Heavy Precipitation Events (HPE) pose increasing risks to infrastructure, public safety, and water management. This study examines the dynamics and the intrinsic predictability of HPE in Portugal, focusing on the role of Atmospheric Rivers (AR). Using reanalysis data, objective weather pattern classification, and dynamical systems metrics, we show that the average AR-linked HPE exhibit 36% higher precipitation intensity than non-AR events. Primarily attributed to stronger low-level winds that increase moisture fluxes into the region, rather than a greater total column water vapor content. We employ a dynamical systems framework to evaluate the intrinsic predictability of HPE, analyzing the evolution of upper- and lower-level atmospheric fields. This allows a systematic classification of events based on their synoptic and dynamic signatures. Our findings reveal that high-predictability events are typically associated with well-defined, deep extra-tropical cyclones near 50°N, 15°W, whose mean sea-level pressure anomaly is roughly twice that of low-predictability systems. These events also exhibit enhanced jet stream interaction and more coherent Rossby wave patterns, with average precipitation intensities approximately 80% greater than those of low-predictability events. A detailed case study of the extreme mid-December 2022 event, which caused widespread flooding and damage in Western Portugal, exemplifies the interplay between AR, synoptic dynamics, and forecast confidence. Our results emphasize the value of integrating AR diagnostics with a dynamical systems perspective to improve understanding and prediction of HPE, providing a process-based foundation for enhanced forecasts in Portugal and similar mid-latitude coastal regions. • Intrinsic predictability of heavy precipitation linked with Integrated Water Vapor transport. • The intrinsic predictability of extreme events is quantified using dynamical systems metrics. • Events associated with atmospheric rivers are potentially more predictable. • High-predictability extreme events are associated with well-defined synoptic features.
Bartfeld et al. (Wed,) studied this question.