Study On Correlation Between Particulate Concentration And Spatial Distribution Of Vegetation In Beijing

In recent years,the serious frequent fog and haze weather and particles pollution threat human health and daily life,the study of particles can not be ignored.The regulating function of forest to particulate matter has caused academic attention.The study of the correlation between forest vegetation and particulate matter and influence factors research become the hot and difficult point of scientific research,which has guiding significance for the prevention and treatment of haze,and has significant ecological,social and economic significance.This paper analysis the temporal and spatial distribution characteristics of particulate matter using Kriging interpolation method in Beijing,2016.Uus Spectral analysis and remote sensing method to inverting the particles PM2.5 and PM10 concentration,study the regulation characteristics of forest vegetation on atmospheric particulates,analysis the correlation between various factors and particles concentrations,and establishe multivariate statistical model.The main conclusions are as follows:(1)The temporal and spatial distribution characteristics of particulate matter in BeijingThe particle concentration at night is higher than the day(8:00am-20:00).The 24h concentration change curve of four types of ground monitoring stations are different in different seasons,change of environmental evaluation station and regional background station is common,whoes distribution pattern is shuangfeng type,change of traffic control station and urban clean control station is complex,whoes distribution pattern is wavy distribution with more than 3 peaks.In the four seasons,the concentration of particulate matter performance as winter>autumn>spring>summer,the highest concentration value in December,the lowest value appeared in August,which referent to heating season particulate pollution,meteorological climate.The spatial distribution performance as southern higher than northern and western regions,and this region difference is the most prominent in winter and the least in summer.(2)Tow kind inversion model of PM2,5,PM10.concentrationThis paper inversion the atmospheric aerosol optical thickness(AOD,750m)use Dark Target based on VIIRS_SDR data,build AOD-PM2.5?AOD-PM10 concentration model.The R2 of PM2.5 validation model in four seasons are 0.865?0.644?0.831?0.543.Database of particulate matter concentration in I km grid was built in Beijing.This paper proposes a new method of PM2.5Concentration inversion,obtain high spatial resolution(30m)result with very simple inversion process.(3)Effects of vegetation on PM2.5,PM10.Forest vegetation coverage and biomass is negative correlation with the particles concentration,The correlation coefficient between PM2.5and.FVC is highest in summer(0.534),and PM10.is highest(0.568).in spring.The correlation coefficient between biomass and PM2.5,PM10.is not significant in summer and autumn(?0.500.).The most notable decline of forest landtyp is sparse woodland(19.72%,20.99%)and coniferous forests(18.78%,18.22%),with a decrease of more than 18.00%.The decline order for PM2.5 is sparse>coniferous forest>young afforested>mingled forest>broad-leaved forest>shrubbery>auxiliary production forest>non-stumpage land>suitable land for forest.The decline order for PM10 is sparse>coniferous forest>broad-leaved forest>young afforested>mingled fores>shrubbery>non-stumpage land>auxiliary production forest>suitable land for forest.(4)Model of forest vegetation and particulate matter reduction rateThis paper respectively fit the vegetation coverage and PM2.5,PM10 educed rate model based on subtractive data of the forest area,The R2 of validation model are 0.649?0.811,It is not a simple linear relationship between the particles concentration and vegetation.The reduction rate of particulate matter increased with the increase of vegetation coverage in a certain range,but if the vegetation coverage is too high,the reduction rate will be decreased.This is due to the dense canopy and complex structures affect the air turbulence in the forest and the moisture is larger in the forest,and the moisture absorption of the particles makes it easier to gather.have good performance.The particles concentration change significantly at forest edge area while weekly in the forest area.(5)Correlation analysis between meteorological,topographic factor and PM2.5,PM10Relative humidity,air pressure and vapor pressure are positively correlated with PM2.5,PM10,while temperature and wind speed are negatively correlated.The hygroscopicity of particles and stable atmospheric state caused by high pressure make the pollution of particles in high pressure and high humidity environment is heavier.Increase of near surface temperature enhance atmospheric turbulence,accelerated particles diffusion.Large wind speed make the propagate and transport of particle faste,and reduce the concentration.The correlation coefficients of water vapor pressure and wind speed are large,which are 0.715(summer)and 0.579(spring)respectively.Altitude and slope are weakly negatively correlated with PM2.5 and PM10,but the correlation between slope position and slope aspect is not strong.Particle gravity and air convective characteristics in the vertical direction result in the higher the altitude,the lower the concentration.The gravity of PM10 is larger than that of PM2.5,and this trend is more obvious.The collision and motion of the air at the slope are frequent.particles spread rapidly with air flow.(6)Establishment of PM2.5,PM10 concentration multiple linear regression modelPrincipal component analysis is carried out on factors affecting PM2.5,PM10 in different seasons.Three principal components are extracted in spring,named terrain and vegetation,humidity and temperature,the normalized weights are 53.812%,26.734%and 19.455%,respectively.In summer,autumn and winter,the number of principal components are 4,4,3,respectively.Fit the concentration mode in each season with stepwise regression method.number of variables are 5(PM2.5 is 4),6,8,4,The precision of the model is better in summer(R2 of PM2.5 prediction model is 0.565)and winter(R2 of PM2.5 PM10 prediction model are0.549?0.565),while it is week in spring and autumn?The main innovation of this paper are:(1)A new method of estimating the FVC Based on red edge slopThis paper analysis the spectrum characterization,find the red edge region(680-760nm)is most sensitive to FVC(foerest vegetation cover),and the correlation between the first derivative of red edge region's spectrum and fractional vegetation cover is the highest(>0.980)and it is steady at the same time.Therefor,chooses the ed-edge slope(k)as the parameter of the FVC estimation instead of classical inversion model based on NDVI.The accuracy of new model is high.(2)A New Method of PM2.5,PM10 Concentration inversion Based on Difference IndexThrough analyzing particular spectrum characterization under different concentration condition,find it make the reflectance increased in red band(620-900nm)and decreased in near-infrared(820-900nm),therefore,the difference index(difference,index,DI)is used to characterize the particles concentration.Fit DI-particles concentration and obtain the particles distribution(30m),This modle has high accurence in moderate and severe particulate pollution and make it possible to inversion high resolution particle concentration.(3)Selection of influence factors of PM2.5,PM10This paper introduce vegetation factors such as forest land,biomass and vegetation coverage comprehensive weather and terrain(totle 12 categories)to anlysis the particles influence.It is not limited to a certain category or a particular factor and has reference significance for the expansion of particle influence factors and the establishment of particle model.