Verifying forecasts for major stratospheric sudden warmings

A

stratospheric sudden warming

is perhaps one of the most radical changes of weather that is observed on our planet. As numerical weather prediction models have improved, including better representation of the stratosphere, an extensive amount of studies have been investigating forecasts for major stratospheric sudden warmings (MSSWs), which affect

all layers of the atmosphere

, changing wind circulation patterns and space weather effects like the aurora.

Whereas most previous studies employed single systems for a limited number of MSSWs, a new study published in


Advances in Atmospheric Sciences


sought to verify multi-system MSSW forecasts using hindcasts of four systems archived in the subseasonal-to-seasonal prediction project database. The study is also featured on the cover of the latest issue of the journal.

“The target hindcast period extended from 1998/99 to 2012/13, including 12 MSSWs.” Said the author, Prof. Masakazu Taguchi, from the Department of Earth Science, Aichi University of Education in Japan, “the results show that all four systems can be judged to be skillful for five-day MSSW forecasts when averaged across all available MSSWs.”

For longer lead times, such as 15 or 20 days, however, some systems are skillful, but others are not. Taguchi found that it is more difficult to forecast MSSWs where the polar vortex splits into two or greatly stretches, as compared to MSSWs where the vortex just shifts away from the pole, although a statistically significant difference was not obtained for almost all cases (systems and verification measures).

“This study could be extended in a future line of research to better unravel the characteristics of MSSW forecasts, e.g., in terms of case-to-case variations in predictable lead time, and their determinant (i.e., source of the predictability),” said Taguchi. “It will also be useful to identify connections between specific MSSWs and anomalous weather conditions in both the real world and in forecasts.”

###