Probability of Atmospheric Circulation Pattern Occurrence in Pre-Industrial, Historical, and Future Climates

Occurrence of extreme weather and climate events is dependent on the spatial and temporal confluence of a host of meteorological ingredients. Primary among these ingredients is the circulation pattern of the atmosphere. Research indicates that synoptic- to regional-scale circulation patterns dictate the atmospheric environment, and that the likelihood of an extreme event changes depending on the circulation pattern type present. It follows then, that changes in the frequency and/or duration of particular circulation patterns may alter the frequency or severity of extreme events. Recent work analyzing reanalysis data identified robust multi-decadal trends in the occurrence and duration of some circulation pattern types over select regional mid-latitude Northern Hemisphere domains. Despite the detection of circulation pattern trends, their attribution to internal variability and/or anthropogenic forcing remains unresolved. Here, in an initial step toward attribution, we examine the likelihood of detecting robust circulation pattern trends in pre-industrial, historical, and future climate simulations using the CESM1 (CAM5) Large Ensemble (LENS) Community Project single-model, multi-realization framework. To identify seasonal atmospheric circulation patterns and determine their temporal variation, we analyze mid-atmospheric geopotential heights using self-organizing map (SOM) cluster analysis. Within the LENS framework, we quantify the probability of detecting seasonal circulation pattern trends in climate systems devoid of human influence, with observed human influence, and with RCP8.5 projected forcing. In addition to pattern trend assessment, for each regional domain, we quantify the likelihood that a particular circulation pattern type, e.g., an anticyclonic blocking circulation, will change in frequency or duration due to increased radiative forcing.