Advances in Climate Change Research (Dec 2022)
Record-breaking rainfall over the middle reaches of the Yangtze River in August 2021: Sub-seasonal perspective and its predictability
Abstract
An extremely heavy rainfall cluster was observed over the middle reaches of the Yangtze River (MYR) in August 2021, breaking the historical record since 1981 and causing severe floods. The controlling circulation regime and possible predictability sources of the un-expected rainfall are still unclear, especially from a sub-seasonal perspective. Our results show that the successive heavy rainfall events in August 2021 had significant intraseasonal oscillations for 10–24 d and 30–60 d, and they were synergistically influenced by tropical and extra-tropical circulation regime. Above all, the East Asia–Pacific teleconnection pattern (EAP) turned into a negative phase in late July and maintained throughout August, providing favorable background conditions for the northward transport of tropical water vapor and the southward intrusion of cold air from mid-high latitudes. The 30–60-d oscillation of precipitation was dominated by the first mode of boreal summer intraseasonal oscillation (BSISO1, with 30–60-d period), which strengthened the tropical moisture transport, while the quasi-biweekly oscillation of the western Pacific subtropical high and the anomalous mid-latitude cyclone of EAP intensified the precipitation on the scale of 10–24 d. In addition, the second mode of BSISO (BSISO2, with 10–30-d period) may also intensify the 10–24-d precipitation in late August. The forecast leading time of the S2S models for the persistent heavy rainfall in August 2021 was basically 1–2 weeks, while the EAP and BSISO1 were revealed as the main predictable sources of the abnormal rainfall events. These results highlight the importance of superposition of different intraseasonal oscillations in forming the extreme rainfall event and demonstrate a potential chance to enhance the prediction skill of extreme event if the ensemble members in S2S models could be reasonably selected according to their performances on key predictability sources.
