Precipitation-driven typology of storms in the Alps

Georgia Papacharalampous , Eleonora Dallan, Moshe Armon, Joydeb Saha, Colin Price, Marco Borga, Francesco Marra

Advances in precipitation science often depend on categorizing storms into homogeneous classes, particularly convective- and stratiform-like. To address this need, this study introduces an Alpine storm typology derived from the pairing of a vast sub-hourly gauge dataset, comprising over 790,000 independent storms, with an objective method, driven solely by precipitation features and optimized for big data clustering. Five dominant classes were identified from this partition, with distinct clustering fingerprints in maximum intensity, total volume, total duration and temporal variability, alongside clear spatial organization. Additional traits (initiation month, peak solar time, lightning count) and comparisons with radar-based benchmarks suggest that one class is likely associated with convective-like extremes (high intensity, summer/afternoon peaks, high lightning) and another with stratiform-like behaviors (medium intensity, large volume, long duration, low lightning), while the remaining classes gather moderate and minor storms. The typology could support applications such as class-specific stochastic simulation, class-informed bias adjustment of climate projections or multi-class extreme value analyses. Climatological investigations revealed, among others, higher convective-like activity in recent years and specific regions, offering direct evidence on evolving hazard risks. We provide the historical occurrences of the classes as an open dataset to facilitate further investigation of Alpine storm dynamics and their implications.