In this study, deep convolutional neural network (CNN) models of stimulated tropical cyclone intensity (TCI), minimum central pressure (MCP), and maximum 2 min mean wind speed at near center (MWS) were constructed based on ocean and atmospheric reanalysis, as well Best Track of tropical hurricane data over 2014–2018. In order to explore the interpretability of the model structure, sensitivity experiments were designed with various combinations of predictors. The model test results show that simplified VGG-16 (VGG-16 s) outperforms the other two general models (LeNet-5 and AlexNet). The results of the sensitivity experiments display good consistency with the hypothesis and perceptions, which verifies the validity and reliability of the model. Furthermore, the results also suggest that the importance of predictors varies in different targets. The top three factors that are highly related to TCI are sea surface temperature (SST), temperature at 500 hPa (TEM_500), and the differences in wind speed between 850 hPa and 500 hPa (vertical wind shear speed, VWSS). VWSS, relative humidity (RH), and SST are more significant than MCP. For MWS and SST, TEM_500, and temperature at 850 hPa (TEM_850) outweigh the other variables. This conclusion also implies that deep learning could be an alternative way to conduct intensive and quantitative research.

Chemical characteristics of fine particulate matter (PM_2.5) in Wuxi at urban, industrial, and clean sites on haze and non-haze days were investigated over four seasons in 2016. In this study, high concentrations of fine particulate matter (107.6 ± 25.3 μg/m³) were measured in haze episodes. The most abundant chemical components were organic matter (OM), SO₄²⁻, NO₃⁻, elemental carbon (EC), and NH₄⁺, which varied significantly on haze and non-haze days. The concentrations of OM and EC were 38.5 ± 5.4 μg/m³ and 12.3 ± 2.1 μg/m³ on haze days, which were more than four times greater than those on non-haze days. Source apportionment using a chemical mass balance (CMB) model showed that the dominant sources were secondary sulfate (17.7%), secondary organic aerosols (17.1%), and secondary nitrate (14.2%) during the entire sampling period. The source contribution estimates (SCEs) of most sources at clean sites were lower than at urban and industrial sites. Primary industrial emission sources, such as coal combustion and steel smelting, made larger contributions at industrial sites, while vehicle exhausts and cooking smoke showed higher contributions at urban sites. In addition, the SCEs of secondary sulfate, secondary nitrate, and secondary organic aerosols on haze days were much higher than those on non-haze days, indicating that the secondary particulate matter formations process was the dominating reason for high concentrations of particles on haze days.

In this study, the concentrations and chemical components of size-fractionated particulate matter (PM) in Nanjing at the ground (Gulou, 20 m) and above the urban canopy (Zifeng, 380 m) were sampled and analyzed from 16 November to 12 December in 2016. Higher concentrations of PM₁₀, PM_10-2.1, and PM_2.1 (108.3 ± 23.4 μg m⁻³, 47.3 ± 10.6 μg m⁻³, and 61.0 ± 18.8 μg m⁻³) were measured at Gulou than those (88.1 ± 21.1 μg m⁻³, 31.4 ± 6.7 μg m⁻³, and 56.7 ± 18.6 μg m⁻³) at Zifeng. The most abundant chemical components for size-fractionated PM were SO₄²⁻, NO₃⁻, organic carbon (OC), NH₄⁺, elemental carbon (EC), and crustal elements such as Al, Ca, Fe, and Mg, varying significantly on different particulate sizes. The concentrations of OC and EC were 7.46–19.60 μg m⁻³ and 3.44–5.96 μg m⁻³ at Gulou and were 8.34–18.62 μg m⁻³ and 2.86–4.11 μg m⁻³ at Zifeng, showing an equal importance in both fine and coarse particles. Nitrate, sulfate, and ammonium were more concentrated in PM_2.1, contributing 11.30–13.76 μg m⁻³, 8.91–9.40 μg m⁻³, and 5.78–6.81 μg m⁻³, which was more than in PM_10-2.1, which contributed 2.73–5.06 μg m⁻³, 2.16–3.81 μg m⁻³, and 0.85–0.87 μg m⁻³. In contrast, the crustal elements were larger in coarse particles and at the ground level, accounting for 18.6% and 15.3% of the total PM at Gulou and Zifeng. Source apportionment using the chemical mass balance (CMB) model EPA showed that the dominant three sources were secondary nitrate (18.2–24.9%), secondary sulfate (14.5–20.4%), and secondary organic aerosols (15.5–19.6%) for PM₁₀, PM_2.1, and PM_1.1 at both Gulou and Zifeng during the entire sampling period. However, for PM_10-2.1, the largest three contributors were secondary organic aerosols (18.3%), the coal-fired power plant (15.6%), and fugitive dust (14.4%), indicating dusts including construction dust, fugitive dust, and soil dust would contribute more at the ground. The results also showed that the concentrations of PM₁₀, PM_2.1, and PM_1.1 were lower than the work carried out in the winter of 2010 at the same sampling site by 41.4%, 26.3%, and 24.8%, confirming the improvement of the air quality and the efficient control of PM pollutants.