The dynamical footprint of year-round North American weather regimes

Weather regimes have been defined over multiple regions and used in a range of practical applications, including subseasonal-to-seasonal forecasting and climate model evaluation. Despite their widespread use, the extent to which regimes reflect physical modes of the atmosphere is seldom investigated. Here, we adopt a year-round classification of four North American weather regimes, with a fifth “no regime” class, and leverage dynamical systems theory to investigate their dynamical properties. We find that when the atmospheric flow is assigned to a regime, it displays persistent characteristics and a lifecycle-like temporal evolution. We further find that, regardless of season, these characteristics are enhanced when the atmospheric flow displays a comparatively strong projection onto the cluster-mean of the regime to which it is assigned (while the reverse is true for a weaker projection). We interpret these results as evidence that the four North American weather regimes are physically-meaningful, with a clear dynamical footprint.