Whole-Atmosphere Predictability Across the 2023 Sudden Stratospheric Warming

Operational space-weather forecasting increasingly depends on whole-atmosphere models, yet we still lack a quantitative picture of its predictability. Sudden stratospheric warmings (SSWs) - which reorganize circulation from the surface to the thermosphere - offer a powerful natural test of this question. Using NOAA's Whole Atmosphere Model–Ionosphere Plasmasphere Electrodynamics (WAM-IPE) Forecast System, which produces four 27-hour cycles per day, we map predictability across the January-February 2023 SSW. For each valid time, we compare the 18-hour-lead zonal-wind forecast against the shortest-lead forecast and compute the latitude-weighted global-mean RMSE at every model level. Low RMSE indicates that forecasts initialized at different times converge on a similar atmospheric state, while higher RMSE reflects forecast drift and reduced predictability.

We find that the SSW does not simply degrade forecasts: it actively creates a window of enhanced predictability. Stratospheric, mesospheric, and lower thermospheric RMSE collapses once the vortex breaks down, because the disruption is dominated by large-scale, strongly forced planetary-wave dynamics that the model captures consistently. This predictability window propagates upward into the MLT with only modest delay. The recovery phase tells the opposite story: an unstable new vortex briefly erodes stratospheric predictability.

Our findings suggest that the atmosphere is therefore most predictable when it is most disturbed- and least predictable in the quieter aftermath. These results reframe SSWs as opportunities, not obstacles, for operational whole-atmosphere forecasting, and highlight the vertical reach of stratospheric predictability into the MLT.