Exploring the Influence of Tropical Cyclones on Regional Air Quality Using Multimodal Deep Learning Techniques

Area of Science:

• Atmospheric Science
• Environmental Science
• Data Science

Background:

• Tropical cyclones (TCs) are severe weather events with significant environmental impacts.
• Understanding the influence of TCs on air quality is crucial for public health and environmental management.

Purpose of the Study:

• To investigate the subtle role of tropical cyclones in air quality index (AQI) variations.
• To analyze air quality changes before, during, and after TCs.
• To provide predictive insights into AQI fluctuations associated with TCs.

Main Methods:

• Employed multimodal data including meteorological data and satellite observations.
• Utilized deep learning models (ConvLSTM, CNN, Real-ESRGAN) combined with regression models.
• Applied Convolutional Neural Network (CNN) for TC classification and Extra Trees Regressor (ETR) for AQI prediction.

Main Results:

• Achieved 92.02% accuracy using CNN for TC classification.
• Obtained an R2 score of 83.33% for AQI prediction with ETR.
• Uncovered complex patterns and non-linear interdependencies between TC features and AQI.

Conclusions:

• TCs play a complex role in regional and global air quality.
• Findings enhance understanding of TC-air quality interactions, informing policymakers and researchers.
• Highlights public health concerns related to climate adaptation and urban renewal post-cyclone.