Search

This paper analyzes the relationships between the circulation regimes of the 500-hPa height (z500) and 250-hPa zonal winds (u250) in the Pacific–North America region during boreal winter, and the 45-day Northern Hemisphere oscillation in z500. The regimes were calculated using a k-means clustering applied to the leading 12 principal components of the combined z500–u250 anomaly fields. We divided the oscillation into eight arbitrary phases. The oscillation phase z500 composite maps are spatially well correlated with regime z500 composites: phases 1–2 are best correlated with the Arctic Low, phases 3–5 are best correlated with the Pacific Trough, phase 6 is best correlated with the Arctic High, and phases 7–8 are best correlated with the Alaskan Ridge. We found that these correlations are generally consistent with the regimes that tend to occur during the individual oscillation phases: the Arctic Low occurs above significance in phases 1–2, the Pacific Trough occurs above significance in phase 3, and Alaskan Ridge occurs above significance in phases 7–8. Therefore, the oscillation has a preferred order with respect to the regimes. The regime transitions indicate a pattern that moves through the Pacific Wavetrain, a regime that appears for k = 5 as a mean state. Transitions out of this regime into different regimes are preferred in different phases of the oscillation. These results imply a possible enhancement to regime prediction using the low-frequency oscillations in combination with regimes. Significance Statement: Subseasonal prediction, weather forecasting in the 2–4-week range, is important for many parts of society, e.g., water managers, emergency response units, and farmers. However, current prediction skill in this time range is low. This paper performs an initial analysis of a possible method to increase weather statistic prediction skill beyond 10 days in the winter for the Pacific–North America region. This is done by combining two ways of looking at large, long-lasting patterns of pressure systems in the atmosphere, which are associated with various weather statistics like precipitation extremes and storminess. The results indicate this method holds potential skill for enhancing subseasonal prediction. Further investigation might yield forecasting improvements in this important time range. [ABSTRACT FROM AUTHOR]