Search

Accurate predictions of monthly extremes assume paramount importance in enabling proactive decision‐making, which however are lacked in skills even for state‐of‐the‐art dynamical models. Taking the extreme precipitation prediction over the mid‐to‐lower reaches of the Yangtze River, China, as an instance, a multi‐predictor U‐Net deep learning approach is designed to enhance the prediction over the European Center for Medium‐Range Weather Forecasts (ECMWF) model, with the single‐predictor U‐Net parallelly examined as the benchmark. Focusing on the precipitation extremes, an extreme associated component is incorporated into the model loss function for optimization. Besides, predictions composed by daily outputs with multiple lead times are imported as a comprehensive set in the training phase to augment the deep learning sample size and to emphasize enhancements in predictions at the monthly timescale as a whole. Results indicate that the multi‐predictor U‐Net effectively improves predictions of extreme summer precipitation frequency, showing distinct superiority to the raw ECMWF and the single‐predictor U‐Net. Multiple evaluation metrics indicate that the model shows a significant positive improvement ratio ranging from 65.1% to 80.0% across all grids compared to the raw ECMWF prediction, which has also been validated through applications in the two extreme summer precipitation cases in 2016 and 2020. Besides, a ranking analysis of feature importance reveals that factors such as humidity and temperature play even more crucial roles than precipitation itself in the multi‐predictor extreme precipitation prediction model at the monthly timescale. That is, in such a deep learning approach, the monthly prediction on extreme precipitation benefits significantly from the inclusion of multiple associated predictors. Plain Language Summary: Predicting extreme precipitation is crucial for disaster prevention. This study introduces a new deep learning model using a multi‐predictor U‐Net approach in the mid‐to‐lower Yangtze River in China, aiming to improve the accuracy of monthly extreme precipitation predictions over existing models from the European Center for Medium‐Range Weather Forecasts (ECMWF). The enhanced model integrates additional variables like humidity and temperature, which are crucial for predicting extreme events. By analyzing summer precipitation extremes from 2016 to 2020, the multi‐predictor model shows a significant improvement, increasing prediction accuracy by 65.1%–80.0% of the total grids over the ECMWF model. This research demonstrates that incorporating multiple variables can significantly boost the effectiveness of monthly extreme precipitation forecasts, providing a valuable tool for better preparing for potential disasters. Key Points: A multi‐predictor U‐Net with extreme‐associated loss component is established to enhance monthly range extreme precipitation predictionsThe multi‐predictor U‐Net outperforms European Center for Medium‐Range Weather Forecasts (ECMWF) with conspicuous improvements by 65.1%–80.0%, which is also demonstrated by case analysesHumidity and temperature are crucial predictors boosting the multi‐predictor U‐Net performance in predicting extreme precipitation [ABSTRACT FROM AUTHOR]

Compared with individual extreme events, compound events have more severe impacts on humans and the natural environment. This study explores the change in severity of compound extreme high temperature and drought/rain events (CHTDE/CHTRE) and associated influencing factors. The CHTDE and CHTRE intensified in most areas of China in summer (June–July August) during 1961–2014. Under global warming, the increased water-holding capacity of the atmosphere and the decreased relative humidity led to an increase in the severity of CHTDE. The severity of CHTRE is increased because of enhanced transient water vapor convergence and convective motion. Anthropogenic climate change, especially greenhouse gas forcing, which contributes 90% to the linear change in the severity of CHTDE and CHTRE, is identified as the dominant factor affecting the severity of CHTDE in China. In addition, the historical natural forcing (hist-NAT) may be related to the interannual-to-decadal variability in the severity of CHTDE/CHTRE. [ABSTRACT FROM AUTHOR]