Storm tracks can be defined based on Eulerian band-pass filtered variance and covariance statistics, or Lagrangian cyclone/anticyclone tracking statistics. It is well known that the North Pacific storm track exhibits a mid-winter minimum based on Eulerian statistics. However, how the mid-winter minimum appears in terms of cyclone/anticyclone statistics is still under debate.
In this study, the seasonal cycle of the statistics of cyclones and anticyclones are examined based on 44 years of ERA5 data. Cyclones and anticyclones are defined based on minima and maxima in sea level pressure (SLP), as well as maxima and minima in relative vorticity at 850 hPa. Tracking is conducted on the full SLP and relative vorticity fields, as well as spatially and/or temporally filtered SLP and relative vorticity perturbation fields.
When cyclones and anticyclones are defined based on the total SLP field, the results suggest a maximum in cyclone frequency in midwinter over the Pacific, and a marked minimum in anticyclone frequency, consistent with a recent study that claims that the midwinter minimum should be explained in terms of a midwinter minimum in the frequency of anticyclones. However, the marked minimum in anticyclone frequency disappears when anticyclones are defined using either spatially or temporally filtered SLP. Instead, based on these alternative cyclone/anticyclone metrics, the midwinter minimum can only be found when the seasonal cycle of cyclone intensity is taken into account. Anticyclones defined using negative vorticity centers also do not show any indication of a marked midwinter minimum. Given these discrepancies, as well as previous results showing that climate model projected changes in cyclone activity strongly depend on how cyclones are defined, it is important to reassess how cyclones and anticyclones should be defined.

