Changing Jet-Stream Waviness Assessed Using Self-Organizing Maps (SOMs)

Amplified jet-stream configurations are known to be associated with slow-moving weather patterns that often lead to extreme weather events (Screen and Simmonds, 2014). An outstanding question, however, is whether the frequency of these configurations is changing in response to human-caused climate change. In this study we use an analysis tool known as Self-Organizing Maps to investigate whether the frequency of highly amplified jet-stream patterns has changed in recent decades, and whether climate models project a change in frequency in the future under business-as-usual increases in greenhouse gas concentrations. The SOM is used to separate daily atmospheric fields into characteristic patterns. In this application, we use single contours of 500-hPa heights from reanalyses as well as from simulations by four of the CMIP5 climate models for the 20th century, using observed forcing, and for the 21st century, forced according to the RCP 8.5 scenario. We calculate a Wave Amplitude Index (WAI) for each day in the reanalysis record. SOMs are used to identify characteristic patterns in the 500-hPa contours over the northern hemisphere. Mean WAI values are calculated for each cluster of days within each pattern, and changes in the frequency of occurrence of each pattern are determined. We find that the patterns with largest mean WAI values correspond with the negative phase of the Northern Annular Mode (aka Arctic Oscillation), that the frequency of high-WAI patterns has increased in recent decades, and that model simulations project further increases in these patterns by the end of the 21st century.