of the atmosphere.So,through the establishment of a model for the relationship between aerosol optical depth and the concentration of PM_2.5,then the space-time distribution of PM_2.5 near the ground in a large scale can be obtained,accordingly make up for the lack of the ground monitoring,which makes it possible to recognize the characteristics of regional atmospheric particulate concentration.This study based on the hourly PM_2.5 concentration data and the MODIS AOD data of the national air quality monitoring points in Beijing-Tianjin-Hebei region from 2013 to 2014,uesd the mixed efect model to establish the relationship model of AOD-PM_2.5,the ten-fold cross-validation method,root mean square error?RMSE?and relative prediction error?RPE?were used to evaluate the accuracy of the model fitting.Subsequently,used the correction factor to revise the average of PM_2.5 concentration in the study period,and then to analyze the space-time distribution characteristics of PM_2.5 concentration in the study area.The main conclusions are as follows:?1?The temporal and spatial calibration of AOD-PM_2.5 can improve the correlation between them by using the mixed effect model,and the model fitting R² is 0.93 and the slope is nearly 0.89.The cross-validated R² is 0.74,RMSE and RPE are 19.11?g/m³ and 30.38%,respectively.There is a certain over-fitting phenomenon in the model.From the model verification results,the fitting effect of the mixed effect model is better,which shows that the fitting effect of the mixed effect model is high after the time and space calibration,which can be well applied to predict the PM_2.5 concentration near the ground within a certain range.?2?Due to the lack of modeling data in time and the limitations of model prediction,while the model deviates in the estimation of long-term PM_2.5 concentration.In order to reduce this estimation deviation,we used the correction factor to correct the estimation result.The average correction factor of PM_2.5 concentration in study area from 2013 to 2014 is 1.54,which means that the non-random loss of AOD value caused the estimation of model to be low.After correction,the decision coefficient between the average of PM_2.5 concentration estimated by the model and the measured value at each monitoring point is 0.89 and the slope is nearly 0.90.The correlation is higher,which indicates that the correction factor can be used to improve the estimation results to achieve the expected results.It shows that using the correction factor to correct the model estimates when estimating the long-term PM_2.5concentration can reduce the prediction bias caused by the non-random loss of AOD.?3?From the estimated spatial distribution of PM_2.5 concentration in the study area from2013 to 2014,the average of PM_2.5.5 concentration is 81.47?g/m³,the concentration of PM_2.5 in Baoding,Shijiazhuang,Xingtai,Handan and other place are higher than 105?g/m³,which is more than three times of the national secondary air quality standard,indicating that the pollution in these areas is more serious.The PM_2.5 concentration in northern areas such as Zhangjiakou,Chengde and Qinhuangdao are low relatively,while the PM_2.5 concentration along Beijing-Guangzhou railway is higher.The spatial distribution of PM_2.5 concentration is higher in the southern,lower in the northern.From the perspective of interannual changes,the average annual concentration of PM_2.5 in 2013 is 86.16?g/m³,while in 2014 is 76.03?g/m³.Compared with 2013,the concentration of PM_2.5 in Beijing-Tianjin-Hebei region showed a declining trend,indicating that the air pollution in Beijing-Tianjin-Hebei region in 2014 was better than last year's.It also shows that the air pollution in the Beijing-Tianjin-Hebei region has been effectively controlled.In addition,the concentration of PM_2.5 also showed obvious seasonal variation.During the study period,the concentration of PM_2.5 in the winter half year is significantly higher than that in the summer half year.This showed that the pollution in the study area is more serious in winter,especially in Baoding,Shijiazhuang,Xingtai,Handan.