Robust Assessment of Atmospheric Circulation Trends over Central Europe Using a Multi-Classification Approach

Changes in atmospheric circulation are key drivers of regional climate variability and long-term trends in the midlatitudes, particularly across Europe. This study investigates the concurrence and temporal trends in the frequency and persistence of atmospheric circulation types (CTs) over Central Europe for the period 1951–2020. We apply five distinct classifications: the large-scale Grosswetterlagen (GWL), and four regional-scale schemes—Lityński (LIT), Niedźwiedź (TN), Péczely (PEC), and the objective Jenkinson–Collinson (JCT) classification adapted to the study area.

The analysis is conducted year-round as well as for the four meteorological seasons, offering a comprehensive perspective on synoptic-scale variability and its evolution over time. Results show substantial methodological differences between individual classification schemes. To assess the consistency and co-occurrence of CTs across methods, we employ a suite of statistical measures including contingency tables, mutual probabilities, Cramér’s V, Adjusted Rand Index (ARI), and Normalized Mutual Information (NMI). These measures highlight both areas of agreement and systematic discrepancies, underscoring the risk of overinterpretation when relying on a single classification.

Despite differences, some robust trends emerge: notably, a consistent increase in westerly types and decline in meridional flow patterns during winter, in agreement with a strengthened and eastward-shifted North Atlantic Oscillation (NAO). In contrast, summer and autumn display weaker or more inconsistent trends. Persistence analysis also reveals classification-dependent behavior, with some methods overestimating the duration of synoptic types.

Our findings support earlier research emphasizing the necessity of multi-classification approaches in climatological analyses. We advocate for using classification ensembles to improve the robustness of circulation trend detection, especially in studies related to climate change attribution, regional modelling, and statistical downscaling. The results contribute to refining synoptic-scale diagnostics and improving the interpretability of long-term atmospheric circulation changes over Europe.