Abstract
Rapid intensification (RI) of tropical cyclones often causes majordestruction to human civilization due to short response time. It is animportant yet challenging task to accurately predict this kind of extremeweather event in advance. Traditionally, meteorologists tackle the task withhuman-driven feature extraction and predictor correction procedures.Nevertheless, these procedures do not leverage the power of modern machinelearning models and abundant sensor data, such as satellite images. Inaddition, the human-driven nature of such an approach makes it difficult toreproduce and benchmark prediction models. In this study, we build a benchmarkfor RI prediction using only satellite images, which are underutilized intraditional techniques. The benchmark follows conventional data sciencepractices, making it easier for data scientists to contribute to RI prediction.We demonstrate the usefulness of the benchmark by designing a domain-inspiredspatiotemporal deep learning model. The results showcase the promisingperformance of deep learning in solving complex meteorological problems such asRI prediction.