Spatio-temporal Statistical Modeling And Analysis For PM_2.5 About Bohai Rim Region

With the development of economic globalization and the increasing pace of urbanization,environmental air quality as one of the important indexes of ecological security system,is brought into focus more and more.PM_2.5 as inhalable particulate now has become a common research focus of spatial study,the temporal variations of PM_2.5 mass concentration analysis has important practical significance.This research uses the 187 monitoring datas of air quality monitoring stations in Bohai rim region on January 1,2014 to December 31,2014,combines with spatio-temporal statistic method,and analyses spatio-temporal variation characteristics of PM_2.5 mass concentration of the Bohai rim region.Respectively using the Spatio-Temporal Ordinary Kriging and Spatio-Temporal Indicator Kriging to forecast in the study,bases on these two methods are chosen in this paper.Applying six different theory of spatio-temporal variation function model,it analysis of different models affect the prediction precision of spatio-temporal,matching precision of model validation,fitting the best spatio-temporal theory model.Also,the study based on the spatio-temporal forecast data in spatio-temporal analysis,obtaines spatio-temporal cube and different periods of spatio-temporal prediction,and analyses uncertainty of comprehensive determination method on the spatio-temporal distribution characteristics and uncertainty evaluation results of the analysis on the spatio-temporal.Finally,the spatio-temporal variation characteristics of PM_2.5 Bohai rim region in 2014 are analyzed in detail.The main conclusions are as follows:（1）Use of 2014 daily average concentration data of PM_2.5 Bohai rim region,on account of the spatio-temporal ordinary kriging and spatio-temporal indicator kriging,spatio-temporal variation function model is set up.The fitting precision of disjunctive model is better than of the separation of model.Based on spatio-temporal ordinary kriging and spatio-temporal indicator kriging model fitting,looking at the predict outcome of six models prediction of spatio-temporal,spatio-temporal ordinary kriging accuracy verification RMSE value in the order: NHA（8.898）> CM（12.247）> DM（15.443）> NHB（15.949）> MM（17.573）> GM（30.950）,MAE value in the order: CM（27.733）> DM（27.853）> MM（27.973）> NHA（29.401）> GM（32.225）> NHB（72.441）;spatio-temporal indicator kriging accuracy verification RMSE value in the order: NHA（0.234）> NHB（0.287）> CM（0.306）> MM（0.314）> DM（0.347）> GM（0.439）;MAE value in the order: NHB（0.489）> NHA（0.493）> CM（0.496）> MM（0.498）> DM（0.500）> GM（0.512）。（2）Using spatio-temporal forecast cube data,based on the spatio-temporal ordinary kriging and spatio-temporal indicator kriging and comprehensive definition method,analysis the uncertainty,the results found that the over standard rate of annual average based on comprehensive definition is 57.3%,that in the overall pollution is serious.Especially,in winter,which amounts to 68.38%;among them,which is as high as 83.54% in February,pollution is very serious.Which lower is summer,and May which may clean in good condition.The range based on comprehensive definition is about 57.4% in polluted areas.The misjudgment rate in May based on comprehensive definition is 7.87%,which is the biggest;the lowest misjudgment rates is only 1.69% in June.Misjudgment rate of annual average is obviously higher than of each season,because time dimension will affect the forecast uncertainty,the greater the uncertainty caused by the greater the time dimension.（3）Analysising of daily average concentration temporal changes of PM_2.5 Bohai rim region in 2014,the results appear to be,time on roughly bimodal trend of fluctuations,the change features of high in autumn and winter,low in spring and summer,and has a small peak in July in the summer,and temporal changes of the ups and downs fluctuation of Beijing and Tianjin and Hebei province are more similar.Uneven distribution of space,the change trend of the west high,low in the east.In general,spring and summer of the daily average concentration changes span is less than the autumn and winter season.Moreover,compared with PM_2.5 pollution in Liaoning province is the lightest of the region.Among them,the Beijing daily average concentration of PM_2.5 in peak time on February 15,the concentrations of 437μg·m^-3,is the national ambient air quality standard secondary concentration limit of 5.8 times;Tianjin daily average concentration of PM_2.5 in peak time on November 21,its concentration is 514μg·m^-3,6.8 times of secondary concentration limit;Daily average concentration peak time of PM_2.5 in Hebei province is on December 28,the concentration of 704μg·m^-3,9.4 times of secondary concentration limit;PM_2.5 daily average concentration in peak period of Shandong province is on January 7,the concentrations of 666μg·m^-3,8.9 times of secondary concentration limit;Daily average concentration in peak period of PM_2.5 in Liaoning province is on October 31,the concentration of 927μg·m^-3,12.4 times of secondary concentration limit;Daily average concentration peak beyond secondary concentration limit 75μg·m^-3 is about 5.8 12.4 times.（4）This standard on study is national standard concentration limit of 75μg·m^-3,gets standardized rate of the daily average concentration of PM_2.5 in Bohai rim region,the results show that standardized rate of the annual daily average concentration in Liaoning province is the highest,are higher than the other four areas.On the whole,standardized rate of the whole summer and May and September is higher;One of the highest,is the Liaoning province on September,standardized rate is as high as 96.5%;Not only that,on May,June,July,August of Liaoning province is higher than 80%,respectively 90.65%,88.33%,82.74% and 88.33%.Other four other districts are not have month higher than 80%,on the contrary,the standardized rate is only 25.59% in January,in Shandong province;Hebei province on January of the standardized rate is only 28.65%;Tianjin in January and November,the standardized rate is 27.1%,27.1%;The months of standardized rate is between 32.14% 78.89% in Beijing.