Integrating Deep Learning Techniques for Enhanced Cyclone Prediction and its Impact on Air Quality Index Utilizing Satellite Imagery

Tropical cyclones (TCs greatly impact ecosystems and weather patterns, often leading to catastrophic outcomes this research explores dynamic relationship between the Air Quality Index (AQI) and Tropical Cyclones (TCs), examining changes in air quality during different stages of TCs. Focusing on Cyclone Biparjoy's impact on India's coastline in June 2023, we uncover its influence on air quality. The study shows good predictive performance by utilizing meteorological data, satellite observations, and machine learning algorithms such as CNN and ConvLSTM. Results show promising outcomes, with CNN achieving an RMSE of about 12.03 and an MAE of approximately 8.523 for TC intensity prediction, and ConvLSTM producing an RMSE of approximately 0.22 and a loss of around 0.04 for TC track prediction. Additionally, regression techniques using meteorological data from Tuetempo for AQI prediction were employed. The ExtraTrees Regression model exhibited strong predictive performance with an RMSE of approximately 19.627, while the Xgboost model had a slightly higher RMSE of around 21.72 our dataset included meteorological data from Tuetempo website for AQI forecasts and datasets like Hursat and INSAT 3D satellite photos for TC-related data, this work offers important insights into the dynamics of air quality during TC events.