Research On Spatial-temporal Evolution Process And Driving Factors Of Pm_2.5 In Yunnan Province In The Recent 20 Years

Fine particulate matter(PM_2.5)stands for aerodynamic particle diameter less than 2.5 um in tiny particles.PM_2.5 tends to attach to toxic heavy metal ions and has a long residence time in the atmosphere.It is very harmful to human health,and be-came the primary pollutants in the air.Because of such heterogeneity,this paper takes Yunnan Province as a case to investigate the PM_2.5 patterns.Data used in this study include PM_2.5 measurements of 41 monitoring stations in 16 major cities from Febru-ary 2013 to December 2018,MODIS AOD raster data of 1°×1°and four meteorolog-ical factors?boundary layer height,temperature,wind speed,wind direction,etc.?from 2000 to 2018.By combining back propagation artificial neural network?BPANN?and?-support vector regression??-SVR?,a novel hybrid estimation model was con-structed to impute the historical PM_2.5.5 missing values in the long time series from 2000to 2012,and to estimate the missing PM_2.5 concentration from 2012 to 2014 in some cities.Based on the model,we analyzed the driving factors of PM_2.5 concentration in Yunnan Province in the past 20 years.Results showed that goodness of fit?R²?,root mean square error?RMSE?and mean absolute percentage error?MAPE?,mean abso-lute error?MAE?,mean square error?MSE?of the hybrid estimation model were0.8101,5.1941 and 19.73%,respectively.The hybrid estimation model produced re-sults superior to single model,with a higher accuracy and a stable performance,providing more accurate estimation of historical concentration.Results of temporal analysis indicated that the annual mean PM_2.5 concentration in Yunnan Province was in a relatively stable trend from 2000 to 2012,with a slightly increase in 2001 and 2007,and a small change in other years.From 2013 to 2018,the PM_2.5 concentration de-clined slowly after an obvious upward trend,and reached lowest in 2018.The PM_2.5 concentration in the four regions of Yunnan Province varies according to different population,industrial structure,climatic conditions and neighboring re-gions.The annual mean PM_2.5 concentration in the northern region is the lowest,the highest in the northeast region,in the western and southern regions varies relatively gently.Cities in the southeast,such as Qujing,Kunming,Honghe and Wenshan,have relatively high PM_2.5 concentrations,while cities in the north,such as Diqing,Lijiang,Nujiang and Dali,have relatively low PM_2.5 concentrations,making them stable areas with good air quality.From 2007 to 2013,there was no significant change in the spa-tial distribution of PM_2.5 concentration,the range of high concentration was slightly expanded,and the area of medium and low concentration was basically unchanged.Compared with 2013,the spatial distribution characteristics of PM_2.5.5 annual mean concentration in 2015 showed that the range of high concentration shrank,while the range of low concentration expanded,and the air quality increased.In 2018,the spa-tial distribution characteristics of PM_2.5 concentration changed significantly,with the high concentration area shifting from east to west.