Evaluation of temperature extremes over central Europe and their links to atmospheric circulation in CORDEX RCMs

Automated classifications of atmospheric circulation are routinely used to link synoptic-scale circulation with temperature variability. Though powerful in general, classifications have considerable limitations regarding their skill to capture synoptic links to temperature extremes.

We evaluate and optimize several parameters of the popular method of self-organizing maps, in order to make the method better suited for studying central European temperature extremes. Furthermore, two methods of discretizing Sammon projections of atmospheric circulation have been developed to complement the image obtained by SOMs, and all methods have been used to analyse ERA5 SLP fields in relation to winter extremes.

Here, we plan to apply the new optimized classifications to daily winter and summer SLP fields from the outputs of evaluation and historical runs by CORDEX RCMs to study the skill of the methods to identify RCM biases in simulated circulation and their links to biases in temperature extremes. Furthermore, we plan to utilize various indices of low-frequency large-scale circulation to assess to what extent the eventual biases propagate from the driving (reanalysis, GCM) data.