Predicting the Intensity of Tropical Cyclones over the Western North Pacific Using Dual-Branch Spatio-Temporal Attention Convolutional Network

Abstract This paper proposes a Spatio-temporal Attention Convolutional Network (STACPred) that leverages deep learning techniques to model the spatio-temporal features of tropical cyclones (TC) and enable real-time prediction of their intensity. The proposed model employs dual branches to concurrently extract and integrate features from intensity heatmaps and satellite cloud imagery. Additionally, a Residual Attention (RA) module is integrated into the three-channel cloud imagery convolution process to automatically respond to high wind speed regions. TC’s longitude, latitude, and radius of winds are injected into the multi-timepoint prediction model to assist in the prediction task. Furthermore, a Rolling Mechanism (RM) is employed to smooth the fluctuation of losses, achieving accurate prediction of TC intensity. We use several TC records to evaluate and validate the universality and effectiveness of the model. The results indicate that STAC-Pred achieve satisfactory performance. Specifically, the STAC-Pred model improves prediction performance by 47.69% and 28.26% compared to the baseline (official institutions) at 3 and 6-hour intervals, respectively.