This work demonstrates a new forecasting approach for wintertime SPV anomalies, suggesting high predictive skill based on early-season data.
In Part I, we represent the annual evolution of the Northern Hemisphere stratospheric polar vortex (SPV) in individual years using an ellipse, defined through six ellipse parameters, for the phase‐space trajectory of the daily time series of MU (mass‐weighted zonally integrated zonal wind at 60°N above 400 K) and M (stratospheric air mass over 60–90°N above 400 K) from July 1 to June 30 of the following year. Building on that foundation, this study puts forward a new paradigm for long‐lead forecasts of wintertime SPV anomalies through predicting the six ellipse parameters before the cold season. A set of 17 predictors for predicting these yearly‐varying ellipse parameters, computed exclusively from data available before October 1, are selected from three primary sources: external factors, isentropic mass circulation indices, and phase‐space trajectory precursors. The regression models exhibit high predictive skill with cross‐year correlations between predicted and observed ellipse parameters ranging from 0.79 to 0.91. When translated through the elliptical representation, yearly predicted ellipse parameters yield skillful forecasts of continuously evolving subseasonal‐to‐seasonal (S2S) anomalies of the SPV from October through the following April. The correlation skill of forecasts for seasonal‐mean SPV anomalies remains above 0.8 at lead times of one to 6 months, and that for two‐weekly anomalies is still above 0.6 over lead times of two to 5 months. The forecasts also capture the amplitude and timing of wintertime SPV extrema. These results demonstrate that the elliptical representation provides a skillful framework for long‐lead forecasting of SPV anomalies at S2S scales.
No takes yet. Share an insight, caveat, or question.
Secor et al. (2026) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: