Unveiling critical rainfall patterns triggering torrential processes

Critical rainfall conditions initiating torrential processes like sediment-laden floods and debris flows in steep headwater catchments represent a multifaceted problem in Alpine communities, necessitating comprehensive adaptation and mitigation strategies to safeguard both society and the environment. Understanding the relationship between rainfall drivers and event occurrence is needed for a mechanistic understanding of the initiation process and therefore represents an important basis for developing adequate risk management strategies in a changing climate.

In this study, we investigate rainfall patterns triggering debris flows, debris floods, fluvial sediment transport and floods based on hourly rainfall time series derived from combined radar-rain gauge data for more than 3,600 torrent events in the Austrian Alps between 2003 and 2022. We consider time periods spanning seven days prior to event occurrence as well as the event day itself. These time series are clustered using longitudinal k-means on the cumulative rainfall sums over the whole time period leading up to the event.

Results reveal different archetypical precipitation patterns. While all of the patterns exhibit some rainfall on the event day, differences emerge with respect to antecedent precipitation. Major patterns include an archetype featuring stepwise increases, several patterns with breakpoints followed by an increase, and patterns characterized mainly by differences in their slope, i.e., overall rainfall magnitude. These patterns are largely consistent across all considered process types. A first analysis of the spatial distribution of patterns indicates that some patterns occur mainly south of the main Alpine ridge while others occur all over the Eastern Alps. The results of this study will help improving early warning systems and guiding model development for the initiation of mass flow processes.