Interdecadal Variations in Northern Hemisphere Winter Storm Track Intensity

In this paper, the interannual variations in the Northern Hemisphere winter storm tracks have been examined based on 51 winters (December–January–February) of NCEP–NCAR reanalysis data. The leading empirical orthogonal function (EOF) corresponds to the simultaneous strengthening/weakening of both the Pacific and Atlantic storm tracks. Interannual and month-to-month variations in the Pacific and Atlantic storm tracks are found to be significantly correlated. The principal component associated with the leading hemispheric EOF exhibits pronounced interdecadal variability. There appears to be a transition during the early 1970s from a weak storm track state prior to 1972/73 to a strong storm track state subsequently. Decadal mean storm track intensity during the 1990s is about 30% stronger than that during the late 1960s and early 1970s. The relationship between variations in storm track intensity and low-frequency (seasonal mean) flow anomalies has also been examined. It is shown that storm track variations and their associated seasonal mean flow anomalies are highly correlated. Relations to several other modes of interdecadal variability have also been explored. It is shown that part of the storm track variations may be related to the Arctic oscillation and part to the interdecadal ENSO-like variability. However, even when storm track variations that are linearly dependent on these other modes have been removed, substantial interdecadal variations still remain. In order to ascertain that the interdecadal variations have not been introduced into the reanalysis data by changes in the observation network, radiosonde observations along the storm track have been examined. Preliminary analyses suggest that the radiosonde observations are largely consistent with the reanalysis data, except for the possibility that biases in the reanalysis data may have led to an overestimation of the interdecadal variability. Unfortunately, since the peaks of the storm tracks lie over the oceans where no radiosonde stations with continuous record can be found, the variations over the storm track peaks cannot be verified.