Differential orographic impact on sub-hourly, hourly, and daily extreme precipitation

Extreme precipitation of multiple durations is responsible for major natural hazards in mountainous regions, such as flash floods and debris flows. Understanding the orographic impact on the statistics of precipitation extremes is therefore crucial for improving hydrological design and risk management strategies. Here, we use a novel statistical approach for the analysis of extremes based on ordinary events to improve our understanding of the orographic impact on extreme precipitation of durations ranging between 5 min and 24 h. We focus on Trentino, a rough orographic region in the eastern Italia Alps, and use data from 78 quality-controlled rain gauges with 5-minute resolution. We show that our framework well reproduces the statistical properties of the observed annual maxima (Nash-Sutcliffe 0.82-0.95, Bias from -4% to 7%) as well as their relation with orog- raphy. We then exploit the reduced uncertainty of this approach to quantify the orographic impact on precipi- tation right-tail statistics and on extreme return levels using a regression analysis. We identify two main modes of orographic relationship: a reverse orographic effect for hourly and sub-hourly durations (10-20% decrease per 1000 m elevation) and an orographic enhancement for durations of ~8 h or longer (7.5-10% increase per 1000 m elevation). We observe that these two modes result from three main precipitation regimes, which show different proportion between extreme and very-extreme events and which emerge at very short durations (~20 min or shorter), mid durations (~30 min-1 hour) and long durations (~2 h or longer). These findings are of interest for risk management applications and climate change impact studies.