The transport and accumulation of moisture played an essential role in the extremely heavy rainfall of July 2020 in Japan. To better understand this event in terms of moisture sources and transport routes, backward particle trajectory analysis was conducted. We found two major moisture sources: transport from the tropics and uptake from the subtropics. A narrow moisture channel was found along the edge of the western Pacific Subtropical High (WPSH), transporting the moisture to the Baiu front. However, most moisture from the tropics was lost due to precipitation, and their contributions were reduced to about 15%. In contrast, the subtropical regions contributed over 80% moisture via evaporation and lower tropospheric convection. Among those regions, the western Pacific contributed the most (>33%). This study highlights the role of WPSH in moisture transport and demonstrated the importance of moisture uptake during transport.

The importance of air‐sea coupling in the simulation and prediction of the Madden‐Julian Oscillation (MJO) has been well established. However, it remains unclear how air‐sea coupling modulates the convection and related oceanic features on the sub‐daily scale. Based on a regional cloud‐permitting coupled model, we evaluated the impact of the air‐sea coupling on the convection during the convectively active phase of the MJO by varying the coupling frequency. The model successfully reproduced the atmospheric and oceanic variations observed by satellite and in situ measurements but with some quantitative biases. According to the sensitivity experiments, we found that stronger convection was mainly caused by the higher sea surface temperatures (SSTs) generated in high‐frequency coupled experiments, especially when the coupling frequency was 1 hour or shorter. A lower coupling frequency would generate the phase lags in the diurnal cycle of SST and related turbulent heat fluxes. Our analyses further demonstrated that the phase‐lagged diurnal cycle of SST suppressed deep convection through a decrease in daytime moistening in the lower troposphere. Meanwhile, in the upper ocean, the high‐frequency air‐sea coupling helped maintain the shallower mixed and isothermal layers by diurnal heating and cooling at the sea surface, which led to a higher mean SST. In contrast, the low‐frequency coupled experiments underestimated the SST and therefore convective activities. Overall, our results demonstrated that high‐frequency air‐sea coupling (1 hour or shorter) could improve the reproducibility of the intensity and temporal variation in both diurnal convection and upper ocean processes.

The prediction of oceanic features is always an important issue in oceanography, where deep learning has been proven to be a useful tool. In this study, we applied the improved U-net model to predict the monthly sea surface salinity (SSS) over the western Pacific (WP) Ocean, and the model was designed to use the SSSs from six consecutive months to predict the SSS in the next month. The monthly satellite-based SSSs in 2015–2020 were used for model training, and the data collected after January 2021 were used to evaluate the model’s predictive abilities. The results showed that the predicted SSSs represented the general patterns of SSSs over the WP region. However, the small-scale features were smoothed out in the model, and the temporal variations were also not well captured, especially over the East China Sea and Yellow Sea (ECS&YS) region. To further evaluate the potential of the U-net model, a more specific model was conducted for the ECS&YS region (Domain 2), which successfully predicted both spatial and temporal variations in the SSSs, including the spreading and retreating of the low-salinity tongue. Based on the comparison between the two domains and sensitivity experiments, we found that the prediction biases were contributed by the spatial distributions of the SSSs, the domain size, and the filter numbers. In addition, further multi-step prediction experiments suggested that our U-net model could also be used for long-time prediction, and we have examined up to five months. Overall, this study demonstrated the great ability and potential of the U-net model for predicting SSS, even though only a few trainable data are available.

In August 2022, the Tohoku region of Japan suffered three major heavy rainfall events, while one was related to an anticyclone, and the others were induced by a “Baiu‐like” stationary front. By using the Lagrangian backward trajectories, we found the moisture sources of the three events varied much due to the rapidly changing atmospheric conditions. In general, moisture from the Sea of Japan contributed the most in all events, while the western Pacific also played a certain role. Surprisingly, moisture from tropical storms and the East China Sea did not contribute much due to the along‐transport precipitation. But, based on forward‐traced trajectories, we confirmed that tropical storms did enhance moisture transport from subtropical regions via their southerly wind. Overall, this study highlights the role of the moisture gains nearby and the outer flows of tropical storms during heavy rainfalls in a northern region.