Self-feedback LSTM regression model for real-time particle source apportionment

Atmospheric particulate matter pollution has attracted much wider attention globally.In recent years,the development of atmospheric particle collection techniques has put forwards new demands on the real-time source apportionments techniques.Such demands are summarized,in this paper,as how to set up new restraints in apportionment and how to develop a non-linear regression model to process complicated circumstances,such as the existence of secondary source and similar source.In this study,we firstly analyze the possible and potential restraints in single particle source apportionment,then propose a novel three-step self-feedback long short-term memory(SF-LSTM)network for approximating the source contribution.The proposed deep learning neural network includes three modules,as generation,scoring and refining,and regeneration modules.Benefited from the scoring modules,SF-LSTM implants four loss functions representing four restraints to be followed in the apportionment,meanwhile,the regeneration module calculates the source contribution in a non-linear way.The results show that the model outperforms the conventional regression methods in the overall performance of the four evaluation indicators(residual sum of squares,stability,sparsity,negativity)for the restraints.Additionally,in short time-resolution analyzing,SF-LSTM provides better results under the restraint of stability.

Compositions and sources of organic acids in fine particles(PM_(2.5)) over the Pearl River Delta region, south China

Organic acids as important constituents of organic aerosols not only influence the aerosols＇ hygroscopic property, but also enhance the formation of new particles and secondary organic aerosols. This study reported organic acids including C14-C32 fatty acids, C4-C9 dicarboxylic acids and aromatic acids in PM2.5 collected during winter 2009 at six typical urban, suburban and rural sites in the Pearl River Delta region. Averaged concentrations of C14-C32 fatty acids, aromatic acids and C4- C9 dicarboxylic acids were 157, 72.5 and 50.7 ng/m3, respectively. They totally accounted for 1.7% of measured organic carbon. C20-C32 fatty acids mainly deriving from higher plant wax showed the highest concentration at the upwind rural site with more vegetation around, while Cl4-C18 fatty acids were more abundant at urban and suburban sites, and dicarboxylic acids and aromatic acids except 1,4-phthalic acid peaked at the downwind rural site. Succinic and azelaic acid were the most abundant among C4-C9 dicarboxylic acids, and 1,2-phthalic and 1,4-phthalic acid were dominant aromatic acids. Dicarboxylic acids and aromatic acids exhibited significant mutual correlations except for 1,4-phthalic acid, which was probably primarily emitted from combustion of solid wastes containing polyethylene terephthalate plastics. Spatial patterns and correlations with typical source tracers suggested that C14-C32 fatty acids were mainly primary while dicarboxylic and aromatic acids were largely secondary. Principal component analysis resolved six sources including biomass burning, natural higher plant wax, two mixed anthropogenic and two secondary sources; further multiple linear regression revealed their contributions to individual organic acids. It turned out that more than 70% of C14-C18 fatty acids were attributed to anthropogenic sources, about 50%-85% of the C20-C32 fatty acids were attributed to natural sources, 80%-95% of dicarboxylic acids and 1,2-phthalic acid were secondary in contrast with that 81% of 1,4-phthalic acid was primary.

Pion emission in α-particle interactions with various targets of nuclear emulsion detector

The behavior of relativistic hadron multiplicity for 4He-nucleus interactions is investigated. The experi- ment is carried out at 2.1 A and 3.7 A GeV （Dubna energy） to search for the incident energy effect on the interactions inside different emulsion target nuclei. Data are presented in terms of the number of emitted relativistic hadrons in both forward and backward angular zones. The dependence on the target size is presented. For this purpose the statistical events are discriminated into groups according to the interactions with H, CNO, Em, and AgBr target nuclei. The separation of events, into the mentioned groups, is executed based on Glauber＇s multiple scattering theory approach. Features suggestive of a decay mechanism seem to be a characteristic of the backward emission of relativistic hadrons. The results strongly support the assumption that the relativistic hadrons may already be emitted during the de-excitation of the excited target nucleus, in a behavior like that of compound-nucleus disintegration. Regarding the limiting fragmentation hypothesis beyond 1 A GeV, the target size is the main parameter affecting the backward production of the relativistic hadron. The incident energy is a principal factor responsible for the forward relativistic hadron production, implying that this system of particle production is a creation system. However, the target size is an effective parameter as well as the projectile size considering the geometrical concept regarded in the nuclear fireball model. The data are analyzed in the framework of the FRITIOF model.

Characterization of submicron particles during autumn in Beijing,China

In this study, we performed a highly time-resolved chemical characterization of nonrefractory submicron particles（NR-PM_1） in Beijing by using an Aerodyne high-resolution time-of-flight aerosol mass spectrometer（HR-ToF-AMS）. The results showed the average NR-PM_1 mass concentration to be 56.4 ± 58.0 μg/m^3, with a peak at 307.4 μg/m^3. Due to the high frequency of biomass burning in autumn, submicron particles significantly increased in organic content, which accounted for 51% of NR-PM_1 on average. Secondary inorganic aerosols（sulfate ＋ nitrate ＋ ammonium） accounted for 46% of NR-PM_1, of which sulfate,nitrate, and ammonium contributed 15%, 20%, and 11%, respectively. To determine the intrinsic relationships between the organic and inorganic species, we used the positive matrix factorization（PMF） model to merge the high-resolution mass spectra of the organic species and NO＋and NO_2~＋ions. The PMF analysis separated the mixed organic and nitrate（NO＋and NO_2~＋） spectra into four organic factors, including hydrocarbon-like organic aerosol（HOA）, oxygenated organic aerosol（OOA）, cooking organic aerosol（COA）, and biomass burning organic aerosol（BBOA）, as well as one nitrate inorganic aerosol（NIA） factor. COA（33%） and OOA（30%） contributed the most to the total organic aerosol（OA） mass, followed by BBOA（20%） and HOA（17%）. We successfully quantified the mass concentrations of the organic and inorganic nitrates by the NO＋and NO2＋ions signal in the organic and NIA factors. The organic nitrate mass varied from 0.01-6.8 μg/m^3, with an average of 1.0 ±1.1 μg/m^3, and organic nitrate components accounted for 10% of the total nitrate mass in this observation.

Origin of high particle number concentrations reaching the St. Louis, Midwest Supersite

Ultrafine particles are associated with adverse health effects. Total Particle Number Concentration（TNC） of fine particles were measured during 2002 at the St. Louis — Midwest supersite. The time series showed overall low level with frequent large peaks. The time series was analyzed alongside criteria pollutant measurements and meteorological observations. Multiple regression analysis was used to identify further contributing factors and to determine the association of different pollutants with TNC levels. This showed the strong contribution of sulfur dioxide（SO2） and nitrogen oxides（NO x） to high TNC levels. The analysis also suggested that increased dispersion resulting from faster winds and higher mixing heights led to higher TNC levels. Overall, the results show that there were intense particle nucleation events in a SO2 rich plume reaching the site which contributed around 29% of TNC. A further 40% was associated with primary emissions from mobile sources. By separating the remaining TNC by time of day and clear sky conditions,we suggest that most likely 8% of TNC are due to regional nucleation events and 23% are associated with the general urban background.