Investigating links between the Arctic and East Asian cold extremes with self-organizing maps

<p>To investigate the connection between the Arctic and East Asia cold events, we categorize the continuum of Eurasian circulation patterns using self-organizing maps (SOMs). The analysis identifies several circulation patterns associated with East Asian extreme temperature occurrences during winter months from November to January (NDJ) 1979-2020, including the warm Arctic-cold Siberia-like pattern (WACS).</p><p>The SOM-derived trend pattern of geopotential height at 500 hPa illuminates a zonal wave train of wavenumber 4 extending from central Eurasia to eastern Pacific. The analysis shows that the observed trend is a result of competing influences from the change of frequency of &#160;several different SOM patterns. &#160;</p><p>SOMs reveals that some patterns undergo statistically significant trend over certain time periods, notably a positive trend of WACS-like pattern since late 1990s that has resulted in more frequent cold events in East Asia. The composites based on the WACS-like SOM pattern show that an anomalous Siberia High leads to cold surface air temperature over a broad region in East Asia, which can provide potential sources of predictability.</p><p>In the associated SST composite, a cold tongue of SST emerges in the southeast Pacific centered around 15o stretching offshore from the coast of Peru, which corresponds closely to the observed NDJ SST trend pattern. We explore the relationship between the WACS-like SOM pattern and El Ni&#241;o-Southern Oscillation (ENSO). We also evaluate the subseasonal impacts of WACS on East Asian cold events in winter using a probabilistic forecast model determined through multiple linear regression with WACS index, ENSO, and long-term trend as predictors.&#160;</p>