[Application of Support Vector Machine Regression in Ozone Forecasting].

Support vector machine regression (SVMr) was proposed to forecast hourly ozone (O3) concentrations, daily maximum O3 concentrations, and maximum 8 h moving average O3 concentrations (O3 8 h) by employing the observations of meteorological variables and O3 and its precursors during the high O3 periods from May 20 to August 15, 2016 at an industrial area in Nanjing. The squared correlation coefficient (R2) of the hourly O3 concentrations forecast was 0.84. The mean absolute error (MAE) and mean absolute percentage error (MAPE) were 3.44×10-9 and 24.48, respectively. The key factors for the hourly O3 forecast were the O3 pre-concentrations, amount of ultraviolet radiation B (UVB), and the NO2 concentration. The main factors for the O3 daily maximum forecast were the NOx concentrations at 07:00 and the UVB level. Temperature and UVB played an important role in predicting O3 8 h. In general, taking precursors into account could increase the accuracy of O3 prediction by 10%-28%. For O3 concentration forecasting, SVMr gave significantly better predictions than multiple linear regression methods.

[Source Analysis of Volatile Organic Compounds in the Nanjing Industrial Area and Evaluation of Their Contribution to Ozone].

Ambient volatile organic compounds (VOCs) were continuously measured during the high ozone (O3) periods from May 1 to May 31 and June 1 to July 16, 2015 at an industrial area in the north suburb of Nanjing. A positive matrix factorization (PMF) model and an observation-based model (OBM) were combined for the first time to investigate the contributions of VOC sources and species to local photochemical O3 formation. The average VOC concentrations in 2014 and 2015 were (36.47±33.44)×10-9 and (34.69±34.08)×10-9, respectively. The VOC sources identified by the PMF model for 2014 and 2015 belonged to 7 source categories, including vehicular emissions, liquefied petroleum gas usage, biogenic emissions, furniture manufacturing industry, chemical industry, chemical coating industry, and chemical materials industry emission sources. The OBM was modified to assess the O3 precursors' relationships. Generally, photochemical O3 production was VOC limited, with positive relative incremental reactivity (RIR) values for VOC species and a negative RIR value for NO. It can be seen that alkenes (1.20-1.79) and aromatics (1.42-1.48) presented higher RIR values and controlling O3 would be the most effective when the VOC emissions from alkenes were reduced by 80%. Vehicle emissions (1.01-1.11), LPG (0.74-0.82), biogenic emissions (0.34-0.42), and furniture manufacturing industry (0.32-0.49) sources were the top four VOC sources making significant contributions to photochemical O3 formation, which suggests that controlling vehicle emissions, biogenic emissions, LPG, and furniture manufacturing industry sources should be the most effective strategy to reduce photochemical O3 formation.

[Variation Characteristics and Health Risk Assessment of BTEX in the Atmosphere of Northern Suburb of Nanjing].

BTEX concentrations were determined by GC5000 online gas chromatography in the atmosphere of the north suburb of Nanjing in March 2013 to February 2014, using the EPA human exposure analysis evaluation method for benzene series compounds of volatile organic compounds (VOCs) in health risk assessment. The results showed that the total amount of BTEX showed the variation characteristics of spring > winter > autumn > summer. BTEX concentration was higher in the periods of 07:00-10:00 and 17:00-20:00, and the lowest was detected between 13:00-15:00; At the weekend, the concentration of BTEX was higher than on the working day. The sources of BTEX included traffic sources, industrial sources and solvent evaporation. The HQ of BTEX in all four seasons showed the order of benzene > xylene > ethylbenzene > toluene, and the HQ risk values were within the safety range in all analysis periods. The distribution of R value was winter > autumn > spring > summer, and R was higher than the safety threshold for all the analyses, indicating the existence of carcinogenic risk.

[Characteristics Analysis of the Surface Ozone Concentration of China in 2015].

Surface ozone concentration data from 189 cities in China in 2015 were processed by ArcGIS software in order to obtain the characteristics of the surface ozone concentration, such as time and space, topographical features, temperature, etc. The trend for surface ozone concentration was a decrease followed by an increase in China in 2015. The concentrations during the four seasons followed the order:summer > autumn > spring > winter, and the maximum appeared in July. The ozone pollution of East China, South China, and North China were more serious than other regions in China. The variation of longitude had a small influence on the ozone concentration, while the influence of latitude is significant. According to the analysis contrasting three different topographies in the same latitude, the influence of topography on ozone concentration was negligible. Furthermore, the research found a significant positive correlation between surface ozone concentration and temperature.

[Modeled Deposition of Fine Particles in Human Airway in Northern Suburb of Nanjing].

The particles number concentrations were determined by Wide-range Particle Spectrometer (WPS) in northern suburb of Nanjing in January and April 2015. The information of size distributions was applied in the multiple-path particle dosimetry model (MPPD) v.3.04 to quantify deposition fractions (DF) and number concentration (NC) depositions of fine particles in different regions of human airway, at different air quality levels, at rest and exercise. DF of nucleation mode and Aitken mode at rest and exercise were similar, while DF of accumulation mode at exercise was 2.49 times of that at rest. DF of nucleation mode and Aitken mode in pulmonary (PUL) was the highest, about 48.17% of total deposition fractions (TDF) at rest and 54.23% of TDF at exercise. DF of accumulation mode in head was the highest, about 41.23% of TDF at rest and 80.47% of TDF at exercise. The particle NC deposition in human airway in winter was lower than that in spring, and the total NC deposition in 3 regions was in the order of PUL > tracheobronchial(TB) > head. Compared with resting, nucleation mode deposition in PUL and accumulation mode deposition in TB and head increased at exercise. The worse the air quality, the higher the deposition growth rate of exercising to resting in head. DF difference among regions was mainly due to the different physiological parameters, while NC deposition difference was mainly due to the different particle NC in the local environment.