Analysis of atmospheric particulate pollutants using multisource remote sensing data during the spring festival in Shenyang

Using multi-source remote sensing data such as FY-2G Satellite detection products (OLR, humidity profile) and wind profile radar data, this study analyzes the atmospheric quality and influencing factors in Shenyang during the Spring Festival period from 2020 to 2024. Combining backward trajectory models and clustering analysis methods, the transmission path characteristics are explored, and pollution prediction research is carried out through machine learning methods. The results showed that the air quality in Shenyang during the Spring Festival was mainly excellent-good, and relatively poor in 2020 and 2024, with severe and above polluted weather, due to the continuous control of the COVID-19 and the impact of cold air. In 2020, the overall atmospheric environment showed high humidity and small wind characteristics, with weak boundary layer wind speed and vertical velocity. In 2024, it was characterized by first cooling and then warming, with negative vertical velocity and mostly sinking motion. The majority of the ground is dominated by northerly winds, reaching 63.2 % in 2020 and 17.5 % in 2024, providing stable conditions for the occurrence of polluted weather. When the pollution is relatively heavy, OLR values are generally small. Backward trajectory simulation and cluster analysis show that pollution transmission trajectories can be divided into four categories. In 2020, the main path was the northerly path, accounting for a cumulative proportion of 85.62 %, while in 2024, the main path was the northwest path, with long-distance paths accounting for 34.38 %. The random forest algorithm has high accuracy in predicting pollutant concentration and AQI. Two types of classification predictions were made for the presence or absence of pollution, with an accuracy rate of 93 %.