Extra predictability from a seamless approach for Asian summer monsoon precipitation from days to weeks

Abstract

AbstractMultiple‐scale atmospheric or oceanic prediction has been a challenging issue due to the different dynamical and thermodynamic processes involved and the uncertainty inherent in highly nonlinear and stochastic forcing, such as the prediction of Asian summer monsoon (ASM) precipitation. Seamless prediction was proposed to address this issue. The idea behind this approach is that temporal averaging reduces the spread of the prediction ensemble but retains the approximate first moment, which effectively increases the signal‐to‐noise ratio of the prediction. In this study, the superiority of the seamless prediction was assessed for the ASM precipitation at time‐scales from days to weeks. The hindcasts from the European Centre for Medium‐Range Weather Forecasts in the Subseasonal to Seasonal ensemble dataset were used for the detailed assessment. Results show significant advantages of the seamless approach in evaluating the ASM precipitation prediction on time‐scales ranging from 2 days to 2 weeks. Furthermore, the source of the average window predictability was analysed for two lead times with significantly superior prediction accuracy: lead times of 4d4d and 3w3w. For 4d4d, the Madden–Julian Oscillation is dominant, while the role of the El Niño–Southern Oscillation phase becomes substantial for 3w3w. These results offer a detailed analysis for the prediction of ASM precipitation within a framework of seamless prediction, benefiting further interpretation of seamless prediction of ASM precipitation.