| 英文摘要 |
This project aims to analyze the influence of different seasonal meteorological conditions on visibility, atmospheric oxidation capacity, and air quality in the Taichung region. In accordance with the results of the project, provide related recommendations for the eight dimensions of the air pollution control plan. The results show a consistent annual increase in visibility of 4.52% and an annual decrease in PM2.5 of 4.94% from 2012 to 2021, indicating a significant improvement in air quality and visibility. The number of days with low visibility and high PM2.5 concentrations has decreased significantly. Furthermore, the period from 2017 to 2022 shows a transition from PM2.5-dominated events to PM1.0-dominated events, highlighting the greater contribution of PM1.0 to visibility degradation. Visibility degradation is found to be associated with an increase in secondary inorganic salts (primarily nitrates) in PM2.5. The inorganic salts, characterized by high hygroscopicity, create a positive feedback loop and are found to be critical species in driving visibility degradation. A comparison of the major chemical constituents of PM2.5 across years shows an increasing proportion of organic matter, suggesting potential shifts in pollution characteristics. Throughout the project, 26 visibility degradation events were observed, 96% of which were associated with stagnant dispersion conditions and 12 of which were associated with secondary particle formation, highlighting the importance of both atmospheric dispersion conditions and secondary formation in degrading visibility.
Regarding the oxidative capacity of the atmosphere, satellite retrieval results show a gradual increase in the VOC-limited area for winter ozone formation in the Taichung region over the past five years. This implies that the ozone formation during this period is mainly controlled by VOCs. This result not only provides guidance for future air quality improvement, but also highlights the importance of further monitoring and control of VOC emissions.
Finally, the project developed a visibility prediction model using a combination of multiple linear regression and machine learning algorithms for improved applicability. The model incorporates aerosol composition, physicochemical parameters, and anthropogenic variability parameters, taking into account seasonal variations. In terms of spatial analysis, a satellite-based model of aerosol optical thickness has been established to provide high spatio-temporal resolution retrieval results, aiming to reduce the uncertainty of satellite retrievals and extend the spatial coverage of measurements.
|