Spatio-temporal Evolution And Influencing Factors Of PM_2.5 Concentration In China

An ensemble estimation model of PM_2.5 concentration was proposed on the basis of the extreme gradient boosting,gradient boosting,random forest model and stacking model fusion technology using the measured PM_2.5 data,MERRA-2 aerosol optical depth?AOD?and PM_2.5 reanalysis data,meteorological parameters,and night light data.On this basis,spatiotemporal evolution features of PM_2.5 concentration in China during 2000?2019 were analyzed at monthly,seasonal and annual temporal scales.Finally,the random forest model,order of feature importance and partial dependency plots were applied to investigate the PM_2.5 influencing factors and their regional difference of PM_2.5 spatial pattern.The results showed that:?1?The monthly PM_2.5 concentration from 2000 to 2019 in China can be estimated reliably by the ensemble model.As the time interval between training period and estimating period of the model lengthened,the reliability of the estimated PM_2.5historical concentration decreased.Dependence of model's complexity on the estimated PM_2.5 historical concentration was nonlinear.1000 hPa Temperature,MERRA-2 PM_2.5,and boundary layer height were the key predictors of model,predictors of AOD and night light were of relatively low importance.?2?PM_2.5 annual concentration changed from rapid increase to remaining stable and then changed to significant decline from 2000 to 2019,with the turning points in2007 and 2014.The monthly variation of PM_2.5 concentration showed a U shape that first decreased then increased,with the minimum value in July and the maximum value in December.Natural geographic conditions and human activities laid the foundation for the annually spatial pattern change of PM_2.5 concentration in China,and main tone of monthly spatial pattern change of PM_2.5 concentration was determined by meteorological conditions.The national PM_2.5 concentration average center of standard deviation ellipse moved eastward from 2000 to 2014 but westward from 2014to 2018 at annual scale.At monthly scale,the average center shifted to the west from January to March,moved northward then southward from March to September,and shifted to the east from September to December.?3?PM_2.5 concentration initially increased and then remained stable with increases in AOD,POP,and GDP,and initially decreased and then stabilized with increases in PRE,WS,and NDVI.The responses of DEM and TEM to PM_2.5concentration changed from decline to ascend and then changed to decline again.AOD had the largest influence on PM_2.5 annual concentrations,whereas PRE had the least influence on PM_2.5 concentrations.The same factor had different spatial influencing magnitudes on PM_2.5 annual concentrations in seven geographical subareas.