Resilience to Air Pollution: A Novel Approach for Detecting and Predicting Aerosol Atmospheric Rivers within Earth System Boundaries

Abstract

The study explores extreme aerosol transport (EAT) events using atmospheric river (ARs) dynamics to identify aerosol atmospheric rivers (AARs). This provides insight into their significance in mitigating aerosol pollution and strengthening resilience within Earth system Boundaries (ESBs). AARs are narrow and long regions with high concentrations of various aerosols, including Black Carbon (BC), Dust (DU), Organic Carbon (OC), Sea Salt (SS), and Sulphate (SU), transported over long distances. Leveraging MERRA-2 re-analysis datasets, this study detects the AARs by applying various boundary conditions and develops a Spatio-Temporal AAR Availability Prediction Model (ST-AARAPM) based on a convolutional autoencoder. The model predicts AAR availability for the next t + 5-time frames using Stochastic Gradient Descent (SGD) optimization, minimizing Mean Squared Error (MSE) loss with a Rectified Linear Unit. Model performance is evaluated using metrics such as Structural Similarity Index (SSIM), Root Mean Squared Error (RMSE), Peak Signal Noise Ratio (PSNR), and MSE. From 2015 to 2022, the study identified 128,261 AARs worldwide with at least 8 AARs present at any given time frame. However, the model evaluation indicates satisfactory results, with SSIM, PSNR, RMSE, and MSE ranging from 0.88 to 0.96, 67.60 to 78.50 dB, 0.0656 to 0.1552, and 0.0043 to 0.0247, respectively. The findings highlight the effectiveness of ST-AARAPM in forecasting AAR availability and enhancing resilience in hotspot regions with significant aerosol loading, including the Indo-Gangetic plains, Eastern China, Japan, Northern Africa, Eastern USA, and South America. The study offers a fresh approach to tackling the effects of severe aerosol pollution via AARs within ESB’s. It stresses the need for policies to curb emissions, encourage sustainable production, and embrace clean energy. It calls on vulnerable systems to shift to cleaner technologies for resilience against aerosol pollution. Graphical Abstract: (Figure presented.) © King Abdulaziz University and Springer Nature Switzerland AG 2024.