RTCC-Net: tropical cyclone generation classification model based on multi-source information fusion

Tropical cyclones are among the most destructive extreme weather phenomena, and accurately determining whether a tropical cloud cluster will develop into a tropical cyclone is crucial for disaster prevention and mitigation. Considering that previous deep learning studies on tropical cyclogenesis have insufficiently extracted many key features from tropical cloud cluster data, this paper proposes the Real Time Tropical Cyclogenesis Classification-Net (RTCC-Net) based on multi-source information fusion. The model computes convective core maps and polar-coordinate representations from infrared images of tropical cloud clusters and employs ResNet together with a self-attention mechanism to extract their spatiotemporal features. Experimental results demonstrate that the proposed model achieves high accuracy and stability: it attains a detection rate of 98.7% and a false alarm rate of 1.3% when classifying the formation of tropical cyclones 24 hours in advance. Notably, the model not only ensures potential of real-time classification capabilities from satellite data but also surpasses the accuracy of models that utilize reanalysis data.