Combined Oscillation-Regime Forecasting

One of the most challenging forecasting problems in atmospheric science is the accurate prediction of weather and climate for sub-seasonal to seasonal predictability time scales. In the past, two of the methods used employ sub-seasonal variability (circulation regimes and low-frequency oscillations). They have been applied separately to investigate this issue, but rarely have they have been considered together. This work identifies and uses the associations of oscillations with regimes to predict the transitions between regimes at time scales longer than 10 days. Previous work (Korendyke and Straus, in review) compared boreal Pacific North American winter regimes and the Northern Hemisphere 45-day intra-seasonal oscillation, providing the foundation for the prediction of circulation regime transitions. This prediction is carried out first using a long historical atmospheric data set, 20CRv3, and subsequently on a set of forecasts made with a Linear Inverse Model. Combining these two methods into a new type of forecasting framework and diagnosing its regime forecasting skill in the 1-4 week range might lead to enhanced extreme weather forecasting in the 3-4 week time scale.