Teleconnection-informed clustering of temporally compound hydroclimatic extremes in Europe: post-drought extreme rainfall probability and long-term trends

Understanding the interplay between droughts and subsequent extreme rainfall is critical for managing flood risk and water resources under climate change. A teleconnection-aware analysis examined compound drought–rainfall extremes across Europe (1976–2023). Daily rainfall and gridded SPEI-3 indices were used to quantify the probability of extreme rainfall events (≥ 95th percentile) occurring from concurrent conditions up to 12 months after drought onset, evaluated across six post-drought windows (0–2, 2–4, …, 10–12 months). Hydroclimatic regimes were identified using multiple clustering algorithms, including conventional (K-means, Hierarchical) and novel approaches (Black Hole Clustering, Spectral Bridges - SB), considering as input the probability of extreme rainfall during drought and across six post-drought windows and correlations between rainfall (and SPEI-3) and teleconnection indices. SB produced the most spatially coherent and interpretable classification, yielding three clusters. In northern Europe (Cluster 1), extreme rainfall probability peaks at 10% within 2–4 months, with significant positive rainfall trends (Z = 2.99) but persistent negative SPEI-3 (Z = − 5.14). Southern Europe (Cluster 2) shows stronger, persistent extremes, rising from 2% during drought to 14% within 8–10 months, modulated by Mediterranean convection and teleconnections, while SPEI-3 trends remain negative (Z = − 5.15). Central Europe (Cluster 3) exhibits delayed probability peaks (13%, 10–12 months) with the strongest negative SPEI-3 trends (Z = − 6.26). Moreover, feature importance analysis indicates that teleconnection indices (e.g., Arctic and Mediterranean oscillations) dominate cluster assignment. These results show that extreme rainfall can occur during or after drought conditions without necessarily implying drought termination, and that large-scale atmospheric circulation plays a key role in shaping the occurrence of compound hydroclimatic extremes. Integrating clustering, trend detection, and teleconnection-informed analysis provides a robust framework for assessing evolving extremes and supports quantitative guidance for flood risk management, water planning, and climate adaptation across Europe.