The term “weather whiplash” was recently coined to describe abrupt swings in weather conditions from one extreme to another, such as from a frigid cold spell to anomalous warmth or from drought to prolonged precipitation. These events are often highly disruptive to agriculture, ecosystems, and daily activities. In this study we propose and demonstrate a novel metric to identify weather whiplash events (WWEs) and track their frequency over time. We define a WWE as a transition from one persistent large-scale circulation regime to another distinctly different one, as determined using an objective pattern cluster analysis called self-organizing maps (SOMs). We focus on the domain spanning North America and the eastern N. Pacific Ocean. A matrix of representative atmospheric patterns in 500-hPa geopotential height anomalies is created. We analyze the occurrence of WWEs originating with long-duration events (defined as lasting 4 or more days) in each pattern, as well as the associated extremes in temperature and precipitation. A WWE is detected when the pattern two days following a long-duration event is substantially different, measured using internal matrix distances and thresholds. Changes in WWE frequency are assessed objectively based on reanalysis and climate model output, and in the future with climate model projections. Temporal changes in the future under RCP 8.5 forcing are more robust than in recent decades, with consistent increases (decreases) in WWEs originating in patterns with an anomalously warm (cold) Arctic.