Exploring Aerosol Vertical Distributions and Their Influencing Factors: Insight from MAX-DOAS and Machine Learning

Area of Science:

• Atmospheric Science
• Environmental Science
• Remote Sensing

Background:

• Aerosol vertical distribution is critical for air pollution management but is poorly understood due to limited data.
• Accurate aerosol profiling is essential for developing effective pollution mitigation strategies.

Purpose of the Study:

• To retrieve high-resolution aerosol optical properties and understand their vertical distribution in Shanghai.
• To investigate the key drivers influencing aerosol variations across different altitudes and seasons.
• To inform targeted pollution control strategies based on a multidimensional understanding of aerosol dynamics.

Main Methods:

• Employed multiaxis differential optical absorption spectroscopy (MAX-DOAS) for aerosol measurements.
• Utilized a coupled radiative transfer model-machine learning (RTM-ML) framework for property retrieval.
• Applied multifactor driving ML models and Shapley additive explanations (SHAP) to identify influencing factors.

Main Results:

• Aerosols generally decrease with altitude, with summer peaks in the upper atmosphere and winter peaks in the lower atmosphere.
• Aerosol hygroscopicity shows seasonal variation and increases with altitude.
• Below 0.5 km, emissions, east-west transport, and oxidation dominate; above 0.5 km, humidity and oxidation are key drivers.
• North-south transport significantly impacts aerosols between 0.5 and 1.6 km.

Conclusions:

• Emission reductions are effective for lowering lower-atmosphere aerosols.
• Enhanced atmospheric oxidation promotes secondary aerosol formation, especially in the upper atmosphere.
• Effective pollution mitigation requires a multidimensional approach considering various factors influencing vertical aerosol distribution.