Predictability of stratospheric sudden warming events in the Northern Hemisphere

We compare predictability of major stratospheric sudden warming events occurring in late December 2001 and early January 2004 using the operational ensemble one-month forecast data provided by the Japan Meteorological Agency. The 2001 warming event was brought about by the amplification of zonal wavenumber (WN) 1 components without pre-conditioning after a cold and undisturbed early winter. On the other hand, the 2004 warming occurred after a pronounced minor warming confined to the upper stratosphere in late December 2003. In this case, the intensification of WN 1 first caused the weakening of the polar vortex and brought about the minor warming. After incomplete reestablishment of the polar vortex, the second intensification of WN 1 gave rise to the major warming, under significant helping of WN 2. Hence, the situation of the latter warming seems to be rather complicated. Concerning the 2001 warming event, the stratospheric warming in the polar region is predictable at least from two weeks in advance: All the ensemble members successfully predicted the temperature change in terms of timing as well as its magnitude. In case of the 2004 warming event, however, almost all ensemble members starting from two weeks in advance failed the temperature change accompanied with the warming event, suffering from the previous minor warming. Moreover, it is found that the minor warming itself was also imperfectly predicted from two weeks in advance. Such shortened predictability could be related to more complicated stratosphere-troposphere dynamical interaction during the 2004 warming period.