Present and future European summer weather types obtained by Self-Organizing Maps and a careful GCM preselection

Climate change represents a pivotal challenge of the current time. It is projected that extreme weather events, including heat waves, droughts, and heavy precipitation, will become more frequent and intense. In order to describe and analyze these developments in mid-latitudes, the classification of atmospheric dynamics into weather types (WTs) is frequently applied. WTs characterize large-scale circulation patterns on a synoptic scale and establish a direct link to regional climate conditions. In Central Europe, the occurrence of different WTs plays a major role in spatiotemporal temperature and precipitation patterns. Within this context, the present research investigates the potential changes in circulation patterns under the influence of climate change. The objective of this thesis is to identify and analyze WTs in the European summer (June-August) for both the recent past (1990-2019) and the future (2020-2099; SSP5-8.5) based on an objective, automated classification process. In this study, Self-Organizing Maps (SOMs) are employed to organize meteorological input data (here: sea level pressure and wind in the upper troposphere) into classes (nodes) that represent the different WTs. The primary methodological refinement is the deliberate deviation from the assumption of WT stationarity for the future analysis while integrating a newly developed model selection procedure that identifies the most reliable global climate models (GCMs) for a specific region (here: Central Europe). This approach has not been pursued in the literature to date. The three selected GCMs (MPI-ESM1-2-LR r29i1p1f1, CanESM5 r1i1p1f1, and MRI-ESM2-0 r5i1p1f1) were evaluated regarding their WT properties by using ERA5 reanalysis data. The GCMs show a realistic reproducibility of reanalysis-driven WTs in terms of frequency, duration, synoptic structure, and climatological trends. This fulfills an essential prerequisite for making projections into the future. In all future GCM-driven WTs a consistent increase in 2m air temperature, with trends reaching up to +0.68 °C per decade is demonstrated, with two of the three models also simulating a decrease in precipitation. The projections also indicate that the mean summer temperature will increase by approximately 5.11 °C by the end of the 21st century, relative to the reference period 1990-2019. The mean summer precipitation total is projected to be 258 mm, representing a decrease of 54 mm (-17.3%) compared to the reference period. Regarding the temporal development, WTs with a significant increase in frequency are characterized by pronounced high-pressure systems over the continental landmass. These are accompanied by positive anomalies in incoming and outgoing longwave radiation, indicating a general increase in energetic activity in these WTs. Conversely, WTs exhibiting a significant decline in frequency are characterized by cyclonic, cool, and precipitation-rich conditions. The identification of qualitatively different future WTs with a tendency of increasing persistence suggests the conclusion of structurally altered circulation dynamics in the future. With regard to extreme summer conditions in Southern Germany, the results indicate a substantial increase in the frequency of hot and dry summers throughout the 21st century. The close relationship between precipitation and temperature is illustrated by an average decrease in precipitation anomalies of 0.43 σ per 1 σ increase in temperature anomalies. In each of the three models, the consideration of extreme summers reveals a model-specific extreme WT with pronounced high-pressure regimes, linking these anomalies to specific WTs. The projected increase in the frequency of these extreme WTs underscores the pivotal role that certain circulation patterns play in the occurrence and duration of extreme summer conditions. Finally, the large-scale circulation conditions are transferred to local scales. For Nuremberg, the 98.7th percentile of the daily maximum temperature, corresponding to 35 °C in the recent past (1990-2019), is projected to shift to 39.3 °C by the end of the century. An area-wide projection for the Franconia region, based on the high-resolution BAYWRF dataset, illustrates not only the significant increase but also the significant spatial variability of the expected temperature changes. The findings of this thesis offer method- and content-related contributions in the field of climate research. They underscore the efficacy of the integrated approach, which combines a GCM selection procedure with a SOM classification, in the context of analyzing present and future WTs in Central Europe, without the assumption of WT stationarity. The results establish a solid foundation for future research investigating the interaction between large-scale atmospheric circulation and climate impacts, as well as regional studies, as exemplary shown by the BayTreeNet project. On a local scale, strategies adapted to future climatic conditions can be analyzed and developed, as exemplified by the REKKE project.