Using Self-Organizing Maps to Classify and Investigate Atlantic Jet Regimes and Their Response to Forcing

Abstract:

The winter jet stream in the North Atlantic has been shown to preferentially occur at three distinct latitudes [Woolings et al., 2010; Woolings et al., 2018], which we will call the three Atlantic “jet regimes.” Distinct physical mechanisms may be responsible for each of the three jet regimes—for example, the northernmost jet regime is strongly linked to the Greenland tip jet [White et al., 2019]. Additionally, recent research by Goss et al. [in review] has shown that sudden stratospheric warmings (SSWs) impact the likelihood that the Atlantic jet will be in any particular jet regime. Analyses of these regimes has mostly focused on the use of a “jet latitude index” (JLI), which is determined by finding the latitude of the peak zonal winds over some latitude range, averaged over some longitude range. However, a more natural classification scheme using an objective pattern-finding algorithm may allow for a better physical understanding and interpretation of the three regimes and their interactions, may clarify their relationship with SSWs, and may help address the question of if and how these regimes may change with increased CO2 forcing.

Therefore, we perform an area-weighted Self Organizing Map (SOM) analysis on the 700-hPa zonal wind field over the North Atlantic basin in November-March. These calculations are performed in the ECMWF Interim Reanalysis (ERAI) data set, and in a subset of CMIP6 models. The SOM analysis is used to investigate aspects of the three Atlantic jet regimes. Specifically, 1) how do the SOM-derived jet regimes compare to JLI-based analyses, 2) what is the climatological frequency of transitions between jet regimes, 3) how do SSWs impact the frequency of these SOM-derived jet regimes, 4) what are the meteorological features of the three jet regimes, and 5) do the jet regimes change in response to CO2 forcing, and if so, do these changes manifest as changes in seasonal frequency, changes in the patterns themselves, or both?