On the reproducibility of the September 2002 vortex splitting event in the Antarctic stratosphere achieved without satellite observations

To highlight the impact of satellite measurements, a comparison between the Japanese 55-year reanalysis (JRA-55) and its equivalent without the assimilation of satellite observations (JRA-55C; C stands for ‘conventional’ observations) was conducted. As an illustrative example of the detectability problem of extreme events, we report on the reproducibility of a stratospheric sudden warming (SSW) event that occurred in late September 2002; this event represents the first observed unique vortex splitting event in the Antarctic stratosphere. Through the data assimilation system of JRA-55, the initial tendency of this warming event and the following recovery process were well captured even when no satellite observations were used. However, the warming in JRA-55C does not satisfy the criteria for a major SSW event besides the lack of splitting behaviour in the polar vortex. A prominent difference between JRA-55 and JRA-55C during the SSW event, which was characterized by the sudden appearance of a nearly barotropic structure from the upper stratosphere to the troposphere, was found over the Western Hemisphere reflecting the geographic distribution of observational sites. Moreover, several differences in the precursory state of the polar vortex and the observational anchoring effect are consistent with the proposal that this SSW was caused by the catastrophic breakdown of a highly deformed polar vortex as suggested by some recent works.