Precipitation-driven storm types and their climatology across the Alpine range

The separation of storms into physically meaningful classes, including the key distinction between convective and non-convective events, is crucial for advancing precipitation science. Indeed, each of these classes may necessitate different modelling strategies, or distinct bias adjustment procedures for climate model simulations. Here, we present a large-scale study that aimed at achieving this separation only based on information from precipitation timeseries. We focused on a vast set of sub-hourly rain gauge records collected from four countries across the Alpine region and extracted hundreds of thousands of storms. We used an unsupervised clustering algorithm based on a small set of features to organize the storms into storm types. Despite the simplicity of the clustering approach, we successfully distinguished convective storms from other types, as validated using independent features that were not involved in the clustering, such as lightning counts. We analyzed the climatology of the storm types, including investigations of their spatial coherence and temporal changes in their occurrence. Overall, we believe that the storm clusters we provide can be used for several purposes, ranging from developing stochastic models tailored on the storm types of interests to improving bias adjustment methods for climate simulations. Given its simplicity and versatility, the framework can be transferred to other regions globally, with marginal adjustments based on the prior knowledge of the regional climatology and on the research objectives.

Our study was carried out within the RETURN Extended Partnership and received funding from the European Union Next-GenerationEU (National Recovery and Resilience Plan – NRRP, Mission 4, Component 2, Investment 1.3 – D.D. 1243 2/8/2022, PE0000005).