Spatial-temporal Heterogeneity And Source Apportionment Of Atmospheric Particulate Matter In Heilongjiang Province

Haze pollution is a typical representative of the pollutions that affect human living environment.Scientists have devoted to explore the causes and laws of haze pollution.Since high concentration of atmospheric particulate matter(PM10,PM2.5)is one of reasons of haze pollution,it has become a hot topic to apportion source and analyze influenced factors,spatial-temporal heterogeneity and diffusion paths in atmospheric and environmental science research.In recent years,many cities across China have repeatedly experienced severe haze pollution in different seasons.The air quality in many cities of Heilongjiang Province were often marked with "red lights".The monthly average concentrations of PM10 and PM2.5 in the provincial capital,that is Harbin,have often exceeded national secondary standards in 2013 and 2014.Although the environmental air quality of Heilongjiang Province has generally improved and the number of days of heavy pollution in major cities has decreased dramatically in recent years,there are still days occasionally showed the air quality exceeding national secondary standards.In order to explore the spatio-temporal distribution of atmospheric particulates,based on the data of six standard pollutants(PM2.5,PM10,SO2,NO2,CO,and O3)from automatic monitoring stations of 13 cities,this study established global and local models,including ordinary least squares regression(OLS),linear mixed models(LMM,geographically weighted regression(GWR),temporally weighted regression(TWR),and geographically and temporally weighted regression(GTWR)to quantify the spatial-temporal relationship among six standard air pollutants in Heilongjiang Province from January 2015 to December 2018.Meanwhile,in order to understand the contributions of pollution sources to atmospheric particulate matters(PM10,PM2.5)and seasonal changes in the composition of atmospheric particulate matter,this study manually monitor and analyze the data of receiver composition and pollution source spectrum.The anlaysis of seasonal changes of the chemical composition and source apportionment of atmospheric particulate matter in 2014 were carried out for the provincial capital-Harbin,which experienced heavy pollution.Due to the long winter time in Harbin(late October to early April),four seasons(spring,summer,autumn and winter)are subdivided into five sampling periods including spring,summer,autumn,early winter,and late winter.Receiver samples were collected at four sampling points in the main urban area of Harbin in each periods.Source samples including soil dust,flying dust,building cement dust,coal-burning dust,motor vehicle exhaust dust,biomass burning dust,and catering oil and dust were irregularly collected.The chemical components analysis of receiver and source samples were subdivided into carbon,water-soluble ions,and inorganic elements analysis.According to the obtained receiver composition and pollution source spectrum data after characteristic analysis of the components,the chemical mass balance model(CMB)was applied to apportion the source of atmospheric particulate matter in different seasons and different points,and rates of pollution source contribuation were finally obtained and reasons of source apportionment were analyzed.Results showed that:(1)The LMM and all GWR-based models(i.e.,GWR,TWR,and GTWR)showed great advaltages over ILS in terms of model fitting,such as producing higher R2 and more desirable model residuals than OLS,especially TWR and GTWR that incorporated temporal variation.The GWR,LMM,and TWR and GTWR improved the explanation of variance in PM2.5 by 3%,5%,12%and 12%,respectively,from 85%of OLS.TWR yielded slightly better model performance,prediction accuracy and uncertainty accuracy than GTWR(smaller AICc,RMSE,MAE and standard deviation of Z score of model residuals),and also reduced RMSE and MAE of model residuals by 67%comparing to OLS model;while GWR only reduced RMSE and MAE by 14%?15%comparing to OLS.The traditional OLS and GWR were inadequate for describing the nonstationary of PM2.5.The LMM slightly performed better than GWR since it considered different locations as random effect and meanwhile handled the repeated measurements using R matrix,which provided an alternative solution besides GWR-based models.The temporal heterogeneity is more obvious than spatial heterogeneity in this case.Thus,the incorporation of temporal information is inevitable for modeling the relationship between PM2.5 and the other air pollutants.This work provides evidence of spatial-temporal heterogeneity in the relationship between PM2.5 and the other standard air pollutants,and also provides possible solutions for modeling PM2.5 with the other air pollutants for Heilongjiang province.(2)The seasonal distribution characteristics and main morphology of atmospheric particulate matter components were obtained in this study.The concentrations of most components were highest in the early winter and lowest in summer,with non-significant seasonal characteristics in spring,autumn,and late winter.The seasonal characteristics of the components are related to the seasonal changes of weather conditions and some pollution sources(such as coal burning sources)during sampling periods.During the sampling period,PM2.5 is the main form of atmospheric particulate matter,and the pollution of fine particulate matter is relatively large.(3)Based on the analysis of pollution source samples,the component spectra of the main emission sources of PM10 and PM2.5 were established in this study.Based on sensitivity matrice and the relevant literature,the characteristic components of each pollution source were determined.For example,the characteristic components of coal dust,flying dust,biomass combustion,building dust,secondary organic carbon were Al,Si,K,Ca and OC,respectively.(4)Variations existed among source contribution rates at different points in the same season.However,the trends of the variety of source contribution rates were basically the same,which showed regional characteristics.Except the uncertain "other sources"(which could include steel dust sources,external transmission sources,etc.),coal-fired sources was the primary source of PM10 and PM2.5 in each season.The contribution rate of coal-fired source in winner was about twice than that in spring and summer.Motor vehicle sources were secondary sources of PM10 and PM2.5 in each season,and its contribution rate in winter was higher than that in the other three seasons.Biomass combustion was an important source of PM10 and PM2.5 in autumn and early winter with obvious seasonal characteristics.For PM10,the contribution rate of coal-fired sources in heating period was 2.14 times than that in non-heating periods,and the contribution rate of dust source in non-heating period was 2.21 times than that in heating period.For PM2.5,the contribution rate of coal-fired sources during heating periods was 2.12 times than that in non-heating periods,and the contribution rate of dust sources in non-heating periods is 1.78 times than that in heating periods.For PM10,the contribution rate of biomass combustion source during the combustion period is 25.8 times than that in non-combustion period;for PM2.5,the contribution rate of biomass combustion source during the combustion period is 24.8 times than that in non-combustion period.After reallocating secondary pollutants(sulfate,nitrate,SOC).Except other sources,the most contribution to PM10 and PM2.5 was coal dust,followed by motor vehicle sources,open sources(dust,soil dust,and building dust)and biomass combustion sources.The research of source apportionment quantifies the contribution rate of major sources to atmospheric particulate matter in different seasons.This study conducted a reseach on spatio-temporal heterogeneity,concentration,chemical components charateristics and source apportionment of atmospheric particulate matter in order to deeply understand the seasonal characteristics,distributions and soursces of atmospheric particulate matter.This study is of great significance for explaining the causes,migration and transformation of atmospheric particulate matter and proposing final countermeasures for the prevention and control of atmospheric pollution.