Abstract
Advances in subseasonal-to-seasonal (S2S) forecasting have broad socioeconomic benefits. Stratospheric variability has been identified as a major source of surface predictability on S2S timescales, yet its influence on forecasting North American cold extremes has not been fully understood. Here, using hindcasts from a coupled S2S forecast system, we identify a wave-1–like pattern among leading sources of predictability in the lower stratosphere during extended winters. This mode, characterized by strong stratospheric wave activity, improves second-week temperature prediction over North America by 15% (2 to 28%). It also increases the likelihood of North American cold extremes by 50%. Prediction skill for these extremes increases by 28% in forecasts conditioned on the wave-1–like stratospheric mode, across regions inhabited by over 69% of the Canadian population and over 30% of the U.S. population. These skill enhancements are attributed to the vertical wave coupling between the stratosphere and troposphere, distinct from the well-known downward influence of sudden stratospheric warmings (SSW). By conditioning forecasts on stratospheric states, our findings highlight strong stratospheric waves as windows of opportunity for skillful winter S2S predictions.