An Interpretable Deep Inference Model With Dynamic Constraints for Forecasting the Evolution of Sea Surface Variables in the South China Sea

An interpretable deep inference forecasting model is designed to improve the forecasting capability of sea surface variables. By incorporating the air‐sea coupling mechanism as a dynamic constraint, the interpretability and forecasting performance of the model are improved. More specifically, our findings underscore the critical role of air‐sea interactions in forecasting sea surface variables, especially sea surface temperature (SST) variations induced by tropical cyclones (TCs). Additionally, Liang‐Kleeman information flow (IF), a causal inference method, is introduced to optimize the selection of predictors. Using satellite remote sensing data, our study demonstrates the model's capability in realizing sea surface multivariate forecasts in the South China Sea (SCS) within 10 days. More importantly, the experimental results prove the applicability of the model in both normal and extreme weather conditions, highlighting its effectiveness in enhancing sea surface variables forecasting. Plain Language Summary There are exchanges of momentum, heat, and mass between the ocean and the atmosphere. The sea surface, as a crucial interface for these exchanges, plays a pivotal role in the earth's climate system. Consequently, accurate prediction of sea surface variables is vital for understanding climate dynamics. Despite the considerable forecasting capabilities demonstrated by intelligent forecasting techniques, they still face issues of poor interpretability and low forecasting skills under extreme conditions compared to numerical models. The main reason is that previous intelligent forecasting models often focus on the evolution of a single variable, or only consider interactions within the ocean, and do not forecast under air‐sea coupling conditions. Such practice leads to incomplete systems and incoordination between air‐sea variables, thus failing to describe the interfacial behaviors of the ocean and atmosphere under extreme conditions such as typhoons. This study constructs an interpretable deep inference forecasting model for sea surface variables within the air‐sea coupling framework, illustrating the importance of considering air‐sea interactions to improve the forecasting performance of sea surface variables. Additionally, the model improves the prediction accuracy of sea surface variables (especially sea surface temperature) under extreme weather conditions. Key Points An interpretable forecasting model is proposed for the evolution of sea surface var